While working through (phonetically) my Japanese reader this morning it occurred to me that I would gladly have all my RTK1 knowledge erased from my brain if in return I were given immediate knowledge of 10,000 (or enough to say just about anything) Japanese words purely phonetically. And then AFTER that I would do RTK1 and practice reading/writing kanji. Also RTK2 would make more sense.

I'm not saying that doing RTK1 a year and a half ago was useless, I just find that the knowledge I gained isn't helping learn/remember Japanese words.

Anyone have similar feelings?

Your counterfactual here.

I learned a few hundred kanji and dropped it in favor of learning vocabulary instead. I learned a few thousand words and drilled a few thousand sentences with furigana to get a feel for grammar usage, etc. I've recently been listening to NHK easy and although I can't follow along at all, I can recognize most of the words. I feel like if I could recognize the words about 3x faster, I could follow along pretty good.

The problem I'm running into is that I would really like to have the script in front of me while I listen, but there's a lot of kanji. I'd also like to start reading native material, but pretty much any native material I'd like to read has some kanji. And I'd like to do subs2srs. Of course I could run everything through a program and get passible furigana and I may still do that. Another thing is that I'm considering taking the N3 in December, so now I'm feverishly catching up on my kanji because my vocabulary is useless for that test without kanji.

The way I see it now is that the language is a complete system. I'm not sorry that I learned a few thousand vocabulary because they are common words that I need to know. But I also need to know about a thousand kanji to do well on the N3 or read anything so I'm learning those now. No matter how you look at it, language learning is a lot of memorization. In my opinion, memorizing A before B is the same as B before A. What makes more sense to me is learning both in parallel so that you could get some use out of it while you are learning. For instance, if you were to learn the vocabulary and kanji required to read a short story and then read the story, it would really drive home what you've learned. Even better if you can listen to a native reading of the short story.

Of course you should take your trade if it were given because 10k vocabulary words would take longer to learn that a 1-2k kanji. But 2k kanji for 2k vocab?? Same diff. I don't see any advantage learning one before the other unless you have no interest in learning to read.

John555, if sounds to me like you're relying on the furigana provided in your workbook and haven't yet begun to see the kanji in their "complete" forms. When I finished RTK, I immediately tried to read a novel with barely any furigana. Of course, I had to look up at least five words a sentence, and I soon realized how pointless it was to be able to recognize kanji but not read them. Most of the time, Heisig's keywords were useless, too.

Now, before I go on, let me say this: I don't regret doing RTK. In fact, it's been one of the most beneficial things I've ever done in studying Japanese.

So, after realizing how futile my attempts to read a book were, I decided to do RTK2. Signal primitives are a clever way of categorizing kanji, and there are some readings I'll never forget because of what I learned. Was I able to memorize every definition of every word in the book? Absolutely not. However, I did learn the various nuances of many kanji. More importantly, I began to seem them as characters to be read rather than symbols to be recognized.

After that, I began the Core 6K with an understanding of kanji most people lack when they start grinding vocabulary. When I review the deck, I hardly ever think of Heisig's keywords. In fact, the only time I do think of them is when I review on the main website. But because of my knowledge of their forms, learned through RTK, I recognize all of them and almost never mistake one for another. Because RTK2 taught me how to apply readings to kanji, I absorb even kunyomi without much trouble.

After I hit 5,000 words, I dived into native material. I'm more than halfway through the JRPG Tales of Zestiria, and I haven't run into any problems reading kanji. Most of my difficulties have come from trying to understand complex sentences.

Either way, I don't think I could have achieved this level of kanji fluency without practicing every day. And more importantly, practicing without furigana.

Anyway, that's been my journey. Maybe reading it will help someone.

john555 Wrote:I'm not saying that doing RTK1 a year and a half ago was useless, I just find that the knowledge I gained isn't helping learn/remember Japanese words.

Perhaps it would if you didn't insist on using inefficient methods

Last weekend I added 450 difficult words all involving new kanji readings: pic. This would have been great deal harder without having studied the kanji meanings in advance.

john555 Wrote:While working through (phonetically) my Japanese reader this morning it occurred to me that I would gladly have all my RTK1 knowledge erased from my brain if in return I were given immediate knowledge of 10,000 (or enough to say just about anything) Japanese words purely phonetically. And then AFTER that I would do RTK1 and practice reading/writing kanji. Also RTK2 would make more sense.

I'm not saying that doing RTK1 a year and a half ago was useless, I just find that the knowledge I gained isn't helping learn/remember Japanese words.

Anyone have similar feelings?

As you might know, I advocate first learning lots of vocabulary using roumaji, before starting on kanji.
But I would also advice not to wait longer then half a year to start with kanji.
I would also advice not to waste too much time on the idea you have to learn radicals first. Or rather: start with radicals, provided the radicals you learn are kanji in their own right. Don't bother about pure radicals in the beginning.

I never learned a "kanji" as such in my life. I learned, and learn, vocabulary with its associated kanji.

I don't recommend using Romaji at all. Learn the kana in the first two weeks or so. Then start on kanji.

Why? Because Romaji does not accurately represent Japanese so you will be learning wrong.

Why kanji? Because they show you how words are made up. If you try to learn purely phonetically you will miss a lot of the things that make the words what they are. At first the kanji will feel strange and meaningless, but as you encounter the same friends in word after word they start to make sense.

I do think awareness of the components ("radicals") is important (but again, learn them in context, not in the abstract) and RTK-style mnemonic stories can be very useful.

I realize people work differently, but I absolutely feel that learning kanji organically, in the words they make up, learning the readings one at a time, as you encounter them is the most natural way.

Some might argue that Japanese children learn kanji as kanji. They do. But they know the words already. From the minute they learn them they are associated with meanings, readings and actual words. In fact the reason for the "anomaly" of some early school kanji being much more complicated than later ones is that children learn the kanji for words they are familiar with, not for words they don't know. So it is the knowledge and use of the word, not just the complexity of the kanji, that determines a kanji's grade level.

That is because kanji without words (literally) make no sense.

yogert909 Wrote:Of course you should take your trade if it were given because 10k vocabulary words would take longer to learn that a 1-2k kanji. But 2k kanji for 2k vocab?? Same diff. I don't see any advantage learning one before the other unless you have no interest in learning to read.

My thought exactly. If you want to read, however you put it, Japanese vocabulary is three-fold (kanji, meaning, pronounciation), and since most of us are only used to two-fold vocabulary (where the visual image is the mirror of the sound image), you'll have to invest extra-time learning that extra kanji layer sooner or later. I know the counter-argument that natives perceive them as "letters", but it's a brain shorthand, even for them in the beginning, kanji never had the same status as our alphabetical letters. Heisig's development in this domain is so obvious in hindsight that it's a mystery that no one before him thought of it.

CureDolly Wrote:Why? Because Romaji does not accurately represent Japanese so you will be learning wrong.

Actually romaji can represent Japanese with perfect accuracy; it's just not the writing system used to represent Japanese in most cases.

To quote one example of the confusion caused by Romaji:

From Wikipedia:

"The eba provisional conditional form is characterized by the final -u becoming -eba for all verbs (with the semi-exception of -tsu verbs becoming -teba)."

~つ verbs becoming ~てば is not an exception, semi or otherwise, unless one is thinking in Romaji and specifically in the Hepburn transliteration system. The ば provisional form is perfectly regular with the endings from the う-row regularly shifting to their equivalents in the え-row.

In my experience Japanese people tend to romanize た、ち、つ、て、と as ta, ti, tu, te, to etc. (which is why the site hukumusume is not fukumusume).

This Romanization does away with the "exception" cited by Wikipedia but introduces other problems which some consider greater, which is presumably why Hepburn has become standard in the West. No Romanization system is a perfect representation of Japanese. There is always a trade-off in representing one sound-system with characters designed to represent a different sound-system.

CureDolly Wrote:There is always a trade-off in representing one sound-system with characters designed to represent a different sound-system.

But kana are just simplified kanji. 津, つ, tsu, and tu all represent the same sound. It's true that romaji can sometimes mislead foreign learners, but so can hiragana. The belief that each hiragana maps to exactly one sound is a widespread misconception that causes long-term harm to many people's pronunciation.

If you don't know how a native speaker pronounces the language, no writing system will solve that problem for you. If you do know how a native speaker pronounces the language, it shouldn't matter whether you're seeing "tsu" or つ.

I am not entirely sure what the fact that the shapes of kana originated with kanji really has to do with this. The syllabic system is actually structurally closer to the South Asian Siddham syllabary, on which I believe it was based. It was chosen and adapted to represent the sounds of Japanese.

What you say is correct, of course, in that no written script can teach pronunciation of itself, but Romaji certainly tends to confuse people about how the language is structured as the Wikipedia piece I quoted demonstrates. In this case, Kana does represent the way the language regularly conjugates and Hepburn Romaji doesn't. It produces apparent "exceptions".

More to the practical point, people who really can't be troubled to learn the writing system used by native Japanese are the people most likely to be misled by false analogies with English structure and pronunciation. But then I suppose the argument is rather academic since someone who does not intend to learn kana is unlikely to proceed very far in Japanese anyway.

The question isn't about whether someone should ever learn kana/kanji. Of course they should. The question is whether limited use of romaji at the beginning of someone's study can help them, or at least not harm them. I managed fine even though I didn't learn kana for 4 months.

This doesn't really have anything to do with the topic, but:

Quote:The syllabic system is actually structurally closer to the South Asian Siddham syllabary, on which I believe it was based. It was chosen and adapted to represent the sounds of Japanese.

Both Chinese and Siddham were syllabic systems, but I think the Japanese were exposed to Chinese several centuries before Siddham. By the time Kukai brought back Siddham, the Kojiki and Man'yoshu had already been written, both of which write part of the text in Chinese characters that are used only for their sound to represent Japanese words. They didn't really consciously choose the kana system as much as it fell into their laps. This is how writing has worked for the majority of the world; except for China and Sumeria, everyone just adapts whatever writing system they discover to suit their own language. But then inevitably the language changes and the writing system doesn't, leaving all writing systems with various degrees of disconnect from the languages they represent. Japanese kana have the benefit of very recent reform (1940s), and so the kana are a near-perfect phonetic system for Japanese.

As you say, the kana are a near-perfect phonetic system for Japanese, so it seems to me a good idea to learn them early on. But I do agree that Romaji in the very early stages won't do any immense harm. I still would not recommend it as best practice though.

Another reason for my recommendation of early kana is that I recommend early kanji too. Now this is purely my way of going about things, and of course it isn't everyone's and probably not that of most folks here if they are doing what Heisig-sensei recommends and learning all the joyo kanji before touching anything else.

As I said, my approach is to learn words, not kanji, but to learn kanji within words from very early. Since one can't very well mix kanji and Romaji with no kana, the first step in this is to learn kana.

I was recently affected by the evils of romaji as it happens.

I used to play 囲碁 a number of years back, and still have an interest in the game. Go terminology in English is almost entirely borrowed from Japanese. One of the commonest terms used is 'joseki', meaning a standard sequence of moves, e.g. from an openings book.

The other day I started reading a j-book on Go and learned that the Japanese term is 定石[じょうせき], so it would be more properly romanised as jouseki.

I guess it could be argued that 'joseki' is a derived form but ... nah. Another romaji fail

john555 Wrote:While working through (phonetically) my Japanese reader this morning it occurred to me that I would gladly have all my RTK1 knowledge erased from my brain if in return I were given immediate knowledge of 10,000 (or enough to say just about anything) Japanese words purely phonetically. And then AFTER that I would do RTK1 and practice reading/writing kanji. Also RTK2 would make more sense.

I'm not saying that doing RTK1 a year and a half ago was useless, I just find that the knowledge I gained isn't helping learn/remember Japanese words.

Anyone have similar feelings?

Yes, certainly similar (although not identical) ones. Instead of RTK1 as it is now (そのまま, as I believe the Japanese would phrase it), I would have preferred:
1. a clear division between usefulness linked to whether you're a beginner, upper beginner, lower intermediate (etc.) learner, so that you can tailor your selection of Kanji to your current phase of learning. This can prevent some of the frustration that can arise from investing time in rather outlandish Kanji (about certain trees, for instance) right in one's beginners' phase.

2. no "single keyword" approach (dito for the name tags for primitives), since this:
a. limits the possibilities for making a useful mnemonic;
b. leads to a false sense of accomplishment when the student has managed to learn Heisig's keyword, but not the (often numerous) other meanings of the Kanji. You haven't really "learned" a Kanji until you've mastered at the very least its most common meanings (and varying pronunciations, although these don't necessarily need to be learned at the same time as the Kanji's meanings). So many people who believe that they have "done" all the Kanji in RTK1 are only fooling themselves. They would have been much better off if they had concentrated on getting all of the most relevant meanings of fewer Kanji. This is the approach I adopted towards Kanji after "finishing" RTK1: based on the Kanji I encounter in the Cooori sentences I use, I build review cards around most or all of the Kanji's meanings using all the meanings of the comprising primitives (as I find them on classic.jisho.org/kanji). Not only is it much more satisfying to get the full range of meanings of a Kanji all in one card, but it's also easier to build a mnemonic when one has more meanings for both the primitives and the Kanji itself. What's more, is that this way it's easier to arrive at the more "logical" kind of mnemonic I try to aim for where the primitives more or less logically add up to the meaning of the whole when you have more meanings to work with.

So, on the whole I wouldn't go as far as wishing for trading in my Kanji knowledge for purely phonetic knowledge. Mixing up both pronunciations, Kanji mnemonic building, doing grammar in the form of sentence fragments and hearing practice sentences being pronounced presently works best for me.

"Romaji does not accurately represent Japanese"
->
so does kanji ...

I'm sorry about that, but one has to be realistic: there is almost not one language in the world which has an accurate spelling.
Nor does Japanese.
The idea that roumaji (not "romaji") does not represent Japanese accurately is therefore an unvalid argument.

By the way: the correct roumaji spelling of roumaji is "roumaji", not "romaji", which might partly explain why you think it's not accurate.
It's accurate enough.
Fact is: Japanese don't use it.

"Some might argue that Japanese children learn kanji as kanji. They do. But they know the words already."
->
That's why I argue there's nothing wrong with using roumaji.

Some people treat roumaji as some kind of viral intrusion, but it's not.
However: don't linger to long with it.

It also happens that it's much easier to type, which explains why I'm still using it a lot. Although I very often would like to be able to type in kana or kanji, it are those bloody keyboards which hinder me.
Don't start about IME. Fact is: I would need a separate computer with IME on it, so that I won't be hindered by the emulator when doing my work. Too often a wrong keystroke puts me in the wrong keyboard system. My work is preferential to learning Japanese.

PMotte Wrote:Fact is: Japanese don't use it.

...except when typing on keyboard.

EDIT: WTF is EMI? Did you mean IME?

anotherjohn Wrote:I was recently affected by the evils of romaji as it happens.

I used to play 囲碁 a number of years back, and still have an interest in the game. Go terminology in English is almost entirely borrowed from Japanese. One of the commonest terms used is 'joseki', meaning a standard sequence of moves, e.g. from an openings book.

The other day I started reading a j-book on Go and learned that the Japanese term is 定石[じょうせき], so it would be more properly romanised as jouseki.

I guess it could be argued that 'joseki' is a derived form but ... nah. Another romaji fail

and ko should be kou (コウ) and moyo should be moyou (模様) but fuseki right as fuseki (布石[ふせき]) XD

Hey, if you want some good japanese listening material for Igo check out
https://www.youtube.com/channel/UCYeX5wv...d1A/videos and
https://www.youtube.com/channel/UChBH0Ks...cQQ/videos
on youtube! They've got tons of NHK tournament games uploaded with the japanese commentary.
Sorry off-topic(´・ω・｀)

"To quote one example of the confusion caused by Romaji:"
->
after which you quote Wikipedia ...

Two remarks:
- Wikipedia is put together by amateurs. Full Stop. Amateurs.
- I've noticed often enough that remarks about Japanese are far too often made with English in the back of the head.

Real roumaji is made with Japanese in the back of the head.

vix86 Wrote:

PMotte Wrote:Fact is: Japanese don't use it.

...except when typing on keyboard.

EDIT: WTF is EMI? Did you mean IME?

Yes, I did. I've corrected it.

10k words with /ə tru fənɛtɪk trænskrɪpʃən/ is obviously better than 0 words with no transcription. However from a zero stand point, in my personal opinion, learning kanji as words, in a semi frequency list order, will give the largest benefit in the lowest time.

vix86 Wrote:

PMotte Wrote:Fact is: Japanese don't use it.

...except when typing on keyboard.

That's something I don't get.
Why don't they use keyboards with kana on it? A set of hiragana as basic layout, and a set of katakana using the SHIFT keys? Combined of course with a computer program to select kanji.
I think those keyboards exist.
In my view it would be very unnatural for a Japanse to type "ki" to get to "き".
There are enough keys on our keyboards to, to put all hiragana on it.

PMotte Wrote:Don't start about IME. Fact is: I would need a separate computer with IME on it, so that I won't be hindered by the emulator when doing my work. Too often a wrong keystroke puts me in the wrong keyboard system. My work is preferential to learning Japanese.

The generic 'cycle languages' shortcut doesn't really offer options that aren't easy to press accidentally, but you can disable it and use (Ctrl or Alt)+Shift+number for switching to specific languages. Or disable those too and just use the language bar.

I have to try that last option you're talking about.

I first have to switch from a French keyboard to an English one, and then choose Japanese IME. It's the first stage which can happen by accident.