Hacker News Reader

Comment by yorwba | Hacker News Reader

Chinese script is a bad fit for any non-Sinitic language. The languages that did adopt the script (mostly Japanese, Korean and Vietnamese) all went through a phase where Chinese was the only literary standard and as a result they also borrowed large amounts of Sino-Xenic vocabulary together with the writing system.

This got me curious about whether Mongolian has enough Chinese loanwords that it could be written in a mixture of Chinese characters and phonetic script (similar to Japanese). That question led me to this article on the difference between Sino-Xenic and other borrowings from Chinese, using Mongolian as an example: http://www.cjvlang.com/Spicks/sinoxenic.html#nonsinoxenic It mentions in passing that some people in Inner Mongolia do mix Chinese and Mongolian scripts in informal writing.

View on HN

In fact, Chinese script is a perfect fit for any language already: you don’t even need to speak Chinese to be able to read it. English, in particular, would especially benefit from switching the writing system to (syllabic) pictograms: Show Fur, Bag Dad, Van Go, etc.

You need some way to associate characters and words and to remember that association. In Chinese languages and languages that heavily borrowed Chinese vocabulary, there are at least some cues that make remembering the pronunciation easier. (Characters that share the same phonetic component are pronounced similarly.) Without those cues, using Chinese characters gets very tedious (e.g. writing native Japanese vocabulary using kanji).

If you think otherwise, why don't you try writing the English text of your comment with Chinese characters?

> Chinese script is a bad fit for any non-Sinitic language. The languages that did adopt the script (mostly Japanese, Korean and Vietnamese)

What does "adopt the script" mean? From what I understand, Korean is an alphabet language designed to look a bit like Chinese if you squint hard, and Vietnamese is straight up a latin alphabet language.

This is their current situation, but historically this has not been the case. Korean for a very long time was written with Chinese characters, as was Vietnamese.

The article posted by GP also clarifies a distinction between adopting Chinese loan words as any given language may (such as the word "tea" in English), versus a particular model of bulk adoption of not only characters but vocabulary that occurred in JKV languages.

The script was such a bad fit that it was abandoned later, but it was in widespread use (among those who were literate at all) in both Korea and Vietnam:

https://en.wikipedia.org/wiki/Hanja

https://en.wikipedia.org/wiki/Ch%E1%BB%AF_N%C3%B4m

> The script was such a bad fit that it was abandoned later

The mish-mash of katakana, hiragana, kanji (& their crutch furigana) plus romaji in Japanese sort of signifies how bad of a fit that script is to their own language and its use. In many ways it feels like throwing rocks before one's own feet.

It's visually extremely inelegant as well. Seeing a kanji character next to a single stroke katakana character in the same word is ridiculous and yet somehow commonplace.

I would be more surprised that the Japanese haven't done more to clean this up if it weren't for the cultural conservatism deeply embedded in its society.

Likely one of the reasons Japan still uses its traditional scripts is that Japan has historically maintained a relatively high literacy rate dating back to the Tokugawa shogunate[1]. There's never been much of a push factor into simplifying the script, since the educational infrastructure is in place to teach the script to the vast majority of the population. By contrast, countries like Turkey, the People's Republic of China, and Vietnam had very low literacy rates when the government started engaging in mass literacy campaigns, creating an incentive to simplify the language since most people didn't know the traditional script anyway. Obviously, in all of those cases, there were ideological reasons for replacing those scripts as well

[1] https://en.wikipedia.org/wiki/History_of_education_in_Japan#...

> There's never been much of a push factor into simplifying the script

No, actually there was two pushes in modern times. One advocating the use of romaji (or even French) which did not succeed.

Then there was a set a post-WWII reforms that aimed to simplify the script by: standardizing kanji use (eg 言う instead of 言う or 云う), standardization of kanji readings (当用漢字), kanji stroke simplification (新字体), dropping the old kana usage (旧仮名遣), standardization of kanas themselves to get rid of 変体仮名. The convention also shifted from using katakana everywhere where kanji are not used to the mix of kanji/hiragana/katakana of today. So yes, the Japanese script changed quite a lot recently.

Why is it ridiculous? Why does it need to be cleaned up? I'm not sure what the problem is (I don't think I would classify you not finding it visually elegant as a problem).

If anything, the Japanese writing system is very elegant in terms of information flow. In a general sense, each script has its use: kanji for content (meaning), hiragana for function (grammar), katakana for foreign words and sounds. Of course, there are many exceptions to that, but the mix of scripts has benefits for parsing text.

Obviously, the Japanese writing system is not a single constructed system like Hangul, but I don't see any reason why it is inefficient (aside from, perhaps, the time required to learn kanji for school children).

Generally writing systems with higher workloads to read/write lead to less native literacy rates due to people with learning disabilities not being able to get over the hump, not to mention less foreigners learning the language due to the higher barrier of entry.

Is there research supporting that (specifically in reference to native literacy rates)? Would be curious to see it.

I remember reading it in something semi academic many, many years ago, so you'll have to google it.

I think it was linked to dyslexia and languages with spelling and maybe japanese, and how china simplified it's characters led to higher literacy, although that might have a few confounding variables.

Ridiculous comment. The Japanese writing system as a lot of issues, but visual elegance is not one of them. On the contrary the mix of different kind of characters is really convenient to easily parse sentences and give limited information about word origins. It also provides endless possibilities of word plays.

Indeed, mixing different shapes as letters makes reading text much easier. Blackletter, on the other hand, while beautiful in its uniformity, is notoriously hard on the eyes.

> On the contrary the mix of different kind of characters is really convenient to easily parse sentences

There is also an advantage that most foreigners do not realise. The mix of Kanji, Katakana, Hiragana create natural delimiters in the sentence and make spacing useless.

Reading Japanese fully written in Hiragana is currently much harder due to the lack of delimiters between words. The mix of hiragana, katakana and Kanji allow a very high density per character.

It wasn’t such a bad fit for Vietnamese, whose syllabic structure is basically the same as Chinese languages. Actually Han and Chu Nom were used for a way longer time than Quoc Ngu is. And the change to a latin based alphabet is more due to colonization than anything. For Korean you right though.

The absurd number of diacritics one must append to Latin letters in order to use that alphabet for Vietnamese is the best evidence that Vietnam should have had its own rationally constructed system like Korea.

Read more about Hangul, it was never designed to look like Chinese. It has many interesting design decisions which are unique and not found in any other script.

https://en.wikipedia.org/wiki/Hangul

Its history will also tell you what it meant for Chinese characters to be adopted by China's neighbors.

> it was never designed to look like Chinese

Not so quick, there are design decisions that deliberately carry over certain fundamental features from Chinese:

1. The script is not linear at its unit of writing, but fills a square with its base components from left to right, top to bottom (「한글」, not 「ㅎㅏㄴㄱㅡㄹ」; analogue 「韓字」, not 「十早韋宀子」). 2. One unit represents precisely one syllable. 3. Vertical writing retains the orientation of units (not rotating 90°).

(1) and (3) also apply to Egyptian hieroglyphics... They're pretty generic design decisions for languages with composing characters. And (2) is just a syllabary, which is certainly not unique. Any composed syllabary that is meant to read left-to-right, top-to-bottom will likely meet (1) and (2).

To add to what has already been said:

Hangul is the script you are thinking of, which was invented and introduced in the 15th century by King Sejong.

However, there are many Chinese loan words in Korean language, especially on an academic level. You can even find Chinese characters in Korea today, usually in newspapers, on street or restaurant signs. For example, the Korean word for "president" is 대통령 (Taet'ongnyeong), but in newspapers it may be written as 大統領, which are the matching Chinese characters.

Also, most Korean parents choose Chinese characters to write for their children's names, which read (more or less, as Korean does not have tones) the same as the Korean characters.

I don't think that is the still the case anymore.

I navigate from time to time to Korean website, very few occasion they actually uses Chinese characters, except:

1. Some countries names. 中(China)/美(US)/日(Japan)/北(North Korea)

2. Blue House (think of it as Korean's White House, representing the headquarters of the administration), sometimes referred as 青(Blue).

3. The president's surname, sometimes referred as 文(Moon)

Except for those very limited cases, I don't think Korean people are actually using Hanja anymore, it almost feel like some sort of emoji for then, that they probably can't read it, just comprehend the meaning.

As to Korean names, yes I think they still have a Chinese name registered somewhere, but no longer required and not shown on their government ID card.

Another connection: the modern Korean alphabet shows some Mongolian influences. See https://en.wikipedia.org/wiki/Origin_of_Hangul

Vietnamese only became Latin alphabet recently (100 years ago). In fact it's always completely jarring for me to see Latin or Cyrillic characters for languages like Vietnamese [1] or Turkish [2]—about as jarring as it would be to see English written in Arabic characters. So much that, if you asked me what alphabets these languages use without giving me much of a chance to think, I would probably still intuitively reply that they use something close to Chinese or Arabic, completely forgetting the whole story changed a hundred years ago.

[1] https://en.wikipedia.org/wiki/Ch%E1%BB%AF_N%C3%B4m

[2] https://en.wikipedia.org/wiki/Ottoman_Turkish_alphabet

I understand the feeling for Vietnamese which is a tonal language, and indeed has a ton of "weird" annotations on the basic characters which make the usage of the latin glyphs pointless.

But Turkish seems to fit a latin script quite well, and the previous Arabic/Persian wasn't a "native" script for a Turkic language either, I think.

Why does it make latin glyphs pointless? Isn't it better to have a closer match to the phonemes that one actually has to say than having ideoglyphs? (Which are harder to learn, because writing and speaking becomes only loosely coupled.)

These are the written vowels:

Aa Ăă Ââ Ee Êê Ii Oo Ôô Ơơ Uu Ưư Yy

On each of which you can add 5 different tone marks: https://vietnamesetypography.com/alphabet/

Plus there are diphthongs and triphthongs (look at the zero coda and off glide coda columns): https://en.wikipedia.org/wiki/Vietnamese_phonology#ref_Notes

Disclaimer: I know stuff all about Tiếng Việt, but it seemed really hard to learn how to hear and say the sounds when I tried for a few days with private lessons. Using so many diacritics modifies the Latin alphabet so much that it becomes something else...

It would be interesting to compare Khmer script with Vietnamese.

The Khmer alphabet seems to have 33 consonant letters, 15 vowels, 16 "dependent vowels" (which look to attach like an accent to a consonant), and 12 diacritics.

Since I don't have any knowledge of either language, I don't know if this is better.

https://en.wikipedia.org/wiki/Khmer_script

Thanks! Since there are already many different languages using latin script I'm not surprised that this very heavily annotated version too differs. (I'm Hungarian, so maybe that's why I prefer explicit markings for pronunciation.)

If weird annotations can work for slightly different phonemes in Orthographia Bohemica and for Zamenhof, I'm sure you would be able to use them for tones as well.

You can, it's how it's done, but it means someone who knows any other language using the Latin script still can't read it.

A speaker of French, Polish, German, Hungarian or Turkish can muddle through trying to pronounce another European language using the Latin script and get 95% of it but will have no idea of how to read Vietnamese.

The "weird" annotations are (almost) all part of regular Latin glyphs, which is unsurprising since the Portuguese who created modern written Vietnamese patterned it after Portuguese.

These are also common ones. It's also transplanted onto a bunch of other less well known non-sinitic languages like Khitan, Jurchen, Okinawan, Zhuang etc.

Reply on news.ycombinator.com