Before I discovered anki, I found that flashcards worked for me.

Heck, they worked for me in grade 1 learning to read my own language, I remember using a leitner box for phonics cards following the same basic principles as anki, but in a less efficient more tedious way.

Over time the amount of flashcards I was making, while great for writing practice, was taking up too much room, and was difficult to manage so I gave it up.

Anki let me pick back up on that habit, but in a more rich way with increased accuracy, control, media etc.

Once I have enough exposure that an SRS isn't necessary, great, til then it's what I use to retain what I've learnt, otherwise I forget it.

There is a whole lot I have forgotten from even 10 years ago in my own language simply from non-use and it's my first language ^^.

The Korean from AskAKorean used boxes and boxes of flash cards, which he had grouped in bundles of 50 in order to help manage review spacing. It can be an effective method but is a lot of work. SRS just makes it easier.

"And I've found that whenever I encounter a Kanji there in its full context, it is much more likely to stick in my memory because I have actually seen it in practical use."

I think Fillanzea has it right. The intermediate stage can be a pain. Once you have more experience with reading and so on though it really shouldn't take that long to maintain 2200 cards.

anotherjohn Wrote:I have a total of 15,192 active cards in Anki with 98.22% accuracy on mature cards and *now* you're saying SRS doesn't work?

Wish you had told me sooner

How can you be sure that SRS is what helped you master these cards? Perhaps your general study method is simply very effective.

cophnia61 Wrote:As for me, when I fail a card, it is because:

1) too generic keyword;
2) two or more similar keywords;
3) the story is too generic and bland;
4) the keyword by itself is strong but the link between the keyword and the kanji, and story thereof, is weak (it's like if I ask you to remember "cat" as "a gray cloud with four rain drops"... the two things are unrelated and you are forcing them to stay together);
5) the story is ok but I haven't visualized it well enough;

Either if one or more of those points are true, I've seen that with SRS sooner or later it sticks, and every time the SRS shows it to me, I'm going to remember it for a longer time. As another user already said, SRS is not miraculous, it is only a more efficient way to review things. For me it's working rather well, it's not perfect by itself, but its usefulness is pretty obvious, at least for me.

Well, I'm glad it's working for some people, at least. Let me ask you something: do you also measure its usefulness in terms of how long a card sticks in your memory *after* you have stopped to regularly review it? Or is the increased period of stickiness only realized as long as a card remains part of the endless cycle of reviewing/relearning, etc.?

cophnia61 Wrote:Can I ask you if you use SRS only for kanji, or also for vocabulary/sentences? And what is the rate of mature cards vs young cards?

No, as of yet I haven't used SRS for anything but Kanji (although I am building some Anki-decks for Kana and grammar, in part because the effort of building them and summarizing the material itself helps memorization, and in part because separate Anki-decks can also be seen as small digital notebooks that I find easier to manage then physical notes or different Word documents).

But why do you ask, I wonder?

Learning sentences via SRS can be quite powerful, as it gives you words, Kanji, grammar patterns, all in context. You might find that if you progress with a sentence deck, your Kanji deck will also feel better.

Incidentally, HeThatShallNotBeNamed has written some good stuff about the SRS as a tool. As he says, its a great servant but a terrible master. You need to be making sure you are bringing in what is interesting, deleting what bores you, and keeping things flowing. Definitely worth a read in my opinion. Then there is that article over at SuperMemo too on 20 rules for formalulating knowledge.

1. In order for the facts/data/vocab to stay in your memory, you need to refresh that memory. Otherwise, those facts will be forgotten. (In a matter of days, weeks, months, for some facts even years - but if unrefreshed, your memory will eventually fail on you.)

2. One way of refreshing the memory is by means of flashcards. The other way is to be exposed to the facts in their natural contexts. Using flash cards is systematic and organised (you are in control of what facts are going to the review process). Exposure, OTOH, is spontaneous and unpredictable (as in, you either will meet this newly learned fact in a sufficient time to be still remembered or not).

3. People were using flashcards before SRS, mainly because of their predictability. The problem with using flashcards, in the absence of SRS, was that reviewing was not optimal in terms of time, i.e. some cards were reviewed to early (the memory for those was strong enough that even without refreshing, the fact would be recoverable), or to late (you missed the point were the fact was still recoverable, and now you don't remember it anymore).

4. SRS is an improvement on the old-type flashcards - it optimises the time you spend on refreshing your memory. And that's really all it is to SRS.

5. If your only method of studying is by means of SRS then you shouldn't expect much. By this I mean that SRS delivers it's promise, which is a given retention rate (80% by default in Anki) on the border of the facts to be forgotten. To make sure that your recall is better than that (i.e. better retention rate, faster recall) you should use other methods of studying alongside SRS (reading, listening, watching movies, etc.)

Inny Jan Wrote:1. In order for the facts/data/vocab to stay in your memory, you need to refresh that memory. Otherwise, those facts will be forgotten. (In a matter of days, weeks, months, for some facts even years - but if unrefreshed, your memory will eventually fail on you.)

2. One way of refreshing the memory is by means of flashcards. The other way is to be exposed to the facts in their natural contexts. Using flash cards is systematic and organised (you are in control of what facts are going to the review process).

4. SRS is an improvement on the old-type flashcards - it optimises the time you spend on refreshing your memory. And that's really all it is to SRS.

That's all fine and good, but approximately *how much* better is retention supposed to be when using SRS to structure your learning? A few days, weeks, months? Because so far, I've seen many other posters expound on the supposed benefits of structuring learning with SRS, without much quantification. (Frankly, it's starting to remind me of the way vitamin supplements are recommended: exactly how benefit might be reaped from them is unclear, but since no harm is supposed to come from taking them, we are encouraged to take one daily.)

Inny Jan Wrote:5. If your only method of studying is by means of SRS then you shouldn't expect much. By this I mean that SRS delivers it's promise, which is a given retention rate (80% by default in Anki) on the border of the facts to be forgotten.

But 80% retention for how long? And how much "maintenance" is required for how long in order to guarantee this percentage for a given size of a deck?

Because if the answer is something along the lines of "multiple reviews of a card a week or a month for an indefinite amount of time until you build up enough familiarity with the information from other sources", then learning structured with SRS really isn't learning at all, but more like a stop-gap measure for when you finally get to *really* learn the material from proper sources.
And then the question becomes: is all the effort needed to keep that stop-gap measure implemented day in, day out really worth it, compared to directly slugging your way through source material that comes with some explanatory notes (or furigana) that can give you access to permanent retention?

And although I may be going somewhat far off-topic here, but isn't there something odd about the apparent (to me, at least...) absence of collections of Japanese source texts (with explanatory notes) that gradually increase in difficulty? Surely it should be possible to start with texts from children's books for ages 6-8 and then increase to 8-12, then 12-14, and so on and so on, until we finally get to full-fledged newspaper articles and novels and such? Then none of this "using SRS-structured learning as a stop-gap measure until you are finally ready to start reading full-blown texts for native speakers" would be necessary. (I've found exactly one book on Amazon UK that is a little like this, but unfortunately many reviewers still qualify it as far too advanced for basically anyone who is not fairly close to a native level of Japanese).

Eminem2 Wrote:But 80% retention for how long? And how much "maintenance" is required for how long in order to guarantee this percentage for a given size of a deck?

Forever. 10-20 times as many reviews per day as you add cards per day. Forever.

Quote:And although I may be going somewhat far off-topic here, but isn't there something odd about the apparent (to me, at least...) absence of collections of Japanese source texts (with explanatory notes) that gradually increase in difficulty? Surely it should be possible to start with texts from children's books for ages 6-8 and then increase to 8-12, then 12-14, and so on and so on, until we finally get to full-fledged newspaper articles and novels and such?

Children's books aren't great for beginning to intermediate learners, actually. I have a vocabulary of about 12,000 words in Japanese and I still encounter a lot of new words when I read elementary-school-level chapter books; I encounter almost as many new words as I do when reading non-specialized contemporary adult novels. An 8-year-old Japanese child is typically someone who already has maybe an 8000-word vocabulary, and who has a full grasp of a lot of points of Japanese grammar that most second language learners don't have until they get to a high-intermediate level.

A lot of research in second language acquisition does suggest that extensive reading of easy material is one of the most powerful things you can do. (A lot of that research actually comes from English learning in Japan and South Korea, because teachers and students are so frustrated by the phenomenon of studying English for years and years and still not being able to deal with the language in the real world). But their solution is typically graded readers with a very controlled vocabulary, written for second language learners rather than native speaker children.

There is a graded reader series for Japanese -- the Japanese Graded Reader series -- but it's expensive (because all the texts come with CDs, which they should, but it does add a cost) and there aren't enough of them. There is a very small market for books for people who want to learn Japanese and of those people, there's a much smaller number who get to the point where they need reading material that's more advanced than the very controlled language in early-level textbooks.

The reason I'm still using SRS (for Chinese, where my level is still low-intermediate) despite really considering it to be a stopgap is that when I do read books for native speakers, even children's books, the volume of unknown words is so high, and I end up looking up dozens of words and immediately forgetting them. And then, it might be months before I see most of those words again! This is what's hard and a little paradoxical about vocabulary: you need a vocabulary of somewhere around 9,000-10,000 word families to be able to read with adequate comprehension (in English, at least) but the 8000th or 9000th most common word isn't common enough to see often enough to remember it unless you're able to get really large amounts of input. No matter how good the 'key' is, I'm not going to be able to look up a new word once and remember it for several months, especially if I'm trying to learn 10-25 new words a day.

So, yes, my ideal is to have a lot of graded readers written for second language learners that go from beginner levels to 'advanced enough to read children's chapter books' levels. In the absence of that ideal, well, SRS. (And slogging through as much native material as possible.)

Incidentally, most of the cards in my Chinese deck are cards with an interval of about one to two months. After the initial learning stage I have very few cards that I see more than once a week. I don't see that as especially burdensome. And I don't believe in reviewing forever -- eventually you get to a point with the language where you can read a lot comfortably, and you can understand media when you listen to it, and then you're reinforcing the vocabulary you need just by using the language. It's certainly possible to get to that point without SRS. But I think it's more of a hurdle to get over that point where you know a lot of words but not enough so that you can comfortably read material for native speakers.

Eminem2 Wrote:That's all fine and good, but approximately *how much* better is retention supposed to be when using SRS to structure your learning? A few days, weeks, months?

For a detailed answer to this question you need to go to:
http://www.supermemo.com/

or more specifically, to:
http://www.supermemo.com/english/ol/beginning.htm

There you will find a quantitative comparison of SRS vs other flashcard schedulings.

If you want to compare SRS vs unstructured learning, then I'm afraid you may be hard pressed to find a conclusive data. There is an element of a “strength of memory” involved in learning, and I think, it was demonstrated that context helps with retention but there is no escape from “memory refresh cycles”.

Vempele Wrote:

Eminem2 Wrote:But 80% retention for how long? And how much "maintenance" is required for how long in order to guarantee this percentage for a given size of a deck?

Forever. 10-20 times as many reviews per day as you add cards per day. Forever.

+1

Mind you that with the spacing effect, the interval between repetitions increases with each successful review, so for the most common facts/vocabulary (provided that you do reading, for example) the cost of that maintenance quickly becomes negligible.

Eminem2 Wrote:Because if the answer is something along the lines of "multiple reviews of a card a week or a month for an indefinite amount of time until you build up enough familiarity with the information from other sources", then learning structured with SRS really isn't learning at all, but more like a stop-gap measure for when you finally get to *really* learn the material from proper sources.

Yes, SRS is a stop-gap measure. It's hardly useful on its own.

Einsteins quote is not even remotely discussable, it states a fact: Repeating something under the same circumstances while expecting a different result is stupid.

I don't see why this would apply to SRSing, where the circumstances are supposed to change - however, the circumstances don't change enough, that's why you only use it to remember something you did learn already. Or you cram something in the hopes that under some hopefully coming up circumstances you WILL learn it, and then it all paid out.

That's why I personally prefer reading. I think that a little SRSing ain't bad, but to overdo will yield lackluster results, compared to reading.

By the way, the brain does store everything. However, it only builds connections when needed, and when something is needed is something the brain decides using unknown measures, so if you could convince your brain to accept the data and make it accessible, you could do what that one autistic guy did: read countless books, as you'll automatically learn them by heart and quote everything at will. This doesn't work well with SRS - thank god, or you could say "and on page 385, paragraph 3, line 4 the 12th word is "and"" - which would be totally nonsensical.

I would like to add something to the topic, not directly related to the Einstein quote and SRSing.

The human brain is a biochemical computer of high complexity, the mechanics are only known partially. There are signs that this or that theory might be true of how it works. This still hasn't changed. There have been countless experiments on the matter.

Take for example psychiatric problems. There are tons of different meds for depression. They all are working for something, but "depression" is not a clearly definable state of mind. The reasons, the mechanics and even the physical circumstances differ from person to person. So what doctors do is, they experiment by giving you different medication while watching the results, until they found whatever works for you.

Isn't it a bit silly to assume that "the author of Supermemo" or "the maker of ANKI" found the philosopher's stone? If humans differ and brains are different even on a physical level, how can one method "work, because it's scientifically proven to work"? Nothing is scientifically proven so far. There are theories, one as good or bad as the other. SRS works to some degree under certain circumstances for certain people, and that's fine. If that's your memory pill, go for it.

But like depression patients, one pill won't work for all of us.

Yes, and neither is a hoe a heavy plow (to continue my analogy from before), but it's better than some stick you found on the ground. All current knowledge of learning suggests that spaced reviews and a good understanding of a concept (lots of context, in a language perspective) are the best ways to intentionally acquire knowledge.
In other words, extensive (to some level) reading and SRS are the best available techniques for most people to use. It doesn't mean that it's the only or best way possible to go about it, but it does work if you use it.
Frankly, if the only tools I knew of to turn earth were a stick and a stick with a sharp flat thing at the end (regular flashcards or an SRS program), I think I'd take the one with the sharp flat thing at the end, because it seems like it would turn soil better. SRS seems like it's a better, more efficient form of review than flashcards, and it is (if only because it takes up less space and is quicker to access), just a hoe is more efficient than a stick.

Fillanzea Wrote:The reason I'm still using SRS (for Chinese, where my level is still low-intermediate) despite really considering it to be a stopgap is that when I do read books for native speakers, even children's books, the volume of unknown words is so high, and I end up looking up dozens of words and immediately forgetting them. And then, it might be months before I see most of those words again! This is what's hard and a little paradoxical about vocabulary: you need a vocabulary of somewhere around 9,000-10,000 word families to be able to read with adequate comprehension (in English, at least) but the 8000th or 9000th most common word isn't common enough to see often enough to remember it unless you're able to get really large amounts of input.

Interesting, although curious that my own experiences (with 2nd language learning in general, I don't have any experience with Chinese) have been somewhat different. When I learned English in grammar school, I had the benefit of also seeing many subtitled English TV-shows and movies. And, at the time, I also started reading detective stories (Sherlock Holmes, Agatha Christie), then spy thrillers (James Bond, John LeCarré, Craig Thomas, Robert Ludlum). Of course I didn't fully understand everything at first, but quite soon I did, and English tests at school became a breeze. (French and German, however, I had to learn with very few sub-titled TV-shows and movies or reading novels and - probably as a result - those languages never really came to life for me.)

I never really wondered just how many word-families I had apparently come to know due to being able to read most English texts with relative ease about half-way through grammar school. And even rarer words (the 8,000th or 9,000th most common words, as you refer to them) did manage to stick in my mind without any further repetition other than simply having encountered them in texts (including having seen their translations in subtitles). I never did anything even remotely resembling SRS for English during that time. Sure, I learned vocab like any other student: cram them once before the test and then ace the test. And then I moved on to the next lesson, be it a reading comprehension text with questions, a listening test, pronunciation test, more vocab, grammar, writing exercise or whatever. And it all went smoothly.

So if Japanese has a similar number of word-families that is needed to understand most texts (like the 8,000 you mention for elementary school children), then the same method might work for me there. *If* there were sufficient graded texts available, as you call them. For fan-subbed anime already provide me with the equivalent of the subtitled English TV-shows and movies I watched during grammar school.

Fillanzea Wrote:No matter how good the 'key' is, I'm not going to be able to look up a new word once and remember it for several months, especially if I'm trying to learn 10-25 new words a day.

Again, curious. Not that I'm doubting your word, of course. For me, finding a good 'key' is synonymous with being able to remember the translation for at least a few months. Then again, doing that at a rate at 10-25 words a day is not something I recall ever having consciously tried. And yet, becoming familiar with that minimum of 8,000 English word families you mentioned, is something I apparently managed during my first 3-3.5 years of grammar school without striving for it or putting in much extra effort. 8,000 / (365*3.5) = 6.26 words per day. So maybe the lower pace also made things more manageable?

Fillanzea Wrote:Incidentally, most of the cards in my Chinese deck are cards with an interval of about one to two months. After the initial learning stage I have very few cards that I see more than once a week.

Then that seems to be a very prominent difference compared to this website's SRS, that hits me with about 50 reviews a day even after not having added any new or relearned cards for weeks. That means the entire 2,200 card deck more or less gets recycled after 44 days. Back when I still re-learned failed cards fairly quickly and more or less managed a 75% retention rate, the daily review stack would generally be around 150-180 cards, with some days surpassing 200. So 2,200 cards in eleven days, even if some of those contained some cards that went wrong a little more often. After 3.5 months of that without hardly any drop in the average daily number of expired cards, I go so frustrated that I started okay-ing entire stacks just to be rid of them. Only recently, after the number of daily reviews finally came down, did I try doing them thoroughly again. Now, my retention rate is somewhat lower (40-60%, with some much poorer days) and I'm no longer in any hurry to re-learn failed cards very quickly.

This brings to mind a bit of advice from some advanced learners (who used Anki) I read on this site back when RTK was still new to me: one of the greatest improvements to their language learning was simply giving up on SRS assisted learning and not bothering with the daily Anki review grind any more, instead just focusing on grammar and other things. Back then, I didn't know what to think of that. Now I do.

Inny Jan Wrote:

Eminem2 Wrote:That's all fine and good, but approximately *how much* better is retention supposed to be when using SRS to structure your learning? A few days, weeks, months?

For a detailed answer to this question you need to go to:
http://www.supermemo.com/

or more specifically, to:
http://www.supermemo.com/english/ol/beginning.htm

That second link is basically a research paper. How about a more concise answer in your own words?

Inny Jan Wrote:If you want to compare SRS vs unstructured learning, then I'm afraid you may be hard pressed to find a conclusive data.

That's what I was afraid of...

Inny Jan Wrote:Mind you that with the spacing effect, the interval between repetitions increases with each successful review, so for the most common facts/vocabulary (provided that you do reading, for example) the cost of that maintenance quickly becomes negligible.

That's just a fancy way of saying that you don't need to re-learn things that you haven't forgotten.

Inny Jan Wrote:Yes, SRS is a stop-gap measure. It's hardly useful on its own.

So basically, there is no way to test SRS independently of whatever basic learning method is being used. A cynic might sum this all up by noting that whenever someone who is using SRS isn't getting the desired results, then there's obviously something wrong with the primary learning method. And when SRS-assisted learning *does* yield the desired results, this is proof of the great effectiveness of SRS...

andikaze Wrote:Einsteins quote is not even remotely discussable, it states a fact: Repeating something under the same circumstances while expecting a different result is stupid.

Well, to play Devil's Advocate for a moment and argue against my own point of view: there are some situations in which doing the same thing over and over *will* eventually yield a different result. Think of the classic image of a bank robber trying to crack a safe by using a stethoscope while turning the dial on the safe: if he does it right, the safe will open after having found the right combination. But this example is somewhat problematic in terms of what constitutes a repetition: is each turn of the wheel a separate try, or can only an average number of tries that is generally needed to crack that type of safe be counted as a single try? Probably the latter.

Or someone trying to bash in a door may need to smash against it 10 or 15 times before it finally gives in. Every bash can have the same amount of force, with only the final one being successful.

andikaze Wrote:That's why I personally prefer reading. I think that a little SRSing ain't bad, but to overdo will yield lackluster results, compared to reading.

Interesting. So you have found that the concept of needing to master at least 8,000 word(-groups) in a language before you can read slightly more advanced texts doesn't always apply? (So have I).

andikaze Wrote:By the way, the brain does store everything. However, it only builds connections when needed, and when something is needed is something the brain decides using unknown measures, so if you could convince your brain to accept the data and make it accessible, you could do what that one autistic guy did: read countless books, as you'll automatically learn them by heart and quote everything at will.

My thoughts exactly: there must be something that makes the brain decide that something is worth remembering in connection with something else. And if that "something" could somehow be discovered and implemented as a part of language learning, then it would become a whole lot easier. And SRS would not even be necessary, whether it works or not.

And autistic brains indeed seem to work rather differently from "normal" brains, which opens up the possibility that there maybe many different types of brains when it comes to language learning, each with their own optimal method. No "one size fits all"...

Just wanted to briefly say that it's been backed up by research that

Quote:If 98% coverage of a text is needed for unassisted comprehension, then a 8,000
to 9,000 word-family vocabulary is needed for comprehension of written text
and a vocabulary of 6,000 to 7,000 for spoken text.

Now obviously that doesn't mean you need to learn those 8,000 word families by SRS or flash cards or anything like that.

Quote:Interesting, although curious that my own experiences (with 2nd language learning in general, I don't have any experience with Chinese) have been somewhat different. When I learned English in grammar school, I had the benefit of also seeing many subtitled English TV-shows and movies.

Precisely! IF you have really large amounts of input, then you will be able to get more exposure to those rarer words. And that is essential, far more essential than doing SRS. But for it to work, you need to have sources of input that are on a level that you can comprehend. (Also, not sure of your native language, but it definitely makes a difference if your language is related to English or has cognates.) At the stage where reading is very slow and TV shows just sound like a stream of noise, though, it's quite a bit harder to get that input.

Quote:Again, curious. Not that I'm doubting your word, of course. For me, finding a good 'key' is synonymous with being able to remember the translation for at least a few months. Then again, doing that at a rate at 10-25 words a day is not something I recall ever having consciously tried. And yet, becoming familiar with that minimum of 8,000 English word families you mentioned, is something I apparently managed during my first 3-3.5 years of grammar school without striving for it or putting in much extra effort. 8,000 / (365*3.5) = 6.26 words per day. So maybe the lower pace also made things more manageable?

And I think, possibly, the lower density of unknown words.

If you read a page and find one unknown word, that one unknown word stands out, and is memorable. If you read a page and find ten unknown words, then it's hard to hold all that new information in your mind. Which is a good argument for graded readers.

(Also, I do think there's a difference between using SRS for vocabulary, where you reinforce that vocabulary whenever you read, and using SRS for Heisig, where reading tends to reinforce the meaning of kanji compounds but not necessarily the English Heisig keyword for an individual kanji. My retention rate for vocabulary is somewhere around 95%. That's why I have long intervals, and that's why it's not frustrating. I doubt I could manage that with Heisig.)

My experience has always been, that some words stick at the first encounter. Others take lots and lots of encounters, sometimes another piece of knowledge is needed, and then one day it clicks and they stick, too.

With that in mind, why not read something and look up whatever bothers you (and that might be a lot in the beginning), because that way, while you may forget 80% of the words, 20% will stick. You expand your knowledge bit by bit.

You don't have to wait until "you're ready" to read if the thing you plan to read interests you and you read it for the content. Just like you should definitely not wait till you're "ready to speak", because that day may never come.

Extensive reading will take over at some point, and from there on acquisition will become semi-automatic. Only the first few books are tough. They teach you a lot flashcards don't, however. Sure, ANKI cards can hold some sentences, but not enough to make for a fascinating environment that makes words memorable.

I don't say ANKI can't work. I'd put sentences like 「そのパンティーは、脱がしてやろうぜぃ」, and not the boring stuff from, say, Core 6000 or Tatoeba.

SRS by itself isn't a bad concept. It's a program, a system, or a helpful tool. The feeling of insanity comes from learning what we (unknowningly) consider pointless.

I agree with the idea of using SRS as a database, adding only what you think you need or want to learn.

A popular approach lately has been:
Get to 2000 kanji
Get to 6000 words
(Why?)*

After 10,000 words we assume we've "learnt Japanese". Whether this is true or not, these kinds of goals are vague.

A better mindset would be:
"Interesting word! I'd like to review it, so I can understand the word the next time I see it."
"I'd like to learn 10 interesting words/kanji today"
"This sentence seems difficult to understand. Oh, there's a new word. Should I learn it? (Yes/No/Maybe)."
"Take things one day at a time"

These all relate closely to S.M.A.R.T. goals.
These goals are specific, measurable, attainable, relevant, and time-bound.

Make goals specific and be certain of what you're doing.
Make goals measureable and be aware of your progress.
Make goals attainable and reach for goals you're confident in.
Make goals relevant and know why you're doing them, and how they relate to your goals.
Make goals time-bound. Take things one day at a time. Maybe use time-boxing.

*There seems to be an unspoken consensus that these methods give us a solid foundation and bring us closer to fluency. Which is all right. But be sure to avoid shallow learning. Don't just do it because other people tell you to do it. Learn rtk or core because you want to.

Eminem2 Wrote:[...]

あなたの向こうからトロールのひどい匂いが出てきている。

（少し日本語の練習をしようと思った　：））

The way I understand the OP, they were talking about SRS with RevTK. As I understand the Heisig system, it's a shortcut to get yourself familiar with the kanji so that you can start reading. That's it. There's no real need to do thousands and thousands of reviews to try and lock down the connection between keyword and kanji.

Some stories just won't connect a keyword to a kanji, and that's fine. As long as it works for the bulk of them. And as long as you learn the different components, so you can recognize kanji when you see them. It's very difficult to remember, or even properly recognize kanji, if you don't know the component parts.

But if you learn the radicals, or Heisig's "primitives", and go through the book attaching the best stories you can to keywords, you're pretty much done. Even if you end the book with only 60% or 70% recall, that's still pretty good.

As Heisig says, there's no need to remember the English keyword when you see the kanji. It's just a shorcut to get you as familiar as possible with the kanji within a limited stretch of time.

For the ones you can't connect to the keyword, don't sweat it too much. You'll eventually just associate them with the words they are used to write.

Inny Jan Wrote:あなたの向こうからトロールのひどい匂いが出てきている。

（少し日本語の練習をしようと思った　：））

A reading exercise! How very thoughtful of you. Thx!

Bortrun Wrote:The way I understand the OP, they were talking about SRS with RevTK. As I understand the Heisig system, it's a shortcut to get yourself familiar with the kanji so that you can start reading. That's it. There's no real need to do thousands and thousands of reviews to try and lock down the connection between keyword and kanji.

I agree with part of what you said. I learned all 2,042 kanji in RTK1 WITHOUT USING SRS AND ANKI and without doing thousands of reviews.

The secret was using effective stories/mnemonics for each kanji. If you carefully craft your stories you shouldn't have to do too many reviews.

Personally I find staring at same flashcards over and over again without effective mnemonics doesn't do anything for me. I tried that once it just doesn't work.

Here's something related (but not about kanji): I found a study on the internet once where the researchers concluded that in learning foreign vocabulary, the most effective method was using mnemonics, in this case what they called the "keyword" method. Here's an example of the keyword method: limosus = muddy in Latin. "my limousine was muddy." A year from today I'll remember that limosus - muddy without staring at flashcards using SRS in the meantime.