Jt0104, I'm looking at it as a means to an ends. Being able to correctly read Japanese means I can get input from books, subtitles, manga, etc.

It's been my assertion (amoung others, I'm making it up as I go along and throwing out what does not seem to work) after RTK we need about 1000 sentences that built a base of grammar AND kanji usage. The problem has been that such books don't seem to exist. If it's basic grammar, the kanji seems to be sparse. If it's kanji usage, the grammar is a mesh. KO2001 seems to be a good choice if combined with some grammar references such as Kodansha books.

After that first 1000 or 2000 sentences, from there it's anywhere the road takes you. Maybe one guy wants to complete kanji usage and inputs the rest of Kanji in Context. Another may start mining dramanote.seesaa.net scripts. Yet another may want all 3000 sentences from KO2001.

In addition to that, now you can take on various movies, tv, manga, books, magazines, newspapers and get your passive inputs while at the same time digesting real culture. Maybe even talk in Japanese?

Back to where I started, most of this is all of us trying to figure out the various ways we can take on the next step. We've been instilled with a powerful tool (Heisig RTK), now we want to use that tool in the most efficient means possible.

Man, did I go off on a Tangent. Anyway, yes, the sentences probably make us sound like a tool if we used them for real. Great point to not depend solely on them.

Quote:I'm just a little surprised there isn't already a collection of sorts out there, so instead, people are doing a lot of the same work over and over.

That's also because, once you're out of the really basic things, people go in different directions... They read different things, watch very different shows... So I guess that really contributes to that fact.

... The basics can be found in lists, that is to say, I know of a few basic lists. Now these obviously don't cover everything... But that's kinda the whole method as well... do your own things... right?

Savara Wrote:

Quote:I'm just a little surprised there isn't already a collection of sorts out there, so instead, people are doing a lot of the same work over and over.

That's also because, once you're out of the really basic things, people go in different directions... They read different things, watch very different shows... So I guess that really contributes to that fact.

... The basics can be found in lists, that is to say, I know of a few basic lists. Now these obviously don't cover everything... But that's kinda the whole method as well... do your own things... right?

At first I was going to reply to this by saying that the method isn't necessarily about doing your own things. The main message is to do (<edit: things in) Japanese as often as possible, whilst having fun doing so. However, whilst typing the aforementioned it struck me that the whole concept of 'fun' differs a great deal from person to person, which is probably why it is hard to turn the All Japanese All The Time method into a group effort and why people indeed go into different directions.

(On a side-note: As far as the rest of the AJATT website is concerned, it seems a lot of people haven't read through it in its entirety or have failed to remember various parts and therefore come up with a large number of different interpretations. It is not a bad thing, but does show a certain carelessness or indifference.)

@phauna
I only go both ways with cards I'm trying to learn kanji with. I've gotten pickier about what I choose to go both ways with, because yeah, otherwise it takes a lot of extra time. Before I started doing it, I found that a lot of kanji I could read, I couldn't necessarily write from just kana, so I tend to focus on that. (That's the whole point of KO, IMO.) Literacy isn't just reading, it's writing, too. With the UBJG sentences, I don't usually go both ways unless I'm trying to relearn something or there's new vocab in the sentence. (Rare, to be honest.)

I add cards in bursts, then make sure I process the material well, then I'll add another chunk. So with UBJG, I can add anywhere from 50-100 cards at a time, because I'm mostly using that to review stuff I learned 5 years ago, and I'm just rusty at. A lot of those cards wind up in then 2-5 month review pile pretty quickly. The KO cards take longer, because of their habit of sticking in 4-5 vocab words a sentence, leading to a high fail rate at first. On the upside, you do get more vocab that way... but on the downside, it can take forever to get through them.

Also, Anki keeps piling on review cards while you're trying to learn new cards, which makes things take much longer than they need to. I don't know if there's some sort of "forced learning" mode, where it suspends all reviews of other cards while you're learning new material, but that would be awesome.

@jt
It can't be helped much at first. True, textbook sentences are going to be a bit odd, but they serve a purpose-- the illustration of linguistic rules. The idea isn't to memorize the sentences, it's to expose yourself to linguistic patterns. The only thing I'm actively trying to memorize are kanji and kanji compounds. The rest I'm just trying to get comfortable seeing/reading.

An SRS could be pretty handy for memorizing useful phrases, though. I've stuck some in from a phrasebook or two, just to have my cliches handy. I emphasize those by going both ways, even when there's no kanji.

You also have to figure out how fast/how far to go. Going from the wading pool to the deep end when you can't swim yet is inefficient. But staying in the wading pool won't help either. *shrug* I don't have any answers yet, either. I'm just trying to figure it out as I go along.

Balancing learning kanji compounds with grammar is tricky. (Just to state the obvious there. )

@roderik
The basis of AJATT is simple. Immersion. The tricky bit is figuring out what to immerse yourself in. People get hung up on sentences because it's the trickiest bit of that. You don't want to learn a lot of useless crap, or expose yourself to language that could make you sound like a tool. (Which is surprisingly easy to do in Japan.) The sentences are useful, but they're not the whole method. The whole method is immersion.

I read the blog, and I see what he did. Doing exactly what he did isn't feasible for everyone, but there are a lot of good ideas there. For example-- I stopped buying English-translated manga, and now I only buy manga in Japanese. (Cheaper, anyway.) Same for light novels. I listen to as much Japanese as I can, and watch what I can. If I lived in Japan, it would be much easier to immerse myself, though.

rich_f Wrote:@roderik
The basis of AJATT is simple. Immersion <...> The whole method is immersion.

It doesn't add a thing but: that was the thought behind my message .

Heh, yeah, but people love to argue the details.

@rich_f

There's no arguing that he stresses doing things in Japanese all the time and the like.

My only thing is that, for me, at least, I know I will take sentences selectively from texts. They're going to need to interest me, or I won't remember them enough to use them as a basis for grammar... memorization? I guess.

@rich

I'm still using mnemosyne exactly because i kept getting the wrong ones thrown at me while i was learning new cards. Mnemosyne alots some for the day, which i do in the morning, and then i add cards whenever i have time.

To do the same thing in Anki you can just alter the fail button to a longer interval. It's 10 minutes by default, but i used to make it an hour, as doing my reviews from previous days took about an hour. Of course cards still trickle in but it's not as annoying as getting the same difficult card every ten minutes.

@Ryuu
The idea isn't so much to focus on memorizing the sentences, it's just another way to build up a critical mass of sentences your brain recognizes as "correct" in the same way babies learn. I'm only striving to memorize certain things, like kanji readings, and set phrases. Those get more attention, so you go both ways with them. The rest can be safely relegated to read-only, I think.

I don't plan on doing it this way forever. Just for a few thousand sentences or so.

Ahh, that makes sense as another way to do it.

However, it does return to that whole "baby method" kind of thing. Unfortunately, I think that only works to an extent. The thing that makes learning a language to "easy" for a baby is because they don't question the language. They just accept that. The method can work for you too, if you don't question the language. However, as an adult with an adult brain, that's very nearly impossible.

Yet, I will give that idea a try, combining with the method of just taking sentences you want to fully know because they interest you.

I personally think "the baby method" is a bit of a fallacy. I think babies are more receptive to language as they have to to survive (evolutionarily speaking). There is no genetic need for our brains to learn another language as adults so I don't think the same method would work.

Babies are incredibly more receptive to learning language, and absolutely anything else. They have to be. They have literally no experience. They don't even know what language IS. Yet despite being steeped in a language 24/7, and having brains set up specifically for this learning, it takes them well over a year to squeek out a word, and much longer before they're reasonably competent (sometimes I still question the competency of the 5 year olds I know!). Because with no basis, there's simply THAT MUCH to learn.

Fortunately, while some of your learning ability has faded (your brain starts pruning away sound receptors not used in your native language by year one), you know EXACTLY what a language is. Not only that, you're fluent in one. You know the SPECIFICS of language. You understand nouns, verbs, adjectives, adverbs, clauses (even if you don't know you know), idioms, prefixes, suffixes, conjugation, pitch, tone, stress, and every other little component of language.

You have a HUGE advantage, a filter through which to understand what's coming at you. Baby methods can work for you, and the fact that you already know one language isn't some disadvantage that gets in the way (unless you make it one) - it's a HUGE advantage. It's the whole basis that allows you to learn another language so very efficiently.

Babies take a long time before they repeat things they understand on hearing. You can give a child verbal instructions well before they can speak a single word, and you can discuss moral ramifications of actions years before they'll use 'moral' or 'ramification' (or 'action'!) in a sentence. They input, input, input, until they finally understand enough to use the same phrases their parents, teachers, and peers use. And they do (nearly) ONLY use those phrases - they're the only ones they know make any kind of sense. When they try to do otherwise, usually at the very start of their learning, as you might have noticed, it's absolute gibberish.

I still remember my niece, who would never say more than two words in a string at this time (always the listener), suddenly asking, "When is my mommy going to come pick me up?"
Learn from their example. Input, input, input, input. Finally, output flawlessly.

I think hitting "Yes" even if you fail is not the way to go. The Leitner system would delay your review of that element, but you don't want to do that if it's not learned yet. The idea being (of course), you don't want to see that item for review until you're 90% sure you're about to forget it. It's only reviewing on the edge of forgetting that is efficient use of your time and committing things to long-term memory. But if you don't have it yet, then you want to see it sooner than a "yes" would imply. You'd have to "break the system" to get it to come back in a better time for review.

If you try to learn it after it was failed, and you hit no, it will come back soon as you say. That's fine. Just review it again. If you remember it, pass it.

This is better than just reviewing it and immediately passing it because, if nothing else, you proved you can remember it and it is, at least, in your short term memory. Then, if it comes up in a day or two and you forget it, just repeat the process.

That's what is meant by "letting the SRS do it's thing."

Also, sorry for the first sentence and anything other thing that sounds stupid. Working on over 40 hours of no sleep now. Damn final papers.

If you're using Anki, you can set different times for "Completely Forgot" and "Made a Mistake." I set the time for "Made a Mistake" to 20 minutes, minimum, so you don't see a new fact you're struggling to learn too soon, but soon enough so you challenge your memory a bit. Then if I get it right, I pass it as "Difficult" until I feel it comes easier. I don't think I ever use "Completely Forgot," to be honest. I found 10 minutes to be too soon, but 20 minutes is about right. It depends on how well my brain is working that day.

Anyway, that lets you skip the whole studying portion and just let the SRS do its thing from the get-go. It makes more sense than passing something when you really *don't* know it.

I'd agree that you should just fail a card you've failed, because that's largely how the SRS is designed to work. There's no harm in reviewing a card you 'kindof' know twice quickly, while there is a degree of harm in not reviewing the same card until you've forgotten it. I think.

By the way, I'm not sure about the idea of reviewing a sentence from kana to kanji. I understand what you're going for, I think, making sure you can remember the kanji from a prompt (similar to this site) so that you're sure you can remember how to write it, while still memorizing sentences. But, the thing is, you'll never actually encounter a 'kana sentence' in real Japanese reading, since they're a $!@# to read even for fluent Japanese adults, and it strikes me as a bit of a waste of time because of that. Mightn't it be better to fill your head with Japanese sentences as they actually appear? If you continue to do some kind of RTK review, you shouldn't ever have to worry about forgetting the writing of kanji, and the associations between the character and it's readings should take care of itself as you read Japanese sentences and learn their vocabulary.

Due to the lack of distinct word-shape in the Japanese kana system, which is the reason for kanji's continued necessity in the first place, it's basically like trying to read e n g l i s h l i k e t h i s o n l y e v e n m o r e a n n o y i n g. It's not something I'd gladly subject myself to, and I just question the benefits. 2cents.

QuackingShoe Wrote:But, the thing is, you'll never actually encounter a 'kana sentence' in real Japanese reading

I'm just beginning so I haven't gotten to this point yet (still doing Heisig), but I thought the idea was to simulate listening to speech. It would be better to go from audio clip to kanji, but it's easier to simulate this by going kana to kanji.

Things I'm finding while using kana->kanji:

-It is MUCH better for memorizing how to write kanji. Even having done Heisig.
-I find it's good to (if using Anki) also have cards that use kanji on the question side. It's easy practice and those cards take 5-10 seconds to read at most.
-It's still easy to read sentences in kanji.
-Kanji provide huge hints to the meaning of a word (beyond the reading) and not having the kanji readily there makes me have to actively recall the meaning of the word through context (as if listening).
-Sometimes it is difficult to read whole strings of kana, but it's quite easy to get used to.

-It's tedious somewhat.
-Words that have been used over and over in cards are crystal clear in my memory now.
-I sometimes skip writing the whole sentence, and aim for words that are challenging to me.
-I sometimes enter the kanji for a word if I know it really well AND it has been entered already a few times.
-Words I thought I knew really well already were sometimes (in the beginning) challenging to go from kana->kanji, but quickly became a piece of cake.

I have almost 200 kana-kanji sentences (from Kanji In Context) so I don't have a TON, but a decent amount. My conclusion is that using kana->kanji is REALLY helpful for mastering the writing of words, but it doesn't have to be used over and over again for words that have become easy.

I think there is a frequency or number at which writing from kana->kanji has diminishing returns, even in an SRS environment. This number probably differs from person to person (their exposure to kanji) and the difficulty of the shape of the kanji itself. After some time, it is probably possible to gauge yourself against words and just start entering them (if repeated in new cards) as their kanji form. Likewise, if you are reviewing there might be some words you do not need to write out.

Thanks for the explanations, all. It's still not something I'd choose to do at the moment, ideally I'd prefer to relegate listening-like comprehension to actual listening, but I'll have to keep it in mind if I ever find myself having trouble with written reproduction in the future.

Why not just use Anki, make a card with kanji and kana, and have Anki set to test you both ways? Kills two birds with one stone, ね?

Nestor, to be fair, Anki has a nice little graphing tool that visually shows how many you have due each day from now, how many you have to review if you don't review x number of days, what the spacing you have (similar to RevTK columns). The only thing missing I think is a "how many you reviewed each day" graph.

For Kana to Kanji, most of it has been covered. I did not even think of doing it until Khatzumoto posited the idea on his blog. It made complete sense the instant I heard it I wondered why I didn't do it from the beginning. Granted, I used Anki, so it was a sinch to have both kana to kanji and kanji to kana.

For kana to kanji, it's all about writing it out correctly. Preferably, I'd read the kana sentence and without refering back to it write it out. Not knowing the kanji to the word is a "completely forgot" so I'll get it again in 10 minutes. In that case, I might avoid writing it down (I know ahead of time if I know all the kanji).

For kanji to kana, it's all about reading it correctly. There, I'd read it then after checking if it's correct, go on to try to say it without looking at the sentence. A wrong pronunciation (well, maybe a little leniant on the voicing now and again) means a "completely forgot" for 10 minute review.

Writing out the sentences is tedious. I'm tempted at times to just do the key kanji, but something tells me I need this writing as it will drill it more and more into my head. However, in the short term I notice the strict "completely forgot" policy helps with the retention. It just means very long review sessions.

QuackingShoe Wrote:Due to the lack of distinct word-shape in the Japanese kana system, which is the reason for kanji's continued necessity in the first place, it's basically like trying to read e n g l i s h l i k e t h i s o n l y e v e n m o r e a n n o y i n g. It's not something I'd gladly subject myself to, and I just question the benefits. 2cents.

I'd like to point out that as a native speaker of English, it's incredibly easy to read that text. I imagine that to a native speaker of Japanese, all-kana sentences, while annoying, aren't that hard to read. Unless they're completely out of context or something.

Also, I question the actual necessity of Kanji in Japanese. Certainly, they can make things easier, but they couldn't possibly be vital to comprehending the language. If they were, the spoken language wouldn't function. There's no Kanji to be found there.

Ok, I'm partially convinced, and will have to experiment with it out once I finish up RTK and get back into heavy sentence drilling. Thanks Nest0r for the link, and thanks everyone for covering the topic pretty thoroughly.

Mcjon01: It's not necessarily difficult to read my example(they still form a sort of shape), but it's less convenient, especially over length, and kana is quite worse. Forcing yourself to sound out any word you encounter slows down what should be a rapid form of communication and.. ah, blah blah blah, http://www.guidetojapanese.org/kanji.html#part4 says a lot of it. Verbal communications give more context and cues and pitch distinctions and etc that make it more easily understandable. There's articles about this sort of thing (about japanese specifically and just the word shape phenomena) scattered all over that you can find...
But, no, they're certainly not VITAL to the language, I wouldn't (and didn't) say that. But there's a reason they exist.

I've been doing kana->kanji on top of kanji->kana for about 3-4 weeks now, and I've found similar results as the others have.

-Writing out is incredibly helpful.
-Writing out can be incredibly tedious.
-Saying the kana out loud helps, and helps you to associate the sound of a word with its kanji.
-I remember kanji a whole lot better now.
-Before I started, there were a number of kanji I could read, but not write (just couldn't put it together.) Now if I can read it, I can probably write it, because I've gone both ways with it.
-It's not an "instant" thing. It takes a few tries to start making good connections.
-Sometimes I write full sentences out, sometimes I don't. Depends on my mood and how busy I am. (And why I'm studying a particular sentence.)

There's probably a weed-out point for certain cards, I'll agree. Kanji like 時、朝、食、今日、昨日、明日、etc. are pretty much glued to my brain for good.

I would recommend less theorizing and more trying things out. In the end, you can spend a month thinking about it, and wind up wasting a month you could have spent trying something different out. So far, I like writing out, even if it is a pain in the butt. The ends justify the means. I suppose at some point I won't need to do it anymore, but I don't see that point coming anytime soon.

Also, I find it very useful for learning any new vocab-- not just kanji, but also verb endings and such. Writing things out makes me process what I'm writing, so I'm more likely to remember it.

killerducky Wrote:... but I thought the idea was to simulate listening to speech.

You forgot that there are words and grammar that are for writing only and not used in speech. Also speech has a rhythm and tone to help separate the homophones.

Mcjon01 Wrote:Certainly, they can make things easier, but they couldn't possibly be vital to comprehending the language.

So what does this sentence mean?
Mcjon01がいっちゃった！

I guess someone forgot about homophones ...

(I actually think the kana->kanji is good a idea though.)