I recently finished RTK1 and now my cards are slowly starting to funnel into the last stack. So of course, I'm wondering what I should do next.

I started making a RTK2 deck for anki. The beginnings of the deck are here: (edit: dead link)

and the spreadsheet I'm importing from is here: (edit: dead link)

(it doesn't open with excel easily...but I found the Sun Office spreadsheet program from google pack handles it well).

I'm including fields for:

RTK2 Frame #
Kanji
Onyomi
Example Compound
Example Compound Reading
Example Compound Meaning
RTK1 Frame #
Example Sentence
Notes

I was thinking that I could follow along to the order of RTK2 and learn the readings and sample compounds from that book. But also add example sentences. So that it is sort of a combo of RTK2 + AJATT.

So questions.

Does this seem like a good way to learn readings + useful japanese?

Has anyone made a deck / spreadsheet similar to this before?
(It's time-consuming making it myself, so if someone has already done this I'd rather not reinvent the wheel)

And if the answers to the preceding questions are "yes" and "no" respectively, does anyone wanna help make the deck? ^__^

Any suggestions, comments, etc are welcome

Hm, I've discovered it works better for me when the cards are very simple. Then they are quick to make(maybe I'll make 50 before I feel like stopping instead of, say, 20), and then I won't sit there trying the remember 10 different things for each card while just thinking, "this is boring now." But then I don't know what works best for you.

For RT2, I put the example compound on one side(in mnemosyne, I can't get anki to work nicely), and the pronouncation and meaning on the other. That way I get the on-yomi in context, and I learn an new word at the same time.

johnzep: Yeah, I think you should try and keep to the "minimum information" principle - see http://supermemo.com/articles/20rules.htm

yukamina: what problems did you have with Anki?

Basically it's just the information from the RTK2 book itself, which includes everything listed above, except i've added an example sentence.

As I'm entering the cards now, most of them have the example sentence blank. But I thought it would be nice to have the field there so I could add sentences for vocab that needs more context.

Yeah, it's not too much information: most of the stuff there is for reference, not memorization (who honestly cares what number it is in RTK2?).

My two cents: Memorizing On yomi and Kun yomi intentionally is like memorizing the cities in the state of Texas and New York. It's possible, but tedious and little reward.

Now, imagine you have a product tour that takes you through these two states. Eventually, you'll learn not only city names, but locations within the states and a little relation to each other while selling your product at the same time.

Very bad analogy to say learn pronunciations via context sentences. If you know that 新しい ( あたらしい) and 新聞 (しんぶん) both have the Kanji for New, you kinda know without studying that the On is しんand the Kun is あたら.

Try this out: Create sentences using the Furigana dictionary or a good online J-dictionary with common compounds for the kanji you'll study. Not sure of the order to go in, but usage or JLPT may be a way to go.

The only critical part of the flashcards are the compounds: these are the only part that is worth learning, as you said.

For example, within the first two pages, two compounds listed are 新美 and 加入. By learning these (にいみ and かにゅう, respectively), you can figure out that 美加 is みか the first time you encounter it. This, admittedly, is very much the point of RTK2. It can't be to arbitrarily memorize readings because that's almost as bad as just writing kanji over and over to memorize them.

I'm a bit confused about the reviewing instructions in RTK2. Does reviewing always from compound to reading mean, that I'm reviewing e.g.:

日本語 --> にほんご　--> Japanese
or rather　
Japanese --> 日本語　-->　にほんご

I'd prefer to review like in the second example, since I'd have to remember the Kanji-compound and the reading. The Anki-stack is not too big at the moment, so it's currently no problem to do both review-versions, but it's steadily growing and I'd like to safe some time on the daily reviews.

Your example seems to imply that you're doing a 2-step review, if indeed each arrow represents a single step. Something I didn't know Anki was capable of.

RtK2's recommended direction is kanji + compound to reading (+ translation). The translation part is optional, it's not a vocabulary exercise after all. The main thing you should score yourself on is on whether or not you remembered (or deducted, if you spotted the signal primitive) the reading for the target kanji. That's why you should include that kanji on the question side too, or mark it in some obvious way in the compound.

For your example: 語 + 日本語 --> にほんご (+ Japanese)

Ah, yes you're right, I don't do a 2-step review.
にほんご (Japanese) would be on the same card-side when 日本語 is the question, and when "Japanese" is the question 日本語　（にほんご） would be on the answer-side.

Somehow I missed the part that stated you should include the signal-primitive/kanji. I've only used 日本語　and not 語　+ 日本語 which seems to be a better way to enforce the signal-primitive/kanji recognition.
That the translation part is optional I also wasn't aware of, I thought that you're supposed to learn both, reading and new vocabulary, with an empasis on the reading bit. I've been not as strikt with the learning the translation as with learning the reading.

Thanks for the answer woelpad, it was very helpful.

I'm a bit confused. Memorizing the compounds which are inside of RTK2 is a less effective way to continue learning after RTK1 than just creating sentences and inferring compound meanings from that? RTK2 is better as just a reference?

The question seems to be a bit off-topic here, but as this thread was dying out, I'll answer it anyway.

You need to ask yourself: What is my next goal? I think a good part of the people that go for sentence mining do so in order to gain vocabulary. They choose a text and they aim to read it from start to finish. They prefer furigana texts and pop-up dictionaries. It's the reactive way of gaining kanji knowledge: Treat them as and when they come.

That's fine at first, but at a certain point in your study, you'll get tired of looking up readings for every new compound you come across (much as you did with kanji as a whole, that's the main reason why even advanced people start RtK1). That's when you will most likely make the switch to a proactive method that so many schools and textbooks try to teach you: Learn the common readings for all the common kanji before encountering them.

You can adapt the sentence method to suit this approach, by actively seeking readings for every kanji you consider viable, but you still need a mechanism on how to decide which reading is good to know, which not. Either RtK2 or Trinity could help you there (as could traditional text books). If you use RtK2 in this sense, then indeed RtK2 becomes a reference. But why not try out RtK2 in the first place? In a sense, a compound provides a mini-context for a kanji, so the two approaches are not diametrical opposites. And the work of selecting compounds has been done for you.

Having said that, Heisig is pretty much a proponent of learning words in isolation, as his troglodyte example shows. If you don't know what the word means, then a random sentence using that word will not help you much further, goes his reasoning. Debatable but not ungrounded.

For me, it's very much a question of efficiency. RtK2 looks more efficient than sentence mining with the time that I like to spend on it, therefore I will try RtK2. If a "Kanji in Context" approach looks more efficient to you, then go for that. Perhaps someone who did both can tell what worked better, but it might well be that the combination of the two is superior.

I also keep in mind Heisig's advice to replace any of his compounds with others that you're more familiar with. From that it's easy to see that the goal of RtK2 is not to expand your vocabulary. And, to come back to Biene's question, translations are only as needed as the words are unfamiliar to you. It's not something that Heisig says, it's my interpretation. Leave familiar words untranslated and focus on your primary task, that of memorizing the readings.

woelpad Wrote:That's fine at first, but at a certain point in your study, you'll get tired of looking up readings for every new compound you come across (much as you did with kanji as a whole, that's the main reason why even advanced people start RtK1). That's when you will most likely make the switch to a proactive method that so many schools and textbooks try to teach you: Learn the common readings for all the common kanji before encountering them.

Sounds like this is what the kanji odyssey group is trying to do, ne?

@woelpad

Actually, I'll agree with you on most points about sentence mining, with the exception of using books like KO and KiC, because the sentences are designed to use vocab that teaches you both on- and kun- reading. RTK2 teaches ONLY on-, and just gives you some tips on how to deal with kun- *on your own*. That was the one part of RTK2 that really turned me off. Kun- readings are equally important, because a lot of verbs are kun-readings. So yeah, if you don't need verbs, then RTK2 is fine. But you're only learning half of the language.

What KO and KiC do (I think KO does this a little better than KiC), is grab some common vocab you'll actually use, put them in a sentence full of words that aren't too hard to learn, and give you an idea of how not to sound like an idiot. And they cover both on- and kun- readings at the same time.

I used to be okay with studying words in isolation, but now not so much. Take "beverage" for example. Do you invite someone over for drinks, or beverages? Yes, beverage is technically just fine, but it sounds affected.

What I have found that helps me is to not just limit myself to one sentence for each new vocab chunk-- I try to find other sources when possible.

RTK2 is very organized if you just want to study on-yomi, but the problem is that even when you're done with it, I wonder just how well you'll be able to read-- you'll nail most of the nouns, but most of the verbs will be mysteries. (Except for the する ones, I guess.)

I consider kun-readings just as important as on-readings. I'm currently copying Index V into a spreadsheet and considering my options on how to study these. For kun-readings that only appear in combinations, I'll insert one. Otherwise I think I will try out the phoneme mnemonic thing for most of the unfamiliar readings. And add a translation or example sentence perhaps. Or switch to something else.

I should add that RtK2 is meant in the first place to aid your reading. Writing, speaking, listening is not going to benefit much. Take that into consideration when answering the "What's your next goal?" question.