This article actually hits on a pet peeve of mine where I feel people sorta “mystify” kanji/hanzi unnecessarily.
The truth is that there’s actually nothing particularly weird about being able to read some kanji but not be able to write them…
You actually get close to my point here:
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
Yes! But have you ever heard of someone forgetting how to spell certain words in a language that uses the Latin alphabet (e.g., English)? I can use myself as an example here: while writing this comment, I forgot how to spell “peeve” in “pet peeve” (I thought it had an ‘a’ in it) and I also forgot how to spell “unnecessarily” (I thought it had one n and two c’s).
The western equivalent of being able to read some kanji but not write them is simply called bad speling. No need to mystify kanji in particular.
Nice comparison. When I was living in China, I'd encounter someone who forgot how to write the characters for a word on a weekly basis. I think the difference is that with Latin alphabets you can still misspell something, and having gotten it down on paper, still rely on phonetics to convey your meaning.
Alphabet is such an underrated invention. It's probably higher in significance compared to the invention of wheel. It's the original "bicycle of the mind". For example, Korea pivoting from Chinese characters to its own alphabet or Hangul is very well documented including the positive effects it has in the much improved Korean literacy and civilization after the conversion. Fun facts anyone can learn Hangul alphabet in a single day if they wanted to but the same cannot be said to Chinese characters. If your mother tongue is Korean (e.g Korean American) that only just started learning, it only take one day turnover from illiterate to literate.
Scripts being the main driver of literacy is a pet peeve of mine. It's not the script, it's the schooling system. The high rates of literacy in modern states are just a result of the school system - Japan has a high literacy rate, for example, and their writing system is either the worst in the world or close to it.
That said, the characters are a whole boatload of unnecessary extra effort, and as a student of the two languages, the artificial illiteracy created by kanji, where I often just can't read words I've known for years, is simply maddening. Not having to wrestle with characters does free up a lot of time for both native and foreign students alike.
>their writing system is either the worst in the world or close to it
Yes, it's probably the worst since even Microsoft until now still struggle to provide proper search solution for Japanese names in their Windows OS due to their multitude of writing systems.
By sheer wills of course you can make everything hard feasible but that does not means it's efficient and effective. I consider Japanese as a unique country with extraordinary people that can collectively overcome adversity, that's include a non intuitive and difficult writing systems.
If your mother tongue is Korean (e.g Korean American) that only just started learning, it only take one day turnover from illiterate to literate.
Heritage speakers (of any language, not just Korean) often have limited vocabulary and limited exposure to complex grammar. Being able to sound out words wouldn't be enough to allow a heritage speaker able to fluently read a newspaper.
How many Korean-Americans know the Korean words for things like 'legislature', 'inflation', or 'geopolitical tensions'?
The term that would cover what you mean here is regular. And that is only in regards to correct spelling. Is obviously complicated when considering that we don't have official pronunciation across all dialects for the same word. Even if we do agree on a spelling.
But it is a complete non-sequitur to lead to the modern idea that English isn't phonetic.
This is not about regular VS irregular, there are aspects of English spelling that are highly non-phonetic. It's not uncommon to have letters in words that are entirely irrelevant to the pronunciation. For example the spellings "programme" and "program" would be read the same by any English reader, and yet both persist in certain places. The s in island is completely unnecessary.
Also, the same English word can be read in very different ways by the same speaker, but in different contexts. This is most proeminent with some of the most common words in English - a, the, there, and many other connective words can be pronounced very differently by the same speaker in the same speech, depending on stress (for example "a" can be pronounced as either ə if unstressed or eɪ if stressed). And yet, there is no version of written English that differentiates these - another sign that English is not a phonetic spelling.
Of course, on the other hand, you can't say that there is no correlation between spelling and pronunciation, like you can in Mandarin and other Chinese languages.
That is what is typically meant by a regular orthography. Wikipedia also calls it deep and shallow. These are legit terms that pre-exist to this odd debate that English isn't phonetic.
Nobody that knows how to read English at a level to be on an internet forum is surprised that English has odd spelling. Many people would be deeply confused to be told that written English doesn't follow a phonetic system. Rightfully so.
This happened with cyrillic to me. During Yugoslavia we had to learn both latin and cyrillic. Since I'm from Croatia, I didn't have a need much for cyrillic (or at all). Today, I can read _at speed_ (including subtitles) cyrillic just fine, but I probably couldn't write a thing if my life depended on it. It's weird when I think about it.
> > Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
When I was in grade school, I took notes for my classes using Tengwar (elvish) runes as a way to alleviate boredom and force myself to pay attention (just taking the table in Appendix E from Lord of the Rings and transliterating English letters into them). I could do this at a speed sufficient to keep up with a teacher talking, so pretty fast. I cannot read any of these notes today, much less write them.
If you don't use this stuff for a few decades, you do forget, even when it's just an alphabet.
I learned how to write cursively in school, but as soon as it was no longer mandatory I switched back to print capital letters and some years later to print letters entirely, as it was just much easier for me to make legible.
At this point, it's fair to say I _have_ forgotten how to properly write cursive capitals, if I tried I'd just end up with print capitals with random tails for most of them.
I don't think anyone is "mysticizing" the language. It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
IMHO spelling is an inapt analogy. Every single word I can communicate orally, I can write. Sure I might mess up "I before E" or some other minor issue, but I'm BUILDING the word from first principles, e.g., syllable phonemes. That's why kids are taught to "sound it out" at a young age.
The closest equivalent you have in logographs might be radicals.
> I don't think anyone is "mysticizing" the language. It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
People absolutely do mystify the operation of kanji, like they're more than scribbles that point to words, or that Japanese would fall to incomprehensibility and ruin if they were done away with.
> It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
This isn't quite true - about 80% of Chinese characters are so-called phonosemantic compounds, where people originally started using the character for one thing for another thing whose word sounded similar (say, emoji for "can", as in able to) and then adding a semantic component to differentiate the character from other similar-sounding ones. In Chinese, they smushed the two components into the space of one character, but in eg. Egypt, they simply wrote whe semantic clarifier and the phonetic hint side by side, full size.
That is, the majority of the characters are primarily sound-based, it's just that the connection between a character and its sound is shoddy, even in Chinese languages.
Japanese kun readings for native words do divorce the characters pretty completely from their sound.
Yes! I've studied Japanese for years and read numerous novels in it each year, yet I couldn't physically write to save my life.
But there's nothing crazy about that, like you said it's similar to spelling (not entirely, as I can spell things fine if I have a phonetic keyboard that "writes" for me).
Writing vs reading to me, is more about the type of memory.
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
I’ve actually done this all the time as a bilingual speaker with my language in Cyrillic. It doesn’t happen often but once in a while I’ll freeze and be unable to remember which glyph makes which sound when switching alphabets/languages.
>> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
> Yes! But have you ever heard of someone forgetting how to spell certain words in a language that uses the Latin alphabet (e.g., English)?
Erm... great that you gave English as an example but I'd argue this mostly applies to English. I remember wondering why "spelling bee" was a thing when watching shows from the USA because that wouldn't make any sense in Polish. Same with Spanish. And a lot of others. There are some minor things to remember and you can do errors but in 99,(9)% of the cases it's "you write what you hear".
It could have been a case for English as well but the reform efforts were killed so here we are with all it's quirkyness :)
The reverse almost happened to me before, I write almost exclusively in cursive and once or twice had to stop and think about what a capital Q was supposed to look like.
Small nuance, but the term ワープロ馬鹿 actually is unrelated to software like MS Word, and refers to the at one point ubiquitous ワープロ (Word Pro) dedicated hardware device that many Japanese people owned in the 80s/90s to write letters. Read more here: https://en.wikipedia.org/wiki/Word_processor#Japanese_word_p...
Interestingly the English Wikipedia page above only mentions Japanese word processor devices in a small section, but the Japanese version of that page is almost entirely dedicated to these hardware devices: https://ja.wikipedia.org/wiki/%E3%83%AF%E3%83%BC%E3%83%89%E3...
I was lucky enough to live with a Japanese family in the early 90s and used one to learn how to type Japanese but also write letters home in my own language. I guess you had to live through this age to understand the difference of how Word Pro is used and the hardware association it has in Japanese.
The other weird thing about that term is that it's somewhat uncommon to see 馬鹿 actually written out in kanji, as those characters mean horse and deer with no direct association with 'idiot' (the proposed etymology is a reference to Chinese history and/or a Sanskrit loanword). I guess in a way the kanji form is a word processor autocomplete tell.
A Toshiba Rupo word processor (found in both links) is safely stored in the backyard shed, together with some floppies and documentation. That'll be my winter project next year.
When learning Japanese, I purposely chose to _not_ learn how to write any of it by hand. As the author notes, writing (by hand) is in fact a separate skill from reading. So I decided I would not invest my limited time, motivation, or brain space to writing.
Overall it's been a successful approach, and I recommend it to new learners unless they have a particular interest in being able to write by hand or they feel strongly that writing the characters helps them remember them.
It's only rarely that I have to write anything other than my own name in Japanese. I've practiced my address but writing it in English is fine in 99% of situations. Being able to write properly would save a little embarrassment, but I still believe my language learning time would have a much higher ROI in other areas.
I went to an old-school language school where I was forced to take tests in handwritten Japanese. I probably still have some of that in my brain, but like you, I almost completely abandoned it as soon as I didn't have to take language school tests anymore.
It's occasionally useful to write out a character, but on the whole, it's completely unnecessary now that we have computers with hiragana keyboards.
As a partial aside, the Heisig anecdote that leads off this piece is painful:
> Japanese children learn the spoken language first, then they learn how to write it in elementary school; Chinese students of Japanese (who tend to be pretty good at it) have pre-existing knowledge of character meanings and forms from their mother tongue, so they only have to learn how to pronounce them. Therefore, a Western learner should first focus only on the meaning and writing of those couple of thousand common characters and, only after having mastered those, should move on to studying the pronunciations.
Going from "Japanese people learn the spoken language first" to "you should spend a big chunk of time learning characters before learning sounds, words or grammar" is a pretty remarkable mental backflip.
The author says he spent eleven months doing this before devoting any time to the spoken language. If I could put the "head exploding" emoji here, I would do it. I spent only slightly more time than that at language school, and came out conversational.
Yeah I agree. I way over-indexed on learning kanji (via WaniKani) at the beginning of my Japanese learning journey. I got about halfway through before realizing it was silly that I could read 健忘症 but didn't know many very basic hiragana-only words. It wasn't timed wasted but it probably wasn't the most efficient approach.
In an ideal world maybe learners could focus exclusively on listening and speaking first, then move on to kanji later. But writing is a very useful tool in learning, and having access to that tool can help speed things up.
Like most things in life, a balanced approach is probably the right one. But you have to know what your goal is. Our brains are lazy, they only get better at what we make them get better at. If your goal is to just read kanji, practice reading kanji. If your goal is to understand and speak the language, practice listening to and speaking the language. But if you want to have a balanced language ability, you'll need to practice it all.
WaniKani at least teaches words. Spending almost a year of your life doing nothing other than learning "meanings" of individual Kanji is...well, I guess some people just really get addicted to that mechanical feeling of progress?
Reading is definitely helpful, but I've found the relative importance of reading, listening and speaking goes in cycles, and especially at the early stages, listening and speaking are far more motivating than anything else. And I'm an introvert!
> In an ideal world maybe learners could focus exclusively on listening and speaking first, then move on to kanji later. But writing is a very useful tool in learning, and having access to that tool can help speed things up.
There is no connection between these two sentences. You can learn to read and write in Japanese without ever learning a single kanji, and that's what everybody does do. Kana serve the purpose flawlessly.
With 0 kanji, it's such a small subset that it's hard to call that a finished job of learning to read or write as you'll be limited to material like kids books or NHK news easy.
Yes, I ignored Kanji completely while learning Japanese. I only learned words in hiragana and katakana from my SRS decks. I regret it because it locked me out of a lot of comprehensible input I could have used to actually progress in ways apart from learning the sounds of lots of words (even the sounds I learned were often wrong because I never heard anyone speak the words I was learning).
Am I the only one reading this thread and thinking, "Gee, this sounds like a job for AI?"
Why in the world should any human not from Japan devote limited time and brain capacity learning to read and write Japanese? It's literally a robot's job, at this point.
It will take a lot of time for your Japanese language skills to get up to par with robot translation, that is true. It was also arguably true with google translate many years ago too.
But isn't this true for most things? It will take a complete beginner years to draw as technically proficiently as an image generation model, or to code as well as Claude. Even before AI, most Japanese media has been available in English translations for years now, and there aren't that many other interests where you could find Japanese speaking peers but not English speaking ones.
If your goal in life is to generate the most economic value per unit of time input, maybe then learning to draw, code, or speak Japanese no longer makes sense. And if that's your priority, you won't choose to do these things. But that's not why people take up these pursuits. So I don't think AI will have a huge impact on how many people start them.
In an ideal world, some kind soul would escort kanji behind a shed and we'd hear a loud bang, then marvel at how we can suddenly just read words we've heard for years.
Yes, I agree that trying to learn kanji upfront is a silly idea.
Heisig says in the introduction to RTK I that he learned 1900 characters "before the month was out". If like him you can do the whole set in a month and then have no further need of formal review or study beyond using them as they turn up, then I can see it not being a terrible idea. But as far as I can tell, almost nobody has a mind that works like Heisig's does: people seem to need longer and to rely more on review via an SRS like Anki.
Personally I found my problem with RTK was that I successfully memorised "English keyword to write the character" for 2000 kanji, but this was not at all linked to my actual use of the language, so I still had the problem I started with of "I want to write the word べんきょう but can't bring to mind the kanji for it", because I had no association between Japanese words and the English keywords for their component kanji...
> I still had the problem I started with of "I want to write the word べんきょう but can't bring to mind the kanji for it", because I had no association between Japanese words and the English keywords for their component kanji...
If you didn't know what べんきょう meant, how did you know it was what you wanted to write?
I knew the word (including how it is spoken and what it means), so I would have no difficulty of understanding if it was said to me in conversation, I could read the word whether in kanji or hiragana, I could compose sentences in my head which used it and use it when speaking, I just didn't always remember how to write it in kanji...
(Here べんきょう is just an example: the same issue applies to essentially every word.)
The characters you copy pasted are in a phonetic script called hiragana, conceptually similar to the English alphabet as each character denotes a sound.
The kanji are pictographic characters that often have different pronunciations in different contexts. The kanji for べんきょう are 勉強 (meaning "to study").
In your model of the problem, pm215 knows the following things:
1. There is a Japanese word pronounced benkyoo which refers to activities like reading books with the intent to learn something, doing practice exercises with the intent to learn something, preparing to take a test, doing things that a school might ask you to do, and other similar endeavors.
2. There is an English word "study" which corresponds to the Japanese word spelled 勉強.
But he doesn't know this:
3. In English, activities performed with the intent to learn or review something are referred to by the general term "studying".
I find this hard to believe. Anyone who chose to learn mappings from English words to kanji spellings must be familiar with the meanings of basic English words. He shouldn't be able to think of the activities he wants to refer to without the word "study" coming into his mind, but that was the problem he described.
No, the problem I describe is "I am thinking of a Japanese sentence (which I know the meaning of), and I could write it in hiragana but not in kanji".
This is analogous to "I want to write an English sentence, but I can't remember how to spell one of the words", except it's worse because at least English words are spelled vaguely in line with their pronunciation.
I think where I may have stated the problem confusingly was my reference to not having a link between the Japanese word and the English RTK keywords. RTK assigns one unique keyword to each kanji, which (a) doesn't always line up with the meaning of a word in which it's the only kanji and (b) doesn't inherently help with multi kanji words. In this case benkyou is 勉強 which is two kanji with the RTK keywords EXERTION and STRONG. Unless you actively learn and memorise a link between the Japanese word and this pair of keywords, RTK is not going to help you with writing the word.
> RTK assigns one unique keyword to each kanji, which (a) doesn't always line up with the meaning of a word in which it's the only kanji and (b) doesn't inherently help with multi kanji words. In this case benkyou is 勉強 which is two kanji with the RTK keywords EXERTION and STRONG.
Ah, you're right. I had no idea this was what you had in mind. That is a system that doesn't make sense. You have to learn the spelling of words per word, not hope that the roots make sense.
(By contrast, 勉强 exists in modern Chinese too, where it mostly means "force; coerce; compel", but can also refer to just barely being able to do something. I was amused to see that it means "study" in Japanese - that implies one of the most sharply negative attitudes towards studying that I've ever heard of.)
I think the thing that makes it a problem for most people is that they can't memorise them as "once and done" the way Heisig says he did. So as well as the initial time spent looking at the kanji and coming up with a good memorable story/image/mnemonic/etc, most people I think also spend time in an SRS (e.g. anki, or kanjikoohii) reviewing the characters they learnt previously. It's the review time that really stacks up, especially where you have particular characters that you have trouble with ("leeches").
You could pull it off in the sense of 'I've seen this one before...', but no way do I believe that he learned >60/day with good recall. That'd be like me saying I learned the fundamentals of spoken Japanese in one afternoon. I think it's a kind of bullshit claim to be honest.
Well, I can say that I've spent more than eleven months learning common characters before learning any Japanese sounds, words, or grammar.
But the reason for that is that I was learning Chinese. I spent zero effort on learning any characters preliminary to that. There is no reason you'd need or want character knowledge.
My character knowledge is decent now, because eventually my learning method became "talk with Chinese people over Wechat", and if you do that you will necessarily learn common characters.
The idea that one should learn kanji first just to start learning Japanese is utterly nonsense. The author has Stockholm syndrome.
If one's mother tongue isn't Chinese/Japanese, I guarantee you that's impossible to understand kanji/hanzi as deeply as a native speaker does just by spending 17 months on memorizing how to write them. It simply never happens. Languages come first and writing systems come later no matter which target language you're trying to acquire.
Even native speakers get confused about their own writing systems. It's vs its. Should've vs should of. Doppelganger vs doppleganger. Being able to wield the writing system like a wordmaster is a big plus and a praiseworthy effort. But it really isn't the essence of language acquisition.
Indeed. I thought the Heisig approach was garbage on the same basis. I did get a kanji reference book that was a bit more focused on learning the Joyo kanji (ie the 2000 or so you need to graduate high school or pass N1) but I did not stick with it because it also used mnemonics as a learning tool and I found them distracting and obnoxious (in the sense that the author kept injecting his personality into them - sorry dude, I came here to learn kanji, not to learn your opinions on things).
In general I just do not get the mnemonic approach. It's literally another layer of stuff to recall that frequently doesn't have anything to do with the meaning, and the book I had even connected mnemonics in simpler characters to build into more complex ones. Worst of all, you're learning to identify characters with a bunch of word association in English.
Instead of treating chinese characters as a bunch of "tangled squiggles" that you have to memorize or make up stories about you could just ...learn the radicals (called bushi in Japanese). There are around 230 of them, and they are fundamentally pictographic, so the meaning and appearance are linked. Every kanji character is either a radical itself or made up of other radicals, like a word spelled in two dimensions. Some are used much more than others so most of the time you'll be combining the same 32 simple ones. Once you get familiar with them kanji become much easier to remember; a character that is made up of 10 strokes like 勉 is actually made up of just 4 parts which are just a few strokes each. In some cases you can even guess the meaning of a kanji character you've never seen before by looking at the components, but even where you can't, being able to see a complex character as just a collection of familiar simpler chunks makes everything way easier.
You can find a kanji radicals deck on Ankiweb and it's small enough to do alongside your vocabulary or listening practice without being a burden. It will induce a little cognitive dissonance because a few very familiar characters have different meanings in their radical form but you will get over that soon enough.
https://en.wikipedia.org/wiki/List_of_kanji_radicals_by_freq... is a great reference, both for the frequency and the simpler table of stroke count. The individual radical pages have additional detail tracing each one back to bone script (the oldest system of writing, etching characters on bones) which make the pictorial evolution very clear.
By the way, a book I do recommend is the Kodansha Kanji Learner's Dictionary by Jack Halperin. This too uses its own schema for looking things up, but it's a simple one, based on the shape (left-right/ top/bottom/ nested/ freestanding) and the number of strokes, which makes looking things up fast. It also contains indexes to look up kanji by radical or pronunciation (eg if you have the furigana) so you are not locked into one way of understanding them like mnemonics. It also has common vocabulary that uses the kanji, stroke order, and all that good stuff. And it has a good reference on the rules of stroke order. There are only 8 of them and (like radicals) once you understand the underlying principle you can look at a new kanji you've never seen before and have a pretty good idea of how to write it. You should still check while you're learning, but the more you write the easier it gets.
jisho.org is also incredibly useful, because you can dig into any word to find the kanji within, and then dig into individual kanji to see which radicals they're composed of, as well as looking things up by radical.
I am doing both because while learning to write the symbol for a sound/meaning and identifying the symbol’s sound/meaning are separate skills, they enhance the fluidity of _thinking_ in Japanese significantly for me. It has a synergistic effect and to me seems to improve the brain’s understanding and efficiency in compressing the knowledge.
But my goal is not just to read and understand but to talk conversationally. While Japanese is very different from my other languages, I’m already multilingual (Norwegian, English, Dutch and German) and this approach has always worked best for me.
I might have to play with writing alongside vocab reviews.
Back in the earlier days of the online Japanese learning sphere (when AJATT was still the big new thing), I tried learning by starting with writing by studying kanji independently, but that went nowhere even after several months in.
More recently I’ve been making a point of audibly speaking the sentences associated with vocab cards and that’s helped a lot with being able to fluidly speak the various long trains of sounds that are common in Japanese but rare in English as well for improving recall and improving reading speed. It would follow that writing might enhance that effect.
My foremost goal is to become conversational too. The rest should follow more naturally if I can achieve that.
I’ve taken your approach as well but I did notice that I retain Kanjis that I learn to write significantly more than ones I can just read. But memorising all the Kanjis is a bother.
I found the sweet spot to be writing on a scratch pad as I go through Anki. And not particularly worry about getting writing right too much. Sometimes I’d be confused in my head but my muscle memory would kick in and automatically write the kanji!
Yeah I made the same decision learning Chinese. It's just not worth the extra time and effort relative to the utility.
Occasionally I still have to sign my name, so I specifically picked a name that a) was written the same in both simplified and traditional and b) had a low number of strokes. Like you, the only time this has bitten me is in hospitals and banks where occasionally they ask you to do stuff like write out your address. I sometimes exercise my dumb foreigner privilege and ask the clerk to help, but since addresses have a formal romanization method it's often fine to write that, and I've seen enough locals struggle I don't feel too bad about it.
I'm learning traditional Chinese and found that writing helps me recognise the components and strokes when reading the same characters.
If I just try and visually pattern match with flash cards, anything that's hand written or in an stylised font will throw me off. If I can sympathise with / recognise the stokes used, I find it easier to tell what character they're trying to show.
The one thing I noticed when I was focusing on learning to write is that it helped me a lot with differentiating between similar characters when reading. I forget which ones now, but there are many characters that differ by a single radical and have similar meanings, knowing how to write each one helped me quite a bit there, but overall I rarely write anything other than my name and address now that I live in Japan.
Even if you cannot recall and have to type in your phone first (as natives do often for unusual ones), at least you can write it much faster because you are not just copying.
I think it’s useful to do RTK with SSR and, once you finishes, you only need about 15 min of maintenance per day to keep it in memory.
I can confirm this. I passed N1 without learning to write. I later learned how to write all of the kanji, and all it does it help you distinguish very similar kanji without context. I tried learning all the compound words (i.e. which kanji to use for every word) but gave up a few thousand Anki cards in. It was time consuming and impractical. (Wanted to pass Kanken 2) Props for anyone who put in the work though.
Don't drill words drill sentences (no Japanese word means an English word, and no English word means a Japanese word.) And don't fall in love with any of the sentences, move on to new sentences. Eventually the knowledge accumulates, but only through volume and variety.
You strategy makes sense. Truth be told, even native Chinese and Japanese tend to forget how to write many characters as they spend more time typing than writing.
I think your advice makes a lot of sense for most learners: prioritize the skills you'll actually use, and don't feel guilty about skipping handwriting unless it personally matters to you
I'm following this approach. One of the most interesting things so far has been observing just how separate recall and production are. There are kanji that I can recognise instantly, and recall meanings and pronunciations, but I can't visualise them at all.
It is a separate skill from reading, but I think it's still useful.
At the very least learn the strokes of common radicals. In my experience things like denshi jishos can be VERY picky about how you input them. It makes word lookup much faster IMO anyways.
> How is it possible for you to "see" the text in your mind and not be able to replicate it with a pen? Even if the mental image is faint and fuzzy, surely you can sketch it out roughly at first, then refine it until it settles into its exact form? Apparently, that is not how mental images work, either.
That's really not how it works. I draw as a hobby and I love to study human anatomy. I can conjure in my mind people in many poses with detail, in different clothes and colour. Still the act of drawing without a reference is a world in itself. If I try to sketch from memory without a reference, it quickly falls apart.
It's not that the mental image is incomplete, wrong or an illusion. Is just that knowing it and reproducing it are two very different things.
This is very common in Chinese now. The older generation, many of whom didn’t learn pinyin, just use voice input to send messages; the younger generations just use pinyin input and similarly can’t handwrite beyond the simplest characters.
The phenomenon of forgetting how to write is called 提笔忘字 (tíbǐwàngzì - to pick up the pen and forget the character). It was previously covered here on HN: https://news.ycombinator.com/item?id=41959256
Not really. Even in English there are only a few possible ways to spell out a given sound. With Chinese characters that mostly goes out the window, making speaking and writing more or less orthogonal skills. Not entirely since there are classes of characters that are related by sound, but to a much greater extent than alphabetic languages. Personally if I forget a spelling in English or a related language I realize that my approximation looks wrong and can correct it by elimination, whereas with Chinese characters it's common for people to simply not have a clue how to even write the first stroke of a word they've heard and read before.
What if we call this 'constructive recall', in which your mind needs to take a concept and generate the concrete manifestation of it, vs 'recognition', which is the obverse. There are many examples of this in life, aren't there?
Imagine a song you used to love as a kid and knew by heart. Now try to write down the lyrics word for word. Typically quite difficult.
Now play the song and sing along. You not only know all the words, your mind provides you a just-in-time recollection of all the nuance in the delivery, the pauses, the details of the music itself, nearly a perfect replay of the entire song is there in your mind...you just couldn't knit it together like you can when it's there in front of you.
I've noticed something similar when listening to podcasts. If I'm out doing something, typically driving, while listening to a podcast episode for the first time, I will have these intrusive photographic recalls of what I was doing at that time if I listen to it again. At least if I do that within the first few weeks.
Just seems like a general characteristic of the brain.
When I reached upper level Japanese classes (N2/N1), my native Japanese teachers would regularly (maybe once a week or so) have to look up a character they were writing on the board during class.
> On the surface, this atypical trait seems to explain quite well why I can draw a blank when asked to write the kanji for "plant" (植) from memory. I don't see the character in my mind, so it makes sense that I can't reproduce it on paper.
While the author's aphantasia may have posed some recall issues - it wouldn't explain why they had ever been able to reproduce 植. Kanji has the concept of radicals AND stroke order. One could make the case that perhaps the author's motor cortex is simply storing the equivalent of LOGO programming language instructions for reproducing the logograph.
Take away your mind's ability to find and chunk (木, 十, 具) by showing them "radical"-less characters and I'm sure it would be even more difficult.
As someone with aphantasia, I'm also not convinced that is the cause. I can draw better than average from memory, and used to be quite good but haven't practiced for many years.
There are also animators and artists with aphantasia[1].
My spatial recollection is particularly good - I can sketch out precise diagrammatic drawings from memory much better than I can do artistic drawings.
So while it's not impossible that not having aphantasia would've been a benefit to him, I don't think there's much evidence that it's has any big effect on the ability to draw - something that also fits Ed Catmull's experience on surveying his old employees in the article mentioned.
My wife has aphantasia, is fluent in Japanese, and writes the language better than many of our friends who still live there. She was just forced to learn the stroke order, etc when learning the language.
Aphantasia causes many odd issues for her but the ability to write complex languages or draw are not one of them.
Writing characters without knowing the radicals is kind of like spelling without really clearly understanding the alphabet. The stroke order is easy to remember, the radicals make the full characters easier to remember.
I spent years in Taiwan studying traditional Chinese and even at the height of my proficiency there were plenty of rarer logographs that I'd frequently stumble over - only able to draw "blurry approximations" of them depending on my familiarity.
Coming from a phonetic language with only 26 letters, it was such a surreal feeling being able to effortlessly read a character but be unable to reproduce it.
My understanding (which may have been in the previous HN discussion on the topic) is that Chinese people just substitute with a homophone if they're really stuck and native readers can guess by context what the writer meant. Much like fudged spelling like when Americans mix up do and due
A similar thing happens with all kinds of iconography, from flags to logos. People can easily recognize many logos, but when asked to draw them they often can't come very close.
This is a fascinating subject, that would merit some input from other languages IMHO.
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
In alphabetic languages with separate pronounciation (French, English etc.) people forget spelling instead. And of course we recognize spellings we know, even if we couldn't recall it cold from memory.
Another point: the article defines kanji as images, but each character can be parsed as a composition of base blocks (radicals), and people primarily remember those. Put in computer terms, kanji are not bitmaps but overlayed vectorial blocks. That I think makes a world of difference for how we handle them, foreigner and native alike.
Learning about radicals was a massive Level Up in my Japaneseability.
THey're like protocols/interfaces on OOP classes ^^ or more accurately, like "tele" and "phone" in "telephone", "icon" in "iconography" and obviously many more words in English made up of bits and pieces of other words.
Once you recognize the parts, you'll be able to able to guess/infer what the word means even if it's your first time seeing it. Same with radicals in Kanji.
I have no trouble reading but writing kanji has become a problem. I never need to do it and I can’t remember how to write kanji I have no trouble reading.
It’s Japanese people too, to a lesser degree. My own Japanese wife has to pause to remember how to write something every now and then.
Not when doing fragments in obvious context like that. For one, tone doesn't disambiguate anyways. There's 10+ common characters with pinyin 'dan', and only 4 (5) tones in Mandarin.
But also when the context is super obvious there, there's no need.
This was happening to people I knew when I lived in Japan 30 years ago. Many people were using wa-puro (word processors that let you type in the phonetic form and choose from the appropriate kanji). I imagine the effect is far more common now.
I remember one time when a university engineering professor couldn't remember how to write the kanji for "police". He didn't seem embarrassed asking someone else. I don't know if they still do, but they would often demonstrate by writing out the character with their index finger like a pen in the other hand's palm.
I went through Heisig's book THREE times, writing out each character and reviewing them multiple times each in Anki. I had several notebooks like the one in the OP article.
I still can't write most kanji.
The fact is that you simply cannot learn to write without practicing writing (in real contexts, outside of the notebooks); there is overlap, but it's the same thing as learning the "theory" of a math problem and then blanking once you're asked to solve a problem on a test.
If I try to visualize a word/character that I know, I often end up with a vague picture of the most salient lines, but missing some of the finer detail that would be required to write it out accurately. And I think this is largely fine, as handwriting characters isn't a very important skill anymore. Especially in Japanese where you can always fall back on kana and be understood.
I do still think "learning" the kanji is helpful because there is a system behind it, and understanding that system helps to learn words. To give an example, learning the 教/kyou in 教育/kyouiku (education) helps you make a connection to 教室/kyoushitsu (classroom), and helps you distinguish it from 勉強/benkyou (study) where you'd really think the kyou would be 教, but it isn't!
But crucially, this is all a visual recognition problem, nothing to do with producing the characters yourself. So I think the ideal learning approach is word-first but with some attention to the characters in the word, especially how they relate to other words with those characters/radicals or with the same readings.
Years ago I asked my teacher what percent of street merchants or high school kids would know how to write some of the more complicated, uncommon characters from the advanced textbooks. She replied street merchants maybe 50%, but high school students 100% -- during high school every student will be at their lifetime peak literacy, because they will be cramming and memorizing how to write every possible character in preparation for the college entrance exams.
I tell locals I can "speak, read, and type", but only have 1st grade writing abilities in Mandarin. The key is typing -- you can chat on Line, WeChat, email, or any mobile apps to accomplish day-to-day tasks. But the typing is phoenetic either via Roman pinyin, zhuyin, romaji, etc.
So yes, this problem only comes up in languages where drawing the characters are mostly independent of the pronunciation. You don't need to type individual words, you can type phrases and the input tool presents choices sorted by most commonly used. Therefore, even sending communication is closer to a reading exercise rather than a writing exercise.
Not born & raised in Japan but went to Saturday school for Japanese in Europe and I was excited to be able to read (and understand) the newspaper for the first time at middle school, because my Kanji caught up. From there it's usually very quick how much Kanji you can learn.
But now, 20 years later? When I need to go to fill forms I flip up my phone and search for Kanji all the time. And I know I'm not alone (although probably very bad by Japanese standards), but I can navigate Japan just fine.
I always struggled to explain this to my European friends. "I can read and talk fluently without any issue, but writing not so much. Unless on keyboard".
The best analogy I've come up with is "If I ask you to imagine an apple or a Motorbike, you can do it right. When you see one, you'll instantly recognize what it is. But if I ask you to paint an apple, or a motorbike, you might not fare too well; people might mistake your painting of a motorbike for a bicycle. It's something like that. Using keyboard is like googling for images and copying it in to your PowerPoint slides"
A lot of Japanese learners do hate katakana (personally, a lot of fonts could stand to be clearer about ツシンソ ), because most writing is in kanji+hiragana so they have less practice with katakana. But kana ability is really just exposure. Use it to get used to it.
Same reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
I have experienced this for simplified Chinese. I studied some Chinese while studying computer science in China. The classes would have us learn writing, reading, speaking and listening (sensibly so for a Chinese language class).
Being able to write characters was handy whenever I came across documents that needed to be filled, but since leaving China I never had the need to write characters again. I now just input them using pinyin on keyboards, and I can easily recognise and read / input the correct characters. It is a strange feeling trying to write the characters I once knew, but now have forgotten, yet being able to read them instantly...
I would like to recommend dong-chinese, a language app I came across when I prepared for my stay over there. It taught things in a very efficient manner.
At this point I would like to recreationally increase my vocabulary so I have started working on a game called LingoRogue. My goal is to make it addictive to play, with a sneaky vocabulary-increasing effect. In other words a game that is "learnified" rather than a learning software that is gamified.
I'm convinced that different people process handwriting (and movement) differently. This is true with latin languages as well.
I think this may explain the difference between recognizing shapes versus drawing them for some people.
I remember when I was in school, some people had really neat handwriting, they could write fast and all their letters looked exactly the same with apparently little effort. On the other hand, I had to focus hard to ensure that my letters were all the same style, shape, size and slope... Also, I didn't have a single 'handwriting style' I could write in a number of different styles. I couldn't have both speed and nice looking, consistent letters; it was one or the other.
The interesting thing though is that I was always quite good at drawing... Conversely, I noticed that the people who had beautiful, effortless handwriting would typically be quite bad at drawing... They were the kinds of people who had to start out every drawing as a bunch of circles, triangles and crosses before joining them together to form the final drawing.
I feel like these people automate their hand to some extent. It's like a reflex to them. It lets them render common shapes without much thinking or effort.
It reminds me of that time I did a drawing class and the teacher kept reminding students to "stop thinking in symbols and just draw the different shapes and shades as they appear."
This probably has parallels in a number of areas like sports (e.g. tennis) where being able to offload certain movements to muscle memory can free up your brain for more strategic aspects of the sport.
This also reminds me of Daniel Kahneman's book "Thinking fast and slow." I suspect it would be interesting to try to categorize people based on what kinds of mental activities they offload to system 1 vs system 2 thinking.
> In other words, what feels like a single, monolithic "literacy" ability is actually two distinct skills, each exercised in different instances and each capable of improving and decaying on its own.
This dissociation has been used to test theories of hemispheric specialization. A good overview is in Neurolinguistic Aspects of the Japanese Writing System by Michel Piradis (1985).
I mean, this is not a politically correct statement, but I think one line of reasoning from this is to say that Chinese characters (which is what kanji are) are not a great way to write down language for practical purposes. A friend of mine in grad school irreverently referred to the Chinese writing system as "a really huge, really inefficient syllabary", and I think there's some truth to that. The characters no doubt have a certain beauty and their history is interesting, but a system where the meanings and pronunciations have to be learned totally separately seems to be inherently cumbersome in some ways. Even in a language like English which abuses the Latin alphabet in a notoriously messy manner, the amount of phonetic information that can be gleaned from the written form is fairly high, which gives two paths to the word (via memorized whole-word recognition or incremental sounding-out).
Overall you are probably correct, but there are certain benefits from the Chinese character system.
Long time ago I studied Japanese in Japan. On the way back to my home country I was sitting next to a bunch of Chinese people on the plane who did not speak any English or Japanese, but we were able to have a small conversation using Kanji/Chinese characters, because the characters' meanings are usually the same, although the languages are quite different. If the people would have been Greek and could not speak any English, no conversation would have been possible at all.
Another thing to mention is the radical system. Many Chinese characters consist of two or more characters, of which one is the "radical". This often helps you understand the broad meaning of the character in case you do not know it. For example, the Japanese character for fish is "魚". If you know that character and see another unknown character that used "fish" as a radical (for example "鮭"), you know that the character probably describes some kind of fish (in this case salmon). So it is not simply a huge list of "syllables".
I know almost nothing of Chinese languages, but some Japanese, and when I see boxes and instructions for some random item from Temu that is all in Chinese I'm amused by being able to understand bits of it.
I saw a face mask box that was all in Chinese and had 非医 on it (which means non-medical) which, depending on your font would be written exactly the same in Japanese: 非医.
> Overall you are probably correct, but there are certain benefits from the Chinese character system.
I mean yeah, it's an ill wind that blows nobody any good. But consider this: We could do the same thing in English as in Japanese -- replace loads of letters with Chinese characters: Write "跑ing" instead of "running", and so on. The French and German and Russians and Spanish could all do the same thing with their languages; and then when traveling, people could at a basic level read the signs and menus, and communicate at a basic level by writing, without having to know anything about the language.
Would you choose, post-facto, to add Chinese characters to English?
For my part, I'd say "no way". English orthography is already hard enough for my son to learn, without having to add characters on top of it.
ETA: Just for kicks, I asked Claude to try its hand at writing the opening lines of Pride and Prejudice in this manner:
It is a 真理 universally 認知ed, that a 単一 男 in 所有 of a good 財産 must be in 欲 of a 妻.
However 少 知n the 感情s or 見解s of such a 男 may be on his 最初 入ing a 近所, this 真理 is so 好 固定ed in the 心s of the 周囲 家族s, that he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
"My 親愛 Mr. Bennet," 言ed his 夫人 to him 一日, "have you 聞ed that Netherfield Park is 貸 at 最後?"
Mr. Bennet 返答ed that he had 不.
"But it is," 戻ed she; "for Mrs. Long has 丁度 been here, and she 告ed me 全 about it."
Mr. Bennet 作 no 答.
"Do 不 you 欲 to 知 who has 取n it?" 叫ed his 妻, 不耐.
"You 欲 to 告 me, and I have no 反対 to 聞ing it."
> and then when traveling, people could at a basic level read the signs and menus, and communicate at a basic level by writing, without having to know anything about the language
Well, you have to be careful; something that actually happened to me was recognizing that the first element of 牛蛙 meant "cow", without recognizing that the second part made it "bullfrog".
> We could do the same thing in English as in Japanese -- replace loads of letters with Chinese characters: Write "跑ing" instead of "running", and so on.
Interesting choice. Japanese doesn't do that - it follows classical Chinese by using the character 走 for the sense "run". 跑 is Chinese-specific.
> Just for kicks, I asked Claude to try its hand at writing the opening lines of Pride and Prejudice in this manner
This comes off weirdly to me because it's so Japanese. But I guess that was the idea. I have some particular questions:
>> this 真理 is so 好 固定ed in the 心s
I haven't checked on the English text, but it's difficult for me not to read this as "this truth is so well fixed in the hearts...". But I'm not sure that that sense of 好 is available in Japanese, where I'd expect it to be a verb meaning "like".
>> he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
This doesn't appear to be a problem, but I still felt the need to call out what looks like the use of an exclusively masculine pronoun (in modern Chinese) to refer to "daughters".
>> Mr. Bennet 返答ed that he had 不.
This seems wrong; I feel sure that 未 would be more appropriate than 不.
>> Mr. Bennet 作 no 答.
This would appear to render the English "Mr. Bennet made no reply." I'm not comfortable with the use of 作; just because the word in the idiom is make doesn't mean that any making is involved.
Do you know whether Japanese use of kanji focuses more on establishing that a kanji corresponds to some Japanese syllables and using that kanji wherever those syllables occur [as suggested by "作 no 答"], or more on using kanji to represent certain semantics however those semantics might be pronounced [as might suggest "言 no 答"]?
>>> he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
> This doesn't appear to be a problem, but I still felt the need to call out what looks like the use of an exclusively masculine pronoun (in modern Chinese) to refer to "daughters".
IIRC, in Japanese, 他 is sometimes used for "hoka", which means "other"; as in, "Do you have this item of clothing in other (他) sizes?" The original text says, "some one or other of their daughters"; so that seems to be Claude's thinking.
> Do you know whether Japanese use of kanji focuses more on establishing that a kanji corresponds to some Japanese syllables and using that kanji wherever those syllables occur [as suggested by "作 no 答"], or more on using kanji to represent certain semantics however those semantics might be pronounced [as might suggest "言 no 答"]?
I know a moderate amount of Mandarin (along with smatterings of Cantonese), and about 3 months' spare-time study of Japanese in preparation for a recent trip there. (Note the shopping theme in my example above.)
My expectation, which matches my (very small) experience, is that Japanese are trying to write Japanese; and that therefore they have Japanese words in mind that they're trying to represent with Chinese characters.
It is absolutely not the case that a given Chinese character will always be pronounced the same way in Japanese. For one, multi-character Chinese words are also munged into Japanese words. For two, even for single-syllable words there's a context: e.g,. Japanese has two different ways to say "one", but they use "一" for both. I'm sure I've run into other Chinese characters that are pronounced differently depending on whether they're in a verb or an adjective.
So for example:
> Mr. Bennet 返答ed that he had 不.
> Do 不 you 欲 to 知 who has 取n it?
Remember that we're discussing a hypothetical universe where English speakers use the Chinese characters, but the vast majority of them don't know Chinese.
In English we use "not" in both cases, even though in Mandarin in the first case I'd probably construct a sentence using "没" in the first instance and "不" in the second.
So what would happen in our hypothetical universe? Given that most speakers don't know Mandarin, using "不" in both cases is probably the simplest, most stable result. One could imagine complicated rules for whether you write "不", "没", or "未", which are taught in school and maintained, even though they're all read as "not", and even though most people don't know Mandarin. But it would have to be one of those things which (like Chinese characters themselves) people decided they liked about their writing system and didn't want to give up.
Similarly:
> I'm not comfortable with the use of 作; just because the word in the idiom is make doesn't mean that any making is involved.
And yet, in English, "made a cake" and "made no reply" have exactly the same verb. Again, one can imagine it going both ways, depending on how things were established and maintained: One could imagine using "作" in both cases; or one could imagine using two different characters for two different shades of meaning, just as in Chinese there's 他, 她, 它, and 祂; or perhaps 的 and 得 (which I'm not sure people would naturally consider different words if they were only exposed to the spoken word).
The real limitation here is that there's no way ready-made to indicate the "make" -> "made" transition, as you can with "makes" ("作s" ) and "making" ("作ing").
> I'm sure I've run into other Chinese characters that are pronounced differently depending on whether they're in a verb or an adjective.
The Japanese writing situation is much, much worse than that. Kanji get all kinds of different pronunciations on the theory that the semantics are the same. So the standard spelling of musume ["daughter"; "girl"] is 娘, but it might also be spelled 女. This gives you an alternative to the standard independent reading of 女, which is onna ["woman"]. All characters will also be pronounced differently when they represent Chinese loanwords than when they represent native words.
> And yet, in English, "made a cake" and "made no reply" have exactly the same verb. Again, one can imagine it going both ways
I agree with that.
> The real limitation here is that there's no way ready-made to indicate the "make" -> "made" transition, as you can with "makes" ("作s" ) and "making" ("作ing").
I don't agree with that; you'd do the same thing you do with every other preterite verb and write 作ed. You wouldn't pronounce that maked, but that's not a problem; we're still assuming that everybody knows English.
> IIRC, in Japanese, 他 is sometimes used for "hoka", which means "other"; as in, "Do you have this item of clothing in other (他) sizes?" The original text says, "some one or other of their daughters"
Yes, I determined something similar by looking the character up in a Japanese dictionary. I don't have the skills to determine what kinds of uses are and aren't natural, so I just decided that I couldn't label the usage wrong, but it still stood out as funny for other reasons. I was able to correctly read the English without referencing the original text.
I'm not sure what a "moderate amount of Mandarin" means; if it's on the lower end, you might be interested to know that the "other" sense does survive in Mandarin, in the words 其他 ["other"; extremely common] and 他人 ["other people"; not so common]. 他 by itself is going to be overwhelmed by the far more common use as a pronoun, I would guess.
> My expectation, which matches my (very small) experience, is that Japanese are trying to write Japanese; and that therefore they have Japanese words in mind that they're trying to represent with Chinese characters.
While I do agree that the situation could shake out in several ways, I don't think this is a total defense of the idea that "the same word" is going to be spelled the same way by people who don't know Chinese. I wouldn't expect, for example, that fire [flame] and fire [eliminate from a job] would get the same character spelling.
> One could imagine using "作" in both cases; or one could imagine using two different characters for two different shades of meaning, just as in Chinese there's 他, 她, 它, and 祂
By my understanding, 祂 is not a part of ordinary mainland usage. But this is a good example, in that there is no distinction between the words in the language, and Chinese people aren't able to make the distinction when learning a foreign language even though they do make it in writing their own. (By contrast, Spanish speakers don't have problems choosing between the English words he and she.)
You might also be interested to know that the character 做, pronounced identically to 作 and meaning the same thing, is in fact derived from 作. It (and not 作) is now the ordinary character used for the sense of making or doing. 作 is best known to me as part of the word 作者 "author". This is a pure spelling distinction that arose by some natural process within Mandarin.
> or perhaps 的 and 得 (which I'm not sure people would naturally consider different words if they were only exposed to the spoken word).
I can answer that; they don't. At least they don't for the 得 that introduces manner or result clauses; they might or might not think of the possibility infix seen in 听得了 "able to listen" as different.
> I'm not sure what a "moderate amount of Mandarin" means; if it's on the lower end, you might be interested to know that the "other" sense does survive in Mandarin, in the words 其他 ["other"; extremely common] and 他人 ["other people"; not so common]. 他 by itself is going to be overwhelmed by the far more common use as a pronoun, I would guess.
I had a snippet about how 他 was made up of 也 and 人, "also a person", but ended up editing it out, as I wasn't sure it was actually connected w/ Japanese using it to mean "other". "他人" is still about people, but I had forgotten 其他, which is clearly not specific to people.
At any rate, there's a lot of ways our hypothetical universe could go, WRT how such a 汉字 writing system would be incorporated into English. The point is, there's always a silver lining: It would certainly have some benefits, like making it possible for literate English-speakers to get around in China and Japan w/o learning anything about the local languages and vice versa (with some of the "false friend" [1] traps you've mentioned above -- but those are issues between European languages as well).
But on the whole, I'd consider the cost not worth the benefits by a long shot.
> I had a snippet about how 他 was made up of 也 and 人, "also a person", but ended up editing it out, as I wasn't sure it was actually connected w/ Japanese using it to mean "other".
Well, I don't think it's true, either. As I understand it, the form 他 is just a graphical simplification of 佗, with no connection to the character 也. (You can think of this as similar to how the left-hand component in 脸 is 肉, not 月.) It isn't clear to me either that the character 也 would have held any sense of "also" when the form 他 appeared, though this is a question I'm agnostic on.
I've watched several videos on youtube in which someone presents Japanese people with uncommon kanji and asks them to read them, or failing that to speculate on what they might say. The Japanese unfailingly speculate that both halves of a compound character are relevant to the meaning, which surprises me - it's very rare for a character to be constructed from two meaningful parts. Far more common are the characters in which one part gestures at the meaning and the other part tells you the pronunciation, and 佗 is one of those. 从人它聲。
>have to be learned totally separately seems to be inherently cumbersome in some ways.
It also has quite big advantages. Because writing systems tend to be highly standardized meaning of most characters has changed relatively little, so Chinese people can read ancient texts with aid of a bit of classical Chinese. And this also works across space, China has a lot of mutually unintelligible spoken dialects but pretty much everyone can understand the meaning of written Chinese. And so can even Japanese or Korean people.
In contrast spoken language tends to change so quickly that phonetic writing systems can rapidly become completely unintelligible. Old English is practically a foreign language. Probably nobody understands that "Hwæt. We Gardena in geardagum" means "Hey! We the Spear Danes, in years gone by".
The problem with these advantages is that most actual uses of words depend on some amount of context, and only a relatively small set of word+context combinations can actually work across languages that way. It is like the equivalent of doing a word-by-word translation.
Another way to put it is that this is essentially no longer "writing language", it's some alternate form of meaning representation. Language is not just a sequence of independent words.
It also affects the learning curve. English-speaking children can read far more advanced books earlier than equivalently educated Japanese.
They probably don't outweigh the disadvantages, but there are some small benefits. Once fluent, I think I remember Japanese to be slightly faster to read because of the more unique shapes. And you can make more flexible and elegant graphic text designs and tables (like in Excel) given the compact words and natural vertical writing.
My children are Japanese-English bilingual and can read far more advanced books in English. Initially I took this as an imbalance and suggested they read the same books (or something very close) in Japanese. But their native Japanese language teachers said, no, because of the different learning curves you can't expect them to read the same level of Japanese texts; the equivalently educated/advanced Japanese reader will be behind, at least in the elementary school years.
Japanese as a whole are extremely avid readers, so I don't think there's a gap at the top, only the shape of the learning curve.
If native speakers are starting to have character amnesia too, does that suggest in the long run you would expect the writing systems to simplify towards the phonetic syllabaries? Or is the fact that we have computers as a mediating tool going to forestall that and just make things weird?
IMEs have actually caused the opposite a little in Japanese, because you type the phonetic pronounciation and get a list of possible kanji, then with computer use "writing" kanji really means "recognising" and so people will use kanji that they never would have bothered with remembering how to handwrite.
I'm just a Chinese student from the West so take it as just my two cents but I don't think it will evolve to phonetic syllabaries. Chinese has a lot of homophones so it's useful while reading to have an extra semantic meaning. They also say that once you're used to it, you read faster but at my level I can't confirm it. So with modern input devices you're basically simplifying the hard part of the characters, which is writing, and keeping the reading part where they're better than pure phonetic systems.
I learned Japanese around the same time (2010~). I even lived in Japan and studied at a Japanese university.
To this day I have a hard time recalling Japanese on the spot. However, when I hear Japanese or see it written, I can read and understand well. Even kanji I haven't seen in a decade. It's kinda strange.
I'm pretty sure if I re-studied Japanese I would learn very quickly as I re-activate that part of my brain... something that has been on my todo list for a long time!
Writing out Chinese characters definitely helped me learn to read them as well (not surprising, since engaging multiple senses helps one retain information better), even though in practice I never had opportunity to write them out because I always used either a phone or computer.
Unlike the author, I found most of the mnemonics as much trouble to learn as the characters themselves, and soon stopped using that approach. It just didn't work for me.
Is ワープロ馬鹿 really a term used by native Japanese speakers? As far as I can tell it only really shows up in Japanese->English dictionaries and English forums (see https://www.google.com/search?q="ワープロ馬鹿"+-a+-the).
> Is ワープロ馬鹿 really a word used by native Japanese speakers?
You probably mean idiom. ワープロ (word processor) and 馬鹿 (baka: idiot) are individually both words used by Japanese speakers. Japanese speakers would be more likely to say 漢字健忘 (kanji amnesia) to refer to the phenomenon though.
I had a friend in college whose mother was an immigrant and whose father had grown up bilingual and spent years in the mother's county. My friend was able to hear and understand the mother's tongue with apparent fluency, but had an extremely difficult time forming real sentences in the language, even though he'd heard and understood it for his entire life.
I think there's a significant difference between consuming language and producing it, whether that's written or spoken. It might explain the number of people I see these days who don't understand basic English grammar and make words plural by slapping on an apostrophe and an s.
Being able to read but not write Kanji is so common that it's a meme amongst Japanese people -- to the point where it's a game. For example, here you can watch some Japanese television people play a game where they compete to write words in 10 seconds or less:
> What confuses me is that other people can form images in their minds. Are all those with character amnesia also aphantasic? That can't be, given that aphantasics amount to less than 5% of the population, while a much larger number of people forget how to write (70% of teenage participants in a Chinese TV show were unable to write the word "toad"!).
They were discussing their aphantasia as a precursor to other very interesting points, e.g. about how "seeing" a character in your mind isn't enough to be able to draw it, --> verbatim traces and gist traces.
Completely off topic but that is the shakiest camera work I have ever seen, and I learned to shoot before gimbals were cheap and easily available. For that matter, it's been easy to automatically correct that sort of thin in post for years now. How were they not embarrassed to upload something so unwatchable?
English is written using an alphabet. Alphabets are by and large phonetic so you do not need to remember how to pronounce every word, you learn to read the alphabet and specific spelling (e.g. 'th' in English) and you are mostly done.
Chinese is written using ideograms ('kanji' in Japanese), which convey a meaning but not a pronounciation. So when you encounter a new character you cannot pronounce it.
> Chinese is written using ideograms ('kanji' in Japanese), which convey a meaning but not a pronounciation. So when you encounter a new character you cannot pronounce it.
This isn't quite true - about 80% of Chinese characters are so-called phonosemantic compounds, where people originally started using the character for one thing for another thing whose word sounded similar (say, emoji for "can", as in able to) and then adding a semantic component to differentiate the character from other similar-sounding ones. In Chinese, they smushed the two components into the space of one character, but in eg. Egypt, they simply wrote whe semantic clarifier and the phonetic hint side by side, full size.
That is, the majority of the characters are primarily sound-based, it's just that the connection between a character and its sound is shoddy, even in Chinese languages.
Japanese kun readings for native words do divorce the characters pretty completely from their sound.
You cannot guess the pronunciation of a character from the way it is written. At most the "phonosemantic compound" might provide a clue to possibly reduce the space of possibilities.
It's not pedantic in that the characters themselves really are sound-based and provide a pretty decent clue. That said, it's still just a clue, and there are multiple similar pronunciations associated with any phonetic component, so it's still guesswork in the end.
And so we are back to square one: it is not possible to know how to pronounce a character by looking at it (and any clues are not "decent" in the majority of cases), as opposed to alphabetic systems...
The difference is that for most words you can easily get 90% there in English and, even if you don't read it properly, the listener can somewhat get what you are trying to say.
And say you don't know how to read "conscientiousness" but you know how to read "con" and "ousness" you can try to go for con-shu-tiouness and at least you're somewhere.
It's not the case with Kanjis, sure you might know one part of a word but you might be wrong. Also similar looking kanji don't read similarly at all, so even that is out of the window.
And don't get me started on the approach of applying memoization techniques purely on the strokes of a kanji, that's gonna hurt more than not.
It actually works the same way in Japanese. Radicals give rough clues about meanings as well as pronunciation. Similar looking kanji do read similarly. It's just... you can read pronounce "brochures" in either French style or in Southern American, and which one is more appropriate depends.
Not always, as I wrote in another comment this is because of the Japanese script reform for "modernising" and simplifying the script. So sometimes kanji look similar but are not. More info at: https://en.m.wikipedia.org/wiki/Japanese_script_reform#Mazeg...
The advantage though is that even if you don't know the word, you can guess the meaning if it contains familiar character(s). Also except for the first one, your examples are easy to distinguish.
Yep and no. That's a common misconception. Like Chinese, old Japanese went through a systematic simplification of its kanji 100 or so years ago. The rules were decided a priori to make the job easier. This meant that several kanjis ended up looking somewhat similar even though they have nothing to do with each other. That's why comparing kanji can work but it can also backfire and lead you astray.
Words in English are built from characters we can pronounce individually and this extends to words so whilst you might not get it perfect you can at least guess the sound of an English word. This isn't the case for characters in Mandarin or Japanese, no sound is encoded in the strokes, the equivalent of a character.
Actually English spelling is pretty complicated and completely different from pronunciation. Why on Earth you Englishmen write "a" (as in skate) but pronounce it as "ei". And why you have so many versions for each vowel and why you have 3 letters for "k" sound and 2 letters (i, y) for the same sound.
Very interesting, definitely the first step towards writing a character for me is picturing it in my head. If it's particularly challenging I might, still in my head, project it onto the page and that seems to give my brain the spatial data to begin translating it to real world movements with the pen.
But even still I also can barely write maybe 5% of the kanji I can read. As well words are often made of multiple kanji, but if you showed me the kanji separately I don't always recognise them as part of a word I do know. Recalling a kanji into my minds eye doesn't seem to be part of the skillset of reading, maybe just a by-product of long term repeat exposure.
This thread is reminding me that I desperately want a way to way to have different font sizes for different languages. I like my English text small, but kanji/Chinese characters are obviously more visually dense. So I default to 10/12 point for latin alphabet but I need somewhere between 20-25 point to read Japanese comfortably. Where the two are mixed, like in this thread or on social media (where maybe half my feed is Japanese) it's a massive pain in the ass.
I've noticed the same thing (not with Japanese), but even after learning to write in Korean, my ability to handwrite faded way faster than my reading skills. I figured it was just laziness, but it makes sense that reading and writing use totally different parts of the brain.
That's why I focus on learning words rather than individual kanji characters. Kanji characters have so many different readings, and the Japanese language is full of exceptions that it's not worth the time unless you're passionate about it, eventually you will pick up the readings of most common ones and you will be able to guess new words automatically, writing them no thank you, I barely write in my own language without a keyboard.
Nice post! Enjoy your blog's overall aesthetic too. Perhaps correlation in sense of style, though, as I also used RTK to learn to write Kanji, loved it, and now, ~15 years after that escapade, am kind of in a similar bucket (can write some characters, but mostly just read). I still think RTK great overall method and would do it again!
Also, shoutout to Fabrice, creator of Kanji Koohii -- that was my first foray into SRS back in ~2007/2008, after which I found Anki (pre-mobile).
I find myself having trouble remembering the spelling of certain words, now that autocomplete is almost ubiquitous. It sometimes takes writing them out and looking at the result ("Is it 'beleive'? No, definitely 'believe'.") to remember the correct spelling. Chinese character recall is probably closesly associated with that than with romaji recall.
What drove me crazy when I started to learn Japanese was that a Kanji characters can have so many different pronunciations. The most egregious example is 煙草, whose pronunciation is the really just tabacco(tabako, or タバコ). I knew the etymologies and the historical context on why Japanese evolved like that, but it's nonetheless hard for me to remember all the pronunciations, at least initially.
Also, I find that knowing Kanji is essential in appreciating part of Japanese culture. Take their addresses, for example: Kinukawa is really meaningless, but 鬼怒川 is such a amazing name. Similarly, Akihabara means little if all we know is the pronunciation, yet its Kanji 秋叶原 is such a beautiful and poetic name that invokes complex emotions.
This is probably why Japan still adamantly emphasizes writing.
Written resumes/ fax machines ... remain the norm, and while this may seem anachronistic for the rest of the world (pretty much all of which uses either (semi-) phonetic scripts derived from Aramaic or from Brahmi), it makes sense after you come across the Chinese characters.
No, it's because Japan is a deeply conservative country, really adverse to change, because change means taking risks. If you ever had to deal with Japanese you would know, layers upon layers of convoluted procedures for no real reason.
But hey, at least now I can get the "my number card" to get my hanko certificate from the convenience store. Maybe in 30 years I will be able to use the card directly to prove my ID without needing the stupid hanko!
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
No but when I last tried to hand write a long text in the latin alphabet, my hand hurt after a while. Skills atrophy when not used, more complex skills may atrophy faster.
This reminds me of one time I mentioned to someone I had aphantasia and their response was “how do you spell!?” Seems wild to me that some people see words in their head to spell them but I guess at least one person does. I do wonder if that means they’d have better kanji recall for writing.
I don't consider myself to have aphantasia. If I close my eyes and try to 'see' with my eyes the letter 'm' or an apple, all I see is the back of my eyelids– pitch black.
If I then try to 'see' with my mind the letter 'm', I can imagine the shape and drawing the shape, but it never appears in a physically visual manner. I can trace its lines with my eyeballs, but try as I may to hallucinate an image, it's still only pitch black always. The closest I come to seeing it is being very confident that I know exactly what 'm' looks like, and that I could take that mental model and draw it exactly on paper immediately.
Do some people have such a clear and strong mental image that they can effectively inspect, zoom in and manipulate a mental image as well as a real visual image, and that's considered not-aphantasia?
> Do some people have such a clear and strong mental image that they can effectively inspect, zoom in and manipulate a mental image as well as a real visual image
yes, 100%, and more than that. Even with eyes open I can overlay a completely different environment and stop "seeing" the real world. When I close my eyes I find it difficult to really see blackness. Example: when running laps, I count laps by seeing a giant number fixed in the sky over the lap marker, each number a different material (flaming, made of ice, a trimmed hedge, etc).
It helps distinguish them. Seeing all numbers as e.g. black vinyl makes them blend together in my memory.
Visualizing something lets you leverage visual and spatial memory, but even then: if I were really running past N real giant numbers which all looked identical, I'd lose track just the same. Distinguishing them visually makes them all unique and memorable. The color infuses the entire track and the sky, actually, so it requires little focus, because it's right there in the background.
Basically like the memory palace from Moonwalking with Einstein, but less work, because they don't have to be consistent over time.
Idk, I just read the Blake Ross' (of Firefox) article [1] on having aphantasia, and identify less, rather than more, with how he describes it:
"1. Can you picture my face? >No. But it’s not personal.
2. So you don’t know what I look like? >I know facts ... If you have radiant blue eyes, I may have stored that information. ... I’m unable to project it visually in my mind because there’s no screen.
4. How about picturing something simpler, like a red triangle, or the table right in front of you? >I can’t even understand the question. I can think about the idea of a red triangle. But it’s blackness behind my eyes. Blackness next to my ears. Blackness in every nook and kindle of my brain.
5. You’re just assuming that others can actually SEE things with their eyes. NOBODY can do that, you hypochondriac.
I get it. It’s a “mind’s eye.” I don’t have it."
It's unconvincing. They sort people into categories by a questionnaire, and then find they perform differently on tests that have something to do with vision or imagination. That's it, that's all they've come up with.
> The (semi-official) Japanese language tests do not test language production in _any_ way, even on the highest levels.
This is simply not true. The Kanji Kentei is probably the most popular Japanese language test Japanese people might take outside of school, and it tests writing heavily.
If you're talking about the JLPT, that's a test meant only for non-native speakers.
> all single-choice answers.
Not even multiple choices? Maybe they should reconsider that one.
>This is simply not true. The Kanji Kentei is probably the most popular Japanese language test Japanese people might take outside of school, and it tests writing heavily.
I would not qualify Kanji Kentei as a "language test" any more I would qualify a Spelling Bee as one; though I can see how someone might.
In any case; I thought that it was relatively clear from the context that I was using "language tests" as a shorthand to "tests for non-native speakers of that language"; but I did not explicitly mention that, you're right.
>Not even multiple choices? Maybe they should reconsider that one.
Is that not how these are called in English?
Where I'm from, there's a distinction for tests where given a list of answers, there can be only one valid one ("single choice") and where there can be multiple valid ones ("multiple choice").
Quick googling on my end indicated that _is_ how they're referred to in English, but the websites were quite AI-sloppy, so if I was mistaken, I'd love for you to let me know what the more widely understood terms are.
Since you used the phrase "highest levels", that made it sound to me like you were talking about native tests.
I apologize for misunderstanding.
Non-native japanese language tests do not ever go to a high level as far as I'm aware (N1 definitely doesn't classify as a high level, it's around "can read the newspaper very slowly and with greater than 50% comprehension", which is well below even the least-studious native-speaking adults).
> Is [single choice answer] not how these are called in English?
It's an unusual term to my american-english ears. I've only heard them called "multiple choice questions" regardless of how many answers you can select. It's unusual enough that it conjures to the mind a multiple-choice test where each question has only a single answer, say "a", and you just circle "a" for every answer.
I have recently had the experience of Labguage School in Japan and they also realy heavily lean into having just circle the correct parts or do somethign like insert correct Particle here.
While I have tremendous respect for the Japanese Language and people. The Japanese in all my exposure to it do not know how ot tech languages. Be it their own to people that do not know it or English to their own people.
> This is also why I believe that language is a bottleneck for thought. Most of what you remember is nothing like an approximate copy of the things you experienced in real life—even in the specific case of text, memory is not even remotely like a paraphrase of previously read words. Many of our thoughts happen in a highly abstracted and distilled form, interacting and connecting with each other as a network that simply cannot be faithfully converted into a sequence of words, however long.
Perhaps the most interesting quote in an interesting article.
you're walking down a beautiful waterside bridge in kyoto during cherry blossom season. there's really delicious smelling grilled eel from the store next to you. there's a swedish lady walking by you with a cute accent. your skin feels slightly sticky from humidity.
what i just wrote doesn't even begin to encapsulate the entirety of that moment. there were a million other details your brain can form about that moment: the style of brick on the bridge, the other people around you, the sun being in your eyes, that you're smelly bc you forgot deodorant that morning.
to put it shorter: a picture is worth a thousand words. and the author is saying that by having to use language to describe pictures, we have a huge bottleneck
Perhaps if there were fewer radicals this would be less of a problem. Many thousands of characters could probably be generated from a small number of radicals.
32 radicals are enough for ~75% of characters though. My biggest problem with writing is that I default to writing fairly small so if I have to do a character with more than 12-15 strokes it's either way too big or hard to read later.
when I learned Japanese in university I learned writing with the exact method by James W. Heisig. So I didn't used Japanese for about 20 Years since, I forgot almost everything I learned, except the Kanji. I can still remember most of them. It's crazy!
Interesting topic but a lot of the articles on this blog reads like undeclared LLM slop:
> This is also why I believe that language is a bottleneck for thought. Most of what you remember is nothing like an approximate copy of the things you experienced in real life—even in the specific case of text, memory is not even remotely like a paraphrase of previously read words. Many of our thoughts happen in a highly abstracted and distilled form, interacting and connecting with each other as a network that simply cannot be faithfully converted into a sequence of words, however long. The fact that people can fail even at something as basic as sketching a kanji or a vehicle they've seen hundreds of times before is just another example of the same phenomenon.
A pet peeve of mine is when someone uses their personal datapoints to generalize for humanity. Every sentence here should say “for me”.
What they said doesn’t even support the idea that language is a bottle-neck for thought, it actually argues against it. If language can’t capture the complexities of thought, then that’s the opposite point as language is a bottle-neck for thought.
“In techie terms, the Chinese script doesn't support the structure of languages like English and Japanese. It doesn't have what it takes.”
That’s not techie that LLM slop
“Sometimes, instead of using them for their meaning, they used them for (gasp!) their pronunciation. By ignoring the original content of a kanji, they could string them together to form almost any sound.”
I’m becoming allergic (gasp!) to this kind of writing
Are you sure? If anything, spelling along makes it hard to read Chinese or Japanese. Examples: can Japanese people correctly understand this: “かんじをせんねんいじょうつかっていたけいいからかんじをはいしするとぶんしょうによるいしのそつうにへいがいがしょうじるからです”？ In fact, the following two different sentences both have the same above kana：
Or how about くさくさくさくらくらくさくさくさくさくさくさくらさくさんさくさんくさくさくらんくらくら？
Or how about こうないしゃせい？Is it 校内写生 or 口内射精?
In all seriousness, even Koreans still debate whether they should continue to use some Chinese characters, and they do so in poems and literatures. For instance, 악유원 means shit, yet 乐游原 has layers of meanings. Chinese and Japanese have too many homephones to use a spelling system like Korean. Unlike Chinese, where individual characters correspond closely to morphemes and carry distinct meanings, Japanese is organized around words as the fundamental linguistic units. Furthermore, because Japanese employs a pitch-accent system rather than the kind of lexical tones found in Chinese, writing some native Japanese words (和語) entirely in kana typically does not create confusion.
Also, Japan had at least three major attempts to remove Kanji. First time: The Tale of Genji, the world’s first novel, was written by a woman entirely in kana, with only a small number of Buddhist terms in kanji—terms which, in theory, could also have been rendered in katakana. This demonstrated that early Japanese could be expressed systematically and fully without using any Chinese characters at all. Nevertheless, the Japanese chose to retain kanji.
Second time: During the Meiji Restoration, slogans such as “Abolish kanji, or East Asia will never grow strong” were popular. In the end, Japan found that it could modernize, prosper, and even become one of the five permanent members of the League of Nations while still keeping kanji. The decision was made to continue using them.
Third time: After World War II, General MacArthur commissioned a team of American education specialists to work with Japanese experts to discuss whether to abolish kanji entirely, and to consider the possibility of fully romanizing the Japanese language. Yet, they decided to keep Japanese.
I don't find the argument about homophones very convincing, since people do speak to each other and it works. People will surely be able to adapt.
But, you're basically throwing away the country's history. In just a few years young people wouldn't be able to read any older text anymore. I can see why you wouldn't want that.
It is substantially harder and slower to read a stream of hiragana than to read a stream of hiragana and kanji. I don't know if you know anything about the language, but it's not because of homophones (though that is certainly improved) -- it's chunking.
Japanese written language has the property that when you see kanji followed by one of a very small number of hiragana patterns, you know you're seeing a verb. Stemming and deriving the meaning of the verb is trivial, because it closely follows from the Kanji (e.g. 見る => 'see', and pretty much every noun or verb involving 見 carries that connotation).
Toss in the few particles (を、で、に、へ、が、は), and you've broken a sentence into semantic chunks with very little mental effort -- and along the way, gained much of the meaning at the same time.
Doing the equivalent with a stream of nothing but hiragana requires a kind of parsing that is like depth-first prefix search, but with ambiguous matching at each terminal. It's incredibly tedious.
I can read Japanese, yes, and obviously also find hiragana-only text hard to parse. I think that would be almost completely solved by using spaces and getting more practice though.
I'm much better at reading than speaking/listening, so the Kanji also help as a clue to the meaning, but that is entirely a non-native problem I think.
> the Kanji also help as a clue to the meaning, but that is entirely a non-native problem I think.
Not sure what you mean by that. It's almost a daily occurrence that someone tries to explain a word's meaning by drawing invisible kanji on their palms.
I guess the ultimate question is whether we can have a spelling system that totally mimics how we communicate verbally, so chunking becomes as easy as when we listen to the the spoken language.
I doubt it will happen though. Historical heritage aside, Kanji does pack enough information to make reading very easy. Unless government interferes hard, the new system needs to be way better than the existing one to get adopted.
It doesn't even need to do that. Korean and Vietnamese are easy examples where all the writing is phonetic despite Korean having basically the same structure and problems with Chinese loans as Japanese, and Vietnamese having a ton of Sinitic loans and general properties that are basically the same as Chinese. Yet phonetic writing just works.
I mean, if they got rid of kanji, presumably they would introduce spaces. It's not an alien concept, it was used in early computers and early computer games, and still get used in games and books aimed at kids:
(Pokemon also has a full kana mode, but due to the number of homonyms in Japanese, I think while it might be easier for actual Japanese children, even with the spaces it's harder for a Japanese learner than the kanji mode unless you're literally in your very first month. There's also things that will be obvious to natives but confusing to learners like that すげー in that screenshot is a slangy すごい)
Yet, that's exactly what you do when you read English - and many of us even read it as a foreign language with different roots and different grammar etc. So I don't understand how it would be more tedious in hiragana; it looks like a "just get used to it" thing to me.
Japanese has fewer sounds and a lot more homophones than English, so purely phonetic writing becomes a lot more ambiguous. Children grow out of it at an early age.
Furthermore, kanji often allows the reader to skip phonetics altogether, because the symbol itself carries the meaning. It’s a bit like how people understand emojis without having to make a sound for them in their heads.
As a result, the Japanese are able to read very fast in their native language, so switching to a different system would carry significant drawbacks.
There's an advocate group for all-kana Japanese that existed since 1920, just decade or two after pseudo-Chinese styles was deprecated and kana-kanji mixed as-spoken text became the standard, before even kana switched from katakana to hiragana, and they don't even use all-kana text on its official propaganda page anymore[0], so there's that.
No you don’t. There’s actually a lot of research showing that native English speakers don’t read letters, they read word shapes. And people who read fastest read in even larger groups of words, to the point of eliding segments of sentences or paragraphs entirely.
Chunking is incredibly important for reading speed, and reading hiragana is much closer to reading letters than words. My reading speed in Japanese is nowhere near native, but the way I’ve gained speed so far is almost exclusively by increasing my minimum comprehension unit: I see word patterns, common grammatical constructions, etc., and I don’t need to read them.
> native English speakers don’t read letters, they read word shapes.
Don't try to teach children to read this way, though: it's a high-level technique that comes with practice and familiarity, with near-instant fallback to lower-level techniques as appropriate, or the resort of dyslexics who cannot read any other way. Teaching children to use the approaches used by struggling readers will tend to produce more struggling readers than necessary.
In the real world nobody is masochistic enough to not adopt spaces if writing without kanji.
Old Japanese videogames couldn't use kanji due to technical limitations. They wrote in all kana but used spaces to make the text easier to read.
Modern Japanese children's books and eg. even Pokemon games still? Same thing, kana and spaces.
When Korean transitioned away from Japanese-style mixed script to purely alphabetic writing, what did they do? They adopted spacing.
The only time "but Japanese doesn't have spaces" comes up, ever, is when people argue against the removal of kanji. It's not a realistic scenario, in light of very recent history and current practice.
In a world where hiragana and katakana are adopted to eliminate characters, it's not that hard to imagine also adopting spacing. Hangul for Korean did not originally have spaces in the language.
It's all solvable. Korean did it. Koreans are fine.
For verbs you could just replace the kanji with katakana so that you still get the pattern recognition of okurigana as a visual aide.
They just don't want to do it, because they don't like to change. A generation educated under a new system wouldn't have difficulty using it. They would however lose access to an abundance of cultural artefacts which play a central role in daily life.
Koreans don't seem entirely fine to me, but besides, Japanese pronunciation is actually more aggressively simplified than the other two CJK languages during Edo era that numbers of homophone is out of control.
There are as many as 50 homophones for koushou due to this, for example[1]. Communication by phonetic transcripts alone just isn't going to work.
I've spoken Japanese for the last 15 years. I'm aware of how many homophones there are. I still think by and large it wouldn't be a problem, and that natives would adapt to route around it. It's not a blocker.
It's mostly about lack of practice, frankly. The chief reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from. No one actually reads these English posts by vocalizing letter by letter to slowly build the words together.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
Similar arguments used to be made in Korea, yet if you look at Koreans today they have no difficulty in reading hangeul - they have spaces so the words have form, and they have mass exposure to the hangeul forms of words. Ergo, their shortcuts are for the phonetic forms and those are what feels natural, solid, meaningful and easy to read. Same as both of us in English or me in say, Finnish or Swedish or French.
Maybe you're right that it's all just hard-headed stubbornness from fluent people. But I (and all other learners of Japanese) started by reading hiragana, and only later did we move to mixed script -- this is by necessity. Yet even when you factor in the difficulty of learning to read Kanji, reading Kanji is vastly easier than reading kana, even as a beginner.
Would spaces magically solve this problem? I guess it would solve some things -- you'd no longer have to guess where to terminate the prefix search, and I think you're right about word shape -- but it would definitely not provide the additional semantic context you get from having the high-bit-density characters in the mix. This makes reading faster.
I suspect that one could make a kana-only writing system that would be functional enough, but it would still be slower to read than mixed script. Also, the Korean comparison isn't exactly valid -- Korean has more sounds than Japanese. It seems minor, but Japanese has a ton of homonyms because of the tiny phoneme. Expanding that, even slightly, would be a benefit to reading.
Korean faces the exact same problems as Japanese, though - the language structure is similar, they have a ton of Chinese loans, and have in general gone through a largely identical history of writing development. They have somewhat fewer homonyms than Japanese, sure, but they still have tons from Chinese loans (hell, "coffee" and "nosebleed" sound the same, as do "blood" and "rain" in many cases).
It's somewhat hard to believe that Japanese sits in some magic spot where a phonetic script wouldn't work just fine when Korean does it fine, and on the Sinitic side people write books in pinyin, Vietnamese is phonetic, and the Dungan people write their 3-tone Mandarin dialect with cyrillic alphabet without even notating tones.
> Maybe you're right that it's all just hard-headed stubbornness from fluent people.
It's not just hard-headed stubbornness - reading kana really is more difficult to proficient readers of today's Japanese, and change is work.
One of the advantages of Chinese characters is that you can express an emotion using a single character - like "laugh" or "tears" and you don't need emoji for that. Western alphabet cannot do this.
Also the text is more dense and you can write more using limited number of characters, for example, when you have a button in UI with limited size.
You can't possibly count the third time as a real attempt. A language reform initiated by a foreigner and recent enemy at war, who burned your cities and even nuked two of them? If someone knows about successful examples, I'd be curious to hear about them.
Latin script for Vietnamese replaced Chinese-based script under the French colonial government and were helped by the Nguyen Emperors going along with it. (Interestingly this did not take off in neighboring Cambodia or Laos.)
Hanja is also mostly gone in Korea, particularly in North Korea.
The big thing is that both shifts happened before rapid literacy growth. It's much easier to teach new writing systems when the majority of the population can't write anyways. 95% of Vietnamese could not write in 1945; only 22% of Koreans were literate in that time period.
---
One interesting thing I learned while researching this comment was that a big reason Hanja disappeared was because Koreans gained literacy during the typewriting era, but before computer auto-suggested keyboards, and it was just substantially easier to make and use a letter-based typewriter.
> One interesting thing I learned while researching this comment was that a big reason Hanja disappeared was because Koreans gained literacy during the typewriting era, but before computer auto-suggested keyboards, and it was just substantially easier to make and use a letter-based typewriter.
If you look in the right places, you can find people complaining about how it's impossible to dynamically render hangul blocks, which means that a Korean font needs to define glyphs for every possible Korean syllable as opposed to just defining the elements of the system and letting a word processor assemble them as appropriate.
If that's true, I don't see how hangul could have had any typewriter-based advantage over hanja. From the typewriter's perspective, there's no difference.
The Chinese used typewriters by defining a typewriter code. Assuming that that was necessary for hanja, and also for hangul, why would it promote the disappearance of hanja?
If a typewriter code wasn't necessary for hangul, how did we forget how to lay out the blocks in between then and now? Hangul have been in continuous use for all that period.
> If that's true, I don't see how hangul could have had any typewriter-based advantage over hanja. From the typewriter's perspective, there's no difference.
There are mechanical hangeul typewriters that, while more complicated than Latin or katakana typewriters, are still completely usable for normal writing. The reason hangeul fonts are hard is that a hangeul syllable occupies a standard-sized block, and in eg. careful handwriting the writer would adjust the sizes and positions of the characters to be aesthetically nice. For example, in 해 he the ㅎ andㅐ letters are both the same size. When you write 핸 hen, see how the h especially becomes smaller? In typewritten hangeul, that first consonant is always that small, so you can use only one size of initial h and so on.
You can see from the output in this video how the sizes of letters are very standard and somewhat disproportionate, eg. in CV type syllables the vowel lines are somewhat giant compared to the quarter-of-the-block sizeish consonants, etc.
That way you can still write by pressing alphabet buttons, with some controls as to where you want the letter to go in the block. It's a bit more complicated, but nothing compared to the nightmare that are proper Chinese character typewriters.
> For example, in 해 he the ㅎ andㅐ letters are both the same size. When you write 핸 hen, see how the h especially becomes smaller? In typewritten hangeul, that first consonant is always that small, so you can use only one size of initial h and so on.
I see the opposite. As those characters render in whatever font my browser picked, the ㅎ in 해 occupies much less vertical space than the ㅐ does.
In the 핸, it still occupies less vertical space, but the difference is smaller. It's about as tall as the left-hand bar of the ㅐ instead of being significantly less tall than that.
No, it is a problem exactly equal to the other one. No typewriter can produce 98,000 different characters. And no typewriter can produce 11,000 different characters.
With zero difference between hangul and hanja, how can the typewriter favor one over the other?
There were Chinese typewriters but they were very large and a lot more annoying to use. Japan also used typewriting. They just look a lot different, with a giant cylinder of tiny keys to facilitate thousands of characters. https://www.youtube.com/watch?v=wQ-SHOsbH4Y
Hangul ones were a lot smaller. I misspoke earlier; they had letter-based keyboards and mostly just did compromises on the shape of the syllabic blocks; a keyset for an initial character, a keyset for medial characters and a keyset for terminal characters. If you just assume the initial character set can be tiny to fit both the characters with and without bottom terminals, then you wind up with slightly odd-looking but perfectly serviceable typed Korean.
I've always wondered how some countries manage to drop Chinese characters, while others can’t—or just don’t want to. And how did Vietnam and Korea manage to understand their historical texts after they stopped using Chinese characters? And how do they create new words nowadays? I guess they just borrow words and pronunciations directly from English or other foreign languages?
From what I understand, Chinese characters carry so much meaning that they’re really hard to replace. Sure, we don’t need them when we’re speaking, so in theory it seems like we could just get rid of them altogether. But in practice, we still rely on them a lot - especially when we’re trying to understand what a certain pronunciation means. And reading is a whole different story. Recognizing characters is just way faster than sounding out spellings. Maybe one reason is that the basic unit in Chinese is not word but morphemes, which mostly are just single characters.
Maybe we could come up with a different writing system, kind of like what Koreans did, instead of sticking with Romanized pinyin.
One thing to note is that for non-Chinese languages, Chinese may have semantic meaning but really has nothing to do with pronunciation; and the Chinese etymology really has nothing to do with how a Korean or Vietnamese speaker independently came up with the word so this is less important.
—-
Another thing to note is that a lot of “Chinese” words in the modern day language are actually Japanese in origin, since Japan was the first country using Chinese script to modernize and adapt Western thought in science and philosophy etc., and the associated terminology. This actually provides a political impetus to replace those words with native-constructed ones since they have a negative historical relation with Japan.
—-
People who replace Chinese script with letters often have dictionaries and whatnot from the transition period to trace back words, and people also still learn Chinese in these countries if they want to, so it’s not as if it’s gone and disappeared; in the same way that modern speakers probably couldn’t read Chaucer in Middle English, or Beowulf in Old English, as it was written on a whim, but there are plenty of scholars who have studied for it. And Modern Chinese has little to do with Literary/Classical Chinese anyways.
—-
Mandarin has a phonetic system, Bopomofo, which was abandoned for political reasons in the PRC. But the problem with replacing Chinese script is political; within the PRC and ROC there are multiple mutually unintelligible languages using Chinese script, and if you pick a phonetic script then it is now Mandarin vs. everybody else.
> And how did Vietnam and Korea manage to understand their historical texts after they stopped using Chinese characters? And how do they create new words nowadays? I guess they just borrow words and pronunciations directly from English or other foreign languages?
The answer to all of these are the same as everyone else. It's not like Chinese people routinely read the old classics in the original, and even then they're literally taught Classical Chinese in school as a separate language, because it is. But other countries have scholars who study the old languages and also work to translate classics into modern language for wider use - I certainly can't read Greek but have Aristotle's Nicomachean Ethics sitting on the bench next to me, translated into modern English.
Creating new words happens the same as it does everywhere - people genuinely coin new concepts, they loan foreign concepts as direct loans or calques, or form compound words from existing ones. "creating new words" with Chinese characters is literally just using foreign words to form compounds, something utterly routine everywhere else. It's just not treated as a magical event elsewhere.
For example, people routinely marvel at the ability to make compound words in kanji. But, may I introduce you to German, a language written with boring old non-mysterious Latin script that's famous worldwide for its heavy use of compound words.
> From what I understand, Chinese characters carry so much meaning that they’re really hard to replace.
They're really just scribbles that point at words or morphemes of the language they're used to write, same as phonetic scripts in that regard. They do let you do some tricks that are hard otherwise, like indicating which nuance of a word you intend by your choice of character, and the common stylistic trick of writing some set of characters and then imposing completely arbitrary readings on them by writing clarifications next to the characters being abused.
For example, in Frieren, in one of the early chapters, Frieren says:
Zorutoraaku wa hito wo korosu mahou dewa nakunatta.
In hiragana:
ゾルトラーク は ひと を ころす まほう でわ なくなった。
The manga writes it like this, however (furigana in parentheses after the kanji):
人を殺す魔法(ゾルトラーク) は 人を殺す魔法(ひと を ころす まほう) でわ なくなった。
人を殺す魔法 should be read "hito wo korosu mahou", ie. "magic that kills people", but the manga instructs us to read it first time as "Zorutoraaku", the name of the spell, and the second time properly, when Frieren's supposed to say its description out loud. There's no clarity issue here, it's just a stylistic trick.
For fun, the same line from the Korean translated version:
졸트라크는 더 이상 인간을 죽이는 마법이 아니게 됐지
Jolteurakeuneun deo isang inganeul jugineun mabeobi anige dwaetji.
J. Marshall Unger's Literacy and Script Reform in Occupation Japan goes into details on this.
There were kanji abolitionists and pro-kanji people in both the American and Japanese administrations, and it definitely wasn't just a popularity contest. Things happened such as one pro-roumaji principal who enthusiastically took part in roumaji feasibility experiments being assigned elsewhere because he was having results, or one American pro-kanji official decreeing that roumaji publications should be published in triplicate since there were three competing romanization systems - Nihon-shiki, Kunrei-shiki and Hepburn - so they wouldn't unduly advantage any particular romanization system.
This of course also just so happened to make roumaji publishing three times more expensive. Whether fairness or limiting roumaji publishing by financial means was the real motivation is left as an exercise to the reader.
I don't think it's that crazy for the language reform to make it through, seeing as the entire political system has been reshaped into a pseudo-American democratic structure for instance. With the emperor going out of fashion and the governance of the country being overthrown, I think it would be the perfect moment for other cultural shifts as well. Had the timing coincided with the introduction of computers (which famously struggled with languages like Chinese and most Japanese) a few decades later, I think the plan would've succeeded.
Most of Europe inherited the Latin script from Christianity, which was spread with less than peaceful means. It took more than a few short years, but it certainly altered culture.
Japan had a parliament since the Meiji era (first elected in 1890). The Macarthur occupation changed quite a few things, but less than most people think.
My understanding is that Japanese are very open to external influences, even when they come from enemies. Case in point, they embraced the legal framework and political reform that was pushed by the Americans. On the other hand, MacArthur was careful to respect Japanese culture (arguably, depending on who looked at it, I guess) when pushing his reforms. Nevertheless, Japan quickly became a pretty well run democratic country. It’s fascinating and respectable that Japanese could embrace western civilization while retaining their beautiful culture
I dont know who amongst the Koreans argues for the use of Chinese Characters. The vast majority of Koreans don't know even the small amount of Characters they still use. Even simple ones.
There were online discussions on abolishing Kanji in Japanese (漢字廃止論). There were tons of interesting examples, and I took some notes. Unfortunately I don't remember the original sources now.
You didn't even try to do it the way Koreans did it, though. When Koreans dropped hanja in favour of a hangeul-only script, they adopted spacing because spaced text is much easier to understand than unspaced text. Likewise, products like old Japanese videogames which couldn't use kanji due to technical limitations used spaced kana because it's just much better and no one's masochistic enough to not use them if they don't have access to kanji. The Japanese mixed script doesn't use spaces largely because kanji already serve as word dividers. If you remove that function, everyone sane will use spaces, and pro-kanji arguers will leave them out to make the proposition seem mad.
Importantly, spaces also make it much easier to recognize the shapes of the words.
It's the chief reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from. No one actually reads these English posts by vocalizing letter by letter to slowly build the words together.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
Meme sentences designed to be hard to read that you'd never see in real life aren't an actual point. By that token, the Buffalo buffalo sentence argues for the urgent adoption of kanji in English.
Likewise, in my native Finnish:
- Kokko, kokoo koko kokko kokoon.
- Koko kokkoko?
- Koko kokko.
ie.
- Kokko, assemble the bonfire.
- The entire bonfire?
- The entire bonfire
It's perfectly readable despite the meme value.
Similarily, "kuusi palaa" can mean:
kuusi palaa = the spruce is on fire
kuusi palaa = the spruce returns
kuusi palaa = the number six is on fire
kuusi palaa = the number six returns
kuusi palaa = six (of them, or six pieces of something) are on fire
kuusi palaa = six (as above) return
kuusi palaa = your moon is on fire
kuusi palaa = your moon returns
kuusi palaa = six pieces
Do we need to urgently adopt kanji in order to avoid homophony?
> Chinese and Japanese have too many homephones to use a spelling system like Korean.
> Third time: After World War II, General MacArthur commissioned a team of American education specialists to work with Japanese experts to discuss whether to abolish kanji entirely, and to consider the possibility of fully romanizing the Japanese language. Yet, they decided to keep Japanese.
J. Marshall Unger's Literacy and Script Reform in Occupation Japan goes into details on this. There were kanji abolitionists and pro-kanji people in both the American and Japanese administrations, and it definitely wasn't just a popularity contest. Things happened such as one pro-roumaji principal who enthusiastically took part in roumaji feasibility experiments being assigned elsewhere because he was having results, or one American pro-kanji official decreeing that roumaji publications should be published in triplicate since there were three competing romanization systems - Nihon-shiki, Kunrei-shiki and Hepburn - so they wouldn't unduly advantage any particular romanization system. This of course also just so happened to make roumaji publishing three times more expensive. Whether fairness or limiting roumaji publishing by financial means was the real motivation is left as an exercise to the reader.
All good points. I really can't explain why it is perfectly okay for natives to listen to Chinese/Japanese yet it is so hard to parse Pinyin or Kana when reading, even when spaces are provided.
Maybe it is because we still think in individual characters when reading. As a previous post mentioned, there are many homophones for a given word, let alone for a single character: https://en.wiktionary.org/wiki/%E3%81%93%E3%81%86%E3%81%97%E.... It's not a problem in a conversation because we track context. Yet when we reading, we, at least psychologically, look for specific meaning per character or per word. Say, when I see こえん, I'll have to figure out if it means a park, a small yard, sound, or something else. Of course, I could figure it out from context, but I'd give up that precision and the warm and fuzzy feeling of seeing the corresponding characters, like 故園, or 子園, or 呼延。We'd solve this problem by giving up learning the individual characters and focus on words or phrases, but then it would be a drastic change.
While typing the above, I also realize that maybe the reason we don't mind homophones during a conversation is because we've already learned the associated characters. Using the example above, when I hear the word "こん” and it means "呼延”, I would know that it's an ancient family name that was associated with many famous generals because of the meaning of the characters. Or if the meaning is "故園”, I would get all the poetic feelings as it is precisely these two characters, not the sound, that deliver the meaning of ancestral land, or childhood home, or place left behind, and etc. And when we study the Chinese and Japanese poems, we focus on the masterful use of characters, and every character matters. Is it 推 or 敲 in 僧敲月下门? Why are the characters in 大漠孤烟直 so compelling and masterful even though each character is so commonly used? It's hard for me to imagine how homophones can differentiate such meanings.
Of course, I'm not saying that removing characters can't be done. I'm just trying to figure out the current state and why many people and I are in favor of learning characters.
> A system fails when its natives don't know how to use it.
They definitely do, it's just that hand writing has become a more niche use case in modern society (regardless of the language).
If you removed all kanji in a block of Japanese text (replaced with kana), I'd expect at least a 50% reduction in reading speed for natives, and some errors in comprehension. They're fundamental to the language.
You wouldn't expect the same 50% reduction if the natives had been educated under that paradigm and suitable adjustments made to the system like adding spaces etc to accommodate it.
That reads as a foreigner suggesting to abandon many thousands years long history and culture just because the characters are too difficult for him.
It doesn't even make sense because today we have computers and ML. If you want to write something, you just type the spelling, and the program automatically converts it into proper characters. And if you don't understand the character, AI can translate it for you. I am sure this can be integrated into electronic glasses.
Japan has had pro-romanization societies since the 1920's, and even during the last attempt at large-scale script reform after the war, it wasn't just the Americans pushing it: Many Japanese were enthusiastic about moving to a phonetic script because they perceived it as more efficient and modern. Likewise, not every American administrator was in favour of reforming away the kanji, far from it.
J. Marshall Unger's Literacy and Script Reform in Occupation Japan is a good treatise on the subject.
I also gave you two examples of cultures who already did it: Koreans did it autonomously, the Vietnamese did it during occupation. Both are successful today.
But yes "preserve history" at the expensive of daily life. I'm sure you speak Latin too, plenty of history lost there.
This is a common misconception until you study the language seriously. Japanese nominally has only 50 sounds, known as mora. In practice there are more because people smush them together, but this tends to be for very common words or dialectical variations by region. There are also spoken pitch markers, but they don't have any written indicia.
Because there is such a small number of basic sounds, homophony (words with the same pronunciation but different meanings) is absolutely rampant. This is not too big of a problem when speaking because the context is usually obvious, but writing anything abstract using phonetic script alone will quickly become problematic. It's also tedious to read long texts in phonetic script in the same way that it's tedious to read lots of upper case text in the Latin alphabet.
Another language feature is that Japanese compound words are often formed by just taking 1 or 2 more from several longer words. This would be like shortening 'American-Russian diplomacy' to 'amru diplomacy'. Very clear when written with specific characters, confusing if it is just phonetic. Then, word pronunciation changes within compounds, eg 川 kawa means river, but 山川 yamagawa is mountain river, and these changes are ubiquitous, probably 95% of words go through this phonetic alteration when folded into another word.
It gets more complex again because there are so many particles, words that indicate grammatical inflection of some kind and can drastically change the meaning of a sentence if you use the wrong one. The most common ones are just a single mora and most others are just two. But there are lots of particles. 10-15 essential ones, ~30 that are used all the time, ~65 you need to know for fluency, and more that 150 you might come across in written form.
I could go on and on. The basic reason for all this is that Japanese is fundamentally different from other languages because it developed on islands, and because the weather patterns around those islands made sailing in and out more difficult than islands in calmer latitudes separated by greater distances. They adopted Chinese characters because they didn't have a written language at the time they encountered them, then adapted those characters to develop the phonetic scripts. There's a huge amount of technical debt that would be impossible to unravel.
Do it like Korea
Korean has more different sounds, eg 14 consonant and 10 vowel sounds, vs ~10 consonant sounds and 5 vowel sounds in Japanese. You could also argue there was a higher level of literacy over a longer period, but that requires a lot of history and geography explanation to support - the point being that it would have been much harder to impose by imperial fiat unless it had been done much earlier.
long reply here bc this is a Special Interest of mine.
I'm going to put aside the "so difficult" thing here mostly, because the perceived difficulty is partly modern teaching practices not updating beyond "just write the kanji 100000 times ok" (this is a failure of people, not the writing system), and people actively choosing to not write anything at all, which isn't necessarily a problem because nowadays recognition is more important, and I don't see people panicing about how many people don't know how to spell many english words anymore.
Alternatives would be nightmarish and culturally destructive to implement in the modern day. Further, people confuse poor implementation with complexity being the cause of the problem. Taiwan and HK both have high literacy rates despite using traditional characters, yet simplified characters were apparently necessary to increase literacy? It doesn't compute.
Just under 50% of Japanese words are loanwords from Chinese. Recognising characters allows a ton of written nuance and extra vocabulary which the Japanese take full advantage of. There's at least 10 separate words for "kou kai". Even when not fully-remembered, Kanji allow mental mapping of multiple homophones without issue.
We're really talking about a vocabulary mass exctinction event. Back in the ancient times Korea didn't have such a wide-ranging and culturally-mixed set of words as now. If the Japanese do this it'll be cutting off the vast majority of their cultural history. Seems unideal considering the butchering the French did, did exactly that to Vietnam's literary culture and basically cut the people off from much of it.
Chinese doing this would be insane. Mao was tempted to use roman characters instead of the simplification they rolled out, which from a system design perspective only made a worse and more difficult/confusing system with more exceptions to rules than before, and poorer phonosemantic consistency/relations with other characters. Further because as research has proven many times over, it's harder to recognise many excessively simple & similar characters compared to more unique and specific forms. Thankfully Stalin advised him against it.
We really are talking about doing something similar to a total spelling reform of english and just throwing the last 1000 years of literature, written records etc away. Only in the case of especially Chinese, you're throwing away a system that was developed and specifically tailored to the languages using it, which have due to its relative robustness to change as compared with latin letters, led to writing surviving many many difficult periods, regime changes, wars, famines etc over 3000 years.
If you've ever compared Beowulf with Gawain, with Shakespeare and then modern English, you'll understand what a total overhaul a language can undergo if unchecked -- Chinese characters have enabled a comparatively stable orthography. Less than 800 years and it becomes gibberish.
> many people don't know how to spell many english words anymore.
Yes but that's not really comparable.
You can misspell "litrally" but how close can you get if you can't remember any stroke of the glyph? There's an inherent advantage in spelling words closely to how they sound.
I'm not saying English is great at this either, but I can still write "Kernel Sanders" and you know what it means, without using a whole other writing system as fallback (pinyin)
Mw out here in Japan being barely able to rrad any kanji but being able to hold reasonable small talk in Japanese. Spoken Communication yes written forget it.
Not sure if someone else has already commented on this, but this is the fate of people who primarily use LLMs to code. They’ll be like second-generation children of immigrants, who can vaguely understand the meaning of their parents’ native tongue, but lack the ability to produce utterances on their own.
The ultimate goal of AI companies imo is to disincentivize people from pursuing programming as a career by loss-leading with low LLM prices until nobody actually knows how to code without an LLM, and then jack up the prices.
"I used to know how to write code in ${language} but not anymore, though I can still read it. There is even a term for it: LLM baka (LLM馬鹿), meaning LLM idiot"..
I enjoyed this read, but I am noticing that people who claim to have aphantasia seem to write about themselves and their experiences an awful lot. I doubt the phenomenon is real.
Some people doubt that sun-sneezing is real, so I can entertain the possibility of being wrong. But sun-sneezing is trivial to demonstrate to doubters, and it doesn’t confer any “I’m special points.” No one would pretend to have it, unlike aphantasia.
I would find it more surprising if the brain did not have quirky versions like this, given how complex it is and given how often our genes are not perfectly copied.
This article actually hits on a pet peeve of mine where I feel people sorta “mystify” kanji/hanzi unnecessarily.
The truth is that there’s actually nothing particularly weird about being able to read some kanji but not be able to write them…
You actually get close to my point here:
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
Yes! But have you ever heard of someone forgetting how to spell certain words in a language that uses the Latin alphabet (e.g., English)? I can use myself as an example here: while writing this comment, I forgot how to spell “peeve” in “pet peeve” (I thought it had an ‘a’ in it) and I also forgot how to spell “unnecessarily” (I thought it had one n and two c’s).
The western equivalent of being able to read some kanji but not write them is simply called bad speling. No need to mystify kanji in particular.
Nice comparison. When I was living in China, I'd encounter someone who forgot how to write the characters for a word on a weekly basis. I think the difference is that with Latin alphabets you can still misspell something, and having gotten it down on paper, still rely on phonetics to convey your meaning.
Alphabet is such an underrated invention. It's probably higher in significance compared to the invention of wheel. It's the original "bicycle of the mind". For example, Korea pivoting from Chinese characters to its own alphabet or Hangul is very well documented including the positive effects it has in the much improved Korean literacy and civilization after the conversion. Fun facts anyone can learn Hangul alphabet in a single day if they wanted to but the same cannot be said to Chinese characters. If your mother tongue is Korean (e.g Korean American) that only just started learning, it only take one day turnover from illiterate to literate.
Scripts being the main driver of literacy is a pet peeve of mine. It's not the script, it's the schooling system. The high rates of literacy in modern states are just a result of the school system - Japan has a high literacy rate, for example, and their writing system is either the worst in the world or close to it.
That said, the characters are a whole boatload of unnecessary extra effort, and as a student of the two languages, the artificial illiteracy created by kanji, where I often just can't read words I've known for years, is simply maddening. Not having to wrestle with characters does free up a lot of time for both native and foreign students alike.
>their writing system is either the worst in the world or close to it
Yes, it's probably the worst since even Microsoft until now still struggle to provide proper search solution for Japanese names in their Windows OS due to their multitude of writing systems.
By sheer wills of course you can make everything hard feasible but that does not means it's efficient and effective. I consider Japanese as a unique country with extraordinary people that can collectively overcome adversity, that's include a non intuitive and difficult writing systems.
How many Korean-Americans know the Korean words for things like 'legislature', 'inflation', or 'geopolitical tensions'?
This is an area where the modern insistence that English isn't phonetic baffles me.
It’s probably more phonetic than Chinese but significantly less phonetic than Dutch.
The term that would cover what you mean here is regular. And that is only in regards to correct spelling. Is obviously complicated when considering that we don't have official pronunciation across all dialects for the same word. Even if we do agree on a spelling.
But it is a complete non-sequitur to lead to the modern idea that English isn't phonetic.
This is not about regular VS irregular, there are aspects of English spelling that are highly non-phonetic. It's not uncommon to have letters in words that are entirely irrelevant to the pronunciation. For example the spellings "programme" and "program" would be read the same by any English reader, and yet both persist in certain places. The s in island is completely unnecessary.
Also, the same English word can be read in very different ways by the same speaker, but in different contexts. This is most proeminent with some of the most common words in English - a, the, there, and many other connective words can be pronounced very differently by the same speaker in the same speech, depending on stress (for example "a" can be pronounced as either ə if unstressed or eɪ if stressed). And yet, there is no version of written English that differentiates these - another sign that English is not a phonetic spelling.
Of course, on the other hand, you can't say that there is no correlation between spelling and pronunciation, like you can in Mandarin and other Chinese languages.
That is what is typically meant by a regular orthography. Wikipedia also calls it deep and shallow. These are legit terms that pre-exist to this odd debate that English isn't phonetic.
Nobody that knows how to read English at a level to be on an internet forum is surprised that English has odd spelling. Many people would be deeply confused to be told that written English doesn't follow a phonetic system. Rightfully so.
This happened with cyrillic to me. During Yugoslavia we had to learn both latin and cyrillic. Since I'm from Croatia, I didn't have a need much for cyrillic (or at all). Today, I can read _at speed_ (including subtitles) cyrillic just fine, but I probably couldn't write a thing if my life depended on it. It's weird when I think about it.
> > Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
When I was in grade school, I took notes for my classes using Tengwar (elvish) runes as a way to alleviate boredom and force myself to pay attention (just taking the table in Appendix E from Lord of the Rings and transliterating English letters into them). I could do this at a speed sufficient to keep up with a teacher talking, so pretty fast. I cannot read any of these notes today, much less write them.
If you don't use this stuff for a few decades, you do forget, even when it's just an alphabet.
I learned how to write cursively in school, but as soon as it was no longer mandatory I switched back to print capital letters and some years later to print letters entirely, as it was just much easier for me to make legible.
At this point, it's fair to say I _have_ forgotten how to properly write cursive capitals, if I tried I'd just end up with print capitals with random tails for most of them.
I don't think anyone is "mysticizing" the language. It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
IMHO spelling is an inapt analogy. Every single word I can communicate orally, I can write. Sure I might mess up "I before E" or some other minor issue, but I'm BUILDING the word from first principles, e.g., syllable phonemes. That's why kids are taught to "sound it out" at a young age.
The closest equivalent you have in logographs might be radicals.
> I don't think anyone is "mysticizing" the language. It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
People absolutely do mystify the operation of kanji, like they're more than scribbles that point to words, or that Japanese would fall to incomprehensibility and ruin if they were done away with.
> It's just that logographic languages in general present a uniquely interesting problem whereby the auditory component is largely divorced from the written component.
This isn't quite true - about 80% of Chinese characters are so-called phonosemantic compounds, where people originally started using the character for one thing for another thing whose word sounded similar (say, emoji for "can", as in able to) and then adding a semantic component to differentiate the character from other similar-sounding ones. In Chinese, they smushed the two components into the space of one character, but in eg. Egypt, they simply wrote whe semantic clarifier and the phonetic hint side by side, full size.
That is, the majority of the characters are primarily sound-based, it's just that the connection between a character and its sound is shoddy, even in Chinese languages.
Japanese kun readings for native words do divorce the characters pretty completely from their sound.
Yes! I've studied Japanese for years and read numerous novels in it each year, yet I couldn't physically write to save my life.
But there's nothing crazy about that, like you said it's similar to spelling (not entirely, as I can spell things fine if I have a phonetic keyboard that "writes" for me).
Writing vs reading to me, is more about the type of memory.
Recognition vs recall.
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
I’ve actually done this all the time as a bilingual speaker with my language in Cyrillic. It doesn’t happen often but once in a while I’ll freeze and be unable to remember which glyph makes which sound when switching alphabets/languages.
>> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
> Yes! But have you ever heard of someone forgetting how to spell certain words in a language that uses the Latin alphabet (e.g., English)?
Erm... great that you gave English as an example but I'd argue this mostly applies to English. I remember wondering why "spelling bee" was a thing when watching shows from the USA because that wouldn't make any sense in Polish. Same with Spanish. And a lot of others. There are some minor things to remember and you can do errors but in 99,(9)% of the cases it's "you write what you hear".
It could have been a case for English as well but the reform efforts were killed so here we are with all it's quirkyness :)
[dead]
It's sort of the same like being able to recognize a bunch of people, but not being able to draw their faces to a good resemblance.
Recognition and production are separate skills.
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
Printing? No.
Cursive? Yes.
This is probably a better argument than my original comment (and much simpler!)
I actually can't write cursive at all, but I can usually read it fine. This is because I went to school when they stopped teaching cursive.
But alas, an article entitled "I used to know how to write in cursive" probably won't be very interesting to HN readers...
The reverse almost happened to me before, I write almost exclusively in cursive and once or twice had to stop and think about what a capital Q was supposed to look like.
Small nuance, but the term ワープロ馬鹿 actually is unrelated to software like MS Word, and refers to the at one point ubiquitous ワープロ (Word Pro) dedicated hardware device that many Japanese people owned in the 80s/90s to write letters. Read more here: https://en.wikipedia.org/wiki/Word_processor#Japanese_word_p...
Interestingly the English Wikipedia page above only mentions Japanese word processor devices in a small section, but the Japanese version of that page is almost entirely dedicated to these hardware devices: https://ja.wikipedia.org/wiki/%E3%83%AF%E3%83%BC%E3%83%89%E3...
I was lucky enough to live with a Japanese family in the early 90s and used one to learn how to type Japanese but also write letters home in my own language. I guess you had to live through this age to understand the difference of how Word Pro is used and the hardware association it has in Japanese.
The other weird thing about that term is that it's somewhat uncommon to see 馬鹿 actually written out in kanji, as those characters mean horse and deer with no direct association with 'idiot' (the proposed etymology is a reference to Chinese history and/or a Sanskrit loanword). I guess in a way the kanji form is a word processor autocomplete tell.
A Toshiba Rupo word processor (found in both links) is safely stored in the backyard shed, together with some floppies and documentation. That'll be my winter project next year.
Thanks for sharing this part of your life. That’s so cool.
When learning Japanese, I purposely chose to _not_ learn how to write any of it by hand. As the author notes, writing (by hand) is in fact a separate skill from reading. So I decided I would not invest my limited time, motivation, or brain space to writing.
Overall it's been a successful approach, and I recommend it to new learners unless they have a particular interest in being able to write by hand or they feel strongly that writing the characters helps them remember them.
It's only rarely that I have to write anything other than my own name in Japanese. I've practiced my address but writing it in English is fine in 99% of situations. Being able to write properly would save a little embarrassment, but I still believe my language learning time would have a much higher ROI in other areas.
I went to an old-school language school where I was forced to take tests in handwritten Japanese. I probably still have some of that in my brain, but like you, I almost completely abandoned it as soon as I didn't have to take language school tests anymore.
It's occasionally useful to write out a character, but on the whole, it's completely unnecessary now that we have computers with hiragana keyboards.
As a partial aside, the Heisig anecdote that leads off this piece is painful:
> Japanese children learn the spoken language first, then they learn how to write it in elementary school; Chinese students of Japanese (who tend to be pretty good at it) have pre-existing knowledge of character meanings and forms from their mother tongue, so they only have to learn how to pronounce them. Therefore, a Western learner should first focus only on the meaning and writing of those couple of thousand common characters and, only after having mastered those, should move on to studying the pronunciations.
Going from "Japanese people learn the spoken language first" to "you should spend a big chunk of time learning characters before learning sounds, words or grammar" is a pretty remarkable mental backflip.
The author says he spent eleven months doing this before devoting any time to the spoken language. If I could put the "head exploding" emoji here, I would do it. I spent only slightly more time than that at language school, and came out conversational.
Yeah I agree. I way over-indexed on learning kanji (via WaniKani) at the beginning of my Japanese learning journey. I got about halfway through before realizing it was silly that I could read 健忘症 but didn't know many very basic hiragana-only words. It wasn't timed wasted but it probably wasn't the most efficient approach.
In an ideal world maybe learners could focus exclusively on listening and speaking first, then move on to kanji later. But writing is a very useful tool in learning, and having access to that tool can help speed things up.
Like most things in life, a balanced approach is probably the right one. But you have to know what your goal is. Our brains are lazy, they only get better at what we make them get better at. If your goal is to just read kanji, practice reading kanji. If your goal is to understand and speak the language, practice listening to and speaking the language. But if you want to have a balanced language ability, you'll need to practice it all.
WaniKani at least teaches words. Spending almost a year of your life doing nothing other than learning "meanings" of individual Kanji is...well, I guess some people just really get addicted to that mechanical feeling of progress?
Reading is definitely helpful, but I've found the relative importance of reading, listening and speaking goes in cycles, and especially at the early stages, listening and speaking are far more motivating than anything else. And I'm an introvert!
> In an ideal world maybe learners could focus exclusively on listening and speaking first, then move on to kanji later. But writing is a very useful tool in learning, and having access to that tool can help speed things up.
There is no connection between these two sentences. You can learn to read and write in Japanese without ever learning a single kanji, and that's what everybody does do. Kana serve the purpose flawlessly.
With 0 kanji, it's such a small subset that it's hard to call that a finished job of learning to read or write as you'll be limited to material like kids books or NHK news easy.
Yes, I ignored Kanji completely while learning Japanese. I only learned words in hiragana and katakana from my SRS decks. I regret it because it locked me out of a lot of comprehensible input I could have used to actually progress in ways apart from learning the sounds of lots of words (even the sounds I learned were often wrong because I never heard anyone speak the words I was learning).
Am I the only one reading this thread and thinking, "Gee, this sounds like a job for AI?"
Why in the world should any human not from Japan devote limited time and brain capacity learning to read and write Japanese? It's literally a robot's job, at this point.
It will take a lot of time for your Japanese language skills to get up to par with robot translation, that is true. It was also arguably true with google translate many years ago too.
But isn't this true for most things? It will take a complete beginner years to draw as technically proficiently as an image generation model, or to code as well as Claude. Even before AI, most Japanese media has been available in English translations for years now, and there aren't that many other interests where you could find Japanese speaking peers but not English speaking ones.
If your goal in life is to generate the most economic value per unit of time input, maybe then learning to draw, code, or speak Japanese no longer makes sense. And if that's your priority, you won't choose to do these things. But that's not why people take up these pursuits. So I don't think AI will have a huge impact on how many people start them.
In an ideal world, some kind soul would escort kanji behind a shed and we'd hear a loud bang, then marvel at how we can suddenly just read words we've heard for years.
Wanikani did add some common hiragana-written words into various levels. Not many, but at least nobody hits 60 without learning これ anymore.
Yes, I agree that trying to learn kanji upfront is a silly idea.
Heisig says in the introduction to RTK I that he learned 1900 characters "before the month was out". If like him you can do the whole set in a month and then have no further need of formal review or study beyond using them as they turn up, then I can see it not being a terrible idea. But as far as I can tell, almost nobody has a mind that works like Heisig's does: people seem to need longer and to rely more on review via an SRS like Anki.
Personally I found my problem with RTK was that I successfully memorised "English keyword to write the character" for 2000 kanji, but this was not at all linked to my actual use of the language, so I still had the problem I started with of "I want to write the word べんきょう but can't bring to mind the kanji for it", because I had no association between Japanese words and the English keywords for their component kanji...
> I still had the problem I started with of "I want to write the word べんきょう but can't bring to mind the kanji for it", because I had no association between Japanese words and the English keywords for their component kanji...
If you didn't know what べんきょう meant, how did you know it was what you wanted to write?
I knew the word (including how it is spoken and what it means), so I would have no difficulty of understanding if it was said to me in conversation, I could read the word whether in kanji or hiragana, I could compose sentences in my head which used it and use it when speaking, I just didn't always remember how to write it in kanji...
(Here べんきょう is just an example: the same issue applies to essentially every word.)
The characters you copy pasted are in a phonetic script called hiragana, conceptually similar to the English alphabet as each character denotes a sound.
The kanji are pictographic characters that often have different pronunciations in different contexts. The kanji for べんきょう are 勉強 (meaning "to study").
I know that.
In your model of the problem, pm215 knows the following things:
1. There is a Japanese word pronounced benkyoo which refers to activities like reading books with the intent to learn something, doing practice exercises with the intent to learn something, preparing to take a test, doing things that a school might ask you to do, and other similar endeavors.
2. There is an English word "study" which corresponds to the Japanese word spelled 勉強.
But he doesn't know this:
3. In English, activities performed with the intent to learn or review something are referred to by the general term "studying".
I find this hard to believe. Anyone who chose to learn mappings from English words to kanji spellings must be familiar with the meanings of basic English words. He shouldn't be able to think of the activities he wants to refer to without the word "study" coming into his mind, but that was the problem he described.
No, the problem I describe is "I am thinking of a Japanese sentence (which I know the meaning of), and I could write it in hiragana but not in kanji".
This is analogous to "I want to write an English sentence, but I can't remember how to spell one of the words", except it's worse because at least English words are spelled vaguely in line with their pronunciation.
I think where I may have stated the problem confusingly was my reference to not having a link between the Japanese word and the English RTK keywords. RTK assigns one unique keyword to each kanji, which (a) doesn't always line up with the meaning of a word in which it's the only kanji and (b) doesn't inherently help with multi kanji words. In this case benkyou is 勉強 which is two kanji with the RTK keywords EXERTION and STRONG. Unless you actively learn and memorise a link between the Japanese word and this pair of keywords, RTK is not going to help you with writing the word.
> RTK assigns one unique keyword to each kanji, which (a) doesn't always line up with the meaning of a word in which it's the only kanji and (b) doesn't inherently help with multi kanji words. In this case benkyou is 勉強 which is two kanji with the RTK keywords EXERTION and STRONG.
Ah, you're right. I had no idea this was what you had in mind. That is a system that doesn't make sense. You have to learn the spelling of words per word, not hope that the roots make sense.
(By contrast, 勉强 exists in modern Chinese too, where it mostly means "force; coerce; compel", but can also refer to just barely being able to do something. I was amused to see that it means "study" in Japanese - that implies one of the most sharply negative attitudes towards studying that I've ever heard of.)
It's ~63-64 kanjis per day.
I do imagine someone dedicated would be able to pull that off. But that's still 3-4 hours per day, I guess?
I think the thing that makes it a problem for most people is that they can't memorise them as "once and done" the way Heisig says he did. So as well as the initial time spent looking at the kanji and coming up with a good memorable story/image/mnemonic/etc, most people I think also spend time in an SRS (e.g. anki, or kanjikoohii) reviewing the characters they learnt previously. It's the review time that really stacks up, especially where you have particular characters that you have trouble with ("leeches").
You could pull it off in the sense of 'I've seen this one before...', but no way do I believe that he learned >60/day with good recall. That'd be like me saying I learned the fundamentals of spoken Japanese in one afternoon. I think it's a kind of bullshit claim to be honest.
Well, I can say that I've spent more than eleven months learning common characters before learning any Japanese sounds, words, or grammar.
But the reason for that is that I was learning Chinese. I spent zero effort on learning any characters preliminary to that. There is no reason you'd need or want character knowledge.
My character knowledge is decent now, because eventually my learning method became "talk with Chinese people over Wechat", and if you do that you will necessarily learn common characters.
The idea that one should learn kanji first just to start learning Japanese is utterly nonsense. The author has Stockholm syndrome.
If one's mother tongue isn't Chinese/Japanese, I guarantee you that's impossible to understand kanji/hanzi as deeply as a native speaker does just by spending 17 months on memorizing how to write them. It simply never happens. Languages come first and writing systems come later no matter which target language you're trying to acquire.
Even native speakers get confused about their own writing systems. It's vs its. Should've vs should of. Doppelganger vs doppleganger. Being able to wield the writing system like a wordmaster is a big plus and a praiseworthy effort. But it really isn't the essence of language acquisition.
The time to grind Kanji is when you're ready to really push the breadth of your vocabulary.
While you're still picking up grammar it makes more sense to just pick up kanji here and there as you encounter them.
Indeed. I thought the Heisig approach was garbage on the same basis. I did get a kanji reference book that was a bit more focused on learning the Joyo kanji (ie the 2000 or so you need to graduate high school or pass N1) but I did not stick with it because it also used mnemonics as a learning tool and I found them distracting and obnoxious (in the sense that the author kept injecting his personality into them - sorry dude, I came here to learn kanji, not to learn your opinions on things).
In general I just do not get the mnemonic approach. It's literally another layer of stuff to recall that frequently doesn't have anything to do with the meaning, and the book I had even connected mnemonics in simpler characters to build into more complex ones. Worst of all, you're learning to identify characters with a bunch of word association in English.
Instead of treating chinese characters as a bunch of "tangled squiggles" that you have to memorize or make up stories about you could just ...learn the radicals (called bushi in Japanese). There are around 230 of them, and they are fundamentally pictographic, so the meaning and appearance are linked. Every kanji character is either a radical itself or made up of other radicals, like a word spelled in two dimensions. Some are used much more than others so most of the time you'll be combining the same 32 simple ones. Once you get familiar with them kanji become much easier to remember; a character that is made up of 10 strokes like 勉 is actually made up of just 4 parts which are just a few strokes each. In some cases you can even guess the meaning of a kanji character you've never seen before by looking at the components, but even where you can't, being able to see a complex character as just a collection of familiar simpler chunks makes everything way easier.
You can find a kanji radicals deck on Ankiweb and it's small enough to do alongside your vocabulary or listening practice without being a burden. It will induce a little cognitive dissonance because a few very familiar characters have different meanings in their radical form but you will get over that soon enough.
https://en.wikipedia.org/wiki/List_of_kanji_radicals_by_freq... is a great reference, both for the frequency and the simpler table of stroke count. The individual radical pages have additional detail tracing each one back to bone script (the oldest system of writing, etching characters on bones) which make the pictorial evolution very clear.
By the way, a book I do recommend is the Kodansha Kanji Learner's Dictionary by Jack Halperin. This too uses its own schema for looking things up, but it's a simple one, based on the shape (left-right/ top/bottom/ nested/ freestanding) and the number of strokes, which makes looking things up fast. It also contains indexes to look up kanji by radical or pronunciation (eg if you have the furigana) so you are not locked into one way of understanding them like mnemonics. It also has common vocabulary that uses the kanji, stroke order, and all that good stuff. And it has a good reference on the rules of stroke order. There are only 8 of them and (like radicals) once you understand the underlying principle you can look at a new kanji you've never seen before and have a pretty good idea of how to write it. You should still check while you're learning, but the more you write the easier it gets.
jisho.org is also incredibly useful, because you can dig into any word to find the kanji within, and then dig into individual kanji to see which radicals they're composed of, as well as looking things up by radical.
I am doing both because while learning to write the symbol for a sound/meaning and identifying the symbol’s sound/meaning are separate skills, they enhance the fluidity of _thinking_ in Japanese significantly for me. It has a synergistic effect and to me seems to improve the brain’s understanding and efficiency in compressing the knowledge.
But my goal is not just to read and understand but to talk conversationally. While Japanese is very different from my other languages, I’m already multilingual (Norwegian, English, Dutch and German) and this approach has always worked best for me.
I might have to play with writing alongside vocab reviews.
Back in the earlier days of the online Japanese learning sphere (when AJATT was still the big new thing), I tried learning by starting with writing by studying kanji independently, but that went nowhere even after several months in.
More recently I’ve been making a point of audibly speaking the sentences associated with vocab cards and that’s helped a lot with being able to fluidly speak the various long trains of sounds that are common in Japanese but rare in English as well for improving recall and improving reading speed. It would follow that writing might enhance that effect.
My foremost goal is to become conversational too. The rest should follow more naturally if I can achieve that.
I’ve taken your approach as well but I did notice that I retain Kanjis that I learn to write significantly more than ones I can just read. But memorising all the Kanjis is a bother.
I found the sweet spot to be writing on a scratch pad as I go through Anki. And not particularly worry about getting writing right too much. Sometimes I’d be confused in my head but my muscle memory would kick in and automatically write the kanji!
Yeah I made the same decision learning Chinese. It's just not worth the extra time and effort relative to the utility.
Occasionally I still have to sign my name, so I specifically picked a name that a) was written the same in both simplified and traditional and b) had a low number of strokes. Like you, the only time this has bitten me is in hospitals and banks where occasionally they ask you to do stuff like write out your address. I sometimes exercise my dumb foreigner privilege and ask the clerk to help, but since addresses have a formal romanization method it's often fine to write that, and I've seen enough locals struggle I don't feel too bad about it.
I'm learning traditional Chinese and found that writing helps me recognise the components and strokes when reading the same characters.
If I just try and visually pattern match with flash cards, anything that's hand written or in an stylised font will throw me off. If I can sympathise with / recognise the stokes used, I find it easier to tell what character they're trying to show.
The one thing I noticed when I was focusing on learning to write is that it helped me a lot with differentiating between similar characters when reading. I forget which ones now, but there are many characters that differ by a single radical and have similar meanings, knowing how to write each one helped me quite a bit there, but overall I rarely write anything other than my name and address now that I live in Japan.
Even if you cannot recall and have to type in your phone first (as natives do often for unusual ones), at least you can write it much faster because you are not just copying.
I think it’s useful to do RTK with SSR and, once you finishes, you only need about 15 min of maintenance per day to keep it in memory.
I can confirm this. I passed N1 without learning to write. I later learned how to write all of the kanji, and all it does it help you distinguish very similar kanji without context. I tried learning all the compound words (i.e. which kanji to use for every word) but gave up a few thousand Anki cards in. It was time consuming and impractical. (Wanted to pass Kanken 2) Props for anyone who put in the work though.
Don't drill words drill sentences (no Japanese word means an English word, and no English word means a Japanese word.) And don't fall in love with any of the sentences, move on to new sentences. Eventually the knowledge accumulates, but only through volume and variety.
If that worked for you cool. My Anki deck is a lovingly curated set of sentences that I enjoy.
You strategy makes sense. Truth be told, even native Chinese and Japanese tend to forget how to write many characters as they spend more time typing than writing.
I think your advice makes a lot of sense for most learners: prioritize the skills you'll actually use, and don't feel guilty about skipping handwriting unless it personally matters to you
I'm following this approach. One of the most interesting things so far has been observing just how separate recall and production are. There are kanji that I can recognise instantly, and recall meanings and pronunciations, but I can't visualise them at all.
It is a separate skill from reading, but I think it's still useful.
At the very least learn the strokes of common radicals. In my experience things like denshi jishos can be VERY picky about how you input them. It makes word lookup much faster IMO anyways.
> It's only rarely that I have to write anything other than my own name in Japanese.
I thought they had stamps for that.
> How is it possible for you to "see" the text in your mind and not be able to replicate it with a pen? Even if the mental image is faint and fuzzy, surely you can sketch it out roughly at first, then refine it until it settles into its exact form? Apparently, that is not how mental images work, either.
That's really not how it works. I draw as a hobby and I love to study human anatomy. I can conjure in my mind people in many poses with detail, in different clothes and colour. Still the act of drawing without a reference is a world in itself. If I try to sketch from memory without a reference, it quickly falls apart.
It's not that the mental image is incomplete, wrong or an illusion. Is just that knowing it and reproducing it are two very different things.
This is very common in Chinese now. The older generation, many of whom didn’t learn pinyin, just use voice input to send messages; the younger generations just use pinyin input and similarly can’t handwrite beyond the simplest characters.
The phenomenon of forgetting how to write is called 提笔忘字 (tíbǐwàngzì - to pick up the pen and forget the character). It was previously covered here on HN: https://news.ycombinator.com/item?id=41959256
To people who think this is surprising, it’s not that different to forgetting how to spell rare English words
Not really. Even in English there are only a few possible ways to spell out a given sound. With Chinese characters that mostly goes out the window, making speaking and writing more or less orthogonal skills. Not entirely since there are classes of characters that are related by sound, but to a much greater extent than alphabetic languages. Personally if I forget a spelling in English or a related language I realize that my approximation looks wrong and can correct it by elimination, whereas with Chinese characters it's common for people to simply not have a clue how to even write the first stroke of a word they've heard and read before.
> it's common for people to simply not have a clue how to even write the first stroke of a word they've heard and read before.
It’s much more common to remember most of the character, or accidentally substitute it for a similar character that sounds the same.
That is a very melancholic name.
> the younger generations just use pinyin input and similarly can’t handwrite beyond the simplest characters
In my experience people often use input methods that aren't pinyin. I tend not to be able to type on other people's phones.
What if we call this 'constructive recall', in which your mind needs to take a concept and generate the concrete manifestation of it, vs 'recognition', which is the obverse. There are many examples of this in life, aren't there?
Imagine a song you used to love as a kid and knew by heart. Now try to write down the lyrics word for word. Typically quite difficult.
Now play the song and sing along. You not only know all the words, your mind provides you a just-in-time recollection of all the nuance in the delivery, the pauses, the details of the music itself, nearly a perfect replay of the entire song is there in your mind...you just couldn't knit it together like you can when it's there in front of you.
I've noticed something similar when listening to podcasts. If I'm out doing something, typically driving, while listening to a podcast episode for the first time, I will have these intrusive photographic recalls of what I was doing at that time if I listen to it again. At least if I do that within the first few weeks.
Just seems like a general characteristic of the brain.
When I reached upper level Japanese classes (N2/N1), my native Japanese teachers would regularly (maybe once a week or so) have to look up a character they were writing on the board during class.
It definitely made me feel better.
From the article:
> On the surface, this atypical trait seems to explain quite well why I can draw a blank when asked to write the kanji for "plant" (植) from memory. I don't see the character in my mind, so it makes sense that I can't reproduce it on paper.
While the author's aphantasia may have posed some recall issues - it wouldn't explain why they had ever been able to reproduce 植. Kanji has the concept of radicals AND stroke order. One could make the case that perhaps the author's motor cortex is simply storing the equivalent of LOGO programming language instructions for reproducing the logograph.
Take away your mind's ability to find and chunk (木, 十, 具) by showing them "radical"-less characters and I'm sure it would be even more difficult.
As someone with aphantasia, I'm also not convinced that is the cause. I can draw better than average from memory, and used to be quite good but haven't practiced for many years.
There are also animators and artists with aphantasia[1].
My spatial recollection is particularly good - I can sketch out precise diagrammatic drawings from memory much better than I can do artistic drawings.
So while it's not impossible that not having aphantasia would've been a benefit to him, I don't think there's much evidence that it's has any big effect on the ability to draw - something that also fits Ed Catmull's experience on surveying his old employees in the article mentioned.
[1] https://www.bbc.co.uk/news/health-47830256
My wife has aphantasia, is fluent in Japanese, and writes the language better than many of our friends who still live there. She was just forced to learn the stroke order, etc when learning the language.
Aphantasia causes many odd issues for her but the ability to write complex languages or draw are not one of them.
Writing characters without knowing the radicals is kind of like spelling without really clearly understanding the alphabet. The stroke order is easy to remember, the radicals make the full characters easier to remember.
I spent years in Taiwan studying traditional Chinese and even at the height of my proficiency there were plenty of rarer logographs that I'd frequently stumble over - only able to draw "blurry approximations" of them depending on my familiarity.
Coming from a phonetic language with only 26 letters, it was such a surreal feeling being able to effortlessly read a character but be unable to reproduce it.
My understanding (which may have been in the previous HN discussion on the topic) is that Chinese people just substitute with a homophone if they're really stuck and native readers can guess by context what the writer meant. Much like fudged spelling like when Americans mix up do and due
Also wild how quickly the writing ability fades without regular use
A similar thing happens with all kinds of iconography, from flags to logos. People can easily recognize many logos, but when asked to draw them they often can't come very close.
https://magazine.adler.co.uk/promotional-idea/we-asked-100-p...
Not just logos, but also shapes of everyday objects. Bikes, famously.
That’s a fun little article, thanks for sharing that.
This is a fascinating subject, that would merit some input from other languages IMHO.
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
In alphabetic languages with separate pronounciation (French, English etc.) people forget spelling instead. And of course we recognize spellings we know, even if we couldn't recall it cold from memory.
Another point: the article defines kanji as images, but each character can be parsed as a composition of base blocks (radicals), and people primarily remember those. Put in computer terms, kanji are not bitmaps but overlayed vectorial blocks. That I think makes a world of difference for how we handle them, foreigner and native alike.
Learning about radicals was a massive Level Up in my Japaneseability.
THey're like protocols/interfaces on OOP classes ^^ or more accurately, like "tele" and "phone" in "telephone", "icon" in "iconography" and obviously many more words in English made up of bits and pieces of other words.
Once you recognize the parts, you'll be able to able to guess/infer what the word means even if it's your first time seeing it. Same with radicals in Kanji.
I have no trouble reading but writing kanji has become a problem. I never need to do it and I can’t remember how to write kanji I have no trouble reading.
It’s Japanese people too, to a lesser degree. My own Japanese wife has to pause to remember how to write something every now and then.
This happens in Chinese too
Grocery lists will be a mish mash of characters and pinyin
https://languagelog.ldc.upenn.edu/~bgzimmer/jiaozi.gif
鸡, get halfway through writing 蛋, forget how to do it without a computerized pinyin input, give up, scribble it out and write dan
Thanks for that example. Would a Chinese person ever write the tone in pinyin when writing something for themselves?
Not when doing fragments in obvious context like that. For one, tone doesn't disambiguate anyways. There's 10+ common characters with pinyin 'dan', and only 4 (5) tones in Mandarin.
But also when the context is super obvious there, there's no need.
This was happening to people I knew when I lived in Japan 30 years ago. Many people were using wa-puro (word processors that let you type in the phonetic form and choose from the appropriate kanji). I imagine the effect is far more common now.
I remember one time when a university engineering professor couldn't remember how to write the kanji for "police". He didn't seem embarrassed asking someone else. I don't know if they still do, but they would often demonstrate by writing out the character with their index finger like a pen in the other hand's palm.
I went through Heisig's book THREE times, writing out each character and reviewing them multiple times each in Anki. I had several notebooks like the one in the OP article.
I still can't write most kanji.
The fact is that you simply cannot learn to write without practicing writing (in real contexts, outside of the notebooks); there is overlap, but it's the same thing as learning the "theory" of a math problem and then blanking once you're asked to solve a problem on a test.
If I try to visualize a word/character that I know, I often end up with a vague picture of the most salient lines, but missing some of the finer detail that would be required to write it out accurately. And I think this is largely fine, as handwriting characters isn't a very important skill anymore. Especially in Japanese where you can always fall back on kana and be understood.
I do still think "learning" the kanji is helpful because there is a system behind it, and understanding that system helps to learn words. To give an example, learning the 教/kyou in 教育/kyouiku (education) helps you make a connection to 教室/kyoushitsu (classroom), and helps you distinguish it from 勉強/benkyou (study) where you'd really think the kyou would be 教, but it isn't!
But crucially, this is all a visual recognition problem, nothing to do with producing the characters yourself. So I think the ideal learning approach is word-first but with some attention to the characters in the word, especially how they relate to other words with those characters/radicals or with the same readings.
Years ago I asked my teacher what percent of street merchants or high school kids would know how to write some of the more complicated, uncommon characters from the advanced textbooks. She replied street merchants maybe 50%, but high school students 100% -- during high school every student will be at their lifetime peak literacy, because they will be cramming and memorizing how to write every possible character in preparation for the college entrance exams.
I tell locals I can "speak, read, and type", but only have 1st grade writing abilities in Mandarin. The key is typing -- you can chat on Line, WeChat, email, or any mobile apps to accomplish day-to-day tasks. But the typing is phoenetic either via Roman pinyin, zhuyin, romaji, etc.
So yes, this problem only comes up in languages where drawing the characters are mostly independent of the pronunciation. You don't need to type individual words, you can type phrases and the input tool presents choices sorted by most commonly used. Therefore, even sending communication is closer to a reading exercise rather than a writing exercise.
Don't worry it's common for natives too.
Not born & raised in Japan but went to Saturday school for Japanese in Europe and I was excited to be able to read (and understand) the newspaper for the first time at middle school, because my Kanji caught up. From there it's usually very quick how much Kanji you can learn.
But now, 20 years later? When I need to go to fill forms I flip up my phone and search for Kanji all the time. And I know I'm not alone (although probably very bad by Japanese standards), but I can navigate Japan just fine.
I always struggled to explain this to my European friends. "I can read and talk fluently without any issue, but writing not so much. Unless on keyboard".
The best analogy I've come up with is "If I ask you to imagine an apple or a Motorbike, you can do it right. When you see one, you'll instantly recognize what it is. But if I ask you to paint an apple, or a motorbike, you might not fare too well; people might mistake your painting of a motorbike for a bicycle. It's something like that. Using keyboard is like googling for images and copying it in to your PowerPoint slides"
Is it this way for Hiragana and Katakana for you too? I just started and have not moved on to Kanji yet, and I already struggle to write them.
A lot of Japanese learners do hate katakana (personally, a lot of fonts could stand to be clearer about ツシンソ ), because most writing is in kanji+hiragana so they have less practice with katakana. But kana ability is really just exposure. Use it to get used to it.
Same reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
nta but that is just a matter of practice. Just write every day and i will become automatic after a few months.
I have experienced this for simplified Chinese. I studied some Chinese while studying computer science in China. The classes would have us learn writing, reading, speaking and listening (sensibly so for a Chinese language class).
Being able to write characters was handy whenever I came across documents that needed to be filled, but since leaving China I never had the need to write characters again. I now just input them using pinyin on keyboards, and I can easily recognise and read / input the correct characters. It is a strange feeling trying to write the characters I once knew, but now have forgotten, yet being able to read them instantly...
I would like to recommend dong-chinese, a language app I came across when I prepared for my stay over there. It taught things in a very efficient manner.
At this point I would like to recreationally increase my vocabulary so I have started working on a game called LingoRogue. My goal is to make it addictive to play, with a sneaky vocabulary-increasing effect. In other words a game that is "learnified" rather than a learning software that is gamified.
It's like the muscle memory for writing just evaporates, even though the recognition part stays perfectly intact
I'm convinced that different people process handwriting (and movement) differently. This is true with latin languages as well.
I think this may explain the difference between recognizing shapes versus drawing them for some people.
I remember when I was in school, some people had really neat handwriting, they could write fast and all their letters looked exactly the same with apparently little effort. On the other hand, I had to focus hard to ensure that my letters were all the same style, shape, size and slope... Also, I didn't have a single 'handwriting style' I could write in a number of different styles. I couldn't have both speed and nice looking, consistent letters; it was one or the other.
The interesting thing though is that I was always quite good at drawing... Conversely, I noticed that the people who had beautiful, effortless handwriting would typically be quite bad at drawing... They were the kinds of people who had to start out every drawing as a bunch of circles, triangles and crosses before joining them together to form the final drawing.
I feel like these people automate their hand to some extent. It's like a reflex to them. It lets them render common shapes without much thinking or effort.
It reminds me of that time I did a drawing class and the teacher kept reminding students to "stop thinking in symbols and just draw the different shapes and shades as they appear."
This probably has parallels in a number of areas like sports (e.g. tennis) where being able to offload certain movements to muscle memory can free up your brain for more strategic aspects of the sport.
This also reminds me of Daniel Kahneman's book "Thinking fast and slow." I suspect it would be interesting to try to categorize people based on what kinds of mental activities they offload to system 1 vs system 2 thinking.
> In other words, what feels like a single, monolithic "literacy" ability is actually two distinct skills, each exercised in different instances and each capable of improving and decaying on its own.
This dissociation has been used to test theories of hemispheric specialization. A good overview is in Neurolinguistic Aspects of the Japanese Writing System by Michel Piradis (1985).
I mean, this is not a politically correct statement, but I think one line of reasoning from this is to say that Chinese characters (which is what kanji are) are not a great way to write down language for practical purposes. A friend of mine in grad school irreverently referred to the Chinese writing system as "a really huge, really inefficient syllabary", and I think there's some truth to that. The characters no doubt have a certain beauty and their history is interesting, but a system where the meanings and pronunciations have to be learned totally separately seems to be inherently cumbersome in some ways. Even in a language like English which abuses the Latin alphabet in a notoriously messy manner, the amount of phonetic information that can be gleaned from the written form is fairly high, which gives two paths to the word (via memorized whole-word recognition or incremental sounding-out).
Overall you are probably correct, but there are certain benefits from the Chinese character system.
Long time ago I studied Japanese in Japan. On the way back to my home country I was sitting next to a bunch of Chinese people on the plane who did not speak any English or Japanese, but we were able to have a small conversation using Kanji/Chinese characters, because the characters' meanings are usually the same, although the languages are quite different. If the people would have been Greek and could not speak any English, no conversation would have been possible at all.
Another thing to mention is the radical system. Many Chinese characters consist of two or more characters, of which one is the "radical". This often helps you understand the broad meaning of the character in case you do not know it. For example, the Japanese character for fish is "魚". If you know that character and see another unknown character that used "fish" as a radical (for example "鮭"), you know that the character probably describes some kind of fish (in this case salmon). So it is not simply a huge list of "syllables".
I know almost nothing of Chinese languages, but some Japanese, and when I see boxes and instructions for some random item from Temu that is all in Chinese I'm amused by being able to understand bits of it.
I saw a face mask box that was all in Chinese and had 非医 on it (which means non-medical) which, depending on your font would be written exactly the same in Japanese: 非医.
> Overall you are probably correct, but there are certain benefits from the Chinese character system.
I mean yeah, it's an ill wind that blows nobody any good. But consider this: We could do the same thing in English as in Japanese -- replace loads of letters with Chinese characters: Write "跑ing" instead of "running", and so on. The French and German and Russians and Spanish could all do the same thing with their languages; and then when traveling, people could at a basic level read the signs and menus, and communicate at a basic level by writing, without having to know anything about the language.
Would you choose, post-facto, to add Chinese characters to English?
For my part, I'd say "no way". English orthography is already hard enough for my son to learn, without having to add characters on top of it.
ETA: Just for kicks, I asked Claude to try its hand at writing the opening lines of Pride and Prejudice in this manner:
It is a 真理 universally 認知ed, that a 単一 男 in 所有 of a good 財産 must be in 欲 of a 妻.
However 少 知n the 感情s or 見解s of such a 男 may be on his 最初 入ing a 近所, this 真理 is so 好 固定ed in the 心s of the 周囲 家族s, that he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
"My 親愛 Mr. Bennet," 言ed his 夫人 to him 一日, "have you 聞ed that Netherfield Park is 貸 at 最後?"
Mr. Bennet 返答ed that he had 不.
"But it is," 戻ed she; "for Mrs. Long has 丁度 been here, and she 告ed me 全 about it."
Mr. Bennet 作 no 答.
"Do 不 you 欲 to 知 who has 取n it?" 叫ed his 妻, 不耐. "You 欲 to 告 me, and I have no 反対 to 聞ing it."
> and then when traveling, people could at a basic level read the signs and menus, and communicate at a basic level by writing, without having to know anything about the language
Well, you have to be careful; something that actually happened to me was recognizing that the first element of 牛蛙 meant "cow", without recognizing that the second part made it "bullfrog".
> We could do the same thing in English as in Japanese -- replace loads of letters with Chinese characters: Write "跑ing" instead of "running", and so on.
Interesting choice. Japanese doesn't do that - it follows classical Chinese by using the character 走 for the sense "run". 跑 is Chinese-specific.
> Just for kicks, I asked Claude to try its hand at writing the opening lines of Pride and Prejudice in this manner
This comes off weirdly to me because it's so Japanese. But I guess that was the idea. I have some particular questions:
>> this 真理 is so 好 固定ed in the 心s
I haven't checked on the English text, but it's difficult for me not to read this as "this truth is so well fixed in the hearts...". But I'm not sure that that sense of 好 is available in Japanese, where I'd expect it to be a verb meaning "like".
>> he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
This doesn't appear to be a problem, but I still felt the need to call out what looks like the use of an exclusively masculine pronoun (in modern Chinese) to refer to "daughters".
>> Mr. Bennet 返答ed that he had 不.
This seems wrong; I feel sure that 未 would be more appropriate than 不.
>> Mr. Bennet 作 no 答.
This would appear to render the English "Mr. Bennet made no reply." I'm not comfortable with the use of 作; just because the word in the idiom is make doesn't mean that any making is involved.
Do you know whether Japanese use of kanji focuses more on establishing that a kanji corresponds to some Japanese syllables and using that kanji wherever those syllables occur [as suggested by "作 no 答"], or more on using kanji to represent certain semantics however those semantics might be pronounced [as might suggest "言 no 答"]?
>>> he is 考慮ed as the 正当 財産 of some 一 or 他 of their 娘s.
> This doesn't appear to be a problem, but I still felt the need to call out what looks like the use of an exclusively masculine pronoun (in modern Chinese) to refer to "daughters".
IIRC, in Japanese, 他 is sometimes used for "hoka", which means "other"; as in, "Do you have this item of clothing in other (他) sizes?" The original text says, "some one or other of their daughters"; so that seems to be Claude's thinking.
> Do you know whether Japanese use of kanji focuses more on establishing that a kanji corresponds to some Japanese syllables and using that kanji wherever those syllables occur [as suggested by "作 no 答"], or more on using kanji to represent certain semantics however those semantics might be pronounced [as might suggest "言 no 答"]?
I know a moderate amount of Mandarin (along with smatterings of Cantonese), and about 3 months' spare-time study of Japanese in preparation for a recent trip there. (Note the shopping theme in my example above.)
My expectation, which matches my (very small) experience, is that Japanese are trying to write Japanese; and that therefore they have Japanese words in mind that they're trying to represent with Chinese characters.
It is absolutely not the case that a given Chinese character will always be pronounced the same way in Japanese. For one, multi-character Chinese words are also munged into Japanese words. For two, even for single-syllable words there's a context: e.g,. Japanese has two different ways to say "one", but they use "一" for both. I'm sure I've run into other Chinese characters that are pronounced differently depending on whether they're in a verb or an adjective.
So for example:
> Mr. Bennet 返答ed that he had 不.
> Do 不 you 欲 to 知 who has 取n it?
Remember that we're discussing a hypothetical universe where English speakers use the Chinese characters, but the vast majority of them don't know Chinese.
In English we use "not" in both cases, even though in Mandarin in the first case I'd probably construct a sentence using "没" in the first instance and "不" in the second.
So what would happen in our hypothetical universe? Given that most speakers don't know Mandarin, using "不" in both cases is probably the simplest, most stable result. One could imagine complicated rules for whether you write "不", "没", or "未", which are taught in school and maintained, even though they're all read as "not", and even though most people don't know Mandarin. But it would have to be one of those things which (like Chinese characters themselves) people decided they liked about their writing system and didn't want to give up.
Similarly:
> I'm not comfortable with the use of 作; just because the word in the idiom is make doesn't mean that any making is involved.
And yet, in English, "made a cake" and "made no reply" have exactly the same verb. Again, one can imagine it going both ways, depending on how things were established and maintained: One could imagine using "作" in both cases; or one could imagine using two different characters for two different shades of meaning, just as in Chinese there's 他, 她, 它, and 祂; or perhaps 的 and 得 (which I'm not sure people would naturally consider different words if they were only exposed to the spoken word).
The real limitation here is that there's no way ready-made to indicate the "make" -> "made" transition, as you can with "makes" ("作s" ) and "making" ("作ing").
> I'm sure I've run into other Chinese characters that are pronounced differently depending on whether they're in a verb or an adjective.
The Japanese writing situation is much, much worse than that. Kanji get all kinds of different pronunciations on the theory that the semantics are the same. So the standard spelling of musume ["daughter"; "girl"] is 娘, but it might also be spelled 女. This gives you an alternative to the standard independent reading of 女, which is onna ["woman"]. All characters will also be pronounced differently when they represent Chinese loanwords than when they represent native words.
> And yet, in English, "made a cake" and "made no reply" have exactly the same verb. Again, one can imagine it going both ways
I agree with that.
> The real limitation here is that there's no way ready-made to indicate the "make" -> "made" transition, as you can with "makes" ("作s" ) and "making" ("作ing").
I don't agree with that; you'd do the same thing you do with every other preterite verb and write 作ed. You wouldn't pronounce that maked, but that's not a problem; we're still assuming that everybody knows English.
> IIRC, in Japanese, 他 is sometimes used for "hoka", which means "other"; as in, "Do you have this item of clothing in other (他) sizes?" The original text says, "some one or other of their daughters"
Yes, I determined something similar by looking the character up in a Japanese dictionary. I don't have the skills to determine what kinds of uses are and aren't natural, so I just decided that I couldn't label the usage wrong, but it still stood out as funny for other reasons. I was able to correctly read the English without referencing the original text.
I'm not sure what a "moderate amount of Mandarin" means; if it's on the lower end, you might be interested to know that the "other" sense does survive in Mandarin, in the words 其他 ["other"; extremely common] and 他人 ["other people"; not so common]. 他 by itself is going to be overwhelmed by the far more common use as a pronoun, I would guess.
> My expectation, which matches my (very small) experience, is that Japanese are trying to write Japanese; and that therefore they have Japanese words in mind that they're trying to represent with Chinese characters.
While I do agree that the situation could shake out in several ways, I don't think this is a total defense of the idea that "the same word" is going to be spelled the same way by people who don't know Chinese. I wouldn't expect, for example, that fire [flame] and fire [eliminate from a job] would get the same character spelling.
> One could imagine using "作" in both cases; or one could imagine using two different characters for two different shades of meaning, just as in Chinese there's 他, 她, 它, and 祂
By my understanding, 祂 is not a part of ordinary mainland usage. But this is a good example, in that there is no distinction between the words in the language, and Chinese people aren't able to make the distinction when learning a foreign language even though they do make it in writing their own. (By contrast, Spanish speakers don't have problems choosing between the English words he and she.)
You might also be interested to know that the character 做, pronounced identically to 作 and meaning the same thing, is in fact derived from 作. It (and not 作) is now the ordinary character used for the sense of making or doing. 作 is best known to me as part of the word 作者 "author". This is a pure spelling distinction that arose by some natural process within Mandarin.
> or perhaps 的 and 得 (which I'm not sure people would naturally consider different words if they were only exposed to the spoken word).
I can answer that; they don't. At least they don't for the 得 that introduces manner or result clauses; they might or might not think of the possibility infix seen in 听得了 "able to listen" as different.
> I'm not sure what a "moderate amount of Mandarin" means; if it's on the lower end, you might be interested to know that the "other" sense does survive in Mandarin, in the words 其他 ["other"; extremely common] and 他人 ["other people"; not so common]. 他 by itself is going to be overwhelmed by the far more common use as a pronoun, I would guess.
I had a snippet about how 他 was made up of 也 and 人, "also a person", but ended up editing it out, as I wasn't sure it was actually connected w/ Japanese using it to mean "other". "他人" is still about people, but I had forgotten 其他, which is clearly not specific to people.
At any rate, there's a lot of ways our hypothetical universe could go, WRT how such a 汉字 writing system would be incorporated into English. The point is, there's always a silver lining: It would certainly have some benefits, like making it possible for literate English-speakers to get around in China and Japan w/o learning anything about the local languages and vice versa (with some of the "false friend" [1] traps you've mentioned above -- but those are issues between European languages as well).
But on the whole, I'd consider the cost not worth the benefits by a long shot.
[1] https://en.wikipedia.org/wiki/False_friend
> I had a snippet about how 他 was made up of 也 and 人, "also a person", but ended up editing it out, as I wasn't sure it was actually connected w/ Japanese using it to mean "other".
Well, I don't think it's true, either. As I understand it, the form 他 is just a graphical simplification of 佗, with no connection to the character 也. (You can think of this as similar to how the left-hand component in 脸 is 肉, not 月.) It isn't clear to me either that the character 也 would have held any sense of "also" when the form 他 appeared, though this is a question I'm agnostic on.
I've watched several videos on youtube in which someone presents Japanese people with uncommon kanji and asks them to read them, or failing that to speculate on what they might say. The Japanese unfailingly speculate that both halves of a compound character are relevant to the meaning, which surprises me - it's very rare for a character to be constructed from two meaningful parts. Far more common are the characters in which one part gestures at the meaning and the other part tells you the pronunciation, and 佗 is one of those. 从人它聲。
You might be interested in: https://www.zompist.com/yingzi/yingzi.htm
>have to be learned totally separately seems to be inherently cumbersome in some ways.
It also has quite big advantages. Because writing systems tend to be highly standardized meaning of most characters has changed relatively little, so Chinese people can read ancient texts with aid of a bit of classical Chinese. And this also works across space, China has a lot of mutually unintelligible spoken dialects but pretty much everyone can understand the meaning of written Chinese. And so can even Japanese or Korean people.
In contrast spoken language tends to change so quickly that phonetic writing systems can rapidly become completely unintelligible. Old English is practically a foreign language. Probably nobody understands that "Hwæt. We Gardena in geardagum" means "Hey! We the Spear Danes, in years gone by".
The problem with these advantages is that most actual uses of words depend on some amount of context, and only a relatively small set of word+context combinations can actually work across languages that way. It is like the equivalent of doing a word-by-word translation.
Another way to put it is that this is essentially no longer "writing language", it's some alternate form of meaning representation. Language is not just a sequence of independent words.
It also affects the learning curve. English-speaking children can read far more advanced books earlier than equivalently educated Japanese.
They probably don't outweigh the disadvantages, but there are some small benefits. Once fluent, I think I remember Japanese to be slightly faster to read because of the more unique shapes. And you can make more flexible and elegant graphic text designs and tables (like in Excel) given the compact words and natural vertical writing.
Is that so? Deducing definition of an unknown kanji word isn't that difficult. I never had issues with "advanced kanji" as far back as I can recall.
My children are Japanese-English bilingual and can read far more advanced books in English. Initially I took this as an imbalance and suggested they read the same books (or something very close) in Japanese. But their native Japanese language teachers said, no, because of the different learning curves you can't expect them to read the same level of Japanese texts; the equivalently educated/advanced Japanese reader will be behind, at least in the elementary school years.
Japanese as a whole are extremely avid readers, so I don't think there's a gap at the top, only the shape of the learning curve.
If native speakers are starting to have character amnesia too, does that suggest in the long run you would expect the writing systems to simplify towards the phonetic syllabaries? Or is the fact that we have computers as a mediating tool going to forestall that and just make things weird?
IMEs have actually caused the opposite a little in Japanese, because you type the phonetic pronounciation and get a list of possible kanji, then with computer use "writing" kanji really means "recognising" and so people will use kanji that they never would have bothered with remembering how to handwrite.
I'm just a Chinese student from the West so take it as just my two cents but I don't think it will evolve to phonetic syllabaries. Chinese has a lot of homophones so it's useful while reading to have an extra semantic meaning. They also say that once you're used to it, you read faster but at my level I can't confirm it. So with modern input devices you're basically simplifying the hard part of the characters, which is writing, and keeping the reading part where they're better than pure phonetic systems.
I can relate a lot to this post.
I learned Japanese around the same time (2010~). I even lived in Japan and studied at a Japanese university.
To this day I have a hard time recalling Japanese on the spot. However, when I hear Japanese or see it written, I can read and understand well. Even kanji I haven't seen in a decade. It's kinda strange.
I'm pretty sure if I re-studied Japanese I would learn very quickly as I re-activate that part of my brain... something that has been on my todo list for a long time!
Writing out Chinese characters definitely helped me learn to read them as well (not surprising, since engaging multiple senses helps one retain information better), even though in practice I never had opportunity to write them out because I always used either a phone or computer.
Unlike the author, I found most of the mnemonics as much trouble to learn as the characters themselves, and soon stopped using that approach. It just didn't work for me.
Is ワープロ馬鹿 really a term used by native Japanese speakers? As far as I can tell it only really shows up in Japanese->English dictionaries and English forums (see https://www.google.com/search?q="ワープロ馬鹿"+-a+-the).
edit: s/word/term
> Is ワープロ馬鹿 really a word used by native Japanese speakers?
You probably mean idiom. ワープロ (word processor) and 馬鹿 (baka: idiot) are individually both words used by Japanese speakers. Japanese speakers would be more likely to say 漢字健忘 (kanji amnesia) to refer to the phenomenon though.
I had a friend in college whose mother was an immigrant and whose father had grown up bilingual and spent years in the mother's county. My friend was able to hear and understand the mother's tongue with apparent fluency, but had an extremely difficult time forming real sentences in the language, even though he'd heard and understood it for his entire life.
I think there's a significant difference between consuming language and producing it, whether that's written or spoken. It might explain the number of people I see these days who don't understand basic English grammar and make words plural by slapping on an apostrophe and an s.
Being able to read but not write Kanji is so common that it's a meme amongst Japanese people -- to the point where it's a game. For example, here you can watch some Japanese television people play a game where they compete to write words in 10 seconds or less:
https://www.youtube.com/watch?v=pqQQqLno9hw
I (and every other learner) have the same problem. It's not special, and has nothing to do with aphantasia.
Hm, the author explicitly pointed out the same:
> What confuses me is that other people can form images in their minds. Are all those with character amnesia also aphantasic? That can't be, given that aphantasics amount to less than 5% of the population, while a much larger number of people forget how to write (70% of teenage participants in a Chinese TV show were unable to write the word "toad"!).
They were discussing their aphantasia as a precursor to other very interesting points, e.g. about how "seeing" a character in your mind isn't enough to be able to draw it, --> verbatim traces and gist traces.
Completely off topic but that is the shakiest camera work I have ever seen, and I learned to shoot before gimbals were cheap and easily available. For that matter, it's been easy to automatically correct that sort of thin in post for years now. How were they not embarrassed to upload something so unwatchable?
"is point is that learning kanji presents two obstacles: remembering what the shapes mean and remembering how they are pronounced."
How is this different from English words? How is conscientiousness pronounced? Not to mention rendezvous.
English is written using an alphabet. Alphabets are by and large phonetic so you do not need to remember how to pronounce every word, you learn to read the alphabet and specific spelling (e.g. 'th' in English) and you are mostly done.
Chinese is written using ideograms ('kanji' in Japanese), which convey a meaning but not a pronounciation. So when you encounter a new character you cannot pronounce it.
> Chinese is written using ideograms ('kanji' in Japanese), which convey a meaning but not a pronounciation. So when you encounter a new character you cannot pronounce it.
This isn't quite true - about 80% of Chinese characters are so-called phonosemantic compounds, where people originally started using the character for one thing for another thing whose word sounded similar (say, emoji for "can", as in able to) and then adding a semantic component to differentiate the character from other similar-sounding ones. In Chinese, they smushed the two components into the space of one character, but in eg. Egypt, they simply wrote whe semantic clarifier and the phonetic hint side by side, full size.
That is, the majority of the characters are primarily sound-based, it's just that the connection between a character and its sound is shoddy, even in Chinese languages.
Japanese kun readings for native words do divorce the characters pretty completely from their sound.
This is being pedantic.
You cannot guess the pronunciation of a character from the way it is written. At most the "phonosemantic compound" might provide a clue to possibly reduce the space of possibilities.
It's not pedantic in that the characters themselves really are sound-based and provide a pretty decent clue. That said, it's still just a clue, and there are multiple similar pronunciations associated with any phonetic component, so it's still guesswork in the end.
And so we are back to square one: it is not possible to know how to pronounce a character by looking at it (and any clues are not "decent" in the majority of cases), as opposed to alphabetic systems...
The difference is that for most words you can easily get 90% there in English and, even if you don't read it properly, the listener can somewhat get what you are trying to say.
And say you don't know how to read "conscientiousness" but you know how to read "con" and "ousness" you can try to go for con-shu-tiouness and at least you're somewhere.
It's not the case with Kanjis, sure you might know one part of a word but you might be wrong. Also similar looking kanji don't read similarly at all, so even that is out of the window.
And don't get me started on the approach of applying memoization techniques purely on the strokes of a kanji, that's gonna hurt more than not.
Simple examples:
- 末 and 未
- 大 and 犬
- 千 and 干
It actually works the same way in Japanese. Radicals give rough clues about meanings as well as pronunciation. Similar looking kanji do read similarly. It's just... you can read pronounce "brochures" in either French style or in Southern American, and which one is more appropriate depends.
edit: Simple examples:
Not always, as I wrote in another comment this is because of the Japanese script reform for "modernising" and simplifying the script. So sometimes kanji look similar but are not. More info at: https://en.m.wikipedia.org/wiki/Japanese_script_reform#Mazeg...
So, I don't know what to say, but to be honest, I never even thought 千 and 干 can be considered similar...
皮 彼 波
pi, hi, ha?
The advantage though is that even if you don't know the word, you can guess the meaning if it contains familiar character(s). Also except for the first one, your examples are easy to distinguish.
Yep and no. That's a common misconception. Like Chinese, old Japanese went through a systematic simplification of its kanji 100 or so years ago. The rules were decided a priori to make the job easier. This meant that several kanjis ended up looking somewhat similar even though they have nothing to do with each other. That's why comparing kanji can work but it can also backfire and lead you astray.
Words in English are built from characters we can pronounce individually and this extends to words so whilst you might not get it perfect you can at least guess the sound of an English word. This isn't the case for characters in Mandarin or Japanese, no sound is encoded in the strokes, the equivalent of a character.
Actually English spelling is pretty complicated and completely different from pronunciation. Why on Earth you Englishmen write "a" (as in skate) but pronounce it as "ei". And why you have so many versions for each vowel and why you have 3 letters for "k" sound and 2 letters (i, y) for the same sound.
Very interesting, definitely the first step towards writing a character for me is picturing it in my head. If it's particularly challenging I might, still in my head, project it onto the page and that seems to give my brain the spatial data to begin translating it to real world movements with the pen.
But even still I also can barely write maybe 5% of the kanji I can read. As well words are often made of multiple kanji, but if you showed me the kanji separately I don't always recognise them as part of a word I do know. Recalling a kanji into my minds eye doesn't seem to be part of the skillset of reading, maybe just a by-product of long term repeat exposure.
This thread is reminding me that I desperately want a way to way to have different font sizes for different languages. I like my English text small, but kanji/Chinese characters are obviously more visually dense. So I default to 10/12 point for latin alphabet but I need somewhere between 20-25 point to read Japanese comfortably. Where the two are mixed, like in this thread or on social media (where maybe half my feed is Japanese) it's a massive pain in the ass.
agree this would make a nice accessibility setting in browsers. or maybe there exists a browser extension?
I've noticed the same thing (not with Japanese), but even after learning to write in Korean, my ability to handwrite faded way faster than my reading skills. I figured it was just laziness, but it makes sense that reading and writing use totally different parts of the brain.
How do you lose the ability to handwrite in Korean? That sounds weird to me.
That's why I focus on learning words rather than individual kanji characters. Kanji characters have so many different readings, and the Japanese language is full of exceptions that it's not worth the time unless you're passionate about it, eventually you will pick up the readings of most common ones and you will be able to guess new words automatically, writing them no thank you, I barely write in my own language without a keyboard.
I tried this, but I couldn’t really tell the differences between even quite obvious characters until I sat down with them individually.
Nice post! Enjoy your blog's overall aesthetic too. Perhaps correlation in sense of style, though, as I also used RTK to learn to write Kanji, loved it, and now, ~15 years after that escapade, am kind of in a similar bucket (can write some characters, but mostly just read). I still think RTK great overall method and would do it again!
Also, shoutout to Fabrice, creator of Kanji Koohii -- that was my first foray into SRS back in ~2007/2008, after which I found Anki (pre-mobile).
I find myself having trouble remembering the spelling of certain words, now that autocomplete is almost ubiquitous. It sometimes takes writing them out and looking at the result ("Is it 'beleive'? No, definitely 'believe'.") to remember the correct spelling. Chinese character recall is probably closesly associated with that than with romaji recall.
What drove me crazy when I started to learn Japanese was that a Kanji characters can have so many different pronunciations. The most egregious example is 煙草, whose pronunciation is the really just tabacco(tabako, or タバコ). I knew the etymologies and the historical context on why Japanese evolved like that, but it's nonetheless hard for me to remember all the pronunciations, at least initially.
Also, I find that knowing Kanji is essential in appreciating part of Japanese culture. Take their addresses, for example: Kinukawa is really meaningless, but 鬼怒川 is such a amazing name. Similarly, Akihabara means little if all we know is the pronunciation, yet its Kanji 秋叶原 is such a beautiful and poetic name that invokes complex emotions.
It's Kinugawa, and Akihabara is written as 秋葉原.
Yeah, I mixed up my input and chose simplified Chinese
This is probably why Japan still adamantly emphasizes writing.
Written resumes/ fax machines ... remain the norm, and while this may seem anachronistic for the rest of the world (pretty much all of which uses either (semi-) phonetic scripts derived from Aramaic or from Brahmi), it makes sense after you come across the Chinese characters.
No, it's because Japan is a deeply conservative country, really adverse to change, because change means taking risks. If you ever had to deal with Japanese you would know, layers upon layers of convoluted procedures for no real reason.
But hey, at least now I can get the "my number card" to get my hanko certificate from the convenience store. Maybe in 30 years I will be able to use the card directly to prove my ID without needing the stupid hanko!
> Admittedly, I've never heard of someone forgetting how to write a letter from the Latin alphabet.
No but when I last tried to hand write a long text in the latin alphabet, my hand hurt after a while. Skills atrophy when not used, more complex skills may atrophy faster.
This reminds me of one time I mentioned to someone I had aphantasia and their response was “how do you spell!?” Seems wild to me that some people see words in their head to spell them but I guess at least one person does. I do wonder if that means they’d have better kanji recall for writing.
It’s real. I didn’t know it was a thing until a few years ago then suddenly everything made sense.
I’m glad people are talking about it.
What is aphantasia, specifically?
I don't consider myself to have aphantasia. If I close my eyes and try to 'see' with my eyes the letter 'm' or an apple, all I see is the back of my eyelids– pitch black.
If I then try to 'see' with my mind the letter 'm', I can imagine the shape and drawing the shape, but it never appears in a physically visual manner. I can trace its lines with my eyeballs, but try as I may to hallucinate an image, it's still only pitch black always. The closest I come to seeing it is being very confident that I know exactly what 'm' looks like, and that I could take that mental model and draw it exactly on paper immediately.
Do some people have such a clear and strong mental image that they can effectively inspect, zoom in and manipulate a mental image as well as a real visual image, and that's considered not-aphantasia?
> Do some people have such a clear and strong mental image that they can effectively inspect, zoom in and manipulate a mental image as well as a real visual image
yes, 100%, and more than that. Even with eyes open I can overlay a completely different environment and stop "seeing" the real world. When I close my eyes I find it difficult to really see blackness. Example: when running laps, I count laps by seeing a giant number fixed in the sky over the lap marker, each number a different material (flaming, made of ice, a trimmed hedge, etc).
Why do the numbers have to be made of fire and ice and topiary?
It helps distinguish them. Seeing all numbers as e.g. black vinyl makes them blend together in my memory.
Visualizing something lets you leverage visual and spatial memory, but even then: if I were really running past N real giant numbers which all looked identical, I'd lose track just the same. Distinguishing them visually makes them all unique and memorable. The color infuses the entire track and the sky, actually, so it requires little focus, because it's right there in the background.
Basically like the memory palace from Moonwalking with Einstein, but less work, because they don't have to be consistent over time.
(thanks for "topiary" :) beautiful word)
It’s time to update your self concept. The Wikipedia article is pretty good on it and the apple scale is illustrative of the spectrum.
Idk, I just read the Blake Ross' (of Firefox) article [1] on having aphantasia, and identify less, rather than more, with how he describes it:
"1. Can you picture my face? >No. But it’s not personal. 2. So you don’t know what I look like? >I know facts ... If you have radiant blue eyes, I may have stored that information. ... I’m unable to project it visually in my mind because there’s no screen. 4. How about picturing something simpler, like a red triangle, or the table right in front of you? >I can’t even understand the question. I can think about the idea of a red triangle. But it’s blackness behind my eyes. Blackness next to my ears. Blackness in every nook and kindle of my brain. 5. You’re just assuming that others can actually SEE things with their eyes. NOBODY can do that, you hypochondriac. I get it. It’s a “mind’s eye.” I don’t have it."
[1] https://m.facebook.com/nt/screen/?params=%7B%22note_id%22%3A...
It's unconvincing. They sort people into categories by a questionnaire, and then find they perform differently on tests that have something to do with vision or imagination. That's it, that's all they've come up with.
How are you with reading fiction? Easy or a struggle?
It's easy, the experience isn't very visual though. More about the plot and emotions I'd say.
If you're doing an exam in China or Japan, do you write on paper, or use a computer?
Not the question you asked; but something that actually surprised me, as someone who has taken multiple language tests in the EU:
The (semi-official) Japanese language tests do not test language production in _any_ way, even on the highest levels.
You don't need to speak a single word, nor write a single character to pass the exam, it's all single-choice answers.
AFAICT, the equivalent exams for Chinese and Koreans do include writing, at least on the higher levels; but still don't have a speaking portion.
> The (semi-official) Japanese language tests do not test language production in _any_ way, even on the highest levels.
This is simply not true. The Kanji Kentei is probably the most popular Japanese language test Japanese people might take outside of school, and it tests writing heavily.
If you're talking about the JLPT, that's a test meant only for non-native speakers.
> all single-choice answers.
Not even multiple choices? Maybe they should reconsider that one.
>This is simply not true. The Kanji Kentei is probably the most popular Japanese language test Japanese people might take outside of school, and it tests writing heavily.
I would not qualify Kanji Kentei as a "language test" any more I would qualify a Spelling Bee as one; though I can see how someone might.
In any case; I thought that it was relatively clear from the context that I was using "language tests" as a shorthand to "tests for non-native speakers of that language"; but I did not explicitly mention that, you're right.
>Not even multiple choices? Maybe they should reconsider that one.
Is that not how these are called in English?
Where I'm from, there's a distinction for tests where given a list of answers, there can be only one valid one ("single choice") and where there can be multiple valid ones ("multiple choice").
Quick googling on my end indicated that _is_ how they're referred to in English, but the websites were quite AI-sloppy, so if I was mistaken, I'd love for you to let me know what the more widely understood terms are.
Since you used the phrase "highest levels", that made it sound to me like you were talking about native tests.
I apologize for misunderstanding.
Non-native japanese language tests do not ever go to a high level as far as I'm aware (N1 definitely doesn't classify as a high level, it's around "can read the newspaper very slowly and with greater than 50% comprehension", which is well below even the least-studious native-speaking adults).
> Is [single choice answer] not how these are called in English?
It's an unusual term to my american-english ears. I've only heard them called "multiple choice questions" regardless of how many answers you can select. It's unusual enough that it conjures to the mind a multiple-choice test where each question has only a single answer, say "a", and you just circle "a" for every answer.
I have recently had the experience of Labguage School in Japan and they also realy heavily lean into having just circle the correct parts or do somethign like insert correct Particle here.
While I have tremendous respect for the Japanese Language and people. The Japanese in all my exposure to it do not know how ot tech languages. Be it their own to people that do not know it or English to their own people.
Almost always paper. The people who forget are adults who don't write anything by hand anymore except the occasional form or sticky note.
In Japan, paper
In China, paper
> This is also why I believe that language is a bottleneck for thought. Most of what you remember is nothing like an approximate copy of the things you experienced in real life—even in the specific case of text, memory is not even remotely like a paraphrase of previously read words. Many of our thoughts happen in a highly abstracted and distilled form, interacting and connecting with each other as a network that simply cannot be faithfully converted into a sequence of words, however long.
Perhaps the most interesting quote in an interesting article.
Leaves me speechless or something
>> This is also why I believe that language is a bottleneck for thought
I am not sure I agree with this. We think our thoughts using language. I don’t language is the bottleneck.
i think you're missing the point. imagine this:
you're walking down a beautiful waterside bridge in kyoto during cherry blossom season. there's really delicious smelling grilled eel from the store next to you. there's a swedish lady walking by you with a cute accent. your skin feels slightly sticky from humidity.
what i just wrote doesn't even begin to encapsulate the entirety of that moment. there were a million other details your brain can form about that moment: the style of brick on the bridge, the other people around you, the sun being in your eyes, that you're smelly bc you forgot deodorant that morning.
to put it shorter: a picture is worth a thousand words. and the author is saying that by having to use language to describe pictures, we have a huge bottleneck
You'd be surprised how many Japanese people themselves forget. Predictive text changed everything.
Perhaps if there were fewer radicals this would be less of a problem. Many thousands of characters could probably be generated from a small number of radicals.
32 radicals are enough for ~75% of characters though. My biggest problem with writing is that I default to writing fairly small so if I have to do a character with more than 12-15 strokes it's either way too big or hard to read later.
It would make some more of the choices of radicals more arbitrary and hurt recall in other ways though
> I believe that language is a bottleneck for thought.
I think this various from person to person because I don’t think in language.
when I learned Japanese in university I learned writing with the exact method by James W. Heisig. So I didn't used Japanese for about 20 Years since, I forgot almost everything I learned, except the Kanji. I can still remember most of them. It's crazy!
3 years of Japanese in high school and I can still read hiragana 48 years later.
This isn’t about the kana though.
They're joking that despite their three years of study, they've only retained hiragana.
Interesting topic but a lot of the articles on this blog reads like undeclared LLM slop:
> This is also why I believe that language is a bottleneck for thought. Most of what you remember is nothing like an approximate copy of the things you experienced in real life—even in the specific case of text, memory is not even remotely like a paraphrase of previously read words. Many of our thoughts happen in a highly abstracted and distilled form, interacting and connecting with each other as a network that simply cannot be faithfully converted into a sequence of words, however long. The fact that people can fail even at something as basic as sketching a kanji or a vehicle they've seen hundreds of times before is just another example of the same phenomenon.
A pet peeve of mine is when someone uses their personal datapoints to generalize for humanity. Every sentence here should say “for me”.
What they said doesn’t even support the idea that language is a bottle-neck for thought, it actually argues against it. If language can’t capture the complexities of thought, then that’s the opposite point as language is a bottle-neck for thought.
I read the linked article https://aethermug.com/posts/the-beautiful-dissociation-of-th... and it clearly is comprised of a lot of LLM slop.
“In techie terms, the Chinese script doesn't support the structure of languages like English and Japanese. It doesn't have what it takes.”
That’s not techie that LLM slop
“Sometimes, instead of using them for their meaning, they used them for (gasp!) their pronunciation. By ignoring the original content of a kanji, they could string them together to form almost any sound.”
I’m becoming allergic (gasp!) to this kind of writing
A system fails when its natives don't know how to use it.
So time to sunset the system, surely? I don't know why so many countries are so obstinately hanging onto something so difficult.
Do it like Korea if you don't want to go the Vietnamese way.
Are you sure? If anything, spelling along makes it hard to read Chinese or Japanese. Examples: can Japanese people correctly understand this: “かんじをせんねんいじょうつかっていたけいいからかんじをはいしするとぶんしょうによるいしのそつうにへいがいがしょうじるからです”？ In fact, the following two different sentences both have the same above kana：
「監事を専念異常浸かっていた敬意から監事を配しする飛ぶん商による石の疎通に兵が意を招 じる空です。」
「漢字を千年以上使っていた経緯から漢字を廃止すると文章による意思の疎通に弊害が生じる からです。
Or how about くさくさくさくらくらくさくさくさくさくさくさくらさくさんさくさんくさくさくらんくらくら？
Or how about こうないしゃせい？Is it 校内写生 or 口内射精?
In all seriousness, even Koreans still debate whether they should continue to use some Chinese characters, and they do so in poems and literatures. For instance, 악유원 means shit, yet 乐游原 has layers of meanings. Chinese and Japanese have too many homephones to use a spelling system like Korean. Unlike Chinese, where individual characters correspond closely to morphemes and carry distinct meanings, Japanese is organized around words as the fundamental linguistic units. Furthermore, because Japanese employs a pitch-accent system rather than the kind of lexical tones found in Chinese, writing some native Japanese words (和語) entirely in kana typically does not create confusion.
Also, Japan had at least three major attempts to remove Kanji. First time: The Tale of Genji, the world’s first novel, was written by a woman entirely in kana, with only a small number of Buddhist terms in kanji—terms which, in theory, could also have been rendered in katakana. This demonstrated that early Japanese could be expressed systematically and fully without using any Chinese characters at all. Nevertheless, the Japanese chose to retain kanji.
Second time: During the Meiji Restoration, slogans such as “Abolish kanji, or East Asia will never grow strong” were popular. In the end, Japan found that it could modernize, prosper, and even become one of the five permanent members of the League of Nations while still keeping kanji. The decision was made to continue using them.
Third time: After World War II, General MacArthur commissioned a team of American education specialists to work with Japanese experts to discuss whether to abolish kanji entirely, and to consider the possibility of fully romanizing the Japanese language. Yet, they decided to keep Japanese.
I don't find the argument about homophones very convincing, since people do speak to each other and it works. People will surely be able to adapt.
But, you're basically throwing away the country's history. In just a few years young people wouldn't be able to read any older text anymore. I can see why you wouldn't want that.
It is substantially harder and slower to read a stream of hiragana than to read a stream of hiragana and kanji. I don't know if you know anything about the language, but it's not because of homophones (though that is certainly improved) -- it's chunking.
Japanese written language has the property that when you see kanji followed by one of a very small number of hiragana patterns, you know you're seeing a verb. Stemming and deriving the meaning of the verb is trivial, because it closely follows from the Kanji (e.g. 見る => 'see', and pretty much every noun or verb involving 見 carries that connotation).
Toss in the few particles (を、で、に、へ、が、は), and you've broken a sentence into semantic chunks with very little mental effort -- and along the way, gained much of the meaning at the same time.
Doing the equivalent with a stream of nothing but hiragana requires a kind of parsing that is like depth-first prefix search, but with ambiguous matching at each terminal. It's incredibly tedious.
I can read Japanese, yes, and obviously also find hiragana-only text hard to parse. I think that would be almost completely solved by using spaces and getting more practice though.
I'm much better at reading than speaking/listening, so the Kanji also help as a clue to the meaning, but that is entirely a non-native problem I think.
> the Kanji also help as a clue to the meaning, but that is entirely a non-native problem I think.
Not sure what you mean by that. It's almost a daily occurrence that someone tries to explain a word's meaning by drawing invisible kanji on their palms.
Someday I'll figure out how to read that.
> it's chunking.
I guess the ultimate question is whether we can have a spelling system that totally mimics how we communicate verbally, so chunking becomes as easy as when we listen to the the spoken language.
I doubt it will happen though. Historical heritage aside, Kanji does pack enough information to make reading very easy. Unless government interferes hard, the new system needs to be way better than the existing one to get adopted.
It doesn't even need to do that. Korean and Vietnamese are easy examples where all the writing is phonetic despite Korean having basically the same structure and problems with Chinese loans as Japanese, and Vietnamese having a ton of Sinitic loans and general properties that are basically the same as Chinese. Yet phonetic writing just works.
I mean, if they got rid of kanji, presumably they would introduce spaces. It's not an alien concept, it was used in early computers and early computer games, and still get used in games and books aimed at kids:
https://www.famitsu.com/images/000/190/128/y_5e083812bfff9.j...
(Pokemon also has a full kana mode, but due to the number of homonyms in Japanese, I think while it might be easier for actual Japanese children, even with the spaces it's harder for a Japanese learner than the kanji mode unless you're literally in your very first month. There's also things that will be obvious to natives but confusing to learners like that すげー in that screenshot is a slangy すごい)
Yet, that's exactly what you do when you read English - and many of us even read it as a foreign language with different roots and different grammar etc. So I don't understand how it would be more tedious in hiragana; it looks like a "just get used to it" thing to me.
Japanese has fewer sounds and a lot more homophones than English, so purely phonetic writing becomes a lot more ambiguous. Children grow out of it at an early age.
Furthermore, kanji often allows the reader to skip phonetics altogether, because the symbol itself carries the meaning. It’s a bit like how people understand emojis without having to make a sound for them in their heads.
As a result, the Japanese are able to read very fast in their native language, so switching to a different system would carry significant drawbacks.
There's an advocate group for all-kana Japanese that existed since 1920, just decade or two after pseudo-Chinese styles was deprecated and kana-kanji mixed as-spoken text became the standard, before even kana switched from katakana to hiragana, and they don't even use all-kana text on its official propaganda page anymore[0], so there's that.
0: https://www.kanamozi.org/hikari939-0501.html
1: https://ja.wikipedia.org/wiki/%E3%82%AB%E3%83%8A%E3%83%A2%E3...
No you don’t. There’s actually a lot of research showing that native English speakers don’t read letters, they read word shapes. And people who read fastest read in even larger groups of words, to the point of eliding segments of sentences or paragraphs entirely.
Chunking is incredibly important for reading speed, and reading hiragana is much closer to reading letters than words. My reading speed in Japanese is nowhere near native, but the way I’ve gained speed so far is almost exclusively by increasing my minimum comprehension unit: I see word patterns, common grammatical constructions, etc., and I don’t need to read them.
> native English speakers don’t read letters, they read word shapes.
Don't try to teach children to read this way, though: it's a high-level technique that comes with practice and familiarity, with near-instant fallback to lower-level techniques as appropriate, or the resort of dyslexics who cannot read any other way. Teaching children to use the approaches used by struggling readers will tend to produce more struggling readers than necessary.
english has spaces to separate words, japanese doesn't. have you tried reading english without spaces?
In the real world nobody is masochistic enough to not adopt spaces if writing without kanji.
Old Japanese videogames couldn't use kanji due to technical limitations. They wrote in all kana but used spaces to make the text easier to read.
Modern Japanese children's books and eg. even Pokemon games still? Same thing, kana and spaces.
When Korean transitioned away from Japanese-style mixed script to purely alphabetic writing, what did they do? They adopted spacing.
The only time "but Japanese doesn't have spaces" comes up, ever, is when people argue against the removal of kanji. It's not a realistic scenario, in light of very recent history and current practice.
This is a very silly argument.
If one is seriously proposing abolishing kanji, surely "also let's add spaces where they make sense" is a much easier pill to swallow.
In a world where hiragana and katakana are adopted to eliminate characters, it's not that hard to imagine also adopting spacing. Hangul for Korean did not originally have spaces in the language.
It's all solvable. Korean did it. Koreans are fine.
For verbs you could just replace the kanji with katakana so that you still get the pattern recognition of okurigana as a visual aide.
They just don't want to do it, because they don't like to change. A generation educated under a new system wouldn't have difficulty using it. They would however lose access to an abundance of cultural artefacts which play a central role in daily life.
Koreans don't seem entirely fine to me, but besides, Japanese pronunciation is actually more aggressively simplified than the other two CJK languages during Edo era that numbers of homophone is out of control.
There are as many as 50 homophones for koushou due to this, for example[1]. Communication by phonetic transcripts alone just isn't going to work.
1: https://togetter.com/li/2380657?page=2
Guess Japanese people just can't talk to one another without subtitles.
Or there could be cues not captured well in the script. I don't know. But the rightmost column is real.
1: https://en.wiktionary.org/wiki/%E3%81%93%E3%81%86%E3%81%97%E...
I've spoken Japanese for the last 15 years. I'm aware of how many homophones there are. I still think by and large it wouldn't be a problem, and that natives would adapt to route around it. It's not a blocker.
I was just watching random QuizKnock video that came up on YouTube and hey, we're not pronouncing those "homonyms" same, or are we.
This made me spend 20 minutes cleaning up all the drinks from my desk and my screen.
It's mostly about lack of practice, frankly. The chief reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from. No one actually reads these English posts by vocalizing letter by letter to slowly build the words together.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
Similar arguments used to be made in Korea, yet if you look at Koreans today they have no difficulty in reading hangeul - they have spaces so the words have form, and they have mass exposure to the hangeul forms of words. Ergo, their shortcuts are for the phonetic forms and those are what feels natural, solid, meaningful and easy to read. Same as both of us in English or me in say, Finnish or Swedish or French.
Maybe you're right that it's all just hard-headed stubbornness from fluent people. But I (and all other learners of Japanese) started by reading hiragana, and only later did we move to mixed script -- this is by necessity. Yet even when you factor in the difficulty of learning to read Kanji, reading Kanji is vastly easier than reading kana, even as a beginner.
Would spaces magically solve this problem? I guess it would solve some things -- you'd no longer have to guess where to terminate the prefix search, and I think you're right about word shape -- but it would definitely not provide the additional semantic context you get from having the high-bit-density characters in the mix. This makes reading faster.
I suspect that one could make a kana-only writing system that would be functional enough, but it would still be slower to read than mixed script. Also, the Korean comparison isn't exactly valid -- Korean has more sounds than Japanese. It seems minor, but Japanese has a ton of homonyms because of the tiny phoneme. Expanding that, even slightly, would be a benefit to reading.
Korean faces the exact same problems as Japanese, though - the language structure is similar, they have a ton of Chinese loans, and have in general gone through a largely identical history of writing development. They have somewhat fewer homonyms than Japanese, sure, but they still have tons from Chinese loans (hell, "coffee" and "nosebleed" sound the same, as do "blood" and "rain" in many cases).
It's somewhat hard to believe that Japanese sits in some magic spot where a phonetic script wouldn't work just fine when Korean does it fine, and on the Sinitic side people write books in pinyin, Vietnamese is phonetic, and the Dungan people write their 3-tone Mandarin dialect with cyrillic alphabet without even notating tones.
> Maybe you're right that it's all just hard-headed stubbornness from fluent people.
It's not just hard-headed stubbornness - reading kana really is more difficult to proficient readers of today's Japanese, and change is work.
One of the advantages of Chinese characters is that you can express an emotion using a single character - like "laugh" or "tears" and you don't need emoji for that. Western alphabet cannot do this.
Also the text is more dense and you can write more using limited number of characters, for example, when you have a button in UI with limited size.
You can't possibly count the third time as a real attempt. A language reform initiated by a foreigner and recent enemy at war, who burned your cities and even nuked two of them? If someone knows about successful examples, I'd be curious to hear about them.
Latin script for Vietnamese replaced Chinese-based script under the French colonial government and were helped by the Nguyen Emperors going along with it. (Interestingly this did not take off in neighboring Cambodia or Laos.)
Hanja is also mostly gone in Korea, particularly in North Korea.
The big thing is that both shifts happened before rapid literacy growth. It's much easier to teach new writing systems when the majority of the population can't write anyways. 95% of Vietnamese could not write in 1945; only 22% of Koreans were literate in that time period.
---
One interesting thing I learned while researching this comment was that a big reason Hanja disappeared was because Koreans gained literacy during the typewriting era, but before computer auto-suggested keyboards, and it was just substantially easier to make and use a letter-based typewriter.
> One interesting thing I learned while researching this comment was that a big reason Hanja disappeared was because Koreans gained literacy during the typewriting era, but before computer auto-suggested keyboards, and it was just substantially easier to make and use a letter-based typewriter.
If you look in the right places, you can find people complaining about how it's impossible to dynamically render hangul blocks, which means that a Korean font needs to define glyphs for every possible Korean syllable as opposed to just defining the elements of the system and letting a word processor assemble them as appropriate.
If that's true, I don't see how hangul could have had any typewriter-based advantage over hanja. From the typewriter's perspective, there's no difference.
The Chinese used typewriters by defining a typewriter code. Assuming that that was necessary for hanja, and also for hangul, why would it promote the disappearance of hanja?
If a typewriter code wasn't necessary for hangul, how did we forget how to lay out the blocks in between then and now? Hangul have been in continuous use for all that period.
> If that's true, I don't see how hangul could have had any typewriter-based advantage over hanja. From the typewriter's perspective, there's no difference.
There are mechanical hangeul typewriters that, while more complicated than Latin or katakana typewriters, are still completely usable for normal writing. The reason hangeul fonts are hard is that a hangeul syllable occupies a standard-sized block, and in eg. careful handwriting the writer would adjust the sizes and positions of the characters to be aesthetically nice. For example, in 해 he the ㅎ andㅐ letters are both the same size. When you write 핸 hen, see how the h especially becomes smaller? In typewritten hangeul, that first consonant is always that small, so you can use only one size of initial h and so on.
https://www.youtube.com/watch?v=UenaIex_ZXY
You can see from the output in this video how the sizes of letters are very standard and somewhat disproportionate, eg. in CV type syllables the vowel lines are somewhat giant compared to the quarter-of-the-block sizeish consonants, etc.
That way you can still write by pressing alphabet buttons, with some controls as to where you want the letter to go in the block. It's a bit more complicated, but nothing compared to the nightmare that are proper Chinese character typewriters.
https://www.youtube.com/watch?v=PDkR87zHdXk
> For example, in 해 he the ㅎ andㅐ letters are both the same size. When you write 핸 hen, see how the h especially becomes smaller? In typewritten hangeul, that first consonant is always that small, so you can use only one size of initial h and so on.
I see the opposite. As those characters render in whatever font my browser picked, the ㅎ in 해 occupies much less vertical space than the ㅐ does.
In the 핸, it still occupies less vertical space, but the difference is smaller. It's about as tall as the left-hand bar of the ㅐ instead of being significantly less tall than that.
Okay. The point is, the typewriter writes 핸-style ㅎ always. Even if it could be a bit taller, but tends to leave the vowel lines bigger.
There are around theoretical 11,000 possible Korean blocks, which is a lot. But Unicode has ~98,000 Chinese characters.
Nine times the characters is a quaint problem for computers but a very difficult problem for a physical typewriter.
No, it is a problem exactly equal to the other one. No typewriter can produce 98,000 different characters. And no typewriter can produce 11,000 different characters.
With zero difference between hangul and hanja, how can the typewriter favor one over the other?
There were Chinese typewriters but they were very large and a lot more annoying to use. Japan also used typewriting. They just look a lot different, with a giant cylinder of tiny keys to facilitate thousands of characters. https://www.youtube.com/watch?v=wQ-SHOsbH4Y
Hangul ones were a lot smaller. I misspoke earlier; they had letter-based keyboards and mostly just did compromises on the shape of the syllabic blocks; a keyset for an initial character, a keyset for medial characters and a keyset for terminal characters. If you just assume the initial character set can be tiny to fit both the characters with and without bottom terminals, then you wind up with slightly odd-looking but perfectly serviceable typed Korean.
https://www.youtube.com/watch?v=m8lfgxBj440&list=PL7HFg4f79l...
Thanks!
For reference, I talked about a typewriter code before, but I suspect that I was thinking of the telegraph code: https://en.wikipedia.org/wiki/Chinese_telegraph_code
I've always wondered how some countries manage to drop Chinese characters, while others can’t—or just don’t want to. And how did Vietnam and Korea manage to understand their historical texts after they stopped using Chinese characters? And how do they create new words nowadays? I guess they just borrow words and pronunciations directly from English or other foreign languages?
From what I understand, Chinese characters carry so much meaning that they’re really hard to replace. Sure, we don’t need them when we’re speaking, so in theory it seems like we could just get rid of them altogether. But in practice, we still rely on them a lot - especially when we’re trying to understand what a certain pronunciation means. And reading is a whole different story. Recognizing characters is just way faster than sounding out spellings. Maybe one reason is that the basic unit in Chinese is not word but morphemes, which mostly are just single characters.
Maybe we could come up with a different writing system, kind of like what Koreans did, instead of sticking with Romanized pinyin.
One thing to note is that for non-Chinese languages, Chinese may have semantic meaning but really has nothing to do with pronunciation; and the Chinese etymology really has nothing to do with how a Korean or Vietnamese speaker independently came up with the word so this is less important.
—-
Another thing to note is that a lot of “Chinese” words in the modern day language are actually Japanese in origin, since Japan was the first country using Chinese script to modernize and adapt Western thought in science and philosophy etc., and the associated terminology. This actually provides a political impetus to replace those words with native-constructed ones since they have a negative historical relation with Japan.
—-
People who replace Chinese script with letters often have dictionaries and whatnot from the transition period to trace back words, and people also still learn Chinese in these countries if they want to, so it’s not as if it’s gone and disappeared; in the same way that modern speakers probably couldn’t read Chaucer in Middle English, or Beowulf in Old English, as it was written on a whim, but there are plenty of scholars who have studied for it. And Modern Chinese has little to do with Literary/Classical Chinese anyways.
—-
Mandarin has a phonetic system, Bopomofo, which was abandoned for political reasons in the PRC. But the problem with replacing Chinese script is political; within the PRC and ROC there are multiple mutually unintelligible languages using Chinese script, and if you pick a phonetic script then it is now Mandarin vs. everybody else.
The Korean writing system is not different in any sense other than the appearance of the characters from Romanized pinyin.
Creating new words happens the same as it does everywhere - people genuinely coin new concepts, they loan foreign concepts as direct loans or calques, or form compound words from existing ones. "creating new words" with Chinese characters is literally just using foreign words to form compounds, something utterly routine everywhere else. It's just not treated as a magical event elsewhere.
For example, people routinely marvel at the ability to make compound words in kanji. But, may I introduce you to German, a language written with boring old non-mysterious Latin script that's famous worldwide for its heavy use of compound words.
They're really just scribbles that point at words or morphemes of the language they're used to write, same as phonetic scripts in that regard. They do let you do some tricks that are hard otherwise, like indicating which nuance of a word you intend by your choice of character, and the common stylistic trick of writing some set of characters and then imposing completely arbitrary readings on them by writing clarifications next to the characters being abused.For example, in Frieren, in one of the early chapters, Frieren says:
In hiragana: The manga writes it like this, however (furigana in parentheses after the kanji): 人を殺す魔法 should be read "hito wo korosu mahou", ie. "magic that kills people", but the manga instructs us to read it first time as "Zorutoraaku", the name of the spell, and the second time properly, when Frieren's supposed to say its description out loud. There's no clarity issue here, it's just a stylistic trick.For fun, the same line from the Korean translated version:
J. Marshall Unger's Literacy and Script Reform in Occupation Japan goes into details on this.
There were kanji abolitionists and pro-kanji people in both the American and Japanese administrations, and it definitely wasn't just a popularity contest. Things happened such as one pro-roumaji principal who enthusiastically took part in roumaji feasibility experiments being assigned elsewhere because he was having results, or one American pro-kanji official decreeing that roumaji publications should be published in triplicate since there were three competing romanization systems - Nihon-shiki, Kunrei-shiki and Hepburn - so they wouldn't unduly advantage any particular romanization system.
This of course also just so happened to make roumaji publishing three times more expensive. Whether fairness or limiting roumaji publishing by financial means was the real motivation is left as an exercise to the reader.
I don't think it's that crazy for the language reform to make it through, seeing as the entire political system has been reshaped into a pseudo-American democratic structure for instance. With the emperor going out of fashion and the governance of the country being overthrown, I think it would be the perfect moment for other cultural shifts as well. Had the timing coincided with the introduction of computers (which famously struggled with languages like Chinese and most Japanese) a few decades later, I think the plan would've succeeded.
Most of Europe inherited the Latin script from Christianity, which was spread with less than peaceful means. It took more than a few short years, but it certainly altered culture.
Japan had a parliament since the Meiji era (first elected in 1890). The Macarthur occupation changed quite a few things, but less than most people think.
My understanding is that Japanese are very open to external influences, even when they come from enemies. Case in point, they embraced the legal framework and political reform that was pushed by the Americans. On the other hand, MacArthur was careful to respect Japanese culture (arguably, depending on who looked at it, I guess) when pushing his reforms. Nevertheless, Japan quickly became a pretty well run democratic country. It’s fascinating and respectable that Japanese could embrace western civilization while retaining their beautiful culture
I mean... they did just get nuked prior to that. That might have had something to do with it.
I dont know who amongst the Koreans argues for the use of Chinese Characters. The vast majority of Koreans don't know even the small amount of Characters they still use. Even simple ones.
That first example is amazing; where does it come from?
There were online discussions on abolishing Kanji in Japanese (漢字廃止論). There were tons of interesting examples, and I took some notes. Unfortunately I don't remember the original sources now.
> Or how about こうないしゃせい？Is it 校内写生 or 口内射精?
It could be 校内射精.
You didn't even try to do it the way Koreans did it, though. When Koreans dropped hanja in favour of a hangeul-only script, they adopted spacing because spaced text is much easier to understand than unspaced text. Likewise, products like old Japanese videogames which couldn't use kanji due to technical limitations used spaced kana because it's just much better and no one's masochistic enough to not use them if they don't have access to kanji. The Japanese mixed script doesn't use spaces largely because kanji already serve as word dividers. If you remove that function, everyone sane will use spaces, and pro-kanji arguers will leave them out to make the proposition seem mad.
Importantly, spaces also make it much easier to recognize the shapes of the words.
It's the chief reason people say kana-only writing (like in old videogames for example) is hard to read: People competent at reading any language don't spell things out in detail, even when we subvocalize we first recognize the shape of the scribbles and our brain has a shortcut from a certain set of scribbles to certain morphemes/words, where the solid feeling of meaning comes from. No one actually reads these English posts by vocalizing letter by letter to slowly build the words together.
Every competent reader of Japanese is first and foremost used to the kanji-hiragana mixed script, and has shortcuts for the kanji forms of words and the sounds of those words. The hiragana only forms? Not so much. So when they complain about hiragana only being hard to read, they're not lying. It really is harder. But it's not harder due to any inherent defect in a hiragana-only script, it's just about a lack of exposure to form those shortcuts that make reading feel easy.
Meme sentences designed to be hard to read that you'd never see in real life aren't an actual point. By that token, the Buffalo buffalo sentence argues for the urgent adoption of kanji in English.
Likewise, in my native Finnish:
ie. It's perfectly readable despite the meme value.Similarily, "kuusi palaa" can mean:
Do we need to urgently adopt kanji in order to avoid homophony? Curiously, it's possible to write books entirely in pinyin: https://pinyin.info/readings/pinyin_riji_duanwen.htmlSimilarily, I highly doubt people would've printed the Roumaji Zasshi if they believed it to be incomprehensible: https://archive.wul.waseda.ac.jp/kosho/bunko17/bunko17_d0035...
J. Marshall Unger's Literacy and Script Reform in Occupation Japan goes into details on this. There were kanji abolitionists and pro-kanji people in both the American and Japanese administrations, and it definitely wasn't just a popularity contest. Things happened such as one pro-roumaji principal who enthusiastically took part in roumaji feasibility experiments being assigned elsewhere because he was having results, or one American pro-kanji official decreeing that roumaji publications should be published in triplicate since there were three competing romanization systems - Nihon-shiki, Kunrei-shiki and Hepburn - so they wouldn't unduly advantage any particular romanization system. This of course also just so happened to make roumaji publishing three times more expensive. Whether fairness or limiting roumaji publishing by financial means was the real motivation is left as an exercise to the reader.All good points. I really can't explain why it is perfectly okay for natives to listen to Chinese/Japanese yet it is so hard to parse Pinyin or Kana when reading, even when spaces are provided.
Maybe it is because we still think in individual characters when reading. As a previous post mentioned, there are many homophones for a given word, let alone for a single character: https://en.wiktionary.org/wiki/%E3%81%93%E3%81%86%E3%81%97%E.... It's not a problem in a conversation because we track context. Yet when we reading, we, at least psychologically, look for specific meaning per character or per word. Say, when I see こえん, I'll have to figure out if it means a park, a small yard, sound, or something else. Of course, I could figure it out from context, but I'd give up that precision and the warm and fuzzy feeling of seeing the corresponding characters, like 故園, or 子園, or 呼延。We'd solve this problem by giving up learning the individual characters and focus on words or phrases, but then it would be a drastic change.
While typing the above, I also realize that maybe the reason we don't mind homophones during a conversation is because we've already learned the associated characters. Using the example above, when I hear the word "こん” and it means "呼延”, I would know that it's an ancient family name that was associated with many famous generals because of the meaning of the characters. Or if the meaning is "故園”, I would get all the poetic feelings as it is precisely these two characters, not the sound, that deliver the meaning of ancestral land, or childhood home, or place left behind, and etc. And when we study the Chinese and Japanese poems, we focus on the masterful use of characters, and every character matters. Is it 推 or 敲 in 僧敲月下门? Why are the characters in 大漠孤烟直 so compelling and masterful even though each character is so commonly used? It's hard for me to imagine how homophones can differentiate such meanings.
Of course, I'm not saying that removing characters can't be done. I'm just trying to figure out the current state and why many people and I are in favor of learning characters.
> A system fails when its natives don't know how to use it.
They definitely do, it's just that hand writing has become a more niche use case in modern society (regardless of the language).
If you removed all kanji in a block of Japanese text (replaced with kana), I'd expect at least a 50% reduction in reading speed for natives, and some errors in comprehension. They're fundamental to the language.
You wouldn't expect the same 50% reduction if the natives had been educated under that paradigm and suitable adjustments made to the system like adding spaces etc to accommodate it.
How many Americans can write cursive?
I think it's time to abandon the Latin alphabet. It's too inefficient.
Is cursive a requirement to write the English language correctly?
Yes.
(Also, define "correctly".)
That reads as a foreigner suggesting to abandon many thousands years long history and culture just because the characters are too difficult for him.
It doesn't even make sense because today we have computers and ML. If you want to write something, you just type the spelling, and the program automatically converts it into proper characters. And if you don't understand the character, AI can translate it for you. I am sure this can be integrated into electronic glasses.
Japan has had pro-romanization societies since the 1920's, and even during the last attempt at large-scale script reform after the war, it wasn't just the Americans pushing it: Many Japanese were enthusiastic about moving to a phonetic script because they perceived it as more efficient and modern. Likewise, not every American administrator was in favour of reforming away the kanji, far from it.
J. Marshall Unger's Literacy and Script Reform in Occupation Japan is a good treatise on the subject.
EDIT: Also, this was published in 1877 if I did my date conversions right: https://archive.wul.waseda.ac.jp/kosho/bunko17/bunko17_d0035...
I gave you an example: they can't write it.
I also gave you two examples of cultures who already did it: Koreans did it autonomously, the Vietnamese did it during occupation. Both are successful today.
But yes "preserve history" at the expensive of daily life. I'm sure you speak Latin too, plenty of history lost there.
This is a common misconception until you study the language seriously. Japanese nominally has only 50 sounds, known as mora. In practice there are more because people smush them together, but this tends to be for very common words or dialectical variations by region. There are also spoken pitch markers, but they don't have any written indicia.
Because there is such a small number of basic sounds, homophony (words with the same pronunciation but different meanings) is absolutely rampant. This is not too big of a problem when speaking because the context is usually obvious, but writing anything abstract using phonetic script alone will quickly become problematic. It's also tedious to read long texts in phonetic script in the same way that it's tedious to read lots of upper case text in the Latin alphabet.
Another language feature is that Japanese compound words are often formed by just taking 1 or 2 more from several longer words. This would be like shortening 'American-Russian diplomacy' to 'amru diplomacy'. Very clear when written with specific characters, confusing if it is just phonetic. Then, word pronunciation changes within compounds, eg 川 kawa means river, but 山川 yamagawa is mountain river, and these changes are ubiquitous, probably 95% of words go through this phonetic alteration when folded into another word.
It gets more complex again because there are so many particles, words that indicate grammatical inflection of some kind and can drastically change the meaning of a sentence if you use the wrong one. The most common ones are just a single mora and most others are just two. But there are lots of particles. 10-15 essential ones, ~30 that are used all the time, ~65 you need to know for fluency, and more that 150 you might come across in written form.
I could go on and on. The basic reason for all this is that Japanese is fundamentally different from other languages because it developed on islands, and because the weather patterns around those islands made sailing in and out more difficult than islands in calmer latitudes separated by greater distances. They adopted Chinese characters because they didn't have a written language at the time they encountered them, then adapted those characters to develop the phonetic scripts. There's a huge amount of technical debt that would be impossible to unravel.
Do it like Korea
Korean has more different sounds, eg 14 consonant and 10 vowel sounds, vs ~10 consonant sounds and 5 vowel sounds in Japanese. You could also argue there was a higher level of literacy over a longer period, but that requires a lot of history and geography explanation to support - the point being that it would have been much harder to impose by imperial fiat unless it had been done much earlier.
But at the same time, language and script are deeply tied to culture, identity, and history, so it's not as easy as just flipping a switch
long reply here bc this is a Special Interest of mine.
I'm going to put aside the "so difficult" thing here mostly, because the perceived difficulty is partly modern teaching practices not updating beyond "just write the kanji 100000 times ok" (this is a failure of people, not the writing system), and people actively choosing to not write anything at all, which isn't necessarily a problem because nowadays recognition is more important, and I don't see people panicing about how many people don't know how to spell many english words anymore.
Alternatives would be nightmarish and culturally destructive to implement in the modern day. Further, people confuse poor implementation with complexity being the cause of the problem. Taiwan and HK both have high literacy rates despite using traditional characters, yet simplified characters were apparently necessary to increase literacy? It doesn't compute.
Just under 50% of Japanese words are loanwords from Chinese. Recognising characters allows a ton of written nuance and extra vocabulary which the Japanese take full advantage of. There's at least 10 separate words for "kou kai". Even when not fully-remembered, Kanji allow mental mapping of multiple homophones without issue.
We're really talking about a vocabulary mass exctinction event. Back in the ancient times Korea didn't have such a wide-ranging and culturally-mixed set of words as now. If the Japanese do this it'll be cutting off the vast majority of their cultural history. Seems unideal considering the butchering the French did, did exactly that to Vietnam's literary culture and basically cut the people off from much of it.
Chinese doing this would be insane. Mao was tempted to use roman characters instead of the simplification they rolled out, which from a system design perspective only made a worse and more difficult/confusing system with more exceptions to rules than before, and poorer phonosemantic consistency/relations with other characters. Further because as research has proven many times over, it's harder to recognise many excessively simple & similar characters compared to more unique and specific forms. Thankfully Stalin advised him against it.
We really are talking about doing something similar to a total spelling reform of english and just throwing the last 1000 years of literature, written records etc away. Only in the case of especially Chinese, you're throwing away a system that was developed and specifically tailored to the languages using it, which have due to its relative robustness to change as compared with latin letters, led to writing surviving many many difficult periods, regime changes, wars, famines etc over 3000 years.
If you've ever compared Beowulf with Gawain, with Shakespeare and then modern English, you'll understand what a total overhaul a language can undergo if unchecked -- Chinese characters have enabled a comparatively stable orthography. Less than 800 years and it becomes gibberish.
> many people don't know how to spell many english words anymore.
Yes but that's not really comparable.
You can misspell "litrally" but how close can you get if you can't remember any stroke of the glyph? There's an inherent advantage in spelling words closely to how they sound.
I'm not saying English is great at this either, but I can still write "Kernel Sanders" and you know what it means, without using a whole other writing system as fallback (pinyin)
Mw out here in Japan being barely able to rrad any kanji but being able to hold reasonable small talk in Japanese. Spoken Communication yes written forget it.
Not sure if someone else has already commented on this, but this is the fate of people who primarily use LLMs to code. They’ll be like second-generation children of immigrants, who can vaguely understand the meaning of their parents’ native tongue, but lack the ability to produce utterances on their own.
The ultimate goal of AI companies imo is to disincentivize people from pursuing programming as a career by loss-leading with low LLM prices until nobody actually knows how to code without an LLM, and then jack up the prices.
"I used to know how to write code in ${language} but not anymore, though I can still read it. There is even a term for it: LLM baka (LLM馬鹿), meaning LLM idiot"..
"I still do. But I used to, too." -Mitch Hedberg
A favorite joke of mine with a similar structure was Dave Attell's
"I used to do drugs waaaay baaaack - there!" And he points to the side of the stage.
[dead]
If you're a beginner (like me), I built a little space-invaders-like game to help me practice the kana https://xenodium.com/mochi-invaders-now-on-the-app-store It's free (no ads nor funny business).
nice little game!
I enjoyed this read, but I am noticing that people who claim to have aphantasia seem to write about themselves and their experiences an awful lot. I doubt the phenomenon is real.
Some people doubt that sun-sneezing is real, so I can entertain the possibility of being wrong. But sun-sneezing is trivial to demonstrate to doubters, and it doesn’t confer any “I’m special points.” No one would pretend to have it, unlike aphantasia.
I would find it more surprising if the brain did not have quirky versions like this, given how complex it is and given how often our genes are not perfectly copied.
Given the amount of literature on the matter, what durable evidence is there that it actually doesn't exist?