Yes, I know, Heisig points out you need to review Keyword to Kanji only. In fact, I think I chimed into the "Anki?" thread which discussed this during the subject of suppressing the reverse review part of the program. Still, I sort of realized walking this morning I do sometimes not know what a particular Kanji is though the keyword springs it to mind. I then realized Heisig was talking about paper flashcards only (it was in the 70's and early 80's after all).

So: Considering we use Leitner as our review method, would there be harm and would it actually be beneficial to do a Kanji to Keyword review? Yes, I don't think RevTK would incorporate it but those that use anki or other programs can do it with relative ease. If the Kanji to Keyword is easy, it soon gets scheduled far more distantly than the more difficult Keyword to Kanji card (and vice versa in some cases).

In summary the question is: if you used traditional Flash Cards, just do Keyword to Kanji as Heisig suggests. If you use a computerized Leitner system, it's safe to do both sides so long as they're treated as seperate cards (I mention this since Kanji Gold program connects the two).

This concerned me too when I looked at familiar kanji and couldn't recall the keyword. I also considered reviewing the other way around, but I thought about it a bit more and came to the opinion that it was a bad idea.

In a way the keywords in RTK are markers. When you have finished the book you really want the markers to fall away as quickly as possible and be replaced with Japanese. If you review from Kanji to marker then you will make this process all the more difficult because you are training your mind to recall an English (or whatever) keyword when the kanji is encountered. I think that it could therefore be a handicap to future development.

The exception, I guess, would be reviewing from Japanese keyword where you really want to enforce that link back - but I've not finished the book yet, so that is currently ahead of me at the moment.

I don't really see a danger in reviewing from kanji to keyword, aside from not getting as much practice writing from memory. It seems I'm not the only one who has had trouble remembering what a kanji means when seen in context, even though it's easy doing keyword to kanji. Though I don't review the kanji anymore(just words and sentences), if I did, I would go kanji to keyword. I'd say the English fades away over time like the stories do.

I don't see a point in doing kanji to keyword practice. We get that when we read, and I've never had a problem not recognizing a character that I've previously studied, even if it's one I'd potentially get wrong going keyword->kanji.

If YOU do have problems recalling the meaning of a kanji when you see it, maybe it would be helpful to do some. I wouldn't recommend doing it just because, since you are putting the kanji back into your short term memory when you get it as a question, so it spoils when you're going from keyword to kanji in your normal studies.

Jarvik7 Wrote:If YOU do have problems recalling the meaning of a kanji when you see it, maybe it would be helpful to do some. I wouldn't recommend doing it just because, since you are putting the kanji back into your short term memory when you get it as a question, so it spoils when you're going from keyword to kanji in your normal studies.

I'm finally doing the Kanji to Keyword reviews. I used the plug-in that allowed my RevTK scheduling to be inputted into the Anki Heisig list.

Now, I think that since we are using the Leitner system, it protects against the warning Heisig gave about reviewing Kanji to Keyword. The concern you gave about ruining the learning process would go away as the cards gets spaced differently and the review times get further and further out. We would be doing both directions. I think there's legitimate reason to incorporate the option to do reverse review in RevTK (but not solely reverse review). Of course, there has to be some sort of spacing between seeing a normal review and reverse review of a card. Anki does 24 hours, but 48 hours sounds more reasonable.

[ファブリス]

You'll get plenty of Kanji to keyword review in Trinity. There's no need for explicit kanji to keyword reviews, and there's far too many flashcard reviews already in RtK1, imho. Plus, Trinity or not, the principle is the same : you just get to review from kanji to keyword by actually using them. You see a new compound, likely you'll want to have a guess at what it means, so you rack your brain for the keywords. You may want to bring back the mental image that you have for the character and do a small adaptation to make the kanji>keyword recall work better next time, but that's about it.

Perhaps what Heisig meant when he said the kanji to keyword recall "takes care of itself", is that the information is there, the keyword to kanji reviews took care of that. You just need to create those paths back from kanji to keywords, this is not the same as learning, and I don't think it warrants doing so much extra reviews. Like I said, in Trinity, every time you lookup compounds by kanji/reading and especially when you work on kanji chains, you have to recall the keywords first. When you review sentences, you can simply hover on a character to see its keyword.

PS: I'll add to that that doing the recall when actually using the characters, that is, when reviewing compounds and learning new ones, is more effective because you can then realise what the meaning does combined with other characters. You'll need to adapt the meanings which came solely from the keyword, to how they actually combine in kanji compounds. Some Heisig keywords are spot on and require no work, others you find that the meaning you thought of does not fit.

Basically what I'm saying is you really only know the meaning once you start seeing the keywords combining into many kanji compounds. So practicing recall of keyword/meaning while seeing the kanji in compounds seems far more effective to me.