(somewhat long post, but it's from the heart)

Einstein once defined insanity as "doing the same thing over and over again and expecting different results."

What I've been wondering, is: does this not also apply to SRS "learning"? Because if a Kanji (or sentence, or capital of a country or whatever) hasn't stuck the first time one tried learning it, why would repeating that failed method of learning it yield any different results?

And yes, I know the philosophy behind SRS is more or less that the more often you see something, the more your brain will start to believe it must be useful to remember. But that would mean Einstein was wrong, and since he wasn't wrong very often that simply doesn't sound very likely. And the philosophy behind SRS also presupposes that the brain has very few intelligent filtering mechanisms and simply stores everything it sees an X number of times. If that were really true, then simple cramming would be the be-all, end-all of learning and there wouldn't be any point at all to making Heisig-style stories to remember Kanji, because simply looking at them in combination with their key meanings an X number of times would be all that was needed to get the job done.

After doing at least 30,000+ SRS reviews (mostly using RevTK), I find it increasingly hard to escape the conclusion that Einstein was right once again. Sure, having seen a card a few days (or perhaps even weeks) ago might mean that it has stuck in my short term memory for a while. And re-learning it after it has slipped my short term memory will be effective for perhaps another few weeks. But given how much time it costs to continually perform maintenance on just the 2,200 RTK1 Kanji in order to keep them in my short term memory, this more and more has begun to feel like an insane undertaking.

So, recently, I've been shifting my efforts to learning and summarizing grammar (JTMW). 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. (And painfully often, Heisig's key meanings turn out to have been inaccurate or at the very least very outlandish. I mean, "petition" for the polite version of "please" i.e. お願 or おねがい? Not very helpful, in my opinion.)

Basically, the conclusion that I have come to with some pain in my heart over all my wasted effort, is that SRS is basically "the emperor's new clothes". The concept may sound convincing on the surface, and all the many bells, whistles and progress charts that digital learning can add to it can look very enticing, but that's about all. In my experience, memorization is like fitting a key to a lock: either the key turns or it does not. If you're lucky, then the key turns on your very first try. If you're not, then you may need to try a large number of different keys. But success is measured in terms of whether or not you have found the right key, not in terms of how many times you have tried to turn the wrong one(s).

So I'm going to use SRS as a database for the keys I've found (or think I've found) and not as a learning tool. In as far as I still do reviews, they have the purpose of identifying ineffective keys and the "failed" stack will become my "keys yet to be improved" stack.

You are taking some guy's uninformed (and hopelessly vague) quote way too seriously. Einstein may have been a genius when it comes to figuring out the structure and rules of the universe, but human mind is really not the same at all. Einstein was also wrong a lot. In this context Einstein is nothing but "some guy with an opinion".

Who decided that a word not sticking the very first time makes the method a failure? That if you have to do something twice/thrice/some random number to make it work it means we are insane? I honestly don't think Einstein meant that, and if he did, he was wrong. Very popular advice here is that if a word/kanji doesn't stick for a number of tries, you move on and try to learn it later or through some other method. In that way we are advocating the principles of Einstein's message even though it may not follow the exact Einstein's guidelines.

Lastly "short-term memory" can hold information for a matter of seconds. Once you can recall the word after an hour has passed, it has already moved to your long-term memory. Further, SRS's function is to help you move new information from your short-term memory to your long-term memory, and for that purpose SRS is an excellent tool. Nothing stops you from using context in SRS too, and the Core decks already come with a context for learning words. Is this mean to be a critique of SRS or Heisig?

The idea isn't that if you see it often it sticks, but that if you have to remember it often it sticks. It's an ages old, tried and true method of learning. If you review everyday, learning vocabulary becomes painless: that's just common sense. In my opinion SRS is just a way of managing your flashcards without having to think about it, so I don't think there are any bells and whistles attached.

I'm not sure about SRS exercising your short-term memory - I seem to have the most failed cards in boxes 3 and 4, not in box 5, and then again I only get about 2 or 3 failed cards every session. Plus when I learn a new word I often realise the kanji are familiar.

I do a traditional kanji book besides Heisig, though. Maybe that helps: it definitely helps with recognition.

But I agree that grammar is very important.
As for context, I think it's essential that vocabulary flashcards come from texts, even if they are textbook conversations.

I think it's a mistake to rely entirely on Heisig. Heisig is a fast way of becoming familiar with the characters so learning vocabulary is less painful. Complaining that "petition" is a misleading keyword (you could've chosen a better example: "negai" does mean "petition", and "onegai suru" literally means "to request", as does "negau") is missing the point. But I do suspect that if I just kept doing Basic Kanji Book with Anki and no Heisig it wouldn't be too bad.

Betelgeuzah Wrote:Who decided that a word not sticking the very first time makes the method a failure?

I didn't quite say that. It's just that the "keep going at it until it sticks in your memory" approach is wrong, in my experience.

Betelgeuzah Wrote:That if you have to do something twice/thrice/some random number to make it work it means we are insane?

Well, the expectation that something that has proved to be ineffective will somehow become effective through a lot of repetition (not just 2 or 3 times) does sound rather insane to me.

Betelgeuzah Wrote:I honestly don't think Einstein meant that, and if he did, he was wrong. Very popular advice here is that if a word/kanji doesn't stick for a number of tries, you move on and try to learn it later or through some other method. In that way we are advocating the principles of Einstein's message even though it may not follow the exact Einstein's guidelines.

And that's fine by me.

Betelgeuzah Wrote:Lastly "short-term memory" can hold information for a matter of seconds. Once you can recall the word after an hour has passed, it has already moved to your long-term memory.

Really? Then there must be some sort of "mid-term" memory for anything between a couple of days and a couple of weeks. Or perhaps there is a difference between entering long term memory and really being fixed in it.

Betelgeuzah Wrote:Further, SRS's function is to help you move new information from your short-term memory to your long-term memory, and for that purpose SRS is an excellent tool.

I haven't found it to be an "excellent tool" at all. If it was, then "cramming" would be the king of memorization. And it isn't, else this whole website might not have existed in the first place.

Betelgeuzah Wrote:Nothing stops you from using context in SRS too, and the Core decks already come with a context for learning words.

True, but why would any context be necessary if the concept of SRS worked so well all by itself?

Betelgeuzah Wrote:Is this mean to be a critique of SRS or Heisig?

Primarily SRS, but also Heisig.

Eminem2 Wrote:

Betelgeuzah Wrote:Lastly "short-term memory" can hold information for a matter of seconds. Once you can recall the word after an hour has passed, it has already moved to your long-term memory.

Really? Then there must be some sort of "mid-term" memory for anything between a couple of days and a couple of weeks. Or perhaps there is a difference between entering long term memory and really being fixed in it.

Long-term memory isn't quite forever, though. For example, do you remember how to calculate a square root by hand, or the phone number of the house you first lived in? You might have a "ohh, that's right" moment if you refresh those things, but off the top of your head...?

SRS is a mean to maintain what you've already learned, not to learn it in the first place. For it to work you must push the reviews forward, maybe forget them, or it's better to say fail to recall them, then the next time your inability to recall them must be smaller. When I forget a kanji, let's see after a week, the next time I "relearn" it, I see I can recall it after a month, then if I forgot it another time, the next time I see I can remember it pretty well after two or three months. I cannot speak about larger intervals, or about vocabulary.

It's the same for english (if it's not your L1), if you learned it by reading, the first time you look the same word again and again and not only you forget what it means, you also forget you have already seen it before. But by persisting in this rote memorization process, you see the next time your forgetting rate of the same word is smaller, and so on by larger and larger intevals, and one day you notice you can recognize that word after years by just looking at it, as like as it happens for your L1. And this is the "SRS" part of the process.

Obviously if you use only SRS it's effect will be weak or nonexisting for some words, it's only a way to review and maintain fresh in memory what you've already learned, and instead of reviewing ad random intervals or worst every day, it optimize your reviewing process.

poblequadrat Wrote:The idea isn't that if you see it often it sticks, but that if you have to remember it often it sticks. It's an ages old, tried and true method of learning.

Frankly, "remembering it" and "sticks" are more or less synonymous in my book. I consider something "remembered" if it has managed to "stick" in memory after about 4-5 weeks time without any further effort.

As for the "tried and true" method of learning that SRS purportedly is, why have people deemed it necessary to develop alternatives for it, such as "immersion" if it was such a flawless method? Or in other words, why doesn't everyone remembers their high school French so well after all these years, after they looked at their text books so often back then?

poblequadrat Wrote:If you review everyday, learning vocabulary becomes painless: that's just common sense.

Then why are study efforts like memorizing dates (history) or vocab through repetition so universally detested by students? Perhaps because facts or words learned this way are so easily forgotten after having given so much trouble to learn initially?

poblequadrat Wrote:I'm not sure about SRS exercising your short-term memory - I seem to have the most failed cards in boxes 3 and 4, not in box 5, and then again I only get about 2 or 3 failed cards every session.

The transition from box 4 to 5 is pretty close to where I have experienced the dividing line between "short" and "long" term memory to be.

poblequadrat Wrote:Plus when I learn a new word I often realise the kanji are familiar.

I'm not sure about what you're trying to say here... Do you mean when a new Kanji is partly made up of primitives that you are already familiar with?

poblequadrat Wrote:I do a traditional kanji book besides Heisig, though. Maybe that helps: it definitely helps with recognition.

What does "traditional" entail, exactly?

poblequadrat Wrote:But I agree that grammar is very important.
As for context, I think it's essential that vocabulary flashcards come from texts, even if they are textbook conversations.

Absolutely, but that rightfully shifts the discussion from whether or not SRS works to what other kinds of learning systems work. If a flashcard based on a text (even with context) works better than one without, then that's a strong indication that it's the quality of the card that is decisive and not how many times is has been SRSed.

poblequadrat Wrote:I think it's a mistake to rely entirely on Heisig. Heisig is a fast way of becoming familiar with the characters so learning vocabulary is less painful. Complaining that "petition" is a misleading keyword (you could've chosen a better example: "negai" does mean "petition", and "onegai suru" literally means "to request", as does "negau") is missing the point.

If the most common usage of "negai" is as part of "please", then that is what every beginning learner of Japanese will want to know. Ignoring that bit of information and only including the (at least in English) far less used word "petition" seems very impractical to me.

Here's the way I think about it, as someone who does use SRS but not so much RTK. Really learning a word means having a lot of connections to it in your memory. You need to have seen it multiple times in multiple contexts. If you just add a word to your SRS and drill it, then you only have one connection, and that's very fragile. You can't learn a word that way -- but you CAN keep it alive in your memory so that it's easier to learn the next time you see it in context.

The problem when you're at an intermediate stage of learning is that it takes massive amounts of reading to get the kind of vocabulary that will let you actually do massive amounts of reading. And one way to get over that Catch-22 is to brute-force memorize quite a lot of words and let them act as weak placeholders in your head until real reading can take over. (Another way is with lots of carefully leveled graded readers. I think that's better, but there aren't enough of them yet in Chinese.)

If anything *does* work just by doing the same thing over and over again, it's memory. There are probably silly commercial jingles that were on TV when you were a kid that you could recite from memory with the proper cue.

But when you've learned the English keyword that corresponds to a particular kanji, you haven't learned any words using the kanji, you haven't learned the full range of meaning associated with the kanji, you haven't learned any readings -- whether that's a helpful stopgap or not is not for me to say, but it really is just a stopgap.

You may be happier with vocabulary SRS than with RTK SRS.

@ Fillanzea

I can really relate to what you're saying. The only way to get more or less "active" vocabulary is indeed by having seen a lot of examples in actual use. Either in the form of reading or in the form of subtitles for languages that aren't spoken so rapidly that the viewer can still make out individual words.

Although, as far as the number of connections to a word goes, I am not sure that it is purely a matter of quantity. Quality also matters a lot, in the sense that a word you have seen used in a memorable scene or by a memorable character will stick in your memory much better than words that you have seen being used in more drab circumstances. Your example of the silly, commercial jingle is one I find to be more illustrative of quality than of quantity, since these tunes are generally selected for their catchiness and then get repeated a lot in the hopes of reaching many potential customers that way, and not so much to make them memorable. (Although I suppose some not-so-catchy tunes have probably etched themselves into memory as well after hearing them perhaps thousands of times, but I doubt if this method could be used to memorize a great many things within a reasonable amount of time because the needed number of repetitions is so huge.)

And that's more or less what I meant with my earlier comments about "keys" fitting instead of having tried many non-fitting keys: language learning in general and learning vocab in particular have so much more to do with things that speak to us on a personal level, with "love" if you will, than with dry concepts such as "techniques" or "numbers of reviews". Even if drilling words or Kanji may provide some sort of weak awareness of them until you find a more memorable occasion where they are used, as you mention. (The only exceptions are perhaps extremely everyday nouns and verbs, like "car", "eat", "sleep" and "house" that probably carry such vivid archetypical images that I had little trouble remembering them (except for "sleep" for some reason, come to think of it...)

Eminem2 Wrote:Well, the expectation that something that has proved to be ineffective will somehow become effective through a lot of repetition (not just 2 or 3 times) does sound rather insane to me.

I don't know who has such expectations here. Usually 2 to 3 times is enough, if you are using SRS correctly. When you get to 4-5 reps and the word doesn't stick, you move on and try later (or try a different method if the information is so essential that it cannot wait). I think this is what the people in koohii.com would advice someone to do. It also doesn't make SRS ineffective because 2 to 3 times is all it takes in a great majority of cases. If that doesn't happen then there is clearly a need to change something, whether it be the information you try to memorize or the method (usually fixing the former will prove to be enough). SRS is only as effective as how it is used.

Quote:Really? Then there must be some sort of "mid-term" memory for anything between a couple of days and a couple of weeks. Or perhaps there is a difference between entering long term memory and really being fixed in it.

You will forget things even if they have been in your long-term memory for years. Information does not suddenly become "forget-proof" if it enters your long-term memory or even if you manage to make it so the information stays there for years. Unless you keep encountering the information it will eventually fade from your memory. But the more you encounter the information the longer it will take to forget it. This is what SRS helps you with: it shows you the information when necessary so that a) you won't forget it and b) you will remember it for longer. Initially the time span is short, but later on it grows exponentially. The algorithm is not perfect but we must remember that imperfect does not always equal ineffective.

Quote:I haven't found it to be an "excellent tool" at all. If it was, then "cramming" would be the king of memorization. And it isn't, else this whole website might not have existed in the first place.

Cramming is the opposite of SRS. SRS shows you the information only when necessary for it to stick in your long-term memory. When you "cram" you study the information much more than would be necessary for it to stick in your long-term memory. That is what the SRS attempts to avoid. Now, do not mistake SRS for helping you understand a new concept. Understanding happens outside SRS, SRS simply helps you memorize what you have understood. In the case of kanji and vocab for example, a complete understanding happens elsewhere, but even memorizing (hooking your mind to the concept to be recalled and understood in different contexts later) is incredibly useful.

Quote:True, but why would any context be necessary if the concept of SRS worked so well all by itself?

It's not necessary but it helps. Context only helps you understand what you are trying to learn. SRS helps you move what you have understood into your long-term memory for increasingly longer time periods. You can do that by encountering the information elsewhere, but in the case of Japanese that means going through native materials when you haven't hooked your mind to the information presented to you (the words and grammar). It is painful and many find SRS to be a great alternative initially.

He's messing with you guys.

Eminem2 Wrote:

Betelgeuzah Wrote:Lastly "short-term memory" can hold information for a matter of seconds. Once you can recall the word after an hour has passed, it has already moved to your long-term memory.

Really? Then there must be some sort of "mid-term" memory for anything between a couple of days and a couple of weeks. Or perhaps there is a difference between entering long term memory and really being fixed in it.

The explanation we were given in class (not in terms of biological function) was basically 'use it or lose it'. Memories are just paths of neural connections; once one is established, it's there, but weak; if you don't access that memory, it gets lost in the noise of the other memories, since the connections aren't that strong; but if you use it, you reinforce the connections of that memory and strengthen it and your ability to access it (stronger connections making for faster 'travel' along them).
EDIT: Also, there is only short-term and long-term memory, no medium-term.

That's the goal of SRS, to systematically review cards so that those connections are strengthened. The reason you fail some cards that have been around for ever is because the system can't read your mind and is going off models of memory decay, so those particular memories, for one reason or another, started decaying earlier than predicted.

Anyway, SRS is not just blind repetition where you're hoping something will be different, it's a method for preventing memory decay. Just to make sure it makes sense to you, the repetition described by Einstein is more akin to opening a door over and over, thinking that, eventually, the door will open to a completely different world, where everything lives on the underside of the crust of some hollow planet, so if you look up, you can see the other side of the world.
A more generic explanation would be following the steps of a process perfectly (no error) and expecting something different to happen. If this process only has one possible outcome, there is no way it could ever be different, so long as the process is carried properly.

SRS is not that kind of repetition. That's like saying, when tossing a coin repeatedly, expecting it to land on tails instead of heads is insane.

Not to mention, you're not supposed to learn things from an SRS deck anyway. You're supposed to review with SRS. Granted, I don't bother learning a word somewhere else first, so long as it sticks easily enough. However, saying that the system doesn't work and is insane, simply because you're using it incorrectly, is invalid.

Or, to take another approach, that saying of Einstein's is incompatible with processes that have multiple possible results. If you are learning new words in Anki, you are going through a process whereby you expose yourself to a new word and, after an interval of time, check to see if you've remembered it; there are two possible simple outcomes, you remember or you don't.
Granted, the memory process isn't as simple as that, but we don't know enough about it to use other large volume study systems at home without some kind of trans-cranial stimulation device (and, frankly, I'd rather use Anki than attempt to build a device to deliberately pass a current through my head, when I don't have the required understanding to do so safely).

cracky Wrote:He's messing with you guys.

No, I really wasn't. But I suppose I may have misunderstood a few things. And I'm not the only one, since there seems to even be dissent about whether SRS is a study method or not.

sholum Wrote:The explanation we were given in class (not in terms of biological function) was basically 'use it or lose it'. Memories are just paths of neural connections; once one is established, it's there, but weak; if you don't access that memory, it gets lost in the noise of the other memories, since the connections aren't that strong; but if you use it, you reinforce the connections of that memory and strengthen it and your ability to access it (stronger connections making for faster 'travel' along them).
EDIT: Also, there is only short-term and long-term memory, no medium-term.

That's the goal of SRS, to systematically review cards so that those connections are strengthened.

Anyway, SRS is not just blind repetition where you're hoping something will be different, it's a method for preventing memory decay.

Hmmm... Frankly, I *have* seen SRS described in such a manner, but I guess the source could have been unreliable...

sholum Wrote:Just to make sure it makes sense to you, the repetition described by Einstein is more akin to opening a door over and over, thinking that, eventually, the door will open to a completely different world, where everything lives on the underside of the crust of some hollow planet, so if you look up, you can see the other side of the world.

As creative as that description is, I'm sticking to my original "key" analogy: either the "key" opens the door (i.e. retrieves the correct answer from memory) or it doesn't. And that does, of course, presuppose that the correct answer had been learned initially, so before it was entered into the SRS in the first place.

sholum Wrote:Not to mention, you're not supposed to learn things from an SRS deck anyway.

And I wasn't, not initially. But after a card fails, all the info needed for restudying it is already present in the "story" field. So re-learning it from the SRS is very well possible and that's what I did every time I went through the "restudy" stack (back when I still did that).

sholum Wrote:You're supposed to review with SRS. Granted, I don't bother learning a word somewhere else first, so long as it sticks easily enough. However, saying that the system doesn't work and is insane, simply because you're using it incorrectly, is invalid.

I'm not convinced that I *was* using it incorrectly to begin with. I came up with the best mnemonic/story for a Kanji I could, learned it for the first time when entering it into the SRS and then the cardwould re-appear at the intervals this website's SRS has programmed to. So then the whole "strengthening the memory pathways" process should have begun, right? Yet what I found is that many cards will pass the first 2, 3 or even 4 times and then came the longer interval for maturing into one of the higher boxes and I would simply draw a blank. So I dutifully re-learned it, sometimes even managing to improve the story a bit, and back into the cycle the card would go. Only to fail again after another 3-4 passes. Rinse and repeat, I thought, until I had cards that had passed some 30 times and failed about 15, or ones that just passed 1 time and would fail the next time (again and again) so that the numbers of fails and passes would be more or less equal for those cards. And I kept this up until I had a total number of reviews that numbers in the tens of thousands, at which point I concluded that the whole "strengthening of the memory pathways" concept doesn't work that well in practice. For me, at least. Barring a number of cards (perhaps 60-70% of the total) for which I came up with really effective stories and that hardly ever failed at all. But the remaining 30-40% with the less-than-stellar stories, which I had hoped to master through intensely repeating them with an SRS simply don't stick in my long term memory.

sholum Wrote:Or, to take another approach, that saying of Einstein's is incompatible with processes that have multiple possible results. If you are learning new words in Anki, you are going through a process whereby you expose yourself to a new word and, after an interval of time, check to see if you've remembered it; there are two possible simple outcomes, you remember or you don't.

Yup, hence my "key" analogy: either the "key" opens the lock, or it doesn't.

sholum Wrote:Granted, the memory process isn't as simple as that, but we don't know enough about it to use other large volume study systems at home without some kind of trans-cranial stimulation device (and, frankly, I'd rather use Anki than attempt to build a device to deliberately pass a current through my head, when I don't have the required understanding to do so safely).

Well, I certainly wasn't suggesting that anyone run electric currents through their heads, either...

Anyhow, the point where our interpretations of what an SRS is supposed to do really differ, is on the definition of success itself. I consider something to have been successfully learned when it remains in long term memory for a fairly long period of time *without* further memorization efforts. If it takes 2 or 3 or even 10 or 20 times to get to that point by "strengthening the memory paths" by using an SRS, then that's what I'll gladly do. But the way I understand some of the OP define SRS, the point where something (some Kanji, to be precise) remains in long term memory for a longer period of time *without* the need for further maintenance is never really reached. Constant reviewing and re-learning remains necessary. Well, until you can really start learning the Kanji in their full context by encountering them in many texts, that is. So if that's all that SRS can do, then it doesn't really meet my defintion of really having "learned" something.

I dunno it works decently for me.

eminem2 Wrote:And I wasn't, not initially. But after a card fails, all the info needed for restudying it is already present in the "story" field. So re-learning it from the SRS is very well possible and that's what I did every time I went through the "restudy" stack (back when I still did that).

Maybe you don't visualize it the right way... 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 efficent 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.

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?

Where I've used 'you', 'one' might be a better word; I forget about using 'one' for hypothetical situations sometimes.

Honestly, I blew through RTK (maybe four months, max, including review) because I only used it to improve my ability to 'see' kanji. Most of my experience using Anki (the only SRS I've used; I never used the one on this site) has been with vocab, so I guess it assumes that kanji are no longer 'things to learn' as much as they are 'components of this thing to learn', which may be part of your problem: the base for your current study is small, so your memories are more fragile.

I don't know if this site's SRS supports it, but you could try adding an extra interval or otherwise increase the number of times you see a card over whatever period of time.

Alternatively, you could look up one or two common words that use the kanji you have difficulties with (or all of them, if you like) and review them alongside your RTK reviews. I've read about doing this somewhere, and it seemed to help the people that did it.
Of course, that'll slow your RTK progress a bit, but it might be worth it.

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

I don't share OP's experience and disappointment with SRS, but I did enjoy reading everyone's thoughtful and interesting replies.

Just FYI, that quote is often attributed to Einstein, but there's no evidence at all that he ever wrote or said it. It first appeared in print in addiction recovery literature in the early 1980's.

codex Wrote:I don't share OP's experience and disappointment with SRS, but I did enjoy reading everyone's thoughtful and interesting replies.

Just FYI, that quote is often attributed to Einstein, but there's no evidence at all that he ever wrote or said it. It first appeared in print in addiction recovery literature in the early 1980's.

Even if he had, he was a physicist, not a psychiatrist or whoever has the credentials to provide diagnostic criteria for mental illness ^^

Smart people can be wrong, especially about that which isn't their field.

But as pointed out, we have no evidence he actually said this anyways.

That all being said, with the original quote, there is only merit to the claim if the beginning scenario is always the same.

Eg. repeatedly dropping a penny into an empty glass, and expecting it to one time full it up.

If we kept dropping pennies into a cup without emptying it, then eventually it will get filled.

Eminem2 Wrote:

cracky Wrote:He's messing with you guys.

No, I really wasn't. But I suppose I may have misunderstood a few things. And I'm not the only one, since there seems to even be dissent about whether SRS is a study method or not.

Oh, well if you're really having trouble with it, I'd check out this: http://www.supermemo.com/articles/20rules.htm

Eminem2 Wrote:... this more and more has begun to feel like an insane undertaking.

I mostly agree with what you said. I no longer use SRS the way I once did.

The difference being, I never Fail anything. If I miss a card, I simply Bury it, then I review all the buried cards again after I'm done. At that point, I'll either mark them Hard or Good. That way, I still get repetition, but I don't get into that insane loop of reviewing the same failed cards over and over. This has been more effective for memorizing vocabulary.

Eminem2 Wrote:...Heisig's key meanings turn out to have been inaccurate or at the very least very outlandish.

Agreed. Actually, Heisig's book is pretty terrible, but the idea of understanding kanji through its components and using mnemonics is sound. (Neither of which did Heisig come up with, of course.) I replaced about 80% of his keywords with my own, which were more accurate, and that helped immensely.

I have never used SRS or anki and I had no trouble learning all the 2,042 kanji in RTK1.

I never really got the point of this spaced repetition thing. I learned the kanji by learning stories to fix the kanji in my memory.

I've always wanted to ask those who push anki and SRS: what did people do in the days before computers? Back in the day before computers people obviously learned thousands of kanji without SRS and anki.

As to the keywords in Heisig: they are useful, because they often represent a meaning of the kanji, and also the keywords are a way of referring to specific characters.

john555 Wrote:I've always wanted to ask those who push anki and SRS: what did people do in the days before computers? Back in the day before computers people obviously learned thousands of kanji without SRS and anki.

They probably learned less, slower, with more mistakes and forgot faster.

john555 Wrote:I have never used SRS or anki and I had no trouble learning all the 2,042 kanji in RTK1.

I never really got the point of this spaced repetition thing. I learned the kanji by learning stories to fix the kanji in my memory.

I've always wanted to ask those who push anki and SRS: what did people do in the days before computers? Back in the day before computers people obviously learned thousands of kanji without SRS and anki.

They used regular flashcards or a physical SRS. They also spent a crap ton more time studying than I do using Anki and generally either had funding or a really good reason to learn something like a language.

I've always wanted to ask people who use this argument: what do you think people did before tractors? Answer, a lot more manual labor and a few oxen and/or horses (depends where and when you are). Sure, you can have a small subsistence farm or maybe one that produces enough surplus for your taxes, provided you have enough people willing to help plant, harvest, and tend the crops (apparently, children are good for this, but I don't think that'd stand today), but oxen don't want to pull the plow straight and require constant attention and 'convincing' (read: shoving) to keep your rows somewhat straight. The speed and ease at which you can plow and harvest with modern tractor and harvesting technology allows for huge stretches of farmland to be managed by a relatively small number of people.

Point being, you can study any number of ways, but the whole point of an SRS, whether as a physical set of cards or a computer program, is to maximize the efficiency of review (granted, physical SRS decks require a lot of extra management and preparation). If people didn't benefit from using it, no one would try to tell others about it.

Of course, that doesn't mean that you can't accomplish the same things with other review methods, just that many people (most, I'd presume) would prefer the ease and efficiency of an SRS program to flashcards or other methods of review for such a large data set.
Besides, after 100 cards or so, flashcards just become too bulky; I'd much prefer a digital deck of 10,000+ cards to a physical one of the same.