Yo chochajin, It gets easier with the time.

I guess the sentences method only gets to its full potential after you have covered all the most coomon readings with vocabulary.

I don't do the sentence method. I just learn vocab from daily exposure (and JLPT vocab list drilling) and end up knowing readings from that. I don't really get what you mean by full potential. Like I said being able to name all the possible readings for a kanji is little more than trivia. With the multiple possible readings per kanji you can't reliably guess the reading of a word anyways.

Knowing a reading for a kanji happens pretty quickly for me, after one or two words.

Do second language learners of english spend time learning all possible morphemes in isolation because it'll help them when they encounter those morphemes in words?

Sometimes divide and conquer is the best plan, sometimes 一石二鳥 is.

You've made your point pretty clearly. Everybody is free to do it that way. In fact, everybody SHOULD do it that way at first. I did. I've tried learning Japanese a few ways and I can compare them:

Without RtK: Sucked
With RtK: Much much better. What was that crap I was learning before this? That wasn't Japanese.
With RtK and a strong onyomi knowledge: EVEN BETTER. New words just seem to enter my brain and stay there. It takes very little effort to learn the meaning and pronunciation of most new compound words, and they jump to the tongue much quicker in conversation. Life is good.

Nobody is saying you have to do it like this. Everyone has their own path. I'm only citing my own experience. I don't understand the point of lecturing people on how they SHOULDN'T learn Japanese.

Calm down shakkun. He did put his opinion clear, maybe he was a bit harsh, but still lets calm down.

Yeas Jarvik, kanjis have many many readings, but knowing 1 reading for every kanji can't be considered trivia. It is basic knowledge. It just is a step further from RTK, and it takes little effort.

The full potetial of the sentences method that I mentioned above is to be free to understand the language. To be free to understand the deep relationship between words that binds the language together, not the shallow one that we call grammar.

Free from readings, free from grammar, free from vocabulary.

Ps, you should try using a SRS or your vocabulary drilling.

err, I do use an SRS for my vocab drilling (would be hard to manage a deck of ~8000 otherwise). I just don't sentence mine etc since flashcards are basically throwaway and I don't want to invest that much time in manual labor when I could be doing more material in less time.

Do you use paper flashcards?

Anki.

alyks Wrote:

Tobberoth Wrote:My personal experience is that there really isn't any good reason to learn kanji readings out of context. The reason being that most kanji don't HAVE a reading out of context because they simply aren't used like that (there are of course tons of exceptions like 水 and 家 for example). Question being, what good does it do to know a kanji reading?

When I'm going through 100-150 sentences a day, it helps not having to memorize the compound readings.

In a sentence that has a new compound word (and probably a new verb and new grammar point) I'll have one less thing to have to memorize. It really adds up fast.

But that was my whole point, you have to remember the compound readings anyway. Most kanji have several on'yomi, you still have to memorize which one is used in any given compound so how does that make it any different? All it takes is to learn one compound with the kanji to learn just as much as this movie method, there is no difference in time invested in memorizing at all.

My whole point as to why I feel one shouldn't learn readings out of context is because in my opinion, you learn them for FREE in context. You don't have to spend any time at all, as soon as you learn that 砂漠 is read "sabaku" you learn that 砂 can be read "sa" and 漠 can be read "baku". I didn't spend an extra minute learning sabaku, I just learned it and because I learned it, I learned the readings automatically. I know the readings because I know the word, not the other way around. Instead of having to put the kanji in movie scenes and advanced mnemonics like that, you have a context and a word, so you won't forget it. In the very few cases where you can't parse the word for which kanji has which reading, all you have to do is check the readings in wwwjdic or jisho.org (which you would of course have to do anyway if you did the movie method, so it's not like there's any difference in time invested.)

Okay, after using the method for 2 days I'm now here to report back:
Basically what I did so far was trying to figure out what works best for me. I really have a headache by now. This is tough!

So what I did until know was using anime characters for each reading and then just created a story with the keywords of the kanji that have that reading. It was nice, a lot of fun, but ..... I think once you have 200+ stories - some of them pretty LONG ... you'll confuse things like WOAH! Although I think it's hard to remember the LONG stories and then still know what were you keywords and what are just words in the story. Know what I mean?

So, yesterday night I read alyks blog again and this whole thread.
Here's what I think:
It's probably way better to connect your RTK1 story with the kanji once again when you place it somewhere. It will be easier and faster when remembering things later (alyks did that right away because he was learning "meaning" and "reading" of a kanji at the same time - but for those of us who have finished RTK1 I think that's the way to go) - although I imagine it might be harder to do if you always have to connect everything: kanji - location - keyword - story. Any feedback? I know some of you have been doing exactly this.

Next problem. As I write stories the kanji are somewhat connected to each other and not really independent. And it's hard to remember sometimes, too. Let's say "give" is the keyword. When I see the kanji and remember the keyword is "give" - then I have to think back ... okay, where was a story where somebody "gave" something .... Uhm .... that might be pretty tough.
Doing what alyks did might be the better solution.

So here's what I want to try next (not knowing if it'll work out this time):

Use anime series (problem: coming up with 200+ series, but I think better than using random locations - I don't play rolegames or videogames at all btw.).
Put each kanji in a location / scene within this series seperately.
Next problem: I think this is way more work than what I did before - but of course I want the method to be as effective as possible.

Also, do you write down anything when using this method? If so how exactly do you write it down?
Alyks, on your blog you have an example movie. Did you write the stuff down exactly like you did on your blog when doing your method?

Quote:KOKU: Pulp Fiction

Revelation: 告
Jules (Samuel L. Jackson) had a revelation. After being shot at, he's convinced it was "divine intervention" that saved their lives. So the scene is when they're in the diner talking about it. Jules is shoving a big cow into his mouth as part of his breakfast.

Cruel: 酷
There are so many scenes to choose from. Let's pick the scene where Vincent (John Travolta) is on the toilet, and Butch (Bruce Willis) finds him there about to shoot him. Normally I'd use the essential image of "shoving a big cow into his mouth", but it doesn't quite fit in here. So instead we replace Vincent here with a cow who's mouth is hanging open in surprise (a stunning display of anthropomorphism, if that helps). Butch splashes the cow with sake.

Country: 国
This one was easy for me. Vincent just got back from Amsterdam - a foreign country. So in the scene where they're talking about the "little differences", Vincent shows Jules something he got overseas: A glass encased jewel that's dripping. Weird.

If you could show/send me your deck that would be really helpful.
Or maybe I can just add the readings to my existing RTK1 deck - slowly. So whenever I review my Heisig kanji then, I can do keyword + reading.
That will be a lot of work though.

shakkun Wrote:About a week, like I said... maybe half that time spent on the first 500 trying to figure out what I was doing/what worked. And writing things down. Cannot stress enough how much that is not worth it.

How were you able to manage all that in only one week?
And you did not write down anything? Not at all? Then how did you do it then? I can't imagine how you would go about it WITHOUT writing anything down. Please fill me in



Uh, that was long.
But see I'm really eager to get the readings down using this kind of method - it just stresses me that I'm obviously too stupid to find a good way that'll work for me. Once I've figured that out I finally can have fun with kanji again

Tobberoth Wrote:

alyks Wrote:In a sentence that has a new compound word (and probably a new verb and new grammar point) I'll have one less thing to have to memorize. It really adds up fast.

But that was my whole point, you have to remember the compound readings anyway. Most kanji have several on'yomi, you still have to memorize which one is used in any given compound so how does that make it any different? All it takes is to learn one compound with the kanji to learn just as much as this movie method, there is no difference in time invested in memorizing at all.

I'm finding that even if I don't know which reading a compound uses, it's still ridiculously easy to remember after seeing it once.

Ok? You should listen to me here. I'm the one who's done it, after all. Trust me when I tell you how much easier it is knowing the compound readings already. A kunyomi reading will take me a couple reps before I get it down. A compound I will hardly think about and be perfectly fine with. Are you really trying to tell me it's just as hard to remember which reading to use as it is to remember a new reading?

Chochajin, I never wrote anything down. If you want my deck send me your email.

alyks Wrote:Chochajin, I never wrote anything down. If you want my deck send me your email.

Done. Thanks again.


While doing my RTK1 reviews in Anki right now I just thought of another way of doing this method. Here we go:
How about if I just use my existing Heisig RTK1 Anki deck for this. I'll open it, go to edit and then work my way through #1 till #2042. I check the most common on-yomi of the kanji (if it has any) and then just modify my Heisig story a little bit to put it in a location or 2,3 (if there are more than one on-yomi).
For example 併 (put together - I'm sorry if that's not the English keyword - I'm using German keywords, so that's just a rough translation to demonstrate what I mean).
My Heisig story goes like this "My person xy likes to put puzzles together."
So what I'll do then is check the on-yomi of this kanji. It is "HEI".
I go to my kanjitown excel file, check if it's in the "hei" group. It is. I color the kanji in the excel file, so I'll know I already did that one.
Then I go back to my Anki deck and modify my story. Let's say "HEI" is a rock concert (location). So the story will be modified like this: My person xy likes to put puzzles together on rock concerts.
If that kanji had 2 readings, then I'd just add another location (then it's not only rock concerts, but also ... a park or whatever - this way I don't have to go back to a kanji and modify the story again when I run into another on-yomi, but I'll do all the on-yomis for a kanji at the same time.)
Now all I have to do is to remember a slightly modified Heisig story I'm already familiar with and then remember the location + reading.

It's pretty much the same as before - just that I'm doing things the other way round, because I think it might be easier.


I hope you get what I mean.

Do you see any problems with doing it this way?

Thanks a lot in advance.

Sounds fine. Personally I found it a lot more convenient to do the readings as groups. But whatever works.

alyks Wrote:Sounds fine. Personally I found it a lot more convenient to do the readings as groups. But whatever works.

Thanks for your deck. It looks pretty much like what I wanna do - just that I'm not going from keyword to kanji + reading this time, but from kanji to keyword + reading.
For the RTK1 kanji I did it both ways so far: keyword > kanji and kanji > keyword, but for the readings I think going from kanji > keyword/reading should be enough.


Basically I think doing on-yomi group for on-yomi group might be nicer, but it will be more work if I want to modify my anki deck - as I'd always have to look for the right kanji in my deck to modify the stories. If I just use the deck's order (which is of course Heisig's order) it might be faster. That's the only reason for my decision

I'll try it out for some days and then report back.
If anybody here already tried this out and ran into problems, please let me know!

chochajin Wrote:So what I did until know was using anime characters for each reading and then just created a story with the keywords of the kanji that have that reading.
[...]
Let's say "give" is the keyword. When I see the kanji and remember the keyword is "give" - then I have to think back ... okay, where was a story where somebody "gave" something .... Uhm .... that might be pretty tough.

Okay, you're making this WAY too hard for yourself! What you're describing sounds like kanji chaining, and I'm convinced it's only useful for small groups of 2-4 kanji.

chochajin Wrote:How about if I just use my existing Heisig RTK1 Anki deck for this. I'll open it, go to edit and then work my way through #1 till #2042. I check the most common on-yomi of the kanji (if it has any) and then just modify my Heisig story a little bit to put it in a location or 2,3 (if there are more than one on-yomi).

This sounds more like what I did. Here are some examples to illustrate:

縁 (affinity): To me, this kanji is Spiderman being followed around by Ms. Piggy (from the Muppets) with a broom. It doesn't matter how this image is tied to the keyword affinity anymore, I've already done that part of RtK. What's left is to associate that image with the park near my house (エン). So I simply close my eyes and focus on that image in a specific place in the park, maybe on the bike path. I take a few moments to do that and move on. I DON'T write anything down, because I don't need to- I'm just using the same image I already made in RtK1.

沿 (run alongside): Somewhere else in the same park, I picture a someone in a dentist chair with water spilling out their mouth (because that was my original image).

査 (investigate): A tree growing out of a bookshelf. Again, it doesn't matter why this means investigate to me. I just take a moment to picture it on the beach near my grandpa's house, washed up in the sand.

That's why you can do it quickly, you don't need to invent anything new. Just take whatever story/image/whatever you used in RtK1 and focus on placing it in a new location.

shakkun Wrote:That's why you can do it quickly, you don't need to invent anything new. Just take whatever story/image/whatever you used in RtK1 and focus on placing it in a new location.

Yes, I really seem to be on the right way now. Finally. What you did sounds pretty much exactly like what I want to do ... just that I write my thoughts down (by modifying my story like described above). I just don't trust my brain that much. And I think it won't cost me that much extra time to type my toughts into my deck.

Just wondering in which order you did it. Grouped by on-yomi or did you use RTK1's order?

I did it grouped by onyomi. I found that easier because it gave me time to really focus on the atmosphere of a particular location. It was a bit of a pain to generate the new cards in Anki though. I basically searched for the reading (eg　チョウ) in the edit cards window and sorted the results by Field: 音読み, and added the missing cards.

Actually, even though you say you don't trust your memory I don't think testing the readings is necessary. My retention has been like 98%, because readings are such simple units of information to remember. I was planning to delete those cards until resolve released Anki for the itouch which makes reviewing so painless. Although I guess if you're already testing kanji>keyword you might as well.

I realized that I can easily search through my anki deck, by just typing in the Heisig number (which is displayed in the excel file I have with on-yomi groups). So doing in in on-yomi order is no problem at all.

So what I do right now:
1. On-yomi group. Look at the first kanji in that group, look at the Heisig number.
2. Go to my anki deck, type in the Heisig number. The right kanji will appear.
3. I've modified my deck, so that I have an extra field for readings now.
4. I type in the reading of the kanji in the reading field and then modify my story within a few seconds.
5. I've modified my whole deck so that it will ask me both ways like this:
Q: keyword
A: Kanji (one field), Reading (second field)
and
Q: kanji
A: keyword (one field), reading (second field)

I hope it'll work out this way.

Up till now it feels way easier and faster than what I did before.
Yay!

alyks Wrote:

Tobberoth Wrote:

alyks Wrote:In a sentence that has a new compound word (and probably a new verb and new grammar point) I'll have one less thing to have to memorize. It really adds up fast.

But that was my whole point, you have to remember the compound readings anyway. Most kanji have several on'yomi, you still have to memorize which one is used in any given compound so how does that make it any different? All it takes is to learn one compound with the kanji to learn just as much as this movie method, there is no difference in time invested in memorizing at all.

I'm finding that even if I don't know which reading a compound uses, it's still ridiculously easy to remember after seeing it once.

Ok? You should listen to me here. I'm the one who's done it, after all. Trust me when I tell you how much easier it is knowing the compound readings already. A kunyomi reading will take me a couple reps before I get it down. A compound I will hardly think about and be perfectly fine with. Are you really trying to tell me it's just as hard to remember which reading to use as it is to remember a new reading?

Chochajin, I never wrote anything down. If you want my deck send me your email.

But it's only a new reading the first time you see the compound. Every other time that kanji is in a compound, you already know the reading from that first compound you learned. You really think it's time effective to spend ages learning every reading of every kanji just so you can spend 10 seconds less learning a new word (but only if the kanji in that compound are new to you)?

And like I said, I learned 砂漠 and I still know the readings without any issue, I didn't spend any extra time and I had never seen those kanji used in a word before. Next time I see either kanji in a compound, I will guess they are read the same way. If they aren't, who cares? I learned that compound, so now I know those new readings as well.

[Katsuo]

Tobberoth Wrote:Most kanji have several on'yomi

Most kanji have one common on'yomi. Here's some data from the Joyo kanji list (total 1,945 kanji):

0 on'yomi 21 kanji
1 on'yomi 1,663 kanji
2 on'yomi 245 kanji
3 on'yomi 16 kanji
5 on'yomi 1 kanji

So 86% of the Joyo kanji that have on'yomi have only one.

As well as that, many of the multiple readings are quite rare. The "Kanji in Context" reference book marks uncommon readings, and by referring to their list the 261 (245 + 16 + 1) kanji above with multiple readings can be reduced to 145.

This suggests that someone who has learned the on'yomi should be able to predict them correctly in unknown words about 90% of the time.

Katsuo Wrote:

Tobberoth Wrote:Most kanji have several on'yomi

Most kanji have one common on'yomi. Here's some data from the Joyo kanji list (total 1,945 kanji):

0 on'yomi 21 kanji
1 on'yomi 1,663 kanji
2 on'yomi 245 kanji
3 on'yomi 16 kanji
5 on'yomi 1 kanji

So 86% of the Joyo kanji that have on'yomi have only one.

As well as that, many of the multiple readings are quite rare. The "Kanji in Context" reference book marks uncommon readings, and by referring to their list the 261 (245 + 16 + 1) kanji above with multiple readings can be reduced to 145.

This suggests that someone who has learned the on'yomi should be able to predict them correctly in unknown words about 90% of the time.

It may or may not be true that most users of japanese will only need to know 1 on'yomi reading for 86% of the jouyou kanji, it really doesn't change my point much.

The difference between learning in context and learning using the movie method is that learning with the movie method teaches you on'yomi, in context teaches you kun'yomi as well. It also teaches you words at the same time as you learn the readings, unlike the movie method where you learn JUST the readings. It teaches you not only how to read the kanji, it teaches you how to use them.

So I ask you, what is easiest? Find one jukugo/word for every kanji reading and learn it by SRS/sentence method, or look up 200 anime series and put individual kanji in scenes in them? What is fastest, is the movie method so much faster that it beats learning more than 2000 new words at the same time? How many of the readings you learn with the movie method will you not use for months or even years because there are no useful words which contain them?

Remember, I'm just arguing which technique is more effective to use, not whether or not the movie method is really effective for learning readings.

Dude, I learned the onyomi at the same time as I did the kanji. No extra time or effort required. 50 days and I finished memorizing both the kanji and the readings. Now I have a massive advantage to learning compounds and it's really easier. For no extra effort, I have a much easier time with compounds and can blast through massive amounts of sentences.

After doing the movie method and spending no extra time doing it, I can go through those 200 anime sentences in probably a lot less time than you.

Addendum:
I never really intended people to use this after RTK, but it can't be helped. I think if a person want's to go through it post-RTK, it's because they want to make things easier, not worry about speed. Some people would think it's easier to do the method and would hate having so many unknowns in a sentence, others would think it's better to do sentences right at first.

I also think that if a person completes the Movie Method after RTK they will be able to go through sentences faster, which will negate the time spend on it.

alyks Wrote:Dude, I learned the onyomi at the same time as I did the kanji. No extra time or effort required. 50 days and I finished memorizing both the kanji and the readings. Now I have a massive advantage to learning compounds and it's really easier. For no extra effort, I have a much easier time with compounds and can blast through massive amounts of sentences.

After doing the movie method and spending no extra time doing it, I can go through those 200 anime sentences in probably a lot less time than you.

Addendum:
I never really intended people to use this after RTK, but it can't be helped. I think if a person want's to go through it post-RTK, it's because they want to make things easier, not worry about speed. Some people would think it's easier to do the method and would hate having so many unknowns in a sentence, others would think it's better to do sentences right at first.

I also think that if a person completes the Movie Method after RTK they will be able to go through sentences faster, which will negate the time spend on it.

What do you mean would go faster? I really don't see it. Do you mean that when you see a new sentence which for some reason didn't have furigana with it (like Kanji odessey etc has), you do not look up compounds, you just guess the readings based on your on'yomi knowledge? Because I mean, whenever you look up a jukugo for meaning, you get the reading as well, so it's not like you to spend any form of extra time on reading.

Personally I don't see what problem one can have with a jukugo just because they don't know the on'yomi, but maybe that's because I've been studying Japanese for a long time.

Yes, Tobberoth. You are studying it for a long time already.
This method was not made to remember ALL the readings. Just 1 reading.

It does help a lot just to "fit" readings instead of "conjuring" them from nowhere.

Remember they are also not forcing no one to follow their methods.
We are doing it because we think it is better than remembering the readings without any mnemonic aid at all.

I am doing it _after_ doing sentences alone for some time, like Katz proposes. And I think it is much better.

Jarvik7 Wrote:Do second language learners of english spend time learning all possible morphemes in isolation because it'll help them when they encounter those morphemes in words?

No, but first language learners do it.

With japanese, we are learning to read again. It is not only a new language. It is a new writing system.

mentat_kgs Wrote:Yes, Tobberoth. You are studying it for a long time already.
This method was not made to remember ALL the readings. Just 1 reading.

It does help a lot just to "fit" readings instead of "conjuring" them from nowhere.

Remember they are also not forcing no one to follow their methods.
We are doing it because we think it is better than remembering the readings without any mnemonic aid at all.

I am doing it _after_ doing sentences alone for some time, like Katz proposes. And I think it is much better.

I know they aren't forcing anyone. I'm not forcing or even asking anyone to quit either. I'm just discussing or arguing whether or not it's a useful technique, just like I bet tons of people argued with Heisig whether or not his technique was good.

Learning readings from context isn't learning without mnemonics. You have a word, hell a whole sentence if you want, as a mnemonic. You have semantics. Not some location in a movie or some atmosphere or a character which really has nothing to do with neither the reading nor the kanji.

IMO, saying "Oh, 公? It's read "kou" as in 公開." makes more sense than saying "Oh, 公? It's read "kou" because when I see it, I think of that hilarious scene in Azumanga Daiou where Chiyo was speaking okinawan." I personally think 公開 is easier to remember, but that's probably where we all differ since you guys prefer the movie method.

I'm pretty confident that most Japanese people know readings from context (in contrast to mentat_kgs idea that first learners "do it", but I might have missunderstood what he meant), even if they may or may not have studied readings directly.