This is a very good idea provided the answers are written as well as said. Although, I would advise against learning kanji and readings together. While you're essentially picking up an equal amount of information, tangible progress will seem much slower.

I'm essentially doing exactly this on two android applications (not stealing thunder here, just sharing some good resources)...

https://play.google.com/store/apps/deta … kanabyhand

Includes both kana scripts and 250 readings of basic words. The moment I found myself coming up with mnemonics for the vocab readings (e.g. "I see you've caught an illness, but you'll びょうき. I promise."), I instantly downloaded the kanji version (with 10,000 words, but it's not free ); before recently, I knew no other interactive learning environment that provided me with the interest to think that hard about remembering words.

https://play.google.com/store/apps/deta … anjibyhand
^ the kanji app

Testing knowledge by writing is very engaging.

Are you learning Kanji readings without any words that reading is used in?

amtrack wrote:

Basically, the card has the kanji, its readings, and the meaning.

What do you mean by "the meaning"? The English keyword Heisig assigned to it? It's important to note that that's not really the meaning of the Kanji (even saying that Kanji have a meaning is a stretch, for the most part).

Japanese words have meanings. But words are usually written with several characters (Kanji and/or Kana).

I would advise against learning anything about Kanji, except how they are written and what their keyword is, without example words. There's nothing wrong with learning readings systematically (though I prefer just learning them as I learn random vocab), but even if you're doing it systematically, you should use common words they are a part of. Don't just memorize thousands of sounds that don't mean anything in Japanese.

amtrack wrote:

Obviously this is for people like me who are early in their kanji study.

I don't know about that. I've done RTK (lite) and struggle with mock JLPT tests when the answers are in kana for a word I know the kanji by heart. This sort of thing could help

RawToast wrote:

I don't know about that. I've done RTK (lite) and struggle with mock JLPT tests when the answers are in kana for a word I know the kanji by heart. This sort of thing could help

Yeah I realized I made a pretty grave assumption lol, so that has been edited.


@uisukii:  I actually prefer to make my own decks.  Making the cards yourselves reinforces learning.  It is more time-consuming, but that simply means I study less kanji per day.  My goal is simply to be done with the joyou kanji within a year.  Even a pace of 6 kanji per day will meet that goal.  All other energy is devoted towards my textbook/vocab/listening/speaking studies.

yudantaiteki wrote:

Just don't be too strict about this.  Very few kanji have a single English word that can express all (or even half) of the uses of that kanji.

I just use whats there in the kanji dictionary.  Everything from RTK to any kanji dictionary you can find will have a "general" meaning listed by the readings.  Obviously, it wont cover everything, but its not supposed to.  Its just there to give you a point of reference.  Basically, there's no point in not learning it.  The more things you can attach to a kanji, the stronger your memory of it becomes.

To put it simply, knowing more is always better than knowing less.  There's a reason why the Japanese education system is rigorous.  You can't have mastery over a language without going through a similar process.  The same goes for English, or anyone's primary language.

People make fun of the education system, but it is the reason your grasp of your primary language is as strong as it is.

amtrack wrote:

uisukii wrote:

This aside, how would this approach fit into the greater study focus? Is this something you are utilizing for the yomi, specifically? That is to say, is this something alongside some form of vocabulary?

The story is originally I was just going to learn kanji/meaning, kinda like the Heisig way.  But I really didn't want to go through 2k kanji without knowing a thing about them in Japanese.  So I switched to the above method.

To be honest, I didn't really understand how your kanji deck is made and what kind of system you devised to be able to say that learning kanji readings in isolation is easy... I would welcome more details. Could you upload the deck maybe?
I wouldn't consider learning isolated readings to be "knowing a thing about kanjis in Japanese", unfortunately... The only end-goal in Japanese is words and sentences, and you really, really don't care about readings as long as you don't know at least one word using it... There's also the problem of the many kanjis with more than one 音読み or 訓読み, you'd still have to learn which one to use for each word using them... Learning readings when learning vocab is supposed to be faster, more contextual and more natural.

amtrack wrote:

Stansfield123 wrote:

Are you learning Kanji readings without any words that reading is used in?

I put example compounds on the back of my cards.  Not to memorize them, but just to get a feel for how the readings are used.

Stansfield123 wrote:

I would advise against learning anything about Kanji, except how they are written and what their keyword is, without example words.

I'm aware of this, which is why I include compounds on all my cards.  But even if someone didn't, auxiliary vocabulary studies will introduce people to kanji compounds using kanji they have learned.

Be careful, anything written on the back side of an Anki card should be treated as information that won't be learned. Either you're skipping this information too fast to be remembered, either you're spending time on it trying to remember it by "brute memory", which defeats the purpose of SRS. All information on the back side of a card should be solely designed to help you rate your own answer, and plug some holes in your understanding for the young cards.

There is a good reason why "many people" say it's a bad thing to do, and you can't just say "I don't care because I like my method". Especially when those many people include lots of advanced learners. Have you found someone who did a method similar to yours and got to an advanced level? If so, reproducing his method would be justified indeed. But to me, what you're doing is a time sink that won't bring much to your Japanese in the long term.

amtrack wrote:

To put it simply, knowing more is always better than knowing less.

Of course, but this is not the point here. It's not about quantity of learned information at a time on a given kanji, but about optimal sorting of this new information to maximize effectiveness (quantity/time), comfort, fun. In terms of efficiency, you are learning 2 or 3 times too much information (readings) compared to what you will really need for the 1 or 2 year(s) following RtK completion...
The real question here is not "with how much depth am I learning kanjis day after day, in the beginning", but "how much will I know within 2, 3 years of study". I'm not sure you're dealing with Japanese as a long-term project here. Because essentially what makes you want to learn readings (too) early on, is that you feel frustrated to not know "usable" Japanese in the first 6 months (?) of study.


If you pursue with this method, give us some updates every now and then. It's always interesting to see people experimenting new stuff. But I believe you're not the first one to try this, look for old threads about this and how people managed this. Don't forget to be critical about both other people's methods and your own, always compare and think hard about the process.

By the way how far did you go with the method when you made the thread? How many kanjis did you complete? Quite often people find something new that they have tried for a mere few days, and feel it could be a nice method and talk about it immediately on this forum. Then they don't understand why others are on the reserve. But actually, if you've tried the method for a too short time, you're enthusiastic because it worked well in that short period and because you're motivated by the novelty. But you need at least one or two months to have a better understanding of a new method, when the initial excitation dropped and your method starts being confronted to the mechanisms of long term retention...

Last edited by Warp2243 (2013 May 24, 10:43 am)

Silty wrote:

Warp2243 wrote:

Be careful, anything written on the back side of an Anki card should be treated as information that won't be learned. Either you're skipping this information too fast to be remembered, either you're spending time on it trying to remember it by "brute memory", which defeats the purpose of SRS. All information on the back side of a card should be solely designed to help you rate your own answer, and plug some holes in your understanding for the young cards.

Thanks for the in-depth response!  I know this is covered elsewhere on the forums, but I'm sorry to say that I've read a bunch of topics about how Core is used, and I still don't quite understand what exactly we're supposed to do.  What do you mean by "won't be learned" - isn't the whole purpose to learn it?

I was differentiating the "question side" ("front side") and the "answer side" (what I called "back side") here. The point of SRS is that you spend most of the time on a card analyzing the information on the question side, and then there are a few seconds of active thinking in order to find the answer (e.g. active recall of the readings of the kanjis displayed in the sentences). Those seconds of active thinking are extremely precious, because they are the only sure way to commit something to memory. Then you show the answer side of the card, which is supposed to contain information to check if you were right (so it would have the same sentence with furigana over all kanjis + an English translation to get the idea).

Nothing in they way Anki works forces you to have an active attitude towards the answer side (contrary to the question side since you need to evaluate yourself on it). That's why it's a very bad idea to put information that you do want to remember on the answer side. And de facto, you will spend only a few seconds checking the info on the answer side for the young cards (< 10 or 20 days), and the better you'll know your cards, the faster you'll skip the answer side. After 1.5 years of sentence mining, I read the English translation for less than 5% of the cards and I have an extremely quick look at the furigana, because most likely there's only one syllable I need to check, or even nothing at all, since the question side only had common kanjis whose readings I'm sure I know perfectly. The first steps in Japanese/sentence mining (when everything takes some time), and this behavior I have nowadays when doing SRS, are somewhat the two extremes. There's a spectra of other behaviors in-between, and the more you learn words/sentences, the less time you'll spend on the answer side.

Silty wrote:

I mean, I tried to start Anki with the optimized Core deck yesterday.  I kept repeating the same cards over and over at least 5 times, because I kept failing to remember them each time.  And that was just simple things like "that" and "one".  Sometimes, I just can't remember something if I don't try to remember it with "brute memory" plus a good mnemonic.

So what is supposed to be on the front and the back of the Anki cards?  Did you just use the defaults, or did you put particular things on the front and particular things on the back?  What worked best for you?

Okay, now on to a more concrete discussion about the Core deck. Unfortunately there's not (any more) one "Core deck", originally it's just raw data that has been arranged in different ways. You're not being very specific about the card format you're using, but I suspect you are using cloze deletion (one word is hidden in the sentence of the question side). A few months ago, some friends who finished RtK got the Core deck themselves, so I thought they had the same as mine (passive reading only), but I discovered recently they were doing cloze deletion. I told them to try passive reading instead and they said it was better/easier and felt more natural. Though their production skills may go down a little of course. So it seems the Core deck people find nowadays is not the same as the one I downloaded 1.5 years ago (on this forum).

There's no way you can fail 5 times reading words like これ and ひとつ since it's full hiragana, so I assume you're using cloze deletion... And your case goes to show that you're actually spending time and effort trying to memorize which Japanese word should fit into which card, rather than actually memorizing the words themselves, which is the main drawback of cloze deletion and the one reason I don't understand how it can work... But I don't want to say too much about cloze deletion since I haven't experimented it myself. It obviously is of some interest to improve production ability. I just don't understand how you make it efficient on thousands of words.

Anyway I can only talk about what I've used since the very beginning (sentence reading). I never felt the need to change my card format or method and it steadily brought me to a quite advanced level (though I'm still far from what I would like). I'm using the same pattern for Mandarin Chinese (there's also an awesome Core deck with 5.5k sentences) and it's working great, exactly how it was with Japanese in the beginner/intermediate stages. At the time I just downloaded the data of Core 10k and made my own card format, though it's really classic.
Question side : Japanese sentence with kanjis. I associate a unique Japanese word to each card, so that I avoid having redundant information in my sentence deck (at the very least each card contains this word as new information).
Self-evaluation criteria : I need to be able to read aloud (actually I do this in my head...) the whole sentence, without mistaking any kanji reading. One mistake, even on a dakuten (e.g. こう vs ごう, せい vs ぜい), I fail the card. I've always been super strict on that, since the beginning and up to now. I also need to understand the meaning of each word (at least vaguely on young cards, and with accuracy on mature cards), and the grammar and sentence structure. I wasn't very strict for the first 5k-6k and it turned out ok. Nowadays grammar or precision in vocab understanding (even for new words, thanks to kanjis) is not so much of a problem anymore.
Answer side : dictionary look-up of the "keyword", English translation of the sentence, the audio of the Japanese sentence is automatically played (and this is super important, it impacted hugely my listening comprehension level over the months), and there's a dictionary look-up of all words in the sentence (only for the first 6k sentences).

Here's a screenshot :

(the only reason I failed this one is because I read とうさい instead of とうざい, eheh)

Silty wrote:

Warp2243 wrote:

But actually, if you've tried the method for a too short time, you're enthusiastic because it worked well in that short period and because you're motivated by the novelty.

Maybe I'm completely off the mark here, but I feel like short term is just as important as long term.  Before committing something to long term, I need to have a good handle on it short term.  For example, before making a kanji stick with me long term, I need to actually be able to completely see it in my mind in the short term (which RTK indubitably makes easier).  I don't see what's wrong with having a separate method for short-term comprehension, and then relying on something else to make it stick.

I was referring to the misleading enthusiasm one may experience when discovering a new method, whereas it could be beneficial to thoroughly investigate it on a 1-month period before sharing it publicly. Presenting actual results would be more interesting/convincing for others who haven't try the method. I was more talking about methodology than actual learning, really.

Last edited by Warp2243 (2013 May 24, 12:56 pm)

amtrack wrote:

Stansfield123 wrote:

Are you learning Kanji readings without any words that reading is used in?

I put example compounds on the back of my cards.  Not to memorize them, but just to get a feel for how the readings are used.

That's good. The more words you use the better.

While I still don't actively try to memorize readings, lately, when I run into a difficult word, I have been trying to add a second word with the Kanji who's on reading I don't know, to my vocab list. That does the trick for making me associate the Kanji with that reading, which then makes it easy to remember both words.

Haven't been doing it for kun readings though. They're much easier to remember just from knowing a single word that has them.

Last edited by Stansfield123 (2013 May 24, 6:50 pm)

Warp2243 wrote:

A few months ago, some friends who finished RtK got the Core deck themselves, so I thought they had the same as mine (passive reading only), but I discovered recently they were doing cloze deletion. I told them to try passive reading instead and they said it was better/easier and felt more natural. Though their production skills may go down a little of course.[...]

Passive is the best way, usually.

If you're going at a relatively slow pace with core, production cards are favorable.
For me, core is a secondary deck and I only learn 10 a day. My core format is as follows:
Front
Sentence Cloze.
Kana.
Hidden definition.
Back
Word
Definition

Unfortunately, I'm at a point where most cards don't have sentences, but I make do without them [most of the time].

Anyway I can only talk about what I've used since the very beginning (sentence reading). I never felt the need to change my card format or method and it steadily brought me to a quite advanced level (though I'm still far from what I would like).
Question side : Japanese sentence with kanjis. I associate a unique Japanese word to each card, so that I avoid having redundant information in my sentence deck (at the very least each card contains this word as new information).

+1 for word association

Last edited by Aspiring (2013 May 24, 8:41 pm)