How many of us remember Heisig's reference to Japanese school children learning kanji by "constant repetition"? How many of us really believe this to be the case and accept this without questioning?

Till recently, I have been one of those people but am now not so sure. I am currently holding in my hands a few Japanese kanji books - one is a Grade 2 kanji text book (borrowed from a library), the other is a Japanese kanji reference book covering all the kanji from Grades 1 to 6 (also from the library), and finally a set of kanji practice workbooks (Grades 1-6, that I bought from our supermarket!).

I would assert that these are typical Japanese textbooks meant for Japanese schoolchildren studying kanji, dating from the mid 80s to newly published.

And none of them rely on "constant repetition" as a learning process. The two textbooks took a very mature approach - for each kanji, the on- and kun-yomi readings are presented, as well as a picture representing the meaning. Then there is a "story" that describes the shape of the kanji in relation to the meaning(s), a set of common compounds, and example sentences using the kanji. Finally, there is a brief history of the etymology of the character.

Now the interesting thing is that both textbooks clearly encourage breaking the character into "primitives" and assigning meaning to primitives. There's even a listing of "primitives" at the beginning of one of the books and the meaning of each primitive. Lastly, both books go into the trouble of grouping characters with the same primitives and showing how the meaning changes. One of the books even shows how some primitives act as semantic markers, and others act as phonetic markers.

So clearly these textbooks are expecting Japanese schoolchildren as young as Grade 2 to be learning kanji through an analytical method, not just by constant repetition.

The kanji practice workbooks for Grades 1 and 2 were the only places where they ask the reader to copy the characters. But even then, there are exercises based on primitives (including "chaining games" where you join strings of kanji together based on the fact that they share primitives). From Grade 3 onwards, the workbooks focus on applying kanji in context, in actual sentences.

Anyway, thought you guys might be interested to know that whilst Japanese schoolchildren may have been taught using constant repetition in the 70s when Heisig first wrote RTK1, that doesn't seem to be the case today.

I like one of the textbooks so much I have ordered it from Amazon.

Edit: First ^__^

From what I've seen from the junior high kids at my school and from the "kanji training" workbooks I bought at the 100yen shop there is a lot of repetition.

In the workbook, there is also reading practice and games. The reading practice is a sentence with a kanji marked and they have to write the reading. Also, every kanji has a writing practice section with five boxes to trace the kanji and five boxes to write it without tracing. Then there are sections to copy compounds words 3 or 4 times each. (they are the same compunds used in the reading practice)

And I see kids at my school with worksheets like: "Kanji Writing Practice #14" which is just a list of kanji in one column and a row of boxes for copying.

I'm not sure, but I wonder if they study primatives backwards from the way we use primatives. i.e. I wonder if they study primatives to learn about kanji and to describe kanji they already know how to write.

As far as grouping kanji, at least in the schools, isn't the government list the primary factor in the order the kanji are learned? My school is really strict about following the English curriculum, so I wouldn't be surprised if Japanese class was similar. 食　is 2nd grade and 良　is fourth grade, but does your book put them together?

But i'm curious about the "stories" in your book. what are they like?

johnzep Wrote:In the workbook, there is also reading practice and games. The reading practice is a sentence with a kanji marked and they have to write the reading. Also, every kanji has a writing practice section with five boxes to trace the kanji and five boxes to write it without tracing. Then there are sections to copy compounds words 3 or 4 times each. (they are the same compunds used in the reading practice)

I think this is true for the Grade 1 and 2 workbooks I have seen, but not true for Grade 3 and higher. I have several of these workbooks (including a set based on the "Peanuts" comic strip that I really like, since the pictures of Snoopy are really cute).

johnzep Wrote:And I see kids at my school with worksheets like: "Kanji Writing Practice #14" which is just a list of kanji in one column and a row of boxes for copying.

I think these are to improve writing neatness and legibility rather than for memorization. Just like in our English classes, we had to copy sentences over and over again to improve our writing, but there's no intent for us to memorize the sentences.

Or at least that's how my Japanese teacher explained to me. He said Kanji handwriting legibility is a real problem because a lot of people have sloppy handwriting.

My Mom, who was a primary school Chinese teacher, also confirmed that the main intent of copying is to improve legibility and calligraphy, not for memorization.

johnzep Wrote:I'm not sure, but I wonder if they study primatives backwards from the way we use primatives. i.e. I wonder if they study primatives to learn about kanji and to describe kanji they already know how to write.

This is a good point, but you have to remember that even a grade 1 student already knows how to speak Japanese and have a relatively large vocabulary. So the primary purpose of teaching kanji is to teach them how to write words they already know.

As to whether Japanese break down kanji the same way as Heisig, no they don't. But then, I don't really like some of the Heisig primitives either. I actually like the Japanese way (and by extension, the Chinese way), which is by phonetic and semantic markers. In other words, some parts of the character convey the meaning, others convey the sound. Heisig doesn't distinguish between the two types.

johnzep Wrote:As far as grouping kanji, at least in the schools, isn't the government list the primary factor in the order the kanji are learned? My school is really strict about following the English curriculum, so I wouldn't be surprised if Japanese class was similar. 食　is 2nd grade and 良　is fourth grade, but does your book put them together?

I think the grade school order simply determines which kanji one *must* be able to *write* (and know the readings). At any point in time, the number of kanji that a school child can *recognise* is always much higher. So with respect to the example you quoted, a good teacher when teaching 食 will probably mention 良 but say you don't need to fully learn the readings for the latter till later. And even then, many students I dare say will have have already encountered 良 and can already recognise it.

Remember, just knowing the shapes of the characters are not enough (despite what Heisig says). It's the readings and the (multiple) meanings that are the key to mastering a specific kanji character.

johnzep Wrote:But i'm curious about the "stories" in your book. what are they like?

Only so-so. They are not as colorful as Heisig, but probably adequate. The difference between the Japanese way and Heisig is that you are not supposed to be memorizing how to write a character from the "story" (although it's useful as a mnemonic aid). The main key to learning how to write a character is actually the etymology. This is what my Japanese teacher also stresses (I think he would be horrified by Heisig's stories! He certainly didn't like mine when I described to him my story for 食 - he kept saying it's not historically accurate nor does it truly reflect the meaning of the character. I wanted to point out to him I didn't really care about the history of the character - it was just to help me remember how to write it!)

But I do appreciate that ultimately etymology is the best way to truly appreciate a kanji character in all it's richness. For example 食 wasn't always written like that, and learning the earlier forms actually help you understand how the present form (and meaning) came into being. And after that, how to write it (including stroke order) becomes very easy to remember.

As you may have guessed, I have pretty much given up learning using Heisig, and I'm learning the "traditional way" (well not really traditional, I found the Heisig primitives that I've picked up useful every now and then). It's much slower, I would say 5-10 times slower than Heisig, but the recall rate is higher than Heisig, and the bonus is you really know the character, including multiple meanings (such as when the character is used as noun, vs a verb - and transitive vs intransitive usage) and readings. I surprised my teacher the other day by reading a complete sentence in Kanji (including correct reading and stress pattern) without resorting to furigana - I even surprised myself at how much I remembered. I only know around 200 kanji intimately at the moment, but I'm hoping to double that by the end of the year. My teacher says there is no need to learn 2000 - many characters are botanical, archaic or specialist terms that are extremely rarely used. He says most Japanese probably only know about 1000-1200 characters well, so that's what I'm aiming for. Advanced students I have spoken to say that when you get to that level, you can pretty much guess the meanings (and sometimes even the readings) of characters you don't know - there's apparently some redundancy in compound words.

I have grown to become very much in favour of a more sound traditional approach but I don't think this has to stand in the way of the speed.

However...

Christine_Tham Wrote:My teacher says there is no need to learn 2000 - many characters are botanical, archaic or specialist terms that are extremely rarely used. He says most Japanese probably only know about 1000-1200 characters well, so that's what I'm aiming for. Advanced students I have spoken to say that when you get to that level, you can pretty much guess the meanings (and sometimes even the readings) of characters you don't know - there's apparently some redundancy in compound words.

...my gut feeling is that all parts of the above statement are very, very untrue...

The books you mentioned sound really interesting - could you tell me the titles?

Quote:I have spoken to say that when you get to that level, you can pretty much guess the meanings (and sometimes even the readings) of characters you don't know - there's apparently some redundancy in compound words.

This may be true for native speakers who have acquired a large vocabulary from speaking and television, but it's a different story for second language learners who usually lack the tools to learn words fast enough without the help of kanji.
I'm pretty sure that in such cases as you mentioned above, people quite frequently guess wrong, too.


If it's any reflection on teaching in Japan, my teacher (日本人, I took one year of 2nd year Japanese in college) taught kanji very much in a "here, learn this, and quickly" kind of way.
She would write the characters on the board, pointing out parts that we were familiar with. Then we would be given practice sheets to write them about 2 times. There was no common thread to the characters, just ones that appeared in the reading and were deemed easy enough for us to read at that point.

I suspect that there may be more focus on the stuff in Christine's books during the early years, when most characters are pictographic or ideographic, and this probably lessens later on, when students are considered more mature and more able to hammer disconnected facts into their heads.

I believe he means repetition as in, they're entire life surrounds them with kanji which they practice for years and years, thus they are repeatedly exposed to it. A Japanese student may not the meaning of individual kanji in a difficult compound for something like a government agency, but once they know it, they "suddenly" see it everywhere. The same goes for any language.

Also, as others mentioned, Japanese students already have the vocabulary. Learning the kanji justs connects the points. But for us, every kanji is a huge excursion of multifaceted learning. And significantly, we don't have the repetition of paced learning. Many of use are cramming 12 years of education into 12 months!

This topic has been covered before somewhere. I distinctly remember a person learning through Heisig trying to help their child with kanji and their Japanese spouse yelling at them 'not to break it down into primitives and confuse them, they just need to learn it.' Japanese can utilize visual memory to full effect because of the pace of their learning, as I mentioned above.

Related to this, as has also been discussed in numerous other threads, is that while most Japanese you show think the Heisig method is interesting, they heavily doubt you've learned the kanji (even after showing them comprehensive knowledge of the meanings of 2000+) because it is so unfamiliar to them. As Heisig explains, he applied outside reasoniong to a logical system to derive a quicker way of consuming the information than is required for native learners. They are suspicious of something they have struggled with their whole lives yet you have mastered in months. The caveat of Heisig is that it ONLY works for adults.

dingomick Wrote:The caveat of Heisig is that it ONLY works for adults.

Or teenagers.

Interesting topic. By this point, I find learning 200 kanji in a year way too slow.
I guess you don't need to know every joyo kanji, but I'm sure Japanese people know more than 2000 kanji on average. I started practicing my Japanese by reading short stories with rikaichan, and I often come across non-joyo kanji. Also, non RTK3 kanji(yet another reason for me not to do RTK3). I guess when Japanese people type, they can use kanji they recognize but haven't memorized, but still.

samuize Wrote:The books you mentioned sound really interesting - could you tell me the titles?

Yeah, I recently ordered these two books from Amazon Japan:
小学生のための漢字をおぼえる (this is the one that summarises Kanji Grades 1-6 in a single reference book)
漢字はむずかしくない？24の法 (this is intended for foreign students learning Japanese - contains "stories" and pictures for 192 kanji)

The other one I mentioned is a Grade 2 textbook - I don't have the title handy but it essentially says "Grade 2 Kanji" in Japanese. I found this in the children's section of our local library.

JimmySeal Wrote:This may be true for native speakers who have acquired a large vocabulary from speaking and television, but it's a different story for second language learners who usually lack the tools to learn words fast enough without the help of kanji.
I'm pretty sure that in such cases as you mentioned above, people quite frequently guess wrong, too.

No, I've had at least two people tell me that [it's possible to guess words you don't know once you know at least 1000 words], and neither of them are native speakers, but they have both passed JLPT1.

It's a lot more than a "guess", a better word would be "infer". You probably need a Japanese teacher to demonstrate to you why this method works, and why it's so effective. In class, we go through exercises where a teacher would write a compound containing a word we know, and a word we don't know, and we have to "guess" the meaning of the compound.

To understand why, you need to appreciate how Chinese as a language evolved. Chinese was originally a monosyllabic language, ie. the reading for each kanji was only one syllable, and even today it is still primarily a monosyllabic language.

The problem with just using one syllable for each word is that you run out of syllables very quickly, even if you assigned different tones (Mandarin uses up to 5 I think).

So the Chinese invented the concept of compound words, where you string together 2 or more kanji to form a polysyllabic word.

Now the interesting thing with compounds that often they are redundant - the first character already conveys the meaning, the second character is mainly used for phonetic value (just to add another syllable to the word to distinguish it from another similar sounding word).

Over time, as kanji compounds became more common, new kanji was invented purely to serve as "extra characters" to tack onto new compounds. These kanji are called "compound ideo phonograms", or, as my teacher calls it, composite semantic/phonetic markers.

Basically, these kanji has one part that represents the meaning of the character (which is used to influence the meaning of the overall compound) and another part that represents the sound of the character (as a hint on how to pronounce it).

Over 80% of Kanji characters are composite semantic/phonetic markers. If you know what you are doing, you can easily "guess" the meaning and readings of these characters by decomposing them into their semantic and phonetic markers.

So, if you already know around 1000 kanji, it's very likely (my teacher says greater than 95% probability) that any new kanji you encounter is a composite marker as part of a compound. So you should be able to guess it with high accuracy. And I mentioned before at least two other people have confirmed to me that basically this works.

JimmySeal Wrote:If it's any reflection on teaching in Japan, my teacher (日本人, I took one year of 2nd year Japanese in college) taught kanji very much in a "here, learn this, and quickly" kind of way.

This is called learning Kanji by "context", and it's actually the most effective way of learning kanji (once you get to intermediate level). The problem is, it's not very effective for beginners (because it comes across as completely random and you don't have the background to effectively learn by context) - I think too many teachers use this method too early in the process.

I am now only starting to use this method, but even then I don't think I have fully mastered the fundamentals (a good working knowledge of radicals is a must). "Kanji in Context" which primarily teaches using this method, recommends that you don't study this way until you have mastered at least 250-300 characters.

JimmySeal Wrote:I suspect that there may be more focus on the stuff in Christine's books during the early years, when most characters are pictographic or ideographic, and this probably lessens later on, when students are considered more mature and more able to hammer disconnected facts into their heads.

I think if you reread the original post, you will find I am saying the opposite - that judging by the evidence of textbooks, Japanese children are learning sophisticated methods of kanji decomposition as early as Grade 2.

dingomick Wrote:I believe he means repetition as in, they're entire life surrounds them with kanji which they practice for years and years, thus they are repeatedly exposed to it.

If you read Heisig's paragraph in the preface of RTK1, I don't think this is what he means.

dingomick Wrote:Also, as others mentioned, Japanese students already have the vocabulary. Learning the kanji justs connects the points. But for us, every kanji is a huge excursion of multifaceted learning. And significantly, we don't have the repetition of paced learning. Many of use are cramming 12 years of education into 12 months!

But this does not mean the task is impossible. In fact, the advantage of Heisig is that he demonstrates that it is in fact relatively easy to memorize the meanings and shapes of over 2000 characters in a short amount of time.

What is more difficult is actually learning the characters properly - including readings, compounds, verb/noun usage, transitive/intransitive usage.

What several people (including teachers) have told me is that the first 500 words are the hardest. After that, it's smoother sailing, because techniques like "compound ideo phonograms" decomposition become a useful tool for learning Kanji in context.

I am currently at the mid way point (around 250 characters) and I'm finally starting to see the light at the end of the tunnel, but it's been a hard slog and will continue to be a hard slog for the next few months.

dingomick Wrote:This topic has been covered before somewhere. I distinctly remember a person learning through Heisig trying to help their child with kanji and their Japanese spouse yelling at them 'not to break it down into primitives and confuse them, they just need to learn it.'

I believe you are referring to Column #23 in Kanji Clinic:

http://www.kanjiclinic.com/kc23final.htm

Mary's husband's comments will make more sense (and, by the way, he is right) if you look at an actual Grade 1 Kanji textbook or attend a Grade 1 Kanji class.

In Grade 1, the focus is on learning hiragana and katakana using audio/visual cues. 80 Kanji is taught, but this number is too small to allow for effective decomposition analysis, so teaching using this method will be counterproductive.

As I mentioned in my original post, analytic decomposition is a technique that is introduced as early as Grade 2.

dingomick Wrote:Related to this, as has also been discussed in numerous other threads, is that while most Japanese you show think the Heisig method is interesting, they heavily doubt you've learned the kanji (even after showing them comprehensive knowledge of the meanings of 2000+) because it is so unfamiliar to them.

I would agree with them, and I started learning kanji using Heisig's system. At the end of the day, Heisig's system is just a memorization technique to get you familiar with the shapes of the 2000 characters, and keywords associated with them. When you complete RTK1, you actually haven't learned any kanji at all, at least not in the sense that a Japanese literate person would consider learning. You don't know the readings, you can't differentiate between noun/verb usage, you will not be able to understand compounds (I know, because I have tried).

What's worse, I discovered Heisig's method is actually counterproductive in the long term, because he does not distinguish between phonetic and semantic markers, hence his method is detrimental to the most effective way of learning Kanji (composite phonetic/semantic marker decomposition).

dingomick Wrote:The caveat of Heisig is that it ONLY works for adults.

As my initial post suggests, this is simply not true. Japanese schoolchildren as early as Grade 2 are being taught graphical decomposition.

Christine, that's fantastic insight into kanji and, something I've noticed frequently in compounds with their semantic markers.

You said you gave up on Heisig though? I think Heisig provides the quickest and most thorough insight into the way primitives/radicals function within kanji because of how he arranges the book. You and your teacher are correct to say that learning the etmology of each kanji is the most effective and solid way to learn a kanji. But it also takes significantly longer. Most Japanese adults don't even know kanji etymology, as is indicated by the kanjikentei which tests it at the highest levels.

I eventually want to be able to dissect kanji and know them inside and out. And I believe their is no better foundation than Heisig to get their quickest. As you said, you just want to remember the kanji. And right now, that's what I need on my path towards mastery. Your teacher may be offended that we're using training wheels on our bikes, but I think it prepares us better for the Tour de France than just throwing us on a bike and pushing us towards the mountain. Japanese children have their whole life to prepare and are already expert bike riders. We've decided on a whim to join in and need as much speedy help as possible.

I finished RTK1 in just 2 months. I may look into those books you've mentioned to now expand my kanji skillz.

I would be happy to scan and post pages of official 小学校 texts to put this debate to rest. There is, I am sure, a big difference between what is available on the commercial market to parents with money, and what is officially taught. Perhaps what you have is more representative of that sort of extracurricular text. As has been said elsewhere, Heisig is not for everyone. Its like wearing socks with sandals it offends common sensibility but the people who do it swear by it. I can't speak for anyone else but if it were not for Heisig I would still be beating my head against the chicken scratches that were JLPT 3 kanji. Now things are far better. Socks with sandals worked for me. By the way, in RTKII, Heisig does teach phonetic markers. At that point the difference between the semantic and phonetic markers becomes a little more apparent.

yukamina Wrote:By this point, I find learning 200 kanji in a year way too slow.
I guess you don't need to know every joyo kanji, but I'm sure Japanese people know more than 2000 kanji on average. I started practicing my Japanese by reading short stories with rikaichan, and I often come across non-joyo kanji. Also, non RTK3 kanji(yet another reason for me not to do RTK3). I guess when Japanese people type, they can use kanji they recognize but haven't memorized, but still.

You should be able to learn a lot quicker than 200/year. I started in late July from zero (well, not exactly zero, I knew the numbers and a few basic characters - typical beginner's stuff). Currently I am at around 250, so that's basically 200 in one month.

It is a lot slower than Heisig, because I found with Heisig I could learn as much as 50-100 a day.

I'm hoping to get to at least 500 by the end of the year, perhaps even 800. But realistically I hope to get to 1000 by this time next year. The problem is that there's no point pushing too hard - Kanji is best learnt by context, and my Japanese isn't good enough (so there's no point learning words I can't apply correctly).

According to my teacher, most Japanese would know around 1000-1200 "intimately." They can recognise a lot more, but typically only in the context of specific compounds and may not be able to write them.

We had an example of this in class the other day when a student asked the teacher the Kanji for a specific word, and he wasn't sure. He said it was either this character or that character, and asked me to look it up in JWPce for him!

dingomick Wrote:You said you gave up on Heisig though? I think Heisig provides the quickest and most thorough insight into the way primitives/radicals function within kanji because of how he arranges the book.

I agree - the main thing I learnt from Heisig was how to identify primitives and how to create stories around them. I still do that, and it was a real breakthrough (my previous method was trying to learn Kanji through "osmosis." I failed miserably).

The reason I gave up on Heisig was I found a lot of his primitives counter productive to my learning process. For example, 里 is *not* composed of 田 and 土 - it's actually a separate radical and the etymology is quite different. And it definitely should not be used as "computer" as a semantic marker! Similarly 黒い is not created from 里 and cooking fire, the etymological derivation is actually soot on top of a 赤 fire.

Then of course you get the radicals that are different but look the same, and that's a whole another story (Heisig of course doesn't separate them).

dingomick Wrote:I eventually want to be able to dissect kanji and know them inside and out.

頑張ってください！

You may want to try my current method for learning Kanji, which is very slow but surprisingly effective. It's detailed in this post:

http://forum.koohii.com/showthread.php?pid=7327#pid7327

I started using this method for the JLPT4 kanji, but have now covered most of JLPT3 and I intend to move to JLPT2 by September. So in one month I moved from knowing virtually no Kanji (I couldn't even recognise the Grade 1 characters) so where I am really comfortable with the JLPT4 characters and know enough JLPT3 characters to be able to pass the kanji section of the JLPT3 test. And I don't even do any spaced repetition. I'm currently going through a few Japanese readers, and I notice I don't even glance at the furigana for the words I know - they just "pop" in my head.

I've also recently extended my method to include studying compounds - so in addition to what I wrote in the post I copy useful compounds from the "Kanji in Context" book.

My next step is to start classifying kanji that I know into semantic and phonetic markers. Hopefully this will allow me to really get into that inference technique for compounds that I described earlier.

dilandau23 Wrote:Heisig is not for everyone. Its like wearing socks with sandals it offends common sensibility but the people who do it swear by it.

I lol'd at this ^__^

Christine, I also have big problems with some aspects of Heisig's method but I think it's important to understand why it has become so popular. After reading your posts carefully, I think you're missing one of the key points. This is that it's important to break your learning down. After completing RTK1, you can write over 2000 characters and you have confidence that you will not forget how. However, that's it, you can't do anything else. You can't read them, you don't know their true meanings, your knowledge of their component primitives is limited, you know nothing about their etymology etc. After reading the method that you are employing, you are trying to master each character one at a time. The steps are too big and you're finding that you're exhausted after covering 16 characters in 2 hours. I think that that approach needs to be broken down and you need to be content with covering a smaller distance with each step. What's more, it's important to test yourself often after covering so many steps and Heisig's method lends itself well to this.

With each kanji, there's so many things to do that if you don't break it down, it's overwhelming. Heisig helps us remember the shape. Henshall helps with the etymology. De Roo and other dictionaries help us understand how each kanji interacts with others. Then there's the different readings, the noun/verb/adjective usage,... The list goes on When I read your method, it looks like you're trying to cover a lot of these for each character one at a time! Why not break it down?

Quote:Remember, just knowing the shapes of the characters are not enough (despite what Heisig says).

I agree, but I don't think Heisig said it was enough, and if he did, no one would believe it. It is a start though.

Quote:The main key to learning how to write a character is actually the etymology.

It depends on what you mean by learning how to write a character. I can tell you that most of my Japanese friends' knowledge of kanji etymology is quite fragmented. It's definitely not thorough.

Quote:As you may have guessed, I have pretty much given up learning using Heisig, and I'm learning the "traditional way" (well not really traditional, I found the Heisig primitives that I've picked up useful every now and then). It's much slower, I would say 5-10 times slower than Heisig, but the recall rate is higher than Heisig

I find this surprising. If you do the Heisig method properly and create vivid imagery and stories to help you remember the shape of each character, the recall rate should be very high (maybe 95%). When I say recall rate, I mean that after being given a prompt of some kind like おおきい or "Big", you can write down 大.

Quote:My teacher says there is no need to learn 2000 - many characters are botanical, archaic or specialist terms that are extremely rarely used. He says most Japanese probably only know about 1000-1200 characters well, so that's what I'm aiming for. Advanced students I have spoken to say that when you get to that level, you can pretty much guess the meanings (and sometimes even the readings) of characters you don't know - there's apparently some redundancy in compound words.

I don't know how many characters I knew "initimately" before starting Heisig, but I think it was around 700. I finished RTK1 about 2 years ago and it took a year to complete it. I don't know how many characters I now know "intimately" (i.e. more than just the shape from Heisig) but I'm sure that it's well over 1200. I can tell you for a fact that it's still not even nearly enough. What's more, anything that I can guess from this point on is nothing more than a guess. I have also been told by many people, including Japanese teachers, who say that around 1200 characters, the level needed for JLPT 2 is enough. It just isn't true.

Quote:I believe he means repetition as in, they're entire life surrounds them with kanji which they practice for years and years, thus they are repeatedly exposed to it. A Japanese student may not the meaning of individual kanji in a difficult compound for something like a government agency, but once they know it, they "suddenly" see it everywhere. The same goes for any language.

I agree. Even if it's not what he meant, it's certainly the case.

Quote:What's worse, I discovered Heisig's method is actually counterproductive in the long term, because he does not distinguish between phonetic and semantic markers, hence his method is detrimental to the most effective way of learning Kanji (composite phonetic/semantic marker decomposition).

Why is it counterproductive? Why can't you distinguish later? Use Heisig to memorise the shape, then learn about phonetic and semantic markers later?

Christine Tham wrote
I agree - the main thing I learnt from Heisig was how to identify primitives and how to create stories around them. I still do that, and it was a real breakthrough (my previous method was trying to learn Kanji through "osmosis." I failed miserably).

I also agree and three years of osmosis in Japan, didn`t work for me either. However I am continuing with Heisig, largely because it is accessible and motivating. In the past I tried to learn each kanji intimately, but had difficulty remembering the details and was often confused.
Now I consider a spiral curriculum approach to learning kanji, where I will be learning each kanji a number of times in the future, adding new information each time. For me RTK1 is merely the first step, an overview. I realise I will have to study a lot more to know kanji "intimately".
I am reading a lot more Japanese, which is helping a lot. I suspect writing more Japanese would be useful too. Incidentally, I also have a non-Japanese friend who regularly infers the pronunciation of kanji, despite never having done a JLPT.

Also thanks for sharing your great stories, I often use them because they are very clear and succinct. Maybe we can meet in Australia, if/when I ever return.

wrightak Wrote:Christine, I also have big problems with some aspects of Heisig's method but I think it's important to understand why it has become so popular.

Actually, I'm not that sure it's popular. It seems to be popular on this site, but you would expect that to be the case.

But I'm currently attending three different Japanese classes (one on kanji, one on grammar, and one on conversation) with 5-6 different teachers, and so far I have not met a single other student using Heisig. I have met someone else using Henshall, though. In fact, he's one of the people I referred to who said he could guess the meanings of new characters.

wrightak Wrote:After reading your posts carefully, I think you're missing one of the key points. This is that it's important to break your learning down.

I haven't missed it. That was one of the reasons why I started learning using RTK1. But I've found out I just don't learn very well when things are broken down. My approach in the past has always been to to attack it from multiple angles simultaneously. That's how I'm currently learning Japanese - attending multiple classes, reading from multiple textbooks simultaneously, and this has carried through my kanji learning as well.

wrightak Wrote:The steps are too big and you're finding that you're exhausted after covering 16 characters in 2 hours.

Well, I'm not sure the steps are too big, and besides I like of kind the total immersion approach. It is exhausting, to be sure, and my brain hurts, but it's also fun, and I find that I learn a lot. I don't consciously try to memorize any of the stuff, but I always find that the more I overload the brain, the better I recall stuff. If you think of the brain as a neural network, my approach is like trying to build as many connections and paths between facts as possible simultaneously. The brains feels fried after that, but surprisingly many of the connections stick.

That's been my approach since high school, and it seems to work for me. I don't want to boast, but I graduated not only top of class in my school, but I was one of the top 5 students entering university, and I won the university medal (which basically mean I was top of class for my major). It's my approach at work too - whenever I don't understand anything, I like to take a deep dive.

Quote:It depends on what you mean by learning how to write a character. I can tell you that most of my Japanese friends' knowledge of kanji etymology is quite fragmented. It's definitely not thorough.

What I meant was using etymology as the basis for remembering how to write a character rather than a Heisig story. That's the primary way our teacher in kanji class is teaching us how to write kanji, and it's surprisingly effective. Basically, it's a kind of "story", but a story rooted in history. The point is not to be a kanji scholar and to be well versed enough to debate a character's origins, but sometimes just having an experienced teacher go through how each character evolved is enough to make me remember how to write it afterwards.

Quote:I find this surprising. If you do the Heisig method properly and create vivid imagery and stories to help you remember the shape of each character, the recall rate should be very high (maybe 95%). When I say recall rate, I mean that after being given a prompt of some kind like おおきい or "Big", you can write down 大.

Heisig taught us the value of stories as a memorization tool, but I find a true story (eg. etymology), rooted in facts, is much more effective than a made up, or contrived, story.

I find with my way, when I see 大 I immediately hear the reading in my head, in the context of the sentence, and that's much more satisfying. In fact, I'm now finding that I can read sentences with kanji faster (at near native speeds) than I can read plain hiragana.

Quote:I don't know how many characters I now know "intimately" (i.e. more than just the shape from Heisig) but I'm sure that it's well over 1200. I can tell you for a fact that it's still not even nearly enough. What's more, anything that I can guess from this point on is nothing more than a guess. I have also been told by many people, including Japanese teachers, who say that around 1200 characters, the level needed for JLPT 2 is enough. It just isn't true.

It sounds like you don't really know the 1200 characters well enough to be able to apply the inference technique for unknown composite characters. That's the problem with breaking down the learning process - it makes it difficult to "grok" a character and make it work for you.

Through attending kanji class, I've seen the technique work with me and my classmates. So I'm pretty convinced it's true, based on personal experience.

One person I know who passed JLPT1 never really studied kanji at all - she learnt most of the characters purely by context (apparently, a lot of it from manga) and guessed the rest. If you forced her to sit down and write characters, I doubt she could write more than a 1000. But it doesn't stop her from being a fluent speaker and reading newspapers etc. Personally, I would rather be her than someone who has mastered RTK1 and can write over 2000 characters flawlessly but can't read a single Japanese sentence.

Quote:Why is it counterproductive? Why can't you distinguish later? Use Heisig to memorise the shape, then learn about phonetic and semantic markers later?

I think I've already explained this earlier. Why bother learning half facts, and then finding you have to unlearn them (see my earlier examples), when everything can be learnt in one go? For example, whenever I see (黒い) I don't think "oh, that's a computer on top of flames", or some convoluted story involving rice field, ground and cooking flames, I find it far easier to remember the character by the etymology: "soot (which of course is black) on top of a modified 赤, which is the red from a fire". But actually, increasingly I'm finding the character's picture just pops into my head, I don't need a story associated with it. So eventually the etymology will disappear, and I'll be as vague on the etymology as the Japanese people you were referring to.

Quote:This is called learning Kanji by "context", and it's actually the most effective way of learning kanji (once you get to intermediate level).

Quote:What's worse, I discovered Heisig's method is actually counterproductive in the long term because he does not distinguish between phonetic and semantic markers, hence his method is detrimental to the most effective way of learning Kanji

You say these things like they are absolute facts, when actually they are highly subjective statements. After learning 250 characters, you are not in a position to draw such conclusions about learning kanji, so anything you say is either hearsay (and some of it, I believe, from native speakers who were indoctrinated to think that way), or the hastily drawn conclusions based on a partial knowledge of the characters.

Phonetic markers are not overly useful in Japanese. I have the statistics somewhere but I think something like 20% of characters have the same pronunciation as their phonetic component. Maybe another 30% are strongly similar to the phonetic component. Let alone the fact that such things are completely useless when it comes to kun-yomi.

I can say from experience that Heisig works, and got me to upper proficiency, after seven years of struggling with other methods.

I hear a lot of people saying "Heisig fails to teach you the correct etymologies for the characters." Since when does everyone learning Japanese need to be an orthographic historian, and why to they need to learn the etymologies from the beginning?

Christine_Tham Wrote:Actually, I'm not that sure it's popular. It seems to be popular on this site, but you would expect that to be the case.

But I'm currently attending three different Japanese classes (one on kanji, one on grammar, and one on conversation) with 5-6 different teachers, and so far I have not met a single other student using Heisig. I have met someone else using Henshall, though. In fact, he's one of the people I referred to who said he could guess the meanings of new characters.

I have to comment on this part. RTK is designed for self study, and I have heard of it from numerous self study online communities. It's not really fair to say a self study course is unpopular because people in a class environment aren't using it.

Christine, my main argument is: I believe RTK is the best, fastest, and most thorough way to lay the foundation and framework for the adult kanji student to build upon. Other than that, I think we're merely disagreeing on nuances.

Here is my analogy: mastery of kanji is like a skyscraper. Everything is connected by floors and elevators and departments etc. It's a beautiful thing of logic. The average learner comes to an empty lot though without any plans and trys to build the skyscraper. They maybe start the interior decorating of the 15th floor. Then they wire a bit of the 3rd floor. Then they put in a few of the exterior windows, Etc. Everything is falling apart since it's not connected, a lot is in the totally wrong place, and it's all exposed to the elements and quickly erodes. The construction is a disaster.

Heisig provides not only the plans for what is to come, but lays the foundation erects the frame from which THEN everything else (readings, compounds, verb/noun usage, transitive/intransitive usage) is hung.

I got a degree in construction management, and even I, like the average non-builder, am always amazed at how fast building progress seems during framing. One day there's nothing, and the next a towering column of steel. It's impressive. But I know the majority of the work then comes inserting everything else. But here's the key: that frame is the most difficult and dangerous work. After it's erected, everyone else works in comfort knowing exactly where and how to place their work upon it. The concept of "framework" is of course derived directly from this.

Heisig has drawn blueprints for what most view as an impossibly complicated process. He says it's not at all and shows the user how to start with the foundation and frame, and everything else just slips inside: "remembering the meaning and the writing of the kanji--perhaps the single most difficult barrier to learning Japanese--can be greatly simplified if the two are isolated and studied apart from everything else.

Christine_Tham Wrote:What is more difficult is actually learning the characters properly - including readings, compounds, verb/noun usage, transitive/intransitive usage.

And he agrees, as he says immediately prior to the quote above. And he provides another amazing tool of logical process with RTK2 to help even further with the construction of our kanji skyscraper.

I believe, from my own experience included, learning through methods other than Heisig require brute force attack that eventually allow the learner through trial and error to eventually see and make the patterns that Heisig already alludes to. After RTK1 my kanji world has exploded. Vocabulary and reading especially is a breeze since I can deduce meaning at a glance. I can infer meanings of impossibly long compounds that were impenetrable before. Phonetic/semantic markers present themselves through the vocab I already. Everything is simply sliding into the framework.

A plain frame has its problems though, as you mentioned with his primitive names. You'll fall to your death if get off on the 里 floor thinking it contains computers. But at least you know where the floor is, and if you understand his visualization process, you know that computer is merely a path to ri, which means measurement. From there, in your future learning, you find that the kanji means village/hometown/etc.

I've read through your method from the other thread and think it's fantastic. Thank you for detailing it! I will probably employ it myself soon. I feel it will be greatly simplified though since I've already finished RTK1.

A couple other things:

I think you simply misunderstood that when I said Heisig is ONLY for adults. Using RTK1 the book requires understanding of and the nuances involved with advanced concepts and vocabulary only available to adults, though teens could work their way through.

Also, I'm sticking to my guns with dilandau and Heisig that the Japanese learn kanji by repetition. Not only do the kids kanji practice books I sometimes use (bought from a 100yen store) practice through repetition, all the worksheets and tests the Japanese language teacher who sits to me grades use some form of repetative production and recall of kanji.

Christine_Tham Wrote:In fact, I'm now finding that I can read sentences with kanji faster (at near native speeds) than I can read plain hiragana.

Credit of Heisig, I have always preferred kanji over kana, the latter of which is still quite a pain to read for me. Pure kana is illegible, even if I would otherwise be able to understand the sentence just fine and I prefer kanji forms even if they aren't used very often (ex: 綺麗).

And I'm a newb, so I'll let everyone fire off artillery for me.

Christine_Tham Wrote:I think I've already explained this earlier. Why bother learning half facts, and then finding you have to unlearn them (see my earlier examples), when everything can be learnt in one go? For example, whenever I see (黒い) I don't think "oh, that's a computer on top of flames", or some convoluted story involving rice field, ground and cooking flames, I find it far easier to remember the character by the etymology: "soot (which of course is black) on top of a modified 赤, which is the red from a fire". But actually, increasingly I'm finding the character's picture just pops into my head, I don't need a story associated with it. So eventually the etymology will disappear, and I'll be as vague on the etymology as the Japanese people you were referring to.

What you have just described is EXACTLY the same process and progression Heisig describes in his book. Whether it's computers and soil, or soot and red, you're employing the exact same method he espouses! I think you're Heisig himself in disguise!!

Heisig, RTK1, 307-308 Wrote:This course has been designed to move in steps from the full bodied story (Part One) to the skeletal plot (Part Two) to the heap of bones we call primitive elements (Part Three). This also happens to be roughly the way memory works. At first the full story is necessary (as a rule, for every kanji, no matter how simple it appears), in that it enables you to focus your attention and your interest on the vivid images of the primitives, which in turn dictate how you write the character. Once the image has strutted through the full light of imagination, it will pass on, leaving its footprints on the interstices of the brain in some mysterious way. And those footrpints are often enough of a clue about the nature of the beast to enable you to reconstruct the plot in broad outlines. Should you need to, you can nearly always follow the tracks back to their source and recall your whole story, but that is generally unecessary. The third stage occurs when even the plot is unecessary, and the keyword by itseld suggests a certain number of primitive meanings; or conversely, when seeing a kanji at once conjurs up a specific keyword. Here again, the plot is still within reach if needed, but not worth bothering with once it has fulfilled its task of providing the proper primitive elements.

There is yet a fourth stage to be reached, as you have probably realized by now, but one you ought not to trust until you have completed the full list of the kanji given here. In this stage, the primitive elements are suggested according to form without any immediate association to meaning. Quite early on, you will recall, we insisted that visual memory is to be discarded in favor of imaginative memory. It may now be clear just why that is so. But it should also be getting clear that visual memory deserves a suitable role of some sort or other, once it has a solid foundation. This is a process not to be rushed, however appealing its rewards in terms of writing fluency.

dingomick Wrote:Christine, my main argument is: I believe RTK is the best, fastest, and most thorough way to lay the foundation and framework for the adult kanji student to build upon. Other than that, I think we're merely disagreeing on nuances.

Perhaps. It may work for some people. I tried it, and don't really like it. The method of course works, it's just that I prefer a different method (which is more painful, but works better for me).

Quote:A plain frame has its problems though, as you mentioned with his primitive names. You'll fall to your death if get off on the 里 floor thinking it contains computers. But at least you know where the floor is, and if you understand his visualization process, you know that computer is merely a path to ri, which means measurement. From there, in your future learning, you find that the kanji means village/hometown/etc.

That's where I have a problem. Why go to the process of saying it's a ri (when this is pretty much a secondary, or even tertiary meaning), and then refer to it as a "computer" when you know it really means "village/hometown"? Seems like a roundabout way to me?

Of course, most of Heisig's keywords are in fact pretty good so we are talking about exceptions. But I don't really want to learn the English keywords at all, if possible. Otherwise I will forever be translating from Japanese to English, and that's a no no when learning a foreign language.

The best way to learn a foreign language is on it's own terms: one needs to *think* in Japanese - so when one encounters a word or sentence, it's the Japanese meaning and reading that comes to mind, not the nearest equivalent English keyword.

So to me Heisig is introducing an unnecessary and counter productive step, by forcing me to associate a character with an English keyword when I actually don't want to relate it to English at all.

Quote:I've read through your method from the other thread and think it's fantastic. Thank you for detailing it! I will probably employ it myself soon. I feel it will be greatly simplified though since I've already finished RTK1.

Great! Let me know how you go, and whether it works for you.

Quote:I think you simply misunderstood that when I said Heisig is ONLY for adults. Using RTK1 the book requires understanding of and the nuances involved with advanced concepts and vocabulary only available to adults, though teens could work their way through.

Okay, I think I know what you mean now. Agree completely.

Quote:Also, I'm sticking to my guns with dilandau and Heisig that the Japanese learn kanji by repetition. Not only do the kids kanji practice books I sometimes use (bought from a 100yen store) practice through repetition, all the worksheets and tests the Japanese language teacher who sits to me grades use some form of repetative production and recall of kanji.

I think my point is: Are you absolutely sure the purpose of the worksheets is to learn kanji by repetition, or is it for another purpose (ie. to improve calligraphy)? It's easy to misunderstand the context of a learning aid.

If you were a Japanese person and you see children in an English class repeatedly copying the same sentence over and over again, would you conclude the purpose was to memorise the sentence?

As I've said before, my mother used to be a primary school Chinese language teacher. She says the copying practice is primarily to improve calligraphy, not for memorization. When she's checking a kid's writing practice, she is mainly looking at how well the strokes have been executed (and in the right order) - she assumes the kid already knows the word. In fact, she recommends that to improve calligraphy, you write a character you already know a hundred times and you will find after the hundredth time your strokes will be much more elegant (providing someone corrects you along the way).

I wished I had studied Chinese from her when I was young (I'm basically illiterate when it comes to Chinese), it would make life easier studying Japanese. But I'm also glad I am learning Kanji as an adult, and I can pick and choose the learning method that best suits me.