5 ms·

This discussion wouldn't be complete without a mention of https://en.wikipedia.org/wiki/Lion-Eating_Poet_in_the_Stone_Den https://en.wikipedia.org/wiki/Lion-Eat

by vitus 2y ago

This discussion wouldn't be complete without a mention of https://en.wikipedia.org/wiki/Lion-Eating_Poet_in_the_Stone_Den https://en.wikipedia.org/wiki/Lion-Eating_Poet_in_the_Stone_..., which AIUI was initially constructed as an argument against Romanization. In short, it's the same nominal sound with varying tones ("shi", which is closer in pronunciation to "shirr" than "she"), repeated about a hundred times, which is of course meaningless in spoken form (since there's not enough context to differentiate between the various forms), but actually conveys a story in written form. With the shift toward typing and (especially mobile) computerization in the recent era, it's really not surprising (to me, at least) that Chinese society is moving in a direction where literacy no longer extends to recall of individual characters, and only encompasses recognition, since recall is no longer as necessary of a skill in day-to-day life.

vlz 2y ago
Thanks for the interesting link! Nitpicking a bit, but if I understand this page (linked from the wikipedia article? see point 3) https://pinyin.info/readings/zyg/what_pinyin_is_not.html https://pinyin.info/readings/zyg/what_pinyin_is_not.html correctly however, the text was not meant as an argument against romanization but as a playful example of how pinyin are unfit for classical rather than modern vernacular chinese.
- vitus 2y ago
  I'd accept that interpretation. To be more precise, I view it as a demonstration of information loss from replacing classical characters entirely with romanization, as opposed to a forceful argument against any form of adoption of romanization.
DonaldFisk 2y ago
The poem is written in Classical Chinese, which was spoken over 2000 years ago, and back then would have been intelligible to a listener because the words would have sounded different. Even today, they sound different in e.g. Cantonese. There's a close relative of Mandarin (Dungan) which is written in the Cyrillic alphabet. The spoken language is tonal, but tones aren't used in the written language because written words are polysyllabic, and if you know how to speak Dungan, you can reliably infer the tones. https://www.omniglot.com/chinese/dungan.htm https://www.omniglot.com/chinese/dungan.htm
- porphyra 2y ago
  The poem uses now-rare characters from classical Chinese but it was written in the 1930s and uses the modern Mandarin pronunciation of said characters. The whole point of the poem is to make everything "shi" in modern Mandarin pronunciation, to argue against switching from Chinese characters to Latin alphabet romanization.
  - DiogenesKynikos 2y ago
    You can also construct ridiculous sentences in English that no native speaker will understand [0]. In normal texts written in modern Chinese, this is not a problem. Nobody writes real texts like the "shi" poem. In cases where something can only be understood in written form, you can rephrase it to avoid homophones. 0. https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo_buffalo_buffalo_Buffalo_buffalo https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffal...
    - DonaldFisk 2y ago
      In normal texts, that's correct. However, written Chinese does contains semantic information which the spoken language and Pinyin lack and, unlike English, has fewer distinct syllables, and seldom borrows words from other languages. So someone who's literate in Chinese would usually be able to infer the meaning of unfamiliar words when written down, as they would already know the meaning of all their component characters, but might struggle if they were written phonetically. This is like having a good knowledge of Classical Greek when encountering words like nephropathy or myocarditis for the first time. It still isn't a very good argument, though. Most English speakers get by without any knowledge of classical languages, and accept having to look up words in a dictionary.
      - DiogenesKynikos 2y ago
        Someone who's literate in Chinese would only be able to infer the meaning of an unfamiliar character if they already knew all of the surrounding characters. Then, you can guess the meaning of the character based on context, and possibly hints from the character itself about pronunciation and/or meaning (though this is very hit-or-miss, because many characters don't contain obvious hints). In order to reliably know all of the context surrounding a character, you need to know about 3000 characters total (that's the point at which you can recognize 99% of characters on a page). This is still a very tall order, which takes years of study to achieve. The Chinese characters do indeed contain semantic information that Pinyin (the standard Romanization) does not, but in practice, you don't need that extra semantic information. If you write down a single word in Pinyin, it may have a few homophones, whereas the same word, written in Chinese characters, would be unambiguous. However, in written Pinyin texts, you would almost always be able to figure out which word is meant from context. In the few cases in which that would not be possible, the author could slightly rephrase the text to make it unambiguous. Most languages on Earth (that have a writing system) are written using alphabets. Chinese is not so special that it could not be written using an alphabet as well. The reason why China hasn't switched to an alphabetic script is because of cultural attachment to the script, not because the Pinyin doesn't work just as well in a practical sense.
adrian_b 2y ago
This argument is also used for Japanese, but I do not consider it valid. This just proves that a phonetic writing is not sufficient, but it does not mean that the phonetic writing must be replaced with traditional writing. To resolve the ambiguity of the phonetic writing, both in Chinese and in Japanese, where the ambiguity is much worse, it is enough to retain at most a couple hundred symbols to be used as semantic classifiers. It is likely that a great part of the traditional radicals would be suitable to be retained as classifiers, with perhaps a part of them omitted if redundant and a few other symbols added, if necessary. Then the writing could be phonetic, but with classifier symbols attached to words, wherever the ambiguity makes them necessary. This is not a new method. The oldest writing systems, like those of Egypt or Mesopotamia, also used classifier symbols (with meanings like: "a kind of human", "a kind of god", "a kind of animal", "a kind of stone", "a kind of wood", "a body part", "a kind of tool" and so on) attached to the words written phonetically, to avoid ambiguities. If one would have to learn only 200 classifier symbols and with lower stroke counts than most symbols used now, that would be a great simplification. Many of the Chinese characters are actually intended to be composed of two parts, a semantic classifier and a phonetic symbol, but this principle is applied too inconsistently and with too many variants, so the system can be greatly simplified by using a simple phonetic writing like Pinyin together with semantic classifiers inserted in the text only if they are necessary.
- latentsea 2y ago
  I think ambiguous homophones aren't actually much of a problem. There's usually only correct option that matches the surrounding context, so the correct inference is easy to make even with no characters at all . After all, there aren't subtitles when you're talking to other people, all the homophones still exist, and yet communication doesn't seem to be impeded.
- inkyoto 2y ago
  > Many of the Chinese characters are actually intended to be composed of two parts […] That is not entirely true in the case of Mandarin, but it is more true in the case of Cantonese (and a few other Chinese languages). Owing to the historical loss of sounds (especially finals) over the course of the Mandarin development, many Mandarin words tend to be longer (3-4 syllables are common) compared to their counterparts in, say, Cantonese where they are most of the time (but not always) are two syllables long due to the fact that Cantonese has retained more sounds from Middle Chinese (plus, the intermingling with the Bat Yue) over the course of its development. Which is why the «Lion eating poet in the stone den» still makes some sense when read out loud in Cantonese (also in Wu, Min) and makes no sense in Mandarin.
dehrmann 2y ago
> Lion-Eating Poet in the Stone Den Sounds like Buffalo buffalo, but it's more like someone being clever than pointing out an actual problem with the language.

by vitus 2y ago

vlz 2y ago

Thanks for the interesting link! Nitpicking a bit, but if I understand this page (linked from the wikipedia article? see point 3) https://pinyin.info/readings/zyg/what_pinyin_is_not.html https://pinyin.info/readings/zyg/what_pinyin_is_not.html correctly however, the text was not meant as an argument against romanization but as a playful example of how pinyin are unfit for classical rather than modern vernacular chinese.

vitus 2y ago
I'd accept that interpretation. To be more precise, I view it as a demonstration of information loss from replacing classical characters entirely with romanization, as opposed to a forceful argument against any form of adoption of romanization.

DonaldFisk 2y ago

The poem is written in Classical Chinese, which was spoken over 2000 years ago, and back then would have been intelligible to a listener because the words would have sounded different. Even today, they sound different in e.g. Cantonese. There's a close relative of Mandarin (Dungan) which is written in the Cyrillic alphabet. The spoken language is tonal, but tones aren't used in the written language because written words are polysyllabic, and if you know how to speak Dungan, you can reliably infer the tones. https://www.omniglot.com/chinese/dungan.htm https://www.omniglot.com/chinese/dungan.htm

porphyra 2y ago
The poem uses now-rare characters from classical Chinese but it was written in the 1930s and uses the modern Mandarin pronunciation of said characters. The whole point of the poem is to make everything "shi" in modern Mandarin pronunciation, to argue against switching from Chinese characters to Latin alphabet romanization.
- DiogenesKynikos 2y ago
  You can also construct ridiculous sentences in English that no native speaker will understand [0]. In normal texts written in modern Chinese, this is not a problem. Nobody writes real texts like the "shi" poem. In cases where something can only be understood in written form, you can rephrase it to avoid homophones. 0. https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo_buffalo_buffalo_Buffalo_buffalo https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffal...
  - DonaldFisk 2y ago
    In normal texts, that's correct. However, written Chinese does contains semantic information which the spoken language and Pinyin lack and, unlike English, has fewer distinct syllables, and seldom borrows words from other languages. So someone who's literate in Chinese would usually be able to infer the meaning of unfamiliar words when written down, as they would already know the meaning of all their component characters, but might struggle if they were written phonetically. This is like having a good knowledge of Classical Greek when encountering words like nephropathy or myocarditis for the first time. It still isn't a very good argument, though. Most English speakers get by without any knowledge of classical languages, and accept having to look up words in a dictionary.
    - DiogenesKynikos 2y ago
      Someone who's literate in Chinese would only be able to infer the meaning of an unfamiliar character if they already knew all of the surrounding characters. Then, you can guess the meaning of the character based on context, and possibly hints from the character itself about pronunciation and/or meaning (though this is very hit-or-miss, because many characters don't contain obvious hints). In order to reliably know all of the context surrounding a character, you need to know about 3000 characters total (that's the point at which you can recognize 99% of characters on a page). This is still a very tall order, which takes years of study to achieve. The Chinese characters do indeed contain semantic information that Pinyin (the standard Romanization) does not, but in practice, you don't need that extra semantic information. If you write down a single word in Pinyin, it may have a few homophones, whereas the same word, written in Chinese characters, would be unambiguous. However, in written Pinyin texts, you would almost always be able to figure out which word is meant from context. In the few cases in which that would not be possible, the author could slightly rephrase the text to make it unambiguous. Most languages on Earth (that have a writing system) are written using alphabets. Chinese is not so special that it could not be written using an alphabet as well. The reason why China hasn't switched to an alphabetic script is because of cultural attachment to the script, not because the Pinyin doesn't work just as well in a practical sense.

adrian_b 2y ago

This argument is also used for Japanese, but I do not consider it valid. This just proves that a phonetic writing is not sufficient, but it does not mean that the phonetic writing must be replaced with traditional writing. To resolve the ambiguity of the phonetic writing, both in Chinese and in Japanese, where the ambiguity is much worse, it is enough to retain at most a couple hundred symbols to be used as semantic classifiers. It is likely that a great part of the traditional radicals would be suitable to be retained as classifiers, with perhaps a part of them omitted if redundant and a few other symbols added, if necessary. Then the writing could be phonetic, but with classifier symbols attached to words, wherever the ambiguity makes them necessary. This is not a new method. The oldest writing systems, like those of Egypt or Mesopotamia, also used classifier symbols (with meanings like: "a kind of human", "a kind of god", "a kind of animal", "a kind of stone", "a kind of wood", "a body part", "a kind of tool" and so on) attached to the words written phonetically, to avoid ambiguities. If one would have to learn only 200 classifier symbols and with lower stroke counts than most symbols used now, that would be a great simplification. Many of the Chinese characters are actually intended to be composed of two parts, a semantic classifier and a phonetic symbol, but this principle is applied too inconsistently and with too many variants, so the system can be greatly simplified by using a simple phonetic writing like Pinyin together with semantic classifiers inserted in the text only if they are necessary.

latentsea 2y ago
I think ambiguous homophones aren't actually much of a problem. There's usually only correct option that matches the surrounding context, so the correct inference is easy to make even with no characters at all . After all, there aren't subtitles when you're talking to other people, all the homophones still exist, and yet communication doesn't seem to be impeded.
inkyoto 2y ago
> Many of the Chinese characters are actually intended to be composed of two parts […] That is not entirely true in the case of Mandarin, but it is more true in the case of Cantonese (and a few other Chinese languages). Owing to the historical loss of sounds (especially finals) over the course of the Mandarin development, many Mandarin words tend to be longer (3-4 syllables are common) compared to their counterparts in, say, Cantonese where they are most of the time (but not always) are two syllables long due to the fact that Cantonese has retained more sounds from Middle Chinese (plus, the intermingling with the Bat Yue) over the course of its development. Which is why the «Lion eating poet in the stone den» still makes some sense when read out loud in Cantonese (also in Wu, Min) and makes no sense in Mandarin.

dehrmann 2y ago

> Lion-Eating Poet in the Stone Den Sounds like Buffalo buffalo, but it's more like someone being clever than pointing out an actual problem with the language.