Suggestion: Phonetic-Semantic scientific Wanikani?

WaniKani Feedback
1

Hi there, would it not be a great idea that learners can switch to a more ‘scientific’ Wanikani where people learn the words with the phonetic-semantic components and elements (‘radicals’) which are more scientifically used and have some historic basis.

Right now, it feels like we are using an effective make-shift solution to learning the kanji, whereas I would have loved to make use of the literature on this topic.

I am not talking about the phonetic-semantic components being displayed - but them being an active part of the SRS and learning.

EDIT: after many useful comments and considerations, I think that perhaps the best way to go about this is as follows:

  1. For Kanji which have a phonetic or semantic component shared with at least 2 other high-frequency kanji
  2. Where there are not too many exception cases

Making sure students are made aware of these components and can actually practice remembering them together and their 1/2 exceptions as they learn the language, could add another memory anchor which is likely to be useful.

This could be a tab that is turned on and off: semantic-phonetic maxing haha.

EDIT 2:

Some useful resources for people with similar attitudes:

2

I think personally that that would be fascinating from a linguistic perspective but on the balance likely much less effective in terms of “lowering the barrier for acquiring comprehension of written Japanese”.

3

That would require building an entirely different system that wouldn’t even be Wanikani at that point. The historic basis and scientific use of the phonetic-semantic components and elements are not necessarily helpful in actually learning the kanji, although it could be interesting. It similarly would not be very helpful to learn the history of how an English word originated from Latin when learning English. It would just be another step.

If Wanikani feels like a bit of a make-shift solution to learning Japanese, the entire language really is a make-shift solution. They didn’t have characters originally and just took the Chinese characters, borrowing some meanings and readings and also attached their own meanings and readings over multiple centuries. The origins of why a kanji is pronounced a certain way or how it got its meanings or why it is pronounced a certain way in one vocab word but then different in another vocab word aren’t always logical or really make any sense at all. That’s partially what makes learning Japanese one of the most difficult languages. It just feels completely random and its really just memorization at the end of the day. Wanikani is there to make that memorization as easy as it can with breaking the kanji into radicals and giving mnemonics that stand out, but its still memorization no matter what.

4

the galaxy-brained take is that it’s not just Japanese which is a make-shift solution to the problem of communication: every language is this way, each of them in its own way

5

Hey, I wasn’t referring to Japanese as a make-shift language. Which it obviously is to some extent. I was referring to WK - it is based on pragmatics (outcome) rather than linguistic knowledge. Which is OK, but it would have been nice to be able to switch.

6

I disagree. Knowing the phonetic and semantic parts in the kanji for which it is relevanr can be very useful and add another memory anchor.

7

and I wasn’t replying to you :slight_smile:

8

I’m surprised people are so resistant to the idea. Phonetic components are the #1 thing I look for when learning a new kanji - it feels like learning on easy mode when I realize I already know the reading. I highly recommend the [Userscript] Keisei 形声 Semantic-Phonetic Composition script to anyone who feels the same way.

9

Oops! Wasn’t clear on my phone!

10

This was actually the reason why I was initially quite skeptical about joining WaniKani. Coming from a physics background, I view mnemonics with a bit of suspicion; I don’t just want to know the answer, but the rules behind it.

I’d look up the etymologies as I went along, but as I did, I found a few issues:

  1. The etymologies are often quite speculative; they are systematic best guesses, but we without access to a time machine, we can’t know most of the etymologies with certainty.
  2. The etymologies are often very unhelpful for language-learning purposes. There are some that are quite memorable (like 休 and 安), and there are some phonetic components that are consistent enough in modern-day Japanese that they are worth bearing in mind (see for example Testing the predictive power of phonetic components in Japanese kanji and Identifying Useful Phonetic Components of kanji for Learners of Japanese), but they’re kind of the exception rather than the rule.

Eventually, I reasoned that perhaps WaniKani’s make-shift solution was better for learning purposes, and that if it could get me to 2,000 kanji, maybe I didn’t need a more general ruleset. Indeed, I found this to be the case; I still enjoy reading up on kanji etymologies on the side, but it’s not really a part of my language-learning process, anymore.

That being said, I do still find phonetic components quite useful at times; there are quite a few kanji where I forego the published mnemonic entirely, because it’s easier to remember that 莉 is pronounced like 利 than it is to remember a story about jasmine flowers growing out of reeds.

11

Yes, fair enough. I try to mix them as well. I was 25 level before, by the way, I restarted after a break.

12

If I recall correctly the radicals were invented centuries after Chinese characters were in common use among the literate. So, they also are mostly made up. Japanese is listed among the hardest language for westerners. That said, I don’t think its too hard. Its the Kanji that make it so unapproachable but with WK studies its not too hard.

Actually we kind of live in a golden era for Japanese and Chinese characters in general. Tons of high quality text available, lots of audio visual media, lots of on demand tutors, WK, elite quality textbooks. Apparently in WW2 in the US apart from 1st/2nd/3rd gen Japanese very few almost none managed to learn Japanese.

With so much available it is straightforward to become fluent. Almost trivial if you show up regularly. They are right. the hardest part is showing up every day.

13

I’ll give it a try, always open to trying new things and possibly changing my mind.

I read a very similar analysis before buying Wanikani when I was deciding whether to go with them or not. You worded it much better than I could have

14

That is true for every language but I would argue its even more true for Japanese than possibly any other language. English for example can seem like it’s make-shift because it borrows words from so many other languages like Latin, French, German, Greek, and keeps the original spelling rules for each and such. It makes pronunciation very complex, but at least there is an alphabet that ties it all together and serves as an engine. Letters and letter combinations generally correspond to sounds. Not perfect, but the core engine is there.

Whereas with Japanese, it uses 3 entire writing systems at once. Kanji are meaning-based and were borrowed from Chinese, whereas Hiragana and Katakana are phonetic and were invented and added later because Kanji couldn’t handle Japanese grammar. And because the Chinese Kanji were forced onto their already existing language, most Kanji have both the original onyomi Chinese reading and the native Japanese kunyomi reading, and there’s often multiple onyomi and kunyomi too. There is no consistent rule on when one reading or meaning is supposed to be used, you just have to memorize it for every single word. Kanji have no fixed meaning, or reading, or even a fixed single kunyomi or onyomi reading.

The result is that Kanji have no fixed identity. For example, a Japanese name can be both pronounced multiple ways and written multiple ways. So if someone tells you their name, you don’t know how to write it, and if you see it written, you don’t know how to pronounce it. There is no other modern language on Earth that does that. With any other language, you would at least have either meaning be consistent or reading be consistent for most words, not inconsistent for BOTH for almost ALL words.

Don’t get me wrong, I love the complexity. It’s like one giant puzzle. I choose to learn it specifically because it is complex and difficult. But I think it is possibly the #1 most Frankenstein language out there.

15

I completely agree with your comment, but I just have to add: I think it’s outdated to say Japanese only uses 3 writing systems. The Latin alphabet is in everyday use as well.

16

I’m not sure this is quite right. Both katakana and hiragana are simplifications of kanji, they don’t appear to have been invented wholesale (in the way that hangeul were, for example).

the fact that hiragana are most often used for common words, particles, and parts of words that conjugate and the fact that katakana are used for emphasis, words whose kanji are a bit more obscure (animals, for instance), and words of foreign origin does not mean that’s why they came into being.

(if you think about it for a second, if “we should invent a syllabary, because we have grammar” is a thought that some hypothetical Japanese person or people had when inventing kana, it doesn’t make sense that they would not go as far as Korean, right?)

you’re free to say this, of course, but it weakens your argument significantly, when your reasoning for saying that English is somehow “not make-shift” despite even words like tycoon coming from Japanese boils down to “it’s all letters, and the letters have sounds mostly”.

in fact i’d say that you are used to English and you are not used to Japanese and this is affecting your judgment.

you do not, in fact, have to memorize which reading to use for every word to be able to read Japanese, in exactly the same way that you do need to memorize the reading for a few words in English (eg, which “ough” sound goes with “tough”, with “through” and with “though”?).

of course you are right that kanji do not have any fixed meaning, in the same way that root words in English may be misleading. at the same time, you can get significant mileage from knowing “mis-“ and “lead” and “-ing” in terms of how to pronounce the word and what it means, in the same way you can get a lot out of knowing 漢 and 字 (or 大 and 君, for my earlier example).

17

That’s true, they were not deliberately created and implemented as an intervention like Hangeul was, but that actually goes to my point. Hiragana and Katana evolved from Kanji specifically because they were trapped by the limitations of the Kanji system they already had implemented. Before Hiragana, some Kanji would be used solely for their readings to express grammar. Using a Kanji just for it’s reading for every single syllable was exhausting and inefficient, and that’s ultimately why the simplifications of Hiragana and Katakana came about. That’s not a design choice, that’s make-shiftness that adds further complexity.

I never claimed that English was not make-shift, the first thing I wrote was agreeing with you that it is true that all languages are make-shift to some extent. But there are levels to it, not all languages are created equally. English is definitely more make-shift than German, French, Spanish, etc. with how much it borrows from multiple other languages and tries to fit everything under one coat. But in the hierarchy of make-shift languages, Japanese is probably on top. Because I think Japanese is like this more than any other language, I think it warrants learning it through made-up radicals and mnemonics like Wanikani does. When a language lacks so much consistency, it makes sense to learn it by breaking all the pieces apart and learning it brick by brick rather than trying to follow millennia-old logic that might not even exist.

An alphabet, even if not phonetic, still helps you 90% of the way to pronunciation. When English borrows a word from another language, even if that language is Japanese, you can still use the English alphabet to make a best guess on how to pronounce any word. If you’re still learning English, you’d have difficulty pronouncing words exactly (although tycoon is phonetically pretty straightforward), but at least you always have the alphabet to help guide you. A word is going to sound more or less like it is written in English. Whereas in Japanese, you’d just have to “know” that 台風 is pronounced たいふう by “knowing” that 台 is pronounced たい here and 風 is pronounced ふう (even though both of those kanji have multiple onyomi readings). There’s some clues that help narrow your options down, like the fact that 2 kanji together usually means onyomi (but then which onyomi?), but no clue is ever going to be as helpful as an alphabet.

If we’re talking about vocabulary, then in order to read Japanese and pronounce the words properly you do need to know the exact pronunciation for every word. Unless you’re reading purely for the meanings and skipping pronunciation I guess. You don’t need to learn all of a Kanji’s onyomi and kunyomi readings up front, but you do need to know individually when to use the proper readings for every vocabulary word. You can’t guess based on an alphabet, you just need to know that 試合 is しあい, even though there is absolutely no reason it couldn’t be しあ, しごう, しがっ, こころあい, こころあ, こころごう, こころがっ, ためあい, ためあ, ためごう, or ためがっ. Whereas in English, depending on your proficiency, you’d maybe pronounce “tough” “through” or “though” in a funny accent, but you’d improve on that over time. You’d wouldn’t accidentally say something not even close because you used kunyomi #4 instead of onyomi #2.

18

I suppose I should have just said more straightforwardly “the claim that Japanese is the most make-shift language is a racist claim” and left well enough alone. sorry for wasting your time.

19

As a general remark I’d like to add that the Japanese system of writing is a mixture of make-shift, ad hoc solutions, and organically emerged and created patterns which are more or less clustered semantically and/or phonetically. There is some possibility of systemization but this will always have to make place for the empiric reality of the language. Think about the phonetic-semantic parts and attempts to go back to etymologies. All in all it is a mix of order and chaos, and I think this goes for all languages in different ratios. I speak 5 languages and I believe the chaos in Japanese is the highest I have encountered in terms of the written to speech (and vice versa) conversions. It makes the language interesting and a real accomplishment to learn. It also takes a lot of time and mental energy. I guess you cannot have it all.

20
quote="Fishfulness, post:17, topic:74464"

to know that 試合 is しあい, even though there is absolutely no reason it couldn’t be しあ, しごう, しがっ, こころあい, こころあ, こころごう, こころがっ, ためあい, ためあ, ためごう, or ためがっ

/quote

Actually the following readings are absolutely NOT possible: しあ, しがっ, , こころあ, こころがっ, ためあ, ためがっ

あ is NOT a reading of 合, that doesn’t exist.

There is the verb あう that can be written with 合, usually the inflectable part is written in kana after it, but not always. But there isn’t a standalone word あ that could be written with that kanji.

Kunyomi are not random sounds, they are lexemes (strings of sounds with a meaning), you cannot cut it as if it where two separate parts. Even if in most writings there is a part “hidden” by the kanji and another not, you still have to think of the lexemes as a whole.

「本を読」, while completely non standard, will most likely be ready as ほんをよむ and not as ほんをよ, which doesn’t make any sense at all.

がっ and similar, even if listed as “readings” in some books, actually are not. They are just a (quite regular) phonetic change when two sounds come together.

The reading of 一 for example is いち, not いっ.

In things like いっか、いってん、etc, it is just that the kana orthography follows the phonetic change, but it could as well have been spelt いちか, いちてん and people would have still pronounced it with the gemination, because it is not due to a “reading” arbitrary chosen, but a regular phonetic rule of the Japanese language.

It seems completely arbitrary to you because you are not familiar with it and don’t see the more generic patterns (and wanikany teaching “あ” as a reading, instead of “あ(う)” doesn’t help, and I think is its biggest flaw).