Hello all,

This is my first post on these forums, although I have been lurking them for a bit and using this site for a few weeks (boy it's better than reviewing cards the manual way!). I'm only up to lesson 17, but I am beginning to wonder if I am going about this the wrong way. I guess I'll start by telling you what I do now when I study:

1) Look at the keyword and then the Kanji, then read the accompanying story.

2) Next I write the kanji while repeating the story.

3) Finally I close my eyes and try to form a image for the kanji using the story.

Now that is kind of what Heisig says to do in his book, yet I feel like I may be doing something wrong. Every time I study a lesson, I usually do it at night. Then I add the cards onto this site, and go to bed or do something else. The next day (usually afternoon or evening) I review the ones I did the day before (my "untested" pile). However, it seems when I do this, I end up getting at least 60% or more of the wrong, thus moving them to my "failed" pile and making me study them again. Sometimes it takes failing them 2 or three times before they get past the first review stack.

Does anyone know what this problem could be attributed to? Or is that normal? Does it just take some time for the kanji to stick and then be able to stay in my memory?

If I am doing something wrong, I'd rather stop it now rather than go through the whole book constantly having a failed pile bigger than I have time to restudy.

How are you failing them? Do you have trouble remembering the story? Are you writing them incorrectly? Primitive placement is incorrect? Do you draw complete blanks or do you tend to get it half-right?

Well, I guess it depends on the kanji, but for the most part, when I review the ones that I have just added the day before, I tend to draw complete blanks. Then I try to remember the primitives that I recently learned, and sometimes I can remember the general idea behind the kanji. However, I feel that is kind of cheating, and if I had put those new cards in a large tin with all the kanji I had learned and picked them out, I would probably not think to try to remember the primitives I just learned.

The ones where the primitive placement is incorrect isn't a big deal, because with those I usually remember the story. So I just alter it slightly to remember the primitive placement.

So I guess, in short, I am just completely drawing blanks.

It's my normal mode. I just let the system handle the timetable to review even those kanji I don't remember at all. Thus I don't meditate on kanji while learning it, but place it in second stack right away and actually learn it during subsequent reviews. I never 'restudy' them as such, just make sure I remember the story for each failed kanji and then immediately place them back in second stack. I can't learn anything by rote repeating while knowledge is in my short-time memory. I have to force remembering in order to learn, and thus delayed review question is a way of studying.

So what you do is just fly through the kanji making up (or reading, depending where you are) kanji and then relying on the reviews to implant them in your memory? Well, I would say that sounds inefficient but then again, I'm the one who constantly forgets the ones he's just learned, so who am I to judge?

But did you also say that if you fail it you just place it in stack two anyway, just to move the review days up and put it in long term? It seems to me that if it isn't in your short term, it wouldn't be able to make it to the long term.

Cool. I'm not a memory specialist so take my advice with a grain of salt but if you're drawing blanks, then your stories aren't sticking around for whatever reason.

For starters, I'd swap step two and three around. After reading the story, go straight to your imaginitive memory and let the brain do its thing, keeping in mind any extra info in regards to primitive placement that Heisig notes. Then once you're happy with it, write it down a couple of times. The story is the most important thing because it's the crutch you can always rely on. Writing the kanji first without the imaginative story behind it I don't think would work that well. Certainly, it's not how I do it.

Edit: Pauline explained this better than I did. You're recreating the kanji from your imagination during review, so do the same thing when learning them.

Secondly, maybe Heisig's stories aren't that useful for you? It's possible that you may get more benefit from simply taking his primitives and generating your own images. You're up to lesson 17 so you've already got a good idea on how it works and you arent far off from having to do that anyway. Maybe if you're having particular trouble with a kanji, ditch Heisig's story and conjure up one of your own.

Thirdly, I'm not sure how many cards you're adding each day, but maybe you could try reducing how many you study a night until you find a comfortable rhythm with it? Even something as low as five a night. If you can nail five down well then you save time in not failing them constantly.

And lastly, (and this may be a bit contentious) stop adding more cards until you're happy with your retention of the existing ones.

Just some thoughts...

My first thought is that the order of the steps you go through when learning a kanji is odd. Should not visualizing the story/image be done before you try to recreate the kanji from the story?

It might help to reread chapter 11 with the steps you should follow with every kanji. I don't follow them to the letter, but still uses the same general order and that works well for me.

Well, thanks for the help. I know the way I do it is kind of odd, but that's because it's how I started doing it when I started the book, and it worked alright then. However, I will try to switch up the steps and maybe that will help.

Also, I will be taking a break from adding new cards, as at the beginning of Lesson 17 Heisig recommended making sure that the previous lesson's primatives (arrow, fiesta, parade, march, etc) should be firm in your memory, which they aren't. So until they are, I will hold off on Lesson 17.

One last question, then. I remember switching up the order of the steps I do occasionally (basically its whatever happens, if I decide to imagine the story before I write it, I do, or vice versa), and I'm not sure that is the whole issue. So if that doesn't work, what would you recommend?

P.S. - I already alter Heisig's stories a significant amount.

When learning a new kanji I often first look at the keyword and see what image already comes to mind, before even looking at the elements or the character. Then I work the elements into that image. For example "recitation" made me think of a recurring sketch from an old TV show, Laugh-In. So I picture that particular guy on that particular set, spewing out lots of words for what seems like an eternity.

Maybe you could try something like that.

It sounds like you are doing it in large batches without reviewing soon enough. It never works for me, I had the same difficulties in cases I did it this way.
Do NOT separate the study and first few reviews. Take a small batch of kanji (5-10, whatever you are comfortable with), do what you are doing for the story, and review immediately, making sure you answer all questions instantly. After you are done for the day, repeat the same until you can answer all kanji in that day's batch without hesitation. If it takes over 3-4 seconds, it is a fail.
It is absolutely crucial, if you hesitate right after learning it you will not be able to recall it next day. Next day review again, and you'll see that you will get ~90% right, and those that you won't get will stick real fast after that too. At least that is the way it worked for me. Neglecting a thorough review on the first day would just make most of the work go to waste for me.

Ryuujin27 Wrote:Well, I would say that sounds inefficient but then again, I'm the one who constantly forgets the ones he's just learned, so who am I to judge?

Point is the only alternative is spending more time on 'study' stage and it doesn't work well for me anyway. Efficiency here is only in selecting most efficient course of action from available ones. On the other hand, 50% fail rate only slows down second-pile-reviews 2 times, so no big deal. Also note that delay imposed by extra review doesn't affect the speed of studying, it just, well, delays.

Ryuujin27 Wrote:But did you also say that if you fail it you just place it in stack two anyway, just to move the review days up and put it in long term? It seems to me that if it isn't in your short term, it wouldn't be able to make it to the long term.

It must be stimulated by trying to remember at the point it's not in short-term memory, as obviously some part of the path to remembering is there and it's the way to strengthen it. Btw, I fully agree with laxxy's point of first review right after studying, I just don't require that review to fully succeed.

What you describe pretty much applies to me. I'm usually closer to the 60-70% than 90%. Often I draw complete blanks on the first revision. Now that I'm in the 600s I feel like I've got into the swing of it and I've stopped worrying too much, I just keep failing them until the image has stuck.

I think this works for me because I always clear my failed stack the next day at the same time I add new ones - after 24 hours running through the failed stories in my head. So badly learned ones will get pushed through the cycle repeatedly in quick succession until they "stick".

I think the "complete blanks" are mostly to do with your state of mind - both when originally learning and when reviewing. Any lack of concentration or fatigue, and effectiveness drops sharply. For example yesterday I had 77 to review. I did about 20 at lunchtime (after a busy morning at work) and when I realised my success rate was below 50% I stopped. I did the remaining 50+ at about 10pm and I got more like 90%.

I guess I have accepted that you can't be in complete control of your learning and reviewing - outside pressures always creep in. I can do as much as possible to optimise it but you're always going to get bad days and unexpected circumstances. And the beauty of Leitner is that it compensates for this automatically.

Ryuujin27 Wrote:Hello all,
Does anyone know what this problem could be attributed to? Or is that normal? Does it just take some time for the kanji to stick and then be able to stay in my memory?

It sounds like you're only reviewing once per day. I recommend reviewing all your new kanji twice per day, every day for 3 or 4 days, before you start into a program where you only review some of the kanji once per day. Also, take a day off before you start such a program, so that you don't give yourself an artificially high grade.

If you want to avoid dissappointment completely, learn them in the morning and review them for the first time in the evening. When you review them the following morning you should be doing far better than 50%.

Personally, I feel there is nothing wrong with failing a new kanji 3 or 4 times. I've gone through RTK1 doing just that. It all depends on whether you prefer quick and dirty with large batches (and fail lots), or do small batches slowly and deliberately, and fail almost none. Failing is not a problem if you can accept it as a normal part of your personal learning proces. For me, it works better to go quick and fail a lot. Others would choose the other path, whatever works for you.

Ahh, thank you all for the many replies during the night.

I have received a lot of good suggestions and affirmations that have helped alleviate my fears that I am doing it wrong. I guess if I do fail them, I will just have to accept that and restudy them from the failed pile, perhaps drawing on the story that someone has posted online or completely inventing my own.

Also, I think I may try the suggestion to review right after I finished. I did that once and when I reviewed the next morning I only got one wrong (out of about 30, I think). So perhaps that is the way to go about it. I only stopped doing that because I thought that was cheating my long term memory.

My main fear is to finish RTK1 and then when I'm done, find that most of the kanji didn't "stick." Maybe I'll be able to remember them for a little bit, but in the long run most of them will leave me, even if I continue the review sessions on this site following the Leitner system (excuse me if I misspelled his name). However, I hope this is not the case as I am currently a 日本語 major, so kanji is more or less reinforced on a daily basis.

Anyway, I've managed to turn what was meant to be a short reply into a rather lengthly one. So, thank you once again for showing me just what a helpful community is here on this site, it truly gives me great hope and makes me thank God that I found this site.

じゃあ、ありがとございます。

ー龍神

I think it comes down to two factors. 1. Whether your story is vivid enough and whether it works well in letting you remember the kanji. 2. Repetition spacing.

I'd say number 2 is your problem as laxxy pointed out. Make sure you test yourself immediately after learning and don't go to bed until you've got them all right. i.e. do a standard flashcard approach where you put cards away when you get them right and you see them again when you get them wrong.

I can recommend supermemo for extremely efficient and effective spacing schedules adapted to your particular memory characteristics.

Well, I started to test my self immediately after studying. However, how I started doing it (I am on lesson 18 now) is reviewing the set of kanji that are presented after learning a certain primative. I'm not sure if this is effective or not yet, but it kind of feels as if it cheating a little because I know they will all relate to one primative. However, I able to perfectly relate the story and call to mind the image I associated with it.

So, should I keep doing this or should I just study all the kanji I am going to that day (20-25/30-40 on a real good night with a lot of interesting kanji) and then review or will doing the way I am doing it now be sufficient?

I figured I might as well ask, since these forums proved so very useful when I first posed this question.

I always try and study between 10 and 25 kanji when I find time to sit down. I think about stories for each kanji, but I try not to get bogged down and spend 10 minutes on each. Once I have a picture in my head I check with the book to see if I'm still thinking along the right lines (occasionally forget a primative) and then write the kanji down.
I always write them down the left side of some paper. Then when I've finished learning I go off and do something else for about an hour, then come back and try and write the keyword next to each kanji. Then I leave them alone again for a while before coming back and this time covering up the kanji and trying to write them next to the keyword. This is where I strengthen stories that need it.
Finally I add them to this site and will review them first during my lunchtime review.
I don't worry too much if I can't fully remember them before adding to the site, just check, review and revise the story.

Hmm, see what I was doing is learning some kanji (say approximately 20, on average) and then adding them on the site. Then, as the people in this thread suggested, I would immediately review them. When I did this, I don't think I ever missed one. However, the problem came when I reviewed them once they expired in stack one. That was when I got 70% of them wrong.

Do you think your method will clear this up? I'll give it a try, anyway. Thanks for the suggestion.

Lot's of people have this issue when they try to go immediately to spaced repetition flashcards. Some ignore it, have faith in the program, and let it work itself out over time. Others review all their new cards every day for several days, until they get consistently good results, then they let the program take over. I review my new stuff twice per day every day for 3 or 4 days because this is my time tested method that always works for me.

I believe that cards put into the second stack are scheduled for review three days later. It's a big jump to go from reviewing straight after learning to three days. That's why I add them to the site, but only review them the next day. That way you get a 1 day review and then the 3 day review.

Ahh, yeah that makes some sense. But that's what I use to do, but then I couldn't remember them the next day.

Perhaps, if I were to review using my actual cards (I brought the box of cards from Heisig) right after, then review them on the site the next day, it will probably work.

Thanks for the advice.

Several months ago I experimented with vocabulary. I tried the following methods:
1. learn the words, do my first review the next day & note my retention
2. learn the words, immediately review them thoroughly, review the next day & note my retention
3. learn the words, review them several hours later the same day, review the next day & note my retention

By "learn the words" I mean study a list until I'm able to go through it, up and down, with no failures. I use mnemonics heavily. By "review heavily" I mean go through the list 3 or 4 times in addition to what it takes to learn the words.

1. & 2. gave the exact same results. The next day I had 60-80% retention. (note - these are words; my RTK percentages were much lower when I did them, maybe 30-50%, but I didn't do this experiment with them)
3. The next day my retention was 90%+ every time.

That big dip in retention is probably nothing to worry about in the long run, but I hate it. So I cover material twice, seperated by several hours, the first day, and there's no dip. I continue to study twice a day after that, because I feel the previous results indicate that it's much more efficient. I prefer to do 2 short reviews rather that 1 long one. After 3 or 4 days, I throw the cards into the program, and stop worrying about them.

leosmith Wrote:1. & 2. gave the exact same results. The next day I had 60-80% retention. (note - these are words; my RTK percentages were much lower when I did them, maybe 30-50%, but I didn't do this experiment with them)
3. The next day my retention was 90%+ every time.

Heh, interesting. It feels the same way for me too, but in my case the RTK kanji were usually much easier to remember and had higher retention rates compared to words -- even when words are compounds of known kanji, even when 'known' includes readings.
perhaps i should try using mnemonics more extensively, I know they do help when I do it.

Yeah, my memory just sucks in gereral. I always got the impression others were retaining RTK stuff better than me. It took me 300 hours to finish the book, which is probably longer than normal. As far as vocabulary goes, mnemonics really help me. I like this one the best:
http://www.linkwordlanguages.com/japanese.htm
(I use this method on my own - I don't buy these products)
During that same testing period, I discovered that I can learn 20 words per day, by breaking it into 2 sessions of 10 words, and of course reviewing with supermemo, etc., etc. That's 20 words, pronunciation and reading, sustained for as long as I want. A very pleasent surprise for me, since I want to add 6000 vocab words beginning in May.