yorkii Wrote:by using appropriate sentences (not just things like "あついほん") like I suggested originally, one will know which kanji they need to write.

This is certainly one way to go but perhaps the sentences area would be better suited for this type of review. Speaking personally, I've got decks in Twinkle similar to what you are discussing. I've less interest in reviewing on the site like that too.

Also, deciding on what sentence to use would also cause more confusion than simply on which word to use.

That is a lot like what JLPT/Kanken books use to drill kanji, but it does make some assumptions about existing knowledge, like the meaning of 温かいどころか, which is a much higher level of knowledge than just knowing that 熱い means a hot object. imo sentences are best avoided.

synewave's examples are slightly better since it pares things down quite a bit to requiring knowledge of what a ほん／なつ／etc is instead of a whole dialogue. While those are pretty basic words, I'm pretty sure there will be some where you would have no choice to use something more complex. I think we should avoid requiring any knowledge other than the keyword itself, and providing the information to narrow down exactly which "のぼる" it is in English. Remember that this isn't vocabulary or grammar or reading comprehension drilling. The keyword is just there to give us the idea of which kanji to draw. Having to read & understand sentence(s) would make kanji drilling take MUCH longer as well - enough so that people who would otherwise use it would avoid it. I can go through a hundred or so cards (writing them on paper instead of mentally) while I eat my breakfast. I can't do that if I had to do sentences. Using sentences would also probably require a total replacement of the kanji section and be potentially redundant with the other features Fabrice is discussing in other sections.

so:

あつい (as in an object)
あつい (as in weather)
あつい (as in a thickness)

Note that I kept "thickness" instead of changing it to "as in a book" like synewave did. That could cause confusion with the first example, because a book is also an object (and a book could conceivably be hot, but it wouldn't be the first thing you'd think of). It doesn't matter if we totally give away the meaning of the character because again, this isn't vocabulary drilling. To make things totally clear we could do:

あつい (as in an object's temperature)
あつい (as in the weather's temperature)
あつい (as in thickness)
のぼる (as in a mountain)
のぼる (as in the sun)

This is actually what I do on my (kana) J->E JLPT vocab study lists

synewave Wrote:Also, deciding on what sentence to use would also cause more confusion than simply on which word to use.

I disagree. I think backing up the word with context from a sentence would be the clearest and best way to go with this.

dilandau23 Wrote:Jarvik: I don't agree that it is not vocabulary drilling. For common words like 大きい and 登る, I could see how you would feel like that but what about something like 冴える or even something like 且つ. And this brings me to one last issue I foresee, with a straight port to Japanese keywords. Unless you change the structure of the cards to take into account inflections, is it correct to think of 小 as 小さい? If you include the inflections then it is most surely vocabulary drilling.

While there will of course be some new words, I think it's Japanese vocabulary drilling just as much as using English keywords is English vocabulary drilling (who knew what the various plants etc were?). While ideally you'd turn on the Japanese keywords on a 1 by 1 basis for words you already knew, that's not realistic I think (although you could use one of the greasemonkey scripts).

My point is on what the focus of the exercise is, which in this case is the kanji itself. Putting too much stuff like vocab drilling, sentence reading, etc bogs it down and, in my opinion, significantly lowers the effectiveness (of what you're targetting) of the exercise. Even if you do consider it a vocabulary drill, I think we can agree that we don't want to give 2+ potentially unknown words to the studier. My examples in the previous posts could surely be massaged to remove the give-away of the word's meaning without causing room for confusion, but I think it's best as I put it.

What I imagine one would do when encountering something they don't know such as 且つ is entering that keyword into whatever they are using to study vocabulary normally and study it separately. Actually, since you'd fail the character you didn't know the keyword for, that probably wouldn't even be necessary. You'd look the word up in a (hopefully non-EDICT) dictionary, and try it again the next day. In effect you learned the word as a side effect of reviewing the kanji, but it was never the point of the drill.

As for 小 vs 小さい, if the kunyomi (onyomi will always be much too general) by itself w/o okurigana is enough to tell what what character it is, and a fairly common word then by all means use it. In your example though it isn't enough to determine what character it is. こ could be 子　小　個 or a number of other things. しょう similarly has many possible kanji. ちいさい has only 小さい and is also the most common word to use the kanji.

you would clearly be studying the word from ひらがな to 漢字, e.g. さる would be what you see and you write 猿.

for words like the above - in short, nouns - we only need to see the ひらがなized version to figure out which kanji to write. For things like verbs however, there are multiple ways to write them, as it has been bought up and thus in those situations, I feel that a sentence is the best way to go about knowing which kanji to write.

If you don't know the keywords such as 且つ, I totally agree with the above post, this should be learnt separately from reviewing writing which is what we will be doing here, right?

anyway, how would you review the above without a sentence to support it??

さる could also be 去る (the best keyword choice for the kanji if we're avoiding compounds), so:

"さる (as in the animal)" to get 猿

While I agree that sentences provide the context in a nice natural way, I think the cons outweigh the pros. Namely that reviews would take much longer and it requires a lot of vocabulary, grammar, and kanji knowledge outside of what you're supposed to be testing. Imagine getting a flashcard for a kanji and the keyword has 4 other kanji you don't know in it.

Just playing the Devil's Advocate... would it not be a safe assumption that the student reviewing using Japanese keywords would be above beginner in ability? With enough grammar and vocab under their belt to understand a sentence-ish keyword cue?

i agree aircawn. a simple sentance is not likely to slow down reviewing at all. the sentance would provide clarity into what you are reviewing much quicker than staring at some ひらがな for any length of time.

Having a 訓読み review area that one could move over to once going through RTK1 sounds good.

We're still left with the problem of what prompts to use.

The sentence route that yorkii proposes, while providing completeness, takes us away from the primary focus of reviewing the writing of the characters. Remember we started a Yahoo group ages ago to make sentences for each character and look how far that's gone. Plus, Fabrice is already working on sentences...

フアブリス Wrote:I'm busy with the vocab/sentence review area. But it would be great if we can put together a common list.

Jarvik7's proposal seems to take us away from keywords if we start using keywords + English explanation.

フアブリス Wrote:the goal is to help moving away from english keywords, help distinguish nuances of the kanji by using directly an associated Japanese word.

So trying to find a Japanese "word" that will replace the English keywords is my take on what is being proposed.

aircawn Wrote:would it not be a safe assumption that the student reviewing using Japanese keywords would be above beginner in ability?

This seems reasonable. Which is why using just a 訓読み reading or 訓読み + strong collocate is appealing to me.

For example, やまのぼり, for 登 etc, seems straight forward enough. The answer presumably would be the kanji as well as, in this case のぼ・り

dilandau23 Wrote:Why change something that worked? Just think about all the wrong ways some of you went about studying kanji until you found RTK.

Jarvik: I don't agree that it is not vocabulary drilling. For common words like 大きい and 登る, I could see how you would feel like that but what about something like 冴える or even something like 且つ. And this brings me to one last issue I foresee, with a straight port to Japanese keywords. Unless you change the structure of the cards to take into account inflections, is it correct to think of 小 as 小さい? If you include the inflections then it is most surely vocabulary drilling.

This sounds a bit too much like the same kind of reasoning that prevents Japanese from making simple improvements to things that have "always been done that way" which I heard mentioned on here some time ago. While I understand your hesitation to tinker with something that worked well for you, at the same time I urge you not to blind yourself to the possibililty of of improvement. RTK is good, but nothing is optimal. Also, I don't believe anyone here advocates creating a new system which could be abused by newcomers. The best way I see to implement this would be to go with the ideas that Fabrice and Wrightak have already mentioned, and use the fact that this website records one's progress in RTK1 to prevent newcomers from even becoming distracted by something they aren't ready for. If possible, it could be made to be something that is 'unlocked' once someone records completion of RTK1. (I realize that even this could be abused, but if someone would go that far, then it is likely that any advice to that person would have been wasted anyway.

About the problem of 'inflections' and whether it changes reviewing to vocab drilling, I don't think this is the case either, if people approach the JP keywords from an RTK mindset, as they should be. We will go through the same stories and visualizations again, except this time, with the japanese words in our sentences. It theoretically has a double benefit, as it should take less time since, at least for me, a japanese translation of my existing English keyword stories would be enough to re-cognitively bind my images, and it should take less time, since many stories (which we already know work) will already exist.

Sorry guys, late night last night and I'm back at work again. I may get around to posting the list this evening but if not it'll definitely be up at the weekend.

wrightak Wrote:Sorry guys, late night last night and I'm back at work again.

Late night studying I hope

In reply to the debate about what to do with keywords with the same hiragana:

Firstly, it's quite possible to choose a unique word for each kanji, thus eliminating this problem.

Jarvik suggested two other options: providing some context in English, or providing context using a Japanese sentence. Using English seems that it defeats the objective of introducing Japanese words in the first place. I'm surprised that several people have backed this option. The motivation for choosing it seems to be the extra work load or complexity that a Japanese sentence would introduce. I have been using the Japanese sentence method to review my RTK stories for many months now and I can say that I think that the added work load is negligible. Reading a sentence is easy, you?re not being asked to produce it.

There is another option: you could include a picture. For instance with 菊 and 聞く, in addition to the hiragana prompt, you could have a picture of the flower or someone with their hand to their ear. Someone suggested replacing all of the keywords with just pictures and I think it's a very interesting proposition. Personally, I think the Japanese sentence approach is the way to go and I'll explain why in another post.

synewave Wrote:Late night studying I hope

The JLPT's coming up and it should have been, my friend. *sighs*

synewave Wrote:Jarvik7's proposal seems to take us away from keywords if we start using keywords + English explanation.

I never did espouse English explanations, just enough information to determine which word it is for the homonyms. Give-aways or hints at the kanji's meaning aren't even a bad thing since as was mentioned earlier (且つ) there will undoubtedly be words new to the studier. A reminder can turn a total blank into a successfully drawn kanji, without spoiling task of remembering the kanji. It's not something I'd say to strive for though.

The impact adding sentences would have varies greatly on the sentences you chose. Your example of "やまのぼる? would probably impact speed only slightly if at all. However I think yorkii's snippet of dialogue example is really overdoing it and takes the focus away from the fact that you are flipping through kanji drilling cards, possibly a hundred of them.

Keep in mind that the students who would be using this feature would ALSO be going through their vocab, grammar, sentence, and whatever else they do in Anki or whatever. The kanji review would be knowledge maintenance with the only new cardsif any probably being RTK3. I think we should try as much as possible to have 1 word keywords, with a bit of clarifying information on the ones required to avoid confusion. For kanji with no homonym problems there should be the word only. Heisig didn't use English sentences to give us a feel for how the kanji is used. Also keep in mind that some of the kanji may have meanings but are not used in any words (人名用漢字 among others). That would be a problem for sentence making. As RevTK expands past RTK (which there are other discussion threads about and the people who use the JP keyword feature would be likely to do) these kanji increase greatly with number.

An example (one of my favorite kanji composition wise) is 嫐. If you search EDICT you'll see that no words use it, however in KANJIDIC there is a meaning identical to 嬲 along with the exact same readings. The KANJIDIC entry is actually incorrect in that ジョウ and なぶ.る are not valid readings of the character (at least according to the much more reliable 漢字源) and the meaning is actually different (if related). The only use is in a historical person's name (うわなり, which is not even one of the listed readings for the character). The only way to make a sentence out of this would require the studier to know something about 17th century Kabuki actors.

While you couldn't just put up a representative word for the kanji either, you could give one of the one word definitions in Japanese, provided it doesn't conflict with another word somewhere else. There could be some way to signify that the word is not written with that kanji and that it's a definition. Heisig did similar things with the English keywords to fight overlap I believe - selecting an English word that the kanji is never actually used in in Japanese but does represent the meaning.

I realize that the character I used for my example is not likely going to be one people will try to memorize due to its uselessness and obscurity, but it does serve to illustrate my point.

The number of characters with homophonic readings that can't be worked around is probably few enough that putting everything into sentences is really overkill. I still don't see what great pros they bring to the act of kanji drilling that overcome their various other cons. I get the feeling that yorkii wants to bring in sentences mainly for AJATT-like goals. Heisig is a proponent of divide and conquer and adding in too much stuff really bring us away from what our focus is trying to be. It fundamentally changes what the exercise is. Single word keywords with something in parentheses where necessary is much more true to the original I think. Hell, Heisig should have put some stuff in parentheses rather than selecting more obscure meanings or causing confusion with stuff like spring (verb) vs spring (season) vs spring (water). I know I've missed the watery kind of spring several times thinking it was the season.

In synewave style sentences they are small and innocuous enough that they would take about the same amount of time as my English clarifications would, and thus wouldn't be entirely opposed to using them where required to clarify between homonyms, but I still wouldn't want to use them for EVERY kanji. If every kanji had a long 1+ sentence "keyword" as with yorkii's proposed method, I'm pretty sure I wouldn't use the feature at all. I like to get through my kanji review quickly and move onto new material. Flashcards by definition are supposed to be done quickly (flash).

Jarvik7 Wrote:In synewave style sentences they are small and innocuous enough that they would take about the same amount of time as my English clarifications would, and thus wouldn't be entirely opposed to using them where required to clarify between homonyms, but I still wouldn't want to use them for EVERY kanji.

I agree that one word keywords is the ideal. However as we both know, sometimes clarification will be required. As the aim though is to switch to Japanese keywords, using a word pair in Japanese seems more appropriate.

Jarvik7 Wrote:Heisig is a proponent of divide and conquer and adding in too much stuff really bring us away from what our focus is trying to be.

I agree with this. I'm sure we're all working on phrases and sentences elsewhere.

wrightak Wrote:There is another option: you could include a picture. For instance with 菊 and 聞く, in addition to the hiragana prompt, you could have a picture of the flower or someone with their hand to their ear. Someone suggested replacing all of the keywords with just pictures and I think it's a very interesting proposition. Personally, I think the Japanese sentence approach is the way to go and I'll explain why in another post.

pictures are more abstract than a sentence and I can see it being more problematic to find 2000 odd pictures rather than 2000 short (if needed) sentences to support a keyword..

Quote:I get the feeling that yorkii wants to bring in sentences mainly for AJATT-like goals.

you couldn't be more wrong. I'm not even talking about sentences really, just sufficient context for one to know which kanji you need to write when there is ambiguity. that's all.

how on earth is looking at a tiny snippet of Japanese, rather than an English description slowing down the process?

granted, I was adding extraneous data on some of the examples that I included. but I think you get the idea. it is in principal the same as synewaves, except he wasn't messing around like me when he wrote his.

I agree with yorkii's two posts above.

yorkii Wrote:I'm not even talking about sentences really, just sufficient context for one to know which kanji you need to write when there is ambiguity. that's all.

So if there is no ambiguity, e.g. こころよい, as far as I am aware, am I right in thinking that you would be happy enough with this single word as the Japanese keyword to prompt you to write 快（い）？

I don't fully understand the "it will become like vocabulary/sentence drilling" and "adding too much stuff makes it confusing" problems.
I gather this is meant to be an addition for the intermediate/advanced learners, not everybody. This means that anybody is free to switch to Japanese whenever she feels ready. If the Japanese keywords/sentences are still too hard and confusing for her, she should be able to revert back to English.

Then someone may want to avoid drilling vocabulary with this method and just focus on the kanji, while someone else may like the idea of learning two things at a time. The latter will therefore switch to Japanese sooner, while the former will wait until he has assimilated enough vocabulary by other means. Am I being correct?

nac_est Wrote:someone may want to avoid drilling vocabulary with this method and just focus on the kanji, while someone else may like the idea of learning two things at a time.

My understanding of what we are discussing is slightly different. It's not meant to be a drilling of vocabulary type thing. More like reviewing as is just now. But substituting the current keywords for Japanese keywords represented by one of the kunyomi readings of the characters (if possible).

Ideally you would want to know the Japanese word prior to starting to use it for reviewing purposes. Otherwise it would as you suggest be like trying to review kanji and learn vocabulary at the same time.

[ファブリス]

Jarvik7 Wrote:Many of the Heisig keywords do not accurately represent the common meaning of the characters, and to fulfill the requirement might mean bringing in more compound words than we'd like. Let's stick with the most common (preferably kunyomi) non-compound word.

I'm concerned about RtK1 learners who have completed the book with english keywords, this is the "default" scenario. That's why I think people transiting to reviews using kunyomi compounds, will translate the word in their mind for some time, and then look for the written character from that.

I think doing this as a kun-yomi review mode, separate from the RtK1 review, will work better. Even as a separate mode I think all the compounds should be Japanese readings, because there will be oither tests for On yomi, and it might create confusion. Typically compounds with Japanese readings are the ones with the kanji at the root and some hiragana endings.

For the example of "noboru" I like your idea #2.

Since this mode would supplant the RtK1 review mode, and allow review of the written characters, I would still avoid information overload and keep sentences, even short ones, for the sentence reviews. That's why I think your solution #2 is great.

I understand what you mean, Synewave. I only have a slightly different point of view, perhaps.

The reason for wanting to switch to Japanese keywords, in my case, is simply to extend my exposure to the language, and to stop relying on English. I'm doing this with everything I do, as much as I can. So, in a sense, Japanese keywords are something I would use to learn a little bit more. That's why, while not being deliberate drilling, it would still be useful to improve my vocabulary.
I don't see this switch as a great benefit for someone who already knows Japanese perfectly.

Then maybe I missed the whole point of the concept. Sorry if that's the case, it must mean that even my English is not good enough yet!

edit:
anyway, apart from the discussion of the whole point of having Japanese keywords, if you give people the freedom to switch back and forth whenever they want, then they'll do what they prefer and feel is the better way in their case, as I was saying in the previous post.

ファブリス Wrote:For the example of "noboru" I like your idea #2.

Since this mode would supplant the RtK1 review mode, and allow review of the written characters, I would still avoid information overload and keep sentences, even short ones, for the sentence reviews. That's why I think your solution #2 is great.

If I'm understanding correctly, the prompt would be:

のぼる (mountain)

and the appropriate answer would be

登る

I'm sorry, but I think that's crazy. People have finally gone to the effort of getting rid of their English keywords and they're tagging some more on the end!!! It's like taking your inflatable arm bands off to swim and then tying them round your legs instead.

Maybe people whose first language isn't English would think differently but if I saw

のぼる (mountain)

in big letters on the screen, the word mountain would jump out at me and lodge itself in my brain. I couldn't help registering it first before reading のぼる. That comes from a lifetime of living and breathing English words so much that my brain recognises them instantly. When you're on the street in Japan and you look at all of the kanji on the signs and billboards, if there are letters or English words somewhere, however small, do your eyes not zoom in on them unconsciously?

I cannot understand why you think 山をのぼる would be so much more of an information overload than のぼる.

EDIT: The whole point of changing to Japanese keywords must be to associate kanji with a Japanese word rather than an English one. I believe that tagging something like "mountain" on the end would work against this.