Quote:Is it more effective than RTK1? Let's compare the difference between you, who have studied RTK1 and passed JLPT1, against an advanced student in my class [note: this is not me, at least, not yet]. You say you know 1200 characters but have difficulty pronouncing (ie. "reading") the other 800 characters. Can I suggest that if you had mastered the technique that I was talking about, you should be able to guess the meaning AND reading of any new characters you encounter (provided they are composite characters, which will be extremely likely if you already know 1200) purely by decomposing the character into semantic and phonetic markers. I have seen people, including myself (with a lot of hints), do this successfully in class, with a high accuracy rate.

This is highly inaccurate and I could go so far as to say that your teacher is being deceptive for making these claims. I've never seen the usefulness of kanji phonetics so entirely blown out of proportion.
While the kanji do share phonetic relationships, most of the time there are nothing but exceptions, and the differences are far too many for it to be of any use.
Recognizing the phonetic component of a character isn't difficult. There may be some details of this "technique" that I don't know, but most of the time I can recognize the phonetic component just fine. The problem is that there's no reasonably reliable way of guessing the pronunciation, making the phonetics essentially worthless to someone who hasn't specifically learned each character. In my experience, they are useful for remembering a reading one has already learned, and not a lot more.
And for that reason, I think little benefit can be gleaned from focusing on them at the beginning, when separating out the meaning and reading from everything else a-la Heisig makes the process rather simpler.

Quote:You could argue that the above technique does not work for kunyomi readings, and that's true. However, once you have studied at least 1000 characters, so my teacher argues, you already know nearly all the kunyomi readings. Additional characters are very likely to be composite characters with only onyomi readings. This is why the technique works AFTER 800-1200 characters, but not BEFORE.

This again is shockingly inaccurate. Flip through a level 3 or level 2 kanji kentei prep book and you will see that the kun-yomi are far from done once a student has learned the 1400 most common kanji.

Christine_Tham Wrote:You could argue that the above technique does not work for kunyomi readings, and that's true. However, once you have studied at least 1000 characters, so my teacher argues, you already know nearly all the kunyomi readings. Additional characters are very likely to be composite characters with only onyomi readings. This is why the technique works AFTER 800-1200 characters, but not BEFORE.

JimmySeal Wrote:This again is shockingly inaccurate. Flip through a level 3 or level 2 kanji kentei prep book and you will see that the kun-yomi are far from done once a student has learned the 1400 most common kanji.

Christine, certainly learning 1,000 or so kanji whilst paying attention to semantic and phonetic markers will give useful clues to the onyomi of kanji you have not studied yet. However, I cannot agree with your instructor's statement that a student will know nearly all of the kunyomi readings after studying at least 1,000 kanji.

You have referred to "Kanji in Context" (a fine book IMHO) on several occasions so perhaps you are using this textbook in your kanji class (or you at least have a copy of it). If you look through the kanji in levels 4-6 of this book (#1201-#1947) you will see that there are still many kunyomi left to be learned. I'm not even sure if one could argue that on average there are less kanji with kunyomi than in the first 1,200 kanji of "Kanji in Context."

BTW, I am still interested in learning more about the study method/concept your instructor is using to teach kanji via learning the bushu radicals together with graphically simple common kanji at/near the beginning of one's kanji studies to prepare his/her students with a strong foundation for further kanji studies. (Please see the thread I started entitled "Bushu radical + "building-block kanji" method of learning kanji".) I look forward to hearing more about that topic from you.

Yes, I too would like more info on how to use the bushu radicals. It's easy to find a list with either the Japanese or English names, but what do I do with them? Should I learn all 214? De Roo lists about the same number of what he calls "graphemes" at the beginning of his book. I am going to read through this part of the book a few times, but I can't see how memorizing this would help. I ordered Kanji ABC on Amazon and it's good to know about Kanji in Context for when I am ready. Thank you all!!!

Chaddoro_K Wrote:You have referred to "Kanji in Context" (a fine book IMHO) on several occasions so perhaps you are using this textbook in your kanji class (or you at least have a copy of it). If you look through the kanji in levels 4-6 of this book (#1201-#1947) you will see that there are still many kunyomi left to be learned. I'm not even sure if one could argue that on average there are less kanji with kunyomi than in the first 1,200 kanji of "Kanji in Context."

Yes, many characters have kunyomi readings, but you have to remember many of these kunyomi readings will not be commonly encountered. "Kanji in Context" is "required" to list all the readings in the official Jouyou list, but many of them are rarely used. Remember, any kanji in Levels 4-6 will be encountered less than 5% of the time, so any kunyomi readings that they have will be very rarely encountered. It is very likely that if you encounter these kanji, they will be in compounds so you won't be using the kunyomi reading.

We don't use Kanji in Context in my class. I bought it only because it seems to share the same teaching philosophy as my teacher.

A good hint to determine in "Kanji in Context" whether a particular onyomi reading is common or not is to scan the compounds listed with each character. You will quickly find that if a character has multiple onyomi readings, only one will be common and the other(s) relatively uncommon. So focus on remembering the common reading and learn the others by exception. Unfortunately there is no way in "Kanji in Context" to easily tell whether a kunyomi reading is common or not - suggest initially you mainly learn the ones you already know about and do the others by exception.

By the way, please have a look in my post in the other thread started by chaddoro_k:

http://forum.koohii.com/edit.php?id=8529

Basically I would recommend "Kanji ABC" if you want an approach similar to Heisig but has the flexibility to address some (most?) of the issues I have described.

Even a simple thumb through this book will allow you to start recognising phonetic markers and start differentiating them from semantic markers. Because the book deliberately groups characters sharing the same phonetic markers so they are adjacent to each other.

The good thing about Kanji ABC is that they only list the most common onyomi and kunyomi readings, rather than esoteric ones. I just flicked through the book very quickly, and it seemed that about half the characters only had on yomi listed, so that would be consistent with the theory that if you learn 1000 characters with kunyomi readings, then the remaining ones are usually used in compounds only.

It's easy to be skeptical about the technique, but I have seen it applied successfully so many times in class that I have no doubt that it works with a reasonably high accuracy rate. Of course, every now and then an exception will crop up but there are way too many students who can vouch for the effectiveness of this method that it's not just a fluke.

Remember: even if the technique is only effective 80% of the time, it's still makes a big difference, because it takes you from a reading ability of 94% (assuming you know 1000 characters) to >99%. And surely the ability to read more than 99% of newspaper content knowing only 1000 kanji is pretty good. By the way, I didn't make up these figures - 94% comes from Kanji in Context, but I did make up the 80% figure. But even if the technique was correct 50% of the time (ie., it's marginal), it would still raise the reading ability from 94% to 97%. So, regardless of whether you believe the technique is effective or not, it should provide a significant boost to reading ability.

Laura Wrote:I am going to read through this part of the book a few times, but I can't see how memorizing this would help.

I probably won't suggest memorizing ALL the graphemes (at least not initially), but having more than a passing familiarity with them will help. If you then go through and learn the characters, you will quickly discover that each grapheme imparts a certain connotation to the meaning and after a while they become very familiar to you.

Good luck in your learning! The important point about learning Kanji "in context" is be relaxed about it. Focus on learning words you already know, and everytime you encounter a new word of character, make the extra effort to look it up. No need to memorise it or anything, but take the trouble to write it down, together with notes, in a Kanji notebook.

You will probably find, as I did, that eventually you remember it naturally anyway, without even using techniques like spaced repetition. Of course, using spaced repetition and flash cards will hasten the memorization process, so if you are keen to progress quickly by all means use them.

I was in a Chinese restaurant today and I was amazed to find out that I could recognise some of the characters in the menu! So even pouring through the menu was kind of a fun exercise! Of course, I only know the Japanese readings for them, so I couldn't order using them, but it was still fun anyway.

Another technique I learnt in class yesterday was group related kanji together, each inside a circle, and draw arrows connecting circles denoting relationships. Ie. group adjectives with their antonyms, group all food related kanji and compounds together, etc. I will try using this technique this weekend and see whether it helps in memorization. I can see some real possibilities here. Basically any technique that strengthens linkages in your brain will help, because our brain remembers by connecting related concepts together.

A quick primer on differentiating between multiple onyomi readings in a character.

Based on my teacher's textbook (Japanese for Bilingualism), there are apparently 3 types of onyomi readings (stemming from the fact that Japanese contact with Chinese characters occured over a long period of time and over different regions).

These three different types are called Go-on, Kan-on and Tou-on readings. A given kanji may have up to all three. In addition, there is sometimes a fourth reading, which is normally a slight modification of either Go-on or Kan-on.

For any kanji, the dominant reading is usually the Kan-on reading, but occasionally it is the fourth type (the modified form).

The problem is, whilst my teacher differentiates between the types of readings in his textbook, it only covers a few hundred characters. I have not being able to find an English kanji textbook that specifically calls out the type of the reading for each character, hence the need to "guess" it by looking at compounds.

Hope this helps. If anyone finds a book that explains this better, please let me know.

[ファブリス]

Thanks for enlightening us, professor.

Maybe you want to read Outline of Japanese Writing System from Jack Halpern.

ファブリス Wrote:Thanks for enlightening us, professor.

Maybe you want to read Outline of Japanese Writing System from Jack Halpern.

Thanks for the link. Yes, I have read this, and I am aware of Jack Halpern's dictionaries. Unfortunately, Halpern does not really analyse or group the characters/compounds by Go-on vs Kan-on readings, which is what I was really looking for.

For example, I realise that Go-on readings are derived from the ancient Cantonese pronunciations, whereas Kan-on readings tend to be derived from the Mandarin or Pinyin pronunciations.

This is quite evident looking at the Go-On readings for the numerals: "ichi, ni, san, shi, go, roku, shichi, hachi, ku, juu"

In Modern Cantonese, the approximate pronunciation for these numerals are "yat, yi, sam, sei, ngm (this is a nasal sound that I can't transcribe accurately), lok, chat, pat, kau, sap"

You can see the Go-On readings are very closely related to the Cantonese readings, and basically represent transcribing the Cantonese readings into the Japanese sound palate, plus some modifications. As an example, "pat" (eight) was originally transcribed into Japanese as "pa-chi" (kind of like the way Japanese pronounce English words) which eventually dropped the palatization and become "hachi". "roku" is transcribed from "lok" without any modifications.

For Kan-On readings, which are dominant for phonetic/semantic composites, if you know Mandarin there are actually a set of rules for transcribing the PinYin pronunciation into the Kan-On pronunciation. These rules will no doubt help a Mandarin speaker immensely in translating compounds based from Mandarin words to Japanese, but unfortunately not very useful for those of us who don't know Mandarin.

Christine, I am very interested in the methods that you are describing. Correct me if I'm wrong but I believe that you are saying that if you learn about 1000 kanji using these methods then you will be able to make good guesses about the meanings and pronunciations of later kanji. Would it be possible for you to give us some concrete examples (maybe around three to five?) so we can get some idea of how the process would work?

For example, given 御 from 御社, assuming that this is not in the first 1000 that you learn, how would you use the component breakdown to deduce it's meaning and reading? (I just chose this kanji at random, feel free to pick other examples)

Hi Christine, thank you for your reply.

Christine_Tham Wrote:Yes, many characters have kunyomi readings, but you have to remember many of these kunyomi readings will not be commonly encountered. "Kanji in Context" is "required" to list all the readings in the official Jouyou list, but many of them are rarely used. Remember, any kanji in Levels 4-6 will be encountered less than 5% of the time, so any kunyomi readings that they have will be very rarely encountered. It is very likely that if you encounter these kanji, they will be in compounds so you won't be using the kunyomi reading.

I did look through levels 4-6 and levels 4-5 (1,832 kanji) have many useful kunyomi.

Also, please note that the levels in "Kanji in Context" are only more or less organized by frequency, they also used other considerations to determine their ordering. And they make no claim that their first three levels will give you 94%, they just call it "a broad, adequate knowledge of kanji" (p.24).

So you cannot make the "less than 5% of the time" argument for kanji in levels 4-6. Not to say you are wrong to make it, I just wanted you and others to be aware that while "Kanji in Context" is a fine book, and one that I recommend, its levels are not strictly organized by newspaper frequency.

You could, however, make this argument about the kanji that fall outside of the 1,000 most frequent kanji listed in Jack Halpern's "Kanji Learner's Dictionary". This dictionary includes data from the most recent (1998) and largest frequency study which gives the very similar figure of 95% of newspaper kanji being covered by the 1,000 most frequent kanji. (This user-friendly dictionary gives the frequency data for each kanji entry, highlights the top 1,000 kanji in red typeface, and and also lists them in an appendix.)

95 of the 1,000 most frequent kanji (as listed in Halpern's dictionary) are not in the first three levels (1,200 kanji) of "Kanji in Context". The breakdown of these missing kanji (which I wrote in my "Kanji in Context" book some time ago when I cross-referenced this sort of thing) is as follows:

51 kanji are in Level 4
23 kanji are in Level 5
5 kanji are in Level 6
16 kanji are not in "Kanji in Context"

Total: 95 high-frequency kanji (according to Halpern) not included first three levels (1,200 kanji) of "Kanji in Context"

So people who want to limit their systematic study of kanji to the 1,000 most frequent newspaper kanji might wish to study the 1,000 most frequent kanji as listed by Halpern's "Kanji Learner's Dictionary". (Or perhaps they could study the first three levels in "Kanji in Context" and then refer back to Halpern's "Kanji Learner's Dictionary" to study the remaining 95 kanji. 1,300 kanji is still a lot less than 2,000.)

Christine_Tham Wrote:The good thing about Kanji ABC is that they only list the most common onyomi and kunyomi readings, rather than esoteric ones. I just flicked through the book very quickly, and it seemed that about half the characters only had on yomi listed, so that would be consistent with the theory that if you learn 1000 characters with kunyomi readings, then the remaining ones are usually used in compounds only.

If this is true, to fully utilize this method it would seem that you would need to base your study on the 1,000 kanji with useful kun-yomi readings and not necessarily by frequency rankings. Is this how your instructor organizes his kanji curriculum? (I'm not saying this is a bad thing, I'm just intrigued.)

Christine_Tham Wrote:Remember: even if the technique is only effective 80% of the time, it's still makes a big difference, because it takes you from a reading ability of 94% (assuming you know 1000 characters) to >99%. And surely the ability to read more than 99% of newspaper content knowing only 1000 kanji is pretty good. By the way, I didn't make up these figures - 94% comes from Kanji in Context, but I did make up the 80% figure. But even if the technique was correct 50% of the time (ie., it's marginal), it would still raise the reading ability from 94% to 97%. So, regardless of whether you believe the technique is effective or not, it should provide a significant boost to reading ability.

I think the idea of using semantic / phonetic markers to bring the 95% up to a higher number is intriguing. Mainly I can see it how it could help one to make an educated guess at the onyomi of unstudied kanji, which would speed-up lookup time in a dictionary. But I don't necessarily agree that it will allow one to guess at meaning with any sort of precision. Knowing just one kanji in a compound word is not usually enough to know the meaning of that compound with precision -- unless it is a compound composed of kanji that are synonyms. At least this was my experience ten years ago when I was in Grad school pursuing a Master's in translation (Japanese-English and English-Japanese) and only had a strong command (writing, reading, vocab) of 1,200 or so kanji. Which was not nearly enough to understand the meaning of Japanese newspaper articles with the precision necessary to do accurate translations.

Now, I never was taught according to the method your instructor is using so that may have made some difference, but in general I think people give themselves more credit than they deserve for guessing at meaning when they come across unknown kanji and vocab in context within newspaper articles, etc. I know I did. And I was shocked to find out how much my perception differed from reality when it came to doing accurate translations. Grammar, usage nuance, etc., was the easy part for me. My weak point was vocabulary. (Unfortunately, during the last year of grad school I developed a life-threatening illness which also severely impaired my cognitive functioning and memory. It has been a long road back to health, reclaiming my memory, and relearning Japanese.)

Certainly, systematic study of the 1,000 most frequent newspaper kanji will provide a good foundation. But a better one might be to do a systematic study of the 1,600 most frequent kanji in Halpern's dictionary -- which covers 99%. By systematic study I mean knowing the writing and important readings of a kanji via study of useful compounds they are in. Granted it is 600 more kanji to learn, but I wish I would have had the 1,600 most frequent kanji under my belt back when I was in Grad school. Especially given the fact that when one is reading a content-dense writing system such as Japanese 4 or 5% can have a big impact. Add to this, the fact that depending on the area of one's reading interests a significantly larger (or smaller) number of kanji may be required for reading comprehension, especially if one aspires to near-native speed.

The old adage that "the more one learns and knows about something, the more one becomes aware of how much much one doesn't yet know" is especially true with foreign language learning. But that is the fun part, we have our entire lives to continue to learn more and get better and better at it.

I am still interested in reading more examples of your instructor's method and reasoning, and I thank you all for reading my two-cents worth.

Christine_Tham Wrote:These three different types are called Go-on, Kan-on and Tou-on readings. A given kanji may have up to all three. In addition, there is sometimes a fourth reading, which is normally a slight modification of either Go-on or Kan-on.
... [snip] ...
I have not being able to find an English kanji textbook that specifically calls out the type of the reading for each character, hence the need to "guess" it by looking at compounds.

Christine, thanks for the onyomi primer.

Unfortunately, I do not know of an English kanji textbook that contains this info, but Japanese monolingual character dictionaries do. For example, my Canon Wordtank G55 contains 漢字源 which includes this info for each reading. It is quick and easy to look up kanji using this electronic dictionary. (But of course the Wordtank G55 itself is a bit pricy.)

BTW, people may also be interested in reading this wiki article which states that the kan-on readings are the most common:
http://en.wikipedia.org/wiki/Kanji

Christine,
I've been lurking on this (and the predecessor thread) for a while and thought I'd share my experience FWIW. I've tried to learn kanji 3 or 4 times over the last 30 years and tried a lot of different systems (self-study and classes) for doing so. My experience is that pretty much any reasonable method is adequate for readily learning a few hundred kanji. However, scalability to thousands of kanji is a whole different ballgame. After passing about 500 kanji, there were always so many kanji that they always started to blend together in any approach that is not highly systematic with rigorous mnemonics.

After I started using Heisig, I got through all the 常用漢字 in about four months and have retained them for a year (I review about 30 per day). The extreme formal rigor of Heisig's approach really seems to have been necessary in my case (I view the strict keyword assignments as the key, BTW). This has formed a nice basis for further pronunciation and vocabulary work. Regardless of intellectual debates on its merits, I know for a fact that Heisig can be used to learn and retain thousands of characters successfully from experience.

Your approach sounds logical and intriguing. However, my understanding is that you have only learned several hundred kanji so far. I tried many logical sounding systems and got to several hundred kanji without trouble. I even evangelized some of them, as you are doing here (to my later regret). I will be very impressed with your approach when you have successfully used it to learn and retain the >1200 (your number) to 2500 kanji (my experience) needed for fluent reading. Until then, I am interested, but skeptical.

I see no reason to rush to a conclusion. If your system really works as you claim it will, you should be able to come back and state as a fact that this system works and scales. We'll be waiting here to see how it turns out.

Good luck,

Mike

wrightak Wrote:Christine, I am very interested in the methods that you are describing. Correct me if I'm wrong but I believe that you are saying that if you learn about 1000 kanji using these methods then you will be able to make good guesses about the meanings and pronunciations of later kanji. Would it be possible for you to give us some concrete examples (maybe around three to five?) so we can get some idea of how the process would work?

Unfortunately, I can't really give good examples, because i'm still very new at this so I haven't mastered enough kanji or know enough about phonetic markers to really apply the technique.

But I have seen the technique applied successfully in class - enough times so that I know it's not a fluke.

What we typically do in class is go through a JLPT1 or JLPT2 paper. The teacher will sometimes, when encountering a compound, ask us to guess what the reading is.

Sometimes what happens is that we may know one of the characters in a compound, but not the other, so essentially what he's asking us to do is guess the reading for the other character, and the overall meaning of the compound.

And what happens is that a student will say "well, the radical is X, and the phonetic marker is probably Y, so the reading is probably Z. X and Y probably implies a meaning of M so the overall meaning for the compound is probably N." And usually, that student will be right.

I did manage to do something similar almost by accident the other day with 旺文社, which is the name of a publisher. I didn't recognise the first character, but I thought the second one is "bun" from 文法 and the third is "sha" from 会社. As for 旺, the radical is obviously hi-hen (sun) and therefore the phonetic marker is "ou" (king). So I guessed the reading for the compound must be "Oubunsha" and when I checked it, I was right!

That's not a very good example because it was clear that was a stroke of luck rather than skill on my part, but it illustrates the kind of thinking that students in the class apply.

What I have now started doing whenever I learn a new character is to highlight the phonetic marker (by circling it with a pencil), and linking the phonetic marker to the relevant onyomi reading. I'm hoping over time this will build up my knowledge of phonetic markers so that I can start applying the technique.

Caution: I suspect this technique only works for relatively uncommon characters which are typically "radical+phonetic marker" type construction, and where the reading is predominantly Kan-on. For the common characters, there are too many exceptions, as JimmySeal pointed out when he tried it.

Actually it gets all sorts of muddy, I would say phonetic markers can help you breeze through about 800-900 kanji 音 readings then you could probably struggle through about 300-400 more using the method, after that I think phonetic markers probably just stop being useful. My source is RTK2 of course. Here is an example of the "struggle stage"

カク
各自　かくじ
性格　せいかく
内閣　ないかく

カク・キャク
お客さん　おきゃくさん
旅客　りょかく

ラク
脈絡　みゃくらく
落語　らくご
酪農　らくのう

リャク
略語　りゃくご

ガク
全額　ぜんがく

Not sure if rote wouldn't just be faster at this point.

Chadokoro_K Wrote:I did look through levels 4-6 and levels 4-5 (1,832 kanji) have many useful kunyomi.

Yes, but the question is how often will you encounter the kunyomi. And if you do, fine, just learn it. That's what "learning by context" really means - you learn it when you need to, not before.

Quote:Also, please note that the levels in "Kanji in Context" are only more or less organized by frequency, they also used other considerations to determine their ordering. And they make no claim that their first three levels will give you 94%, they just call it "a broad, adequate knowledge of kanji" (p.24).

I don't have Kanji In Context in front of me right now, but I'm pretty sure that they say 94% is based on knowing 1200 characters, and whilst they do not specifically say these are the exact 1200 characters in Levels 1-3, I think they strongly hint it.

Quote:So people who want to limit their systematic study of kanji to the 1,000 most frequent newspaper kanji might wish to study the 1,000 most frequent kanji as listed by Halpern's "Kanji Learner's Dictionary". (Or perhaps they could study the first three levels in "Kanji in Context" and then refer back to Halpern's "Kanji Learner's Dictionary" to study the remaining 95 kanji. 1,300 kanji is still a lot less than 2,000.)

Thanks for the information on your comparison between Kanji in Context and the Kanji Learner's Dictionary. Very useful!

I'm not strictly following Kanji in Context either, and specifically I am not going through the characters in their order, so your advice is well worth considering.

Quote:If this is true, to fully utilize this method it would seem that you would need to base your study on the 1,000 kanji with useful kun-yomi readings and not necessarily by frequency rankings. Is this how your instructor organizes his kanji curriculum? (I'm not saying this is a bad thing, I'm just intrigued.)

Not really. His lessons are pretty random, or if they aren't random, I'm not smart enough to figure out what the underlying structure is.

I would think that the first 1000 kanji probably need to cover off all the basic characters (that act as building blocks for the important phonetic and semantic markers), plus as much of the commonly used kunyomi as possible. You are right in that this is not necessarily the most popular 1000 characters. And the minimum number is probably closer to 1200 than 1000. But I have also heard some classmates say the technique started working for them from as low as 600-800 characters.

I think it should be possible to derive such a list by cross-correlating between Kanji ABC and the Kanji Learner's dictionary, and also referencing Kanji in Context. Anyone brave enought to try? :-) Hint: not me!

Quote:I think the idea of using semantic / phonetic markers to bring the 95% up to a higher number is intriguing. Mainly I can see it how it could help one to make an educated guess at the onyomi of unstudied kanji, which would speed-up lookup time in a dictionary. But I don't necessarily agree that it will allow one to guess at meaning with any sort of precision.

I think this is possible if you are very familiar with the Bushu radicals and have a good grasp of how they influence meaning, and also if you are very familiar with the phonetic markers (because they may also have meaning connotations). So effectively, if you look at a character that is, say, a radical+phonetic marker, and you can "intuit" what the meaning is from the combination alone. I don't guarantee that this works 100% of the time, but I have seen it working for my classmates.

Quote:At least this was my experience ten years ago when I was in Grad school pursuing a Master's in translation (Japanese-English and English-Japanese) and only had a strong command (writing, reading, vocab) of 1,200 or so kanji. Which was not nearly enough to understand the meaning of Japanese newspaper articles with the precision necessary to do accurate translations.

Well, you are certainly much more experienced than me. You have to remember, my posts are based really on no more than 4-6 weeks of serious Kanji study, so much of what I say is probably questionable.

You are right in that there is a huge difference between guessing the meaning and knowing precisely the meaning in order to do a translation. Certainly the technique will never be precise enough to do that. However, it should allow someone to read a newspaper, and get a rough context of the meaning of the unknown characters. Over time, the meaning will solidify from repeated exposure, or the person can simply look up in a dictionary to confirm.

That's how I built up my vocabulary in English, primarily through reading lots of novels (at one stage I was reading one a day). I never looked up words in a dictionary, but eventually from constant exposure to a variety of words I picked most of them up. I think all my teacher is saying is that this is also possible with Kanji.

Quote:(Unfortunately, during the last year of grad school I developed a life-threatening illness which also severely impaired my cognitive functioning and memory. It has been a long road back to health, reclaiming my memory, and relearning Japanese.)

I'm sorry to hear that. For what it's worth, your posts are very well written, and you obviously have a deep knowledge of Japanese. Best wishes for the future.

Quote:Certainly, systematic study of the 1,000 most frequent newspaper kanji will provide a good foundation. But a better one might be to do a systematic study of the 1,600 most frequent kanji in Halpern's dictionary -- which covers 99%.

Of course, serious learners will want to study all the Jouyou plus the popular non-Jouyou systematically. Just like some Japanese try to learn English by memorizing an entire dictionary.

But my motivation is a little different - i don't care if I never master Japanese and subtle things will remain beyond my grasp. My goal is to get to "a broad, adequate knowledge" to enable me to read simple stuff and watch a movie. Once I get there I will probably stop. I will never sit for the JLPT and I don't plan to work or live in Japan, so my goals are probably different from others in this forum.

mspertus Wrote:I see no reason to rush to a conclusion. If your system really works as you claim it will, you should be able to come back and state as a fact that this system works and scales. We'll be waiting here to see how it turns out.

I think it's important to realise I don't actually have a "system" as such, or if I do then I'm clearly changing it as I go along. And even if I do have a "system", I'm not promoting it. I am a beginner - my serious Kanji study started around 6-8 weeks ago. I clearly do not know enough to design a "system" for studying Kanji!

I think what others are referring to in this thread as my "system" is a series of posts I wrote a few weeks ago, when I said I was concerned about some aspects of learning kanji, and whether RTK1 would adequately cover these aspects. I then developed an alternative studying "method", but really all it was was a specific process for studying from a number of textbooks simultaneously rather than just one. This is a "beginner's method", suitable for the first few hundred characters. It is clearly not scalable for thousands of characters.

Then I started talking about techniques and advice that I have picked up from class. I have to make it clear that these are things I have picked up, they don't form part of my "system" as such (even if I have one). For example, whilst I'm intrigued by the technique of guessing a kanji's meaning and reading from the semantic and phonetic markers, clearly I don't know enough kanji to actually apply this technique so I am not speaking from personal experience.

I do agree with most of what you are saying. It's too early at this stage for me to claim (even if I wanted to) that I have found an optimal system for studying kanji. The realization I had from the weekend is that the system/approach for studying the first 500 characters probably needs to be different from the next 500, and the 1000 after that etc. For example, exhaustively studying meaning and readings and etymology which is what I am currently doing is probably not sustainable beyond the first 1000. Now that I have reached 300, I am already thinking of varying my approach.

But I don't think I will be doing any systematic study after the first 1000. I'm hoping to cover the rest more or less ad hoc, by context. And if I occasionally get confused between characters, then so what? I don't intend to sit for any exams or work in Japan, so that's fine. And my aim is not to be 100% proficient in Japanese, just enough to stumble by and read newspapers, watch movies and casual conversation is fine by me.

Chadokoro_K Wrote:Unfortunately, I do not know of an English kanji textbook that contains this info, but Japanese monolingual character dictionaries do. For example, my Canon Wordtank G55 contains 漢字源 which includes this info for each reading. It is quick and easy to look up kanji using this electronic dictionary. (But of course the Wordtank G55 itself is a bit pricy.)

Thanks for this useful information. I might think about buying a Japanese monolingual dictionary that distinguishes between Kan-on and Go-on etc. Next time I go to Kinokuniya I will check this out.

I know that Cantonese and Mandarin *sounds* very different, I'm wondering if it's possible to guess a Kan-on vs Go-on reading just from the sound. Probably not, but it may be possible to guess whether a compound has a Mandarin or Cantonese origin.

My teacher says Go-on is not very common and tend to be used in Buddhist texts.

Hi Christine,
Thanks for the thoughtful response.

Christine_Tham Wrote:I think it's important to realise I don't actually have a "system" as such, or if I do then I'm clearly changing it as I go along. And even if I do have a "system", I'm not promoting it. I am a beginner - my serious Kanji study started around 6-8 weeks ago. I clearly do not know enough to design a "system" for studying Kanji!

I think what others are referring to in this thread as my "system" is a series of posts I wrote a few weeks ago, when I said I was concerned about some aspects of learning kanji, and whether RTK1 would adequately cover these aspects. I then developed an alternative studying "method", but really all it was was a specific process for studying from a number of textbooks simultaneously rather than just one. This is a "beginner's method", suitable for the first few hundred characters. It is clearly not scalable for thousands of characters.

Bear in mind that the readership of this board are interested in strategies for learning thousands of kanji, so your techniques are naturally going to be evaluated with that in mind. Unfortunately, even if you just want to struggle through newspapers, watch movies, etc., you may need to know thousands of kanji as well.
For example, the 1000 most common kanji only account for 90% of magazine text. This sounds like a lot, but it basically means that you will be confronted with an unknown kanji on almost every sentence, which isn't much fun (rikaichan could change this).

Heisig makes this point explicitly, saying there is no value to learning a small amount of kanji, which is why he makes no effort to teach the most common kanji before the rarest kanji. The truly remarkable thing about Heisig is that it does scale. In order to accomplish this, Heisig had to make certain tradeoffs, but they were very carefully thought out, and he gives detailed information about how to come back and fill in the holes after completing RTK1 (not to mention RTK2!).

Christine_Tham Wrote:Then I started talking about techniques and advice that I have picked up from class. I have to make it clear that these are things I have picked up, they don't form part of my "system" as such (even if I have one). For example, whilst I'm intrigued by the technique of guessing a kanji's meaning and reading from the semantic and phonetic markers, clearly I don't know enough kanji to actually apply this technique so I am not speaking from personal experience.

Heisig's RTK2 is the most thorough study system for phonetic markers that I am aware of, so I hope you are not implying that Heisig neglects phonetic markers. He does discuss in depth why he thinks it is important to learn the meaning and writing of the characters first, but would be the first to admit that learning them alone is inadequate. He also goes on to say that there are severe limitations on the value of the phonetic markers and reflects this in the title of the book. In general, I have found the phonetic markers to be less useful in practice than in theory (Dilandau's post has some great examples from RTK2).

mspertus Wrote:Unfortunately, even if you just want to struggle through newspapers, watch movies, etc., you may need to know thousands of kanji as well.
For example, the 1000 most common kanji only account for 90% of magazine text. This sounds like a lot, but it basically means that you will be confronted with an unknown kanji on almost every sentence, which isn't much fun (rikaichan could change this).

Actually, 90% is pretty good, provided you use the semantic/phonetic decomposition technique to help you guess the meaning and reading of kanji that you don't know.

And Kanji for Context says 94% is covered by 1200 characters.

The point I was making in earlier posts was that just knowing the context of what you are reading and using the above technique can increase your reading rate to roughly 97-99% (depending on how effective you believe the technique to be).

As an example, we have a saying when I was studying at Sydney University more than 20 years ago: "If you haven't started studying for your final exams by the time the jacaranda tree in the Main Quadrangle blooms it's too late."

Now, assuming that you don't know what "jacaranda" is, you already know from the context that it is a type of tree, and furthermore it blooms at a specific time in the year. You actually don't need to know any further at this stage, and you can continue reading.

Of course, if you do find out more about the jacaranda tree, you will realise that it's a very beautiful tree, and the flowers in particular are a beautiful purple. A few weeks after blooming, coinciding with the end of final exams, the flowers rain down and can cover the ground. So this extra knowledge gives you additional insight into the sentence, because there is the additional poignancy about the flowers falling as if the tree is crying for the students who have failed the exams.

The second word that you may not possibly know is "Quadrangle" but again its meaning is relatively clear from context. I certainly did not know what this word meant when I entered university, but I had no problems picking it up and didn't refer to the dictionary at all.

This is the same for kanji. Furthermore, the technique for semantic and phonetic decomposition allows you to make a stab at guessing meaning and reading, which further helps. No one is saying this technique is perfect, or 100% effective, or that it replaces systematic learning. However, it allows the reader to move on without necessarily looking up a dictionary. Sure, if the sentence was important, you would want to look up a dictionary. But if it was a novel, you may not want to.

Quote:Heisig makes this point explicitly, saying there is no value to learning a small amount of kanji, which is why he makes no effort to teach the most common kanji before the rarest kanji.

But you must remember, when Heisig wrote RTK1, he had only spent a month in Japan and knows no Japanese apart from a conceptual understanding of some grammar points, and some time studying etymology. Heisig was obviously stressed by his teachers that the Jouyou list is very important, it's an official list, and the sooner you learn the whole list the better. And all of that is certainly true.

I have to weigh what Heisig says against what experienced Japanese language teachers and other textbooks say. The point I was making about knowing 1000 characters and be able to guess the rest by context is advice I have heard from a lot of people who should know what they are talking about. Including my mother, who was a Chinese language teacher in primary school before I was born - she also makes this claim (in fact, she says you only need 800), and explains the rationale in the same way. She would know, she has seen lots of children progress from no kanji to reading newspapers. And remember, in those days there was no rikaichan and dictionaries are not that easy to carry around.

Think about it this way: do you really believe the majority of Japanese know the entire Jouyou list really well? That they have a very good recollection of every word they have studied? Clearly this is not true, we see instances of Japanese struggling with kanji every day. A lot of them write difficult words in hiragana. But, this does not stop them from being able to read newspapers. In fact, my teacher asserts that most Japanese will struggle to remember more than about 1200 kanji in the Jouyou list.

There is a difference between having to study Jouyou in order to pass something like JLPT1 and what you actually need to study in order to read newspapers. I know people who have studied all of Jouyou and passed JLPT1, but continue to struggle reading newspapers. It's because they have optimized their study to pass the specific requirements of JLPT1. On the other hand, I know people who have studied kanji completely by context. They probably only know about 1000 characters, but have no problems reading newspapers. I want to align myself with the second group, not the first group.

Quote:The truly remarkable thing about Heisig is that it does scale. In order to accomplish this, Heisig had to make certain tradeoffs, but they were very carefully thought out, and he gives detailed information about how to come back and fill in the holes after completing RTK1 (not to mention RTK2!).

Yes, but so does learning by context. I know you are skeptical about this, but there are millions (billions?) of Chinese and Japanese speakers who have learnt this way. If learning by context does not work, then all these people will be illiterate. And yet Japan has one of the highest literacy rates in the world, despite the complexity of the written form of the language.

I have been examining several Japanese grade school textbooks. The latest I have bought is a set of three books covering Hiragana, Katakana, and Kanji Grade 1-2. Looking at these textbooks, along the ones I mentioned earlier in the thread that was closed down, it seemed to me the Japanese learn kanji through fairly sophisticated techniques. Word games help reinforce the reading, comprehension exercises reinforce the meaning, etymological stories help reinforce how to remember writing the characters, pictures help break down the sub-components and songs/poems also work as memorization aids.

All this work at the subconscious level - the child probably thinks that it's just rote memorization, but when I look at the textbooks with adult eyes it's clear there's a lot more than just rote memorization here. No wonder the literacy rates are so high - I wish I had learnt English this way!

Christine_Tham Wrote:Heisig's RTK2 is the most thorough study system for phonetic markers that I am aware of, so I hope you are not implying that Heisig neglects phonetic markers.

Yes, I have mentioned that I own a copy of RTK2. RTK2 clearly does not neglect phonetic markers, but my point was RTK1 does. In fact, RTK2 does not really build on RTK1 (which many people have commented on), and to me that's a shame. It would have been nicer if RTK1 had included phonetic markers, then RTK2 could leverage the same stories and say: "by the way, you already know all the phonetic markers even though you don't realise it yet." That would have been powerful.

Quote:In general, I have found the phonetic markers to be less useful in practice than in theory (Dilandau's post has some great examples from RTK2).

I actually agree with most of what dilandau was saying, but it doesn't negate the fact that it's a useful technique to reduce the number of dictionary lookups. Dilandau says it probably only works for 700-800 characters, I would even be conservative and say within the Jouyou set the number may even be less than that - my guess simply by scanning Kanji ABC is that it's at least 500 but probably not much more.

But even 500 is very effective, because if you start with a base of 1000, being able to infer meaning (even imprecisely) and reading for an additional 500 is a useful technique. And if it is 700-800 as dilandau claim, then that's even better.

And the good point about this technique is that it scales really well, In fact, the more kanji you know, the more likely it is that you will be able to apply the technique effectively. This is because the less common the kanji, the more likely that it has been imported from China and retains an almost pure "Kan-on" only reading, and the more likely it is to be used purely as part of a compound.

This technique is actually a lot more effective in Chinese, which has a purer application of kanji (or hanzi if you prefer) and a lot less exceptions. That's why my mother can claim "all you need is 800 characters" but for Japanese I think it's probably 1000-1200 due to the exceptions and multiple readings and kunyomi.

Hello Cristine,
Interesting as usual. I'm wondering if part of the difference of opinion about the usefulness of semantic and phonetic markers stems from the different experiences people have had not just with learning kanji, but with encountering kanji in their environment. Is it possible that these markers are a lot more useful for people who grew up with kanji in the environment, but did not study it? One's general knowledge is a huge factor in the ability to predict the meaning of a word when reading. So, for example having an internalized if not fully formed sense (my own case) of the patterns and rthythms of Japanese language and having lived among kanji most of my life probably helps a lot. My Chinese speaking friends, even those who can't read Chinese, seem to have a pretty big advantage, as well. This struck me when someone posted a compound word as a sort of challenge to you. I'm pretty sure the word meant something like "your company." The first kanji was an honorific. How did I know that? It was partly De Roo's graphemes, but it was also partly something else.

The learning in context argument just seems obviously true to me. I have come to understand lots of kanji, by looking at couple of Japanese theater magazines I subscribe to. So, now I am exited to see if I can use phonetic markers to more accurately guess at the pronunciation of the harder words. And I would rather not wait and see if I make through RTK 1 to start.

Laura Wrote:Hello Cristine,
Interesting as usual. I'm wondering if part of the difference of opinion about the usefulness of semantic and phonetic markers stems from the different experiences people have had not just with learning kanji, but with encountering kanji in their environment.

I've been thinking about that myself. I think part of the reason is that people on this forum have invested time and effort studying using RTK1, and therefore more inclined to be skeptical of a system that effectively is advising the opposite. After all, if you are using a system, and it works well, then any system that says the opposite must be wrong, right? :-)

I was almost tempted to say this could be a male/female thing as well, because the technique sounds suspiciously like "intuition" and most men are uncomfortable trusting their intuition. But then again, the most successful practitioners of this technique in my class are all male, so that can't be true.

Knowing Chinese I think is a huge initial advantage. One Chinese student in my class says she can often guess the context of any paragraph just from the kanji contained in the paragraph. But I think knowing Chinese is also a barrier - it can make a student lazy and avoid learning readings, which will be damaging in the longer term. Also, many kanji have different meanings in Chinese compared to Japanese. I found this out on the weekend when I was at a Chinese restaurant and I excitedly pointed out kanji I can recognise on the menu. My mother in law and I then swapped readings - I told her the Japanese reading and she gave me the Mandarin reading and meaning - we discovered quite a few characters have subtle meaning differences, and sometimes the meaning has changed substantially.

Quote:So, now I am exited to see if I can use phonetic markers to more accurately guess at the pronunciation of the harder words. And I would rather not wait and see if I make through RTK 1 to start.

Here's what I'm starting to do now. Every Kanji I learn i'm going to put a circle around the phonetic marker and link it to the reading. Separately i will be building a list of phonetic markers I find, and grouping the kanji that share the same marker. I'm also starting to group kanji I know into radicals. So hopefully I will have the collateral to allow me to use the technique sooner rather than later!

Laura Wrote:This struck me when someone posted a compound word as a sort of challenge to you. I'm pretty sure the word meant something like "your company." The first kanji was an honorific. How did I know that? It was partly De Roo's graphemes, but it was also partly something else.

By the way, thanks for posting on the meaning of 御社. I didn't want to respond to that specific example because it wouldn't be the sort of character you could easily apply the technique to, since the phonetic marker is unclear. But you showed us that you could deduce the meaning anyway from what you know. Brilliant!

But the interesting point is even if someone just totally ignored the 御 and just moved on with the assumption that the compound had something to do with "company" it would probably have been okay. Thus illustrating that not knowing the meaning or reading of a character is not necessarily a barrier! And if that was the only unknown word in a paragraph or sentence I'm pretty sure the context would have made it clear anyway.

Okay, how do you know what the phonetic marker is? Not every kanji has one(assuming I understand what you mean by phonetic marker). For 務, how would you know the marker is 矛 and not 力(assuming this is right)?
And how can you guess the meaning from the radical? Several kanji might have the radical for water on the left, but they have different meanings, so how can you guess? I can see if you know the word and half of the kanji compound it's in (knowing かせん　and then recognizing 河川　as it because you know 川　the reading for 可).
汁 浴 沖 泳 洞 <- most of these are related to water or liquid, but you can't know what they mean by that alone. If you see them in a compound, even if you knew the on-yomi of these kanji(these five are very easy), how would you know what it meant? Why not just learn them?

I personally like that with Heisig/mnemonics, all parts of the kanji will help me remember what it means. It doesn't stop me from learning phonetic markers as well.
The way I see it, if I see a new kanji word, even if I know the kanji in it, I'd have to look it up to be sure. I would hate to assume and then get the wrong reading stuck in my head. The use I see in semantic and phonetic markers(or whatever), is that once I look it up, it really makes sense and I can remember it easily. But before I look it up, my guess is just a guess.

Laura Wrote:This struck me when someone posted a compound word as a sort of challenge to you. I'm pretty sure the word meant something like "your company." The first kanji was an honorific. How did I know that? It was partly De Roo's graphemes, but it was also partly something else.

That someone was me. I hope that you will believe me when I say that the purpose of my post was not to challenge anyone, it was to understand the method better.

The fact that you could guess the meaning was great. Can you please go into more detail about how you did it? Or was there no specific method and you just used intuition, which apparently us men might not be good at. Are you sure you just hadn't met the character before and it was in the back of your memory?

I chose 御社 completely at random and I made it clear in my post that any other word could be used to explain the character break down method with more clarity. The word I chose was a bad example because its reading is irregular but I was curious to see if the meaning could be deduced.

I have to say that the fact that Christine can't give any concrete examples at all, with any kanji, reflects very badly on the credibility of her posts. I was really interested in the phonetic and radical breakdown methods and I wanted to know more about it, which was why I posted. The 旺文社 example was interesting but all it illustrated was that you managed to guess the reading. The meaning of the word doesn't really apply here because it's a name. Christine, I realise that you're just beginning with your kanji studies and this may be a lot to ask, but would it be at all possible for you to ask your advanced classmates for a few solid examples of the guessing techniques that they use? How can you expect anyone to place faith in these methods if they can't see them working?

Christine Wrote:That's how I built up my vocabulary in English

What is your first language? Since your mother is a Chinese teacher, and given the above quote, is it Chinese? If so, I think that the intuitive guessing of Chinese characters will come much more easily to you and to other people whose first language is related to Chinese. When I reached the advanced level Japanese classes at the school I studied at, I found that I was increasingly surrounded by fewer western learners and more Asian learners. Most of them were Korean. All of the Korean and Chinese learners were much quicker to pick up the meanings of new words. I would look up words in a dictionary and be given a pathetic translation into English. When I tried to use the words in sentences, I frequently got it wrong and got strange looks from my teachers and classmates. I'm not talking about simple nouns that translate directly, I'm talking about non-western concepts like 遠慮 or 素直 etc. I distinctly remember trying to create several sentences using the word 素直 and failing every time. However, I'm sure that if you look that word up in a Korean <-> Japanese dictionary then the translation will be exact. The same goes for the fact that I will be able to learn European languages far more quickly than my Japanese friends.

What I'm saying is that even though there are Chinese speaking people who can't read Chinese characters, their knowledge of the spoken language will mean that they are far better equipped to understand the meaning of the characters if they choose to learn. The same goes for Korean people who don't use Chinese characters very much but their language is full of words that come from Chinese character words. e.g. I know that the word 家族 translates to a very similar sounding word in Korean, something like kajoku or something. The concepts are identical between those languages but does 家族 really accurately translate as "family"? I would contend that it doesn't and that they are similar but there are nuances in both "family" in English and "家族" in Japanese, that make their meanings different.

I hope that the above hasn't rambled too much. I believe that any guessing done when reading kanji will be more accurately accomplished by those whose first language is related to Chinese in some way.

Quote:For Kan-On readings, which are dominant for phonetic/semantic composites, if you know Mandarin there are actually a set of rules for transcribing the PinYin pronunciation into the Kan-On pronunciation. These rules will no doubt help a Mandarin speaker immensely in translating compounds based from Mandarin words to Japanese, but unfortunately not very useful for those of us who don't know Mandarin.

I myself am able to speak Mandarin and Cantonese, but I haven't been able to find rules from pinyin->kan-on like you mentioned in the quote above. What I do see is that some words with the same pronunciation in mandarin sometimes have the same kan-on readings. It would be a great help if you could give more details on this set of rule. Thank you in advance.