I am curious to find out what methods people have been using to help learn RTK2. I finished RTK 1 on Monday but I am taking a few weeks break to continue working on my reviews and to determine the best way to approach RTK2.

I am approaching RTK2 with some trepidation because I have read some of the previous threads which seem to indicate that the method that Heisig left us for book 2 leave something to be desired. Then again, someone posted that they learned RTK2 using the Heisig method without difficulties.

I appreciate that learning the first chapter (that of the kanji from which the kana are derived) would be difficult because there is nothing linking them together; however, I would think that from the second chapter on (when there are pure, semi-pure, etc. groups) that share similar primitives, that the method outlined by Heisig would work reasonably well.

Could other members give me some guidance as to how best to proceed? It seems to me that the kanjichain method could become quite cumbersome (at least that is the feeling that I get after looking at the beta review section).

As someone pointed out in another topic, one method may work well for some, while another method may work well for others. Maybe there are other methods that may work well for other people that have not come up in discussion. We really only have two methods at this point: kanjichain or Heisig.

What are your thoughts?

I've just started readings practice. I bought RTK2 and I tried kanjichains as well. For me, kanjichains work better, though I was really afraid of the bigger chains. The drawback of Heisig RTK2 is that his pure groups are just as noticable in the kanjichains, so you have the benefit, but at the same time you go quicker as you lump a lot of reading together. I don't like changing and adding to my stories, so I wanted all the kanji with the first on reading of カ for instance to be together, even if it meant doing 32 kanji at once. It's less work to make one story for 32 kanji than to make one for 6, and add to it later with another 5 and another 4 etc. So, instead of going through Heisig, I prepare a story for a chain in a spreadsheet, check the readings with this site's beta, and make sure all the right kanji are in the story.

I decided on a few things that specifically worked for me. First, contrary to what is mostly suggested, I wanted to quickly study a lot of on-readings quickly in a narrow focus, instead of going more broadly. I didn't feel studying one compound for every kanji was the way to go for me. What I see in real texts is almost never the compound I studied, whereas I did feel a sense of satisfaction of having an on-reading ready to get my wordtank working quickly for me.

So, I ditched the compound learning. I took a deep breath, did a lot of small chains to start, and started doing larger and larger ones. It can take a time to make a large story. It sometimes takes me more than a day to come up with a satisfactory story h ook where the kanji keywords look appropriate. But it works far better than I had imagined. My largest chain is now 32 items. I save really big chains for days when I have time, such as the weekends. In a few weeks, I did over 400 readings, smaller and larger chains, so it's working for me. Well, I haven't tried the horrible 67 kanji chain yet, but I'm confident I can do it, as long as the story makes enough sense or can be compartimentalized properly.

To aid the process, every chain I make I enter every kanji as a Q&A in Mnemosyne. Entering the kanji plus reading really helps to cement looking at the kanji carefully and putting it in the chain. If I want to add more readings later, or kun readings, I can easily edit the database to accomodate it.

astridtops Wrote:The drawback of Heisig RTK2 is that his pure groups are just as noticable in the kanjichains, so you have the benefit, but at the same time you go quicker as you lump a lot of reading together. I don't like changing and adding to my stories, so I wanted all the kanji with the first on reading of カ for instance to be together, even if it meant doing 32 kanji at once. It's less work to make one story for 32 kanji than to make one for 6, and add to it later with another 5 and another 4 etc.

This is a great point. I didn't consider this. I tried writing a kanji chain for the on reading for [kana]in[/kana]. It took a little work but it wasn't impossible. I also found that I didn't get one of the keywords truly embedded in the story so I have to work on it. I think this will take considerably more time than RTK1, but it is really about the journey, not about the goal.

Thanks for your response. It was very helpful to me.1

astridtops Wrote:The drawback of Heisig RTK2 is that his pure groups are just as noticable in the kanjichains, so you have the benefit, but at the same time you go quicker as you lump a lot of reading together.

Astrid, this made 100% sense when I first read your post. Now that I'm getting closer to finishing RTK1, like Jon I have been giving more thoughts to the next step.

Is lumping all kanji with the same On-yomi reading together necessarily a good thing? As in, does it really mean faster progression? I don't know the answer...hence the post.

But having had a quick look at RTK2, the idea of signal primatives looks pretty good. When you make chains do you include RTK1 keywords in the story even if they all have the same signal primative? If so, that would seem less efficient than learning the RTK2 way.

Just as an example...while flicking through RTK2 I happened to notice that 包 is pronounced ホウ. That immediately made me think 'wrap' -> Christmas -> Santa -> Ho Ho Ho. So I've picked up 5 or so readings by simply taking in the one signal primative.

So my thinking at the moment is tending me towards going with RTK2. I can't see any benefit chaining has over the pure groups. Once you start moving towards semi-pure groups perhaps the merit of chaining will become more apparent.

I think both methods might work, but it might depend on individual preferences. As for the kanji chains, I do reasonably well on the 550 kanji I've chained so far. At first, I was careful to include some pointers for the kanji story as well, but that only worked for the small chains. Once I tried the bigger ones, over 10 long, I had enough trouble creating the whole story and image including all the keywords. So, I haven't given up on the method yet as working better for me than RTK2 as far as learning one onyomi reading is concerned.

But, continuing study in other areas, now that I'm also working on the vocab spreadsheet I find that having one reading ready can give me a headstart on some of the entries - but it only helps if I already know the vocab. For instance, when I got 喫茶店, I knew all three readings: キツ，チャ and テン. Kitsu-cha-ten becomes kissaten, coffeeshop (excusez le romaji), good: word learned. But in most cases I have to learn the compound meaning - and oddly enough I have trouble remembering the compounds if I guessed correctly for reading, but didn't know the meaning yet. Somehow going through the entire process of learning reading and meaning together can also be a good thing apparently.

Well, apparently the last thing about what's really effective hasn't been said yet. But I remain quite interested in anyone's progression with learning readings and vocab in any fashion.

I am pretty sure I will be going with Heisig for RTK2, to begin with anyway. 'Pure groups' look a piece of cake. 'Semi-Pure' a bit harder but still pretty straight forward. The 'Mixed groups' do look quite tricky so perhaps chaining will win that round.

There were a few posts on another thread that were talking about learning the kanji for vocabulary you already know. This seems like another good idea from Team KanjiKoohii. Not sure if Heisig championed this but I'm sure he would be proud of it. After RTK1 we (should) know the kanji; vocab you already know is also known (obviously); so the only thing one really learns is the association between word and kanji. Simple but hopefully very effective.

Finally, a few people have introduced a little bit of RTK2 to their RTK1 stories, e.g. Fabrice with JOU （嬢）1527-1530. This is something I would recommend taking advantage of to people who are still going through RTK1.

synewave Wrote:Is lumping all kanji with the same On-yomi reading together necessarily a good thing? As in, does it really mean faster progression? I don't know the answer...hence the post.

I wondered this same point myself. I originally started this thread a while ago. I tried kanjichaining but I found that I was spending so much time on kanjichaining that I was neglecting my vocabulary. I also seriously neglected my vocabulary for the three months I spent working on RTK1. So I decided to put the kanjichains aside and work on vocabulary.

On the advice of Aikou (another site member), I started using my vocabulary learning as a way to learn pronunciation of kanji. Please let me elaborate:

Once you have finished RTK1, you are intimate with each kanji that you see. Now it is easy for you to look at vocabulary that you already know (example ゆうめい、famous). This in kanji is 有名． If one breaksdown the constituate kanji one has possess-name; therefore, one who is famous possesses a famous name. Another example is the word for in front. We know that this is generally thought of as まえ in hiragana. This is a basic JLPT 4 word. Heisig helps us by using the key word "in front" for 前, hence, まえ＝前）。 Now, when one wants to learn what someone's first name is, (or name in front), since we have already learned that 名＝な　and 前＝まえ, then なまえ　must be 名前。

Using this method of relearning my existing vocabulary with correctly used kanji, I have improved my vocabulary considerably. I went from barely having a JLPT 4 kanji/vocabulary level to having a much higher kanji knowledge level. Also, my vocabulary is growing faster. Instead of learning abstract thinks such as ゆうめい for famous, I immediately learn and think of 有名. Thus, I am learning new vocabulary and the readings for new characters when I see them. I can also now, deduce the correct reading of unknown compounds using other readings of kanji that I already know.

I am finding that this is a fast and easy way for me to learn kanji compounds and pronunciations instead of using some of the abstract keywords found in RTK2.

Please, do not think that this means that RTK2 is garbage, quite the contrary. As it is commonly pointed out when discussing the pursuit of kanji learning, each person has his own path, though we all end up at the same outcome. So if RTK2 is good for you, use it. If kanjichain is good for you, use it. If learning readings through vocabulary is good for you, use it. Unfortunately, no one can give you an easy answer to the question: "How should I learn the kanji pronunciations?". I think that anyone approaching this process should try all of these ways for a little while until they find which way makes more sense.

When I started this thread, I expected an easy answer. In a way, the answer is easy but it is a matter of self-discovery, not a one size, fits all answer.

Sorry for the extra long post. I just wanted to get my thoughts on the subject out now since I have learned by doing. Hopefully what I have explained here can help those who finish RTK1 to find the right way for them.

So far the best manual for the vocabulary/readings that I've seen was Kanji in Context:
http://www.iucjapan.org/html/text_e.html
It is organized by kanji: first a character, then a few words with that character, and some examples and exercises on it. The kanji are ordered by frequency. This way, since words with the same character are learned at the same time, learning itself is greatly simplified, and the kanji readings come as a free bonus (plus -- you also realize the difference in meanings that is denoted by readings sometimes, like ちゅう reading for 中 means something a little different from じゅう, and なか is of course something else entirely). I was also quite amazed at how useful their ordering is -- in the first chapter at least, I would learn a few words, then watch an anime episode or read something and immediately hear a few of the words right there.

I've tried to do the same myself earlier, but in the Heisig order (i.e. I took a JPLT3 list of compounds and sorted it by the largest Heisig number among the characters in the compound), in order for it to be usable alongside following Heisig. I now think that it was not an optimal thing to do, and it is much better to just finish the entire RTK1 first and then follow a mixed kanji/frequentist approach like in KiC.

I am definitely going to look at the larger groups in RTK2, but kanjichains/kanjitown/etc etc don't seem to be particularly useful to me.

astridtops Wrote:So, I ditched the compound learning.

What I am confused about is that I see absolutely no benefit whatsoever in knowing a reading of a kanji without knowing any compounds with that character/reading. Even if you can read an unknown word aloud correctly (with some probability, probably not a very high one, as the actual reading may differ), what is the benefit of that if you don't know the word?

I am sure I am missing something.

laxxy Wrote:

astridtops Wrote:So, I ditched the compound learning.

What I am confused about is that I see absolutely no benefit whatsoever in knowing a reading of a kanji without knowing any compounds with that character/reading. Even if you can read an unknown word aloud correctly (with some probability, probably not a very high one, as the actual reading may differ), what is the benefit of that if you don't know the word?

I am sure I am missing something.

I absolutely agree with you laxxy. I feel that without using the reading in the context in which it is meant to be used, learning individual sounds does not seem to be some helpful to me.

jondesousa Wrote:On the advice of Aikou (another site member), I started using my vocabulary learning as a way to learn pronunciation of kanji.

Excellent post jondesousa. I also use this method. I did a good chunk of RTK2 (1380 frames) before I got frustrated with it and quit. It was brutal, but I reviewed it like crazy, and finally got to the majic 90%+ after well over 100 hrs. I was using kanjichains to help me.

Right afterwards, I started using Aikou's method. From that point on, I noticed that when I saw a character that I knew through RTK2 and vocabulary, I would remember the reading through the vocabulary word, not my RTK2 mnemonic. That told me vocabulary was working better for me. And vocabulary learning was so much easier.

jondesousa Wrote:What I am confused about is that I see absolutely no benefit whatsoever in knowing a reading of a kanji without knowing any compounds with that character/reading. Even if you can read an unknown word aloud correctly (with some probability, probably not a very high one, as the actual reading may differ), what is the benefit of that if you don't know the word?

You may never need to study individual readings seperately, but if you study compounds only, you'll learn the individual readings that way, and you'll end up using them to bail you out. The benefit is huge. Knowing the individual readings bail me out all the time when I read.

leosmith Wrote:Knowing the individual readings bail me out all the time when I read.

I agree. It's almost unbelievable how everything seems to come together. I'm sometimes impressed with myself how I can deduce the correct reading of a compound without having seen it before.

Keep up the hard work leo. We will all eventually all be reading Japanese without much difficulty.

Nice post Jon. Learning kanji through your existing vocabulary seems such an obvious idea...that I hadn't actually thought about!

I think my first steps will be to go through RTK2 pure and semi-pure groups then hit the vocab.

Not that I am against kanji chaining per se, but nearing the end of RTK1 I feel I have been neglecting other aspects of my Japanese, e.g. vocab + grammar. This is what I want to get back into sooner rather than later.

Cheers,
Andy

synewave Wrote:Not that I am against kanji chaining per se, but nearing the end of RTK1 I feel I have been neglecting other aspects of my Japanese, e.g. vocab + grammar. This is what I want to get back into sooner rather than later.

Hi Andy, this is exactly why I approached the learning of kanji pronunciation this way. I found that I was spending all of my time learning kanji and none of my time learning Japanese. I changed that by learning pronunciations as I learned vocabulary.

As I mentioned earlier, it is different for everyone, but for me, this was the best way. I don't regret it and unless something truly better comes up, I will probably never change the way I approach it.

Good luck and keep up the good work.

I'm almost done with RTK1. Something I've chosen to do instead of going through RTK2 is to studying for the Kanken (Kanji Kentei Nouryoku Shiken, 漢字検定能力試験) and use
it as my way of learning pronunciations and vocab.

While the JLPT is aimed at foreign learners of Japanese, the Kanken is for Japanese
people and is meant to test their Kanji ability. There are 10 levels(10 being the lowest), with the first 6 levels corresponding to the first 6 grades. The official textbook concentrates on the kanji for each level and provides many different exercises for each kanji. It groups the kanji for each levels in chunks of 4-6 kanji at a time.

Also, unlike the JLPT, you actually have to WRITE the kanji,
which means it's great practice for those who finished Heisig Vol 1.
I think it's a great vocabulary builder for someone who has finished
Heisig volume 1, since it gives you a concrete list of kanji to
study for each level and the list of vocab for each kanji is kept short.

The name of the textbook series is "漢字学習ステップ".

Note that from levels 7 to 10, the exam ONLY tests Kanji for that levels.
From levels 6 and up it tests you on ALL kanji up to that level.
Heisig Vol 1. covers all the kanji up to the pre-level 2 exam.

For those in America, you can take the exam in New York, and Los Angeles.
If you're lucky enough to live in LA, they have the computer version of the exam, which you're able to take 6 days a week, all-year round.

Here's a link to the Wikipedia article:
http://en.wikipedia.org/wiki/Kanken

The paper exam is offered 3 times a year(February, June, and October).

At the end of this month, I'll take the exam in New York. I will be taking
the level 8 exam, which is equivalent to the 3rd grade level. The classroom
will likely be filled with little 3rd grade Japanese children and ME(an adult).

HAHA.....can't wait....

NOTE: The contact info on the Official Website is outdated. If you're interested in taking the February exam in New York, please email enausa@msn.com (Harrison, NY) or nishii@jolnet.com (Hartsdale, NY). Note that there's a sizeable Japanese community in Hartsdale. So that might be the better choice if you want to walk around and do stuff(they have an Book-Off Japanese Bookstore).

Neat. I'm taking level 6 at the end of this month at the school where I work (with my 8th & 9th grade students who will be taking level 2 and 3). I'm tempted to recommend those step up books to other kanji learners but I still have mixed feelings about them as general kanji study books.

chamcham Wrote:If you're lucky enough to live in LA, they have the computer version of the exam, which you're able to take 6 days a week, all-year round.

Do you have a link to that? I'm wondering how they do the written part.

Address is:
南加日系商工会議所
Japanese Chamber of Commerce of Southern California
244 South San Pedro Street, #504,
Los Angeles, California 90012 USA

Tel:+1-213-626-3067
Fax:+1-213-626-3070

Here's the link to the CBT(Computer-based Test) page:
http://www.kanken.or.jp/cbt/index.html

CBT Exam Explanation here:
http://www.kanken.or.jp/cbt/cbtsetsumei.pdf

CBT Q&A FAQ:
http://www.kanken.or.jp/qa/cbt_qa.html

These used pen tablets(like the Wacom tablets)
for input. I heard somewhere that it's really crappy
and annoying if you need to correct your handwritten
kanji.

I'm considering just going to Los Angeles once a month
to take it there. At all the paper test locations in the US,
they do ALL of the exams at the same time. So you can't
take more than one level in a day.

So if you want to do more than on exam at time(6 days
a week all year round), then this CBT in LA is the only
option.

I'm taking a break from RTK1 becuase I'm trying to sort myself out so I can pass JLPT2 this year. So to try and cram as many readings for the test into my head in as short space a time as I can, I'm going through the RTK2 Pure Groups.

All I'm doing is taking one exemplary compound for each signal primative and writing it on the front of a word card along with the signal primative(s) from the compound. On the back I put the reading, meaning and what type of signal primatives they are (pure; semi-pure or mixed). These aren't the same cards as for RTK1 because I stopped making them once I found this site!

Then I basically just try to memorise the readings, which I try to make as easy for myself as I can by using exemplary compounds that I already know the word for, just maybe not the word/kanji relationship. For anyone not going down the chaining route, is there anything more than brute force memory that you are using?

synewave Wrote:For anyone not going down the chaining route, is there anything more than brute force memory that you are using?

It is funny that I originally thought that one must use mneumonics to prevent brute force memorization from occuring; however, now that I have begun reading stories, etc. I find that learning new readings in context occurs quite naturally and does not take as much "brute force" as one would expect.

Heisig's RTK volume I uses 2 techniques. The first is chunking. This technique uses the fact that human memory is very efficient at pattern matching. I have found to date complex kanji far easier to remember than some simple ones.

The second technique is a mnemonic story which links the distinctive shapes or position of a kanji's primitive elements to a keyword meaning in your English.

The problem and challenge of re-assigning meaning to a Japanese reading is the goal of RTK volume II. There seems 2 approaches might work. Heisig has adopted what I term the Breadth First Approach. This uses a signal kanji having the same sound across several other kanji which share the signal kanji as a part. A limitation of this for me is that it is deconstructing the chunked kanji I so painstakingly learned before. A second problem is more severe. The approach seems to neglect compounds.

The alternative approach is what I term a Depth First Strategy. Here you concentrate on the compounds of which a signal kanji is the head. This approach encourages chunking so that compounds can be seen as single entities.

The problem of assigning meaning still remains. I am building a bestiary. My kanji world contains many creatures whose names are the Onyomi and Kunyomi. The meaning is contained in the description of the beast and it's behaviour. Kunyomi are subspecies bred to order by the mad genetic scientist who oversees the world.

The world is infinitely extensible as new compounds or single kanji are encountered.

That bestiary sounds like a cool approach! Is this your own variation of the KanjiChain or KanjiTown? Or is this something different? It sounds cool! Could you compare/contrast it to the other techniques mentioned?

graham6565 Wrote:A limitation of this for me is that it is deconstructing the chunked kanji I so painstakingly learned before.

Sorry, I don't understand. Could you explain using an example?

graham6565 Wrote:The approach seems to neglect compounds.

If you read the instructions carefully, you will see that you are to review from compound to reading. So compounds are not neglected. I think some people assume that you are to review from kanji to kanji reading, but this is not what the author says to do. I'm not a huge fan of RTK2, but I think one should be fair when they slam it.

To me, learning to read the compounds that I already know how to pronounce seems much more efficient than learning a new list of words. It's less work, which makes it a smaller step. Remember "what about Bob?" - baby steps, baby steps, baby steps.........

leosmith Wrote:To me, learning to read the compounds that I already know how to pronounce seems much more efficient than learning a new list of words. It's less work, which makes it a smaller step. Remember "what about Bob?" - baby steps, baby steps, baby steps.........

That is the exact reason that I am working with compounds I already know. There are two great reasons for this:

1. You learn to read compounds you know in kanji instead of always relying on hiragana.
2. You learn onyomi readings for individual kanji since you learned those readings when
studying compounds you already know.

This is as they say, "一石二鳥?。　　Ha. I can put some of my onyomi learning to use. I know that this compound is いっせきにちょう；　therefore, I have learned onyomi readings for the four kanji above. Now, when I see those four kanji, I know at least one potential onyomi reading for them. That is how I have been learning readings.

Me too, and I like this method. I'd recommend Kanji in Context as a very useful book for anyone who would follow this method:
http://www.iucjapan.org/html/text_e.html#kic
Amazon.co.jp seems to have it.