It's known that Literary Chinese (or Classical; wényán ), the language of historical Chinese texts, differs completely from modern Mandarin as well as from other spoken Chinese languages, not only in vocabulary but even syntax, grammatical particles, etc.

One difference is that Literary has many single-syllable words, while Mandarin has a lot more polysyllables (especially disyllables). The usual theory is that the spoken language, too, used to have the same short words as Literary, only with more phonemes available to distinguish syllables; as it lost phonemes, the ambiguity was compensated by adding more syllables. This process (the theory goes) ended up creating a gulf between Literary and the vernacular (spoken).

However, Henry Rosemont argues that written Literary Chinese might have been a special, artificial register from the start—that the gulf was always there. In the same vein, the linguist Victor Mair has found evidence of longer words (like taolu "path, way") being very archaic (at least 463 BCE in this case), and proposes that the short Literary words (like tao "path, way") would be truncations of the spoken language, not their predecessors. According to this theory, Literary would be like a telegraphic notation abbreviated from the spoken Old Chinese, and never was anyone's mother tongue.

I find this idea very intriguing, but not knowing Chinese, I'm having trouble finding more resources for or against it. Could anyone suggest research or essays on this subject?

  • 6
    It would not be at all surprising if literary Chinese were a special, organised form of the spoken language, and not directly the ancestor of the modern languages. That is the case for both Latin and Sanskrit.
    – Colin Fine
    Commented May 17, 2013 at 16:28
  • 1
    @Colin Fine. I don't understand your reference to a "special, organized form" of a spoken language. "Special" in what way? Are there "unorganized" languages? More generally, how artificial could Classical Chinese be? Commented May 19, 2013 at 3:48
  • 5
    Most languages, now and in the past, are just what people speak: nobody makes an attempt to fix or codify them. A moderate number of languages today are regarded as "standard" or "national", and there is a more or less formal agreement as to their form. A small number of languages have been explicitly codified by scholars or literati to be a medium for literature: that is what I mean by "organised". Certainly the classical Latin language which most of the authors we know about used, and which grammars were written about, was rather different from the Latin in the street. (continued ... )
    – Colin Fine
    Commented May 19, 2013 at 17:49
  • 3
    ... and Sanskrit referred to a fixed, refined language (the word 'smskrta' means something like "organised" or "manufactured") which continued to be used as a literary medium long after the vernacular from which it was shaped had fallen out of use.
    – Colin Fine
    Commented May 19, 2013 at 17:51
  • 2
    For what it’s worth, I have never seen anyone (except perhaps 徐通锵, whom I would in general not trust too much) claim that Classical Chinese isn’t a codified mishmash of different registers and dialects. I have never seen any evidence that it represents any natural language spoken at any point in time. Commented Jul 27, 2013 at 19:33

3 Answers 3


This question is really old, but I guess I'll still answer.

One main counterpoint against the commonality of polymorphemic words in, say, Old Chinese, is the fact that Old Chinese monosyllables are actually reconstructed to be quite complicated. "施氏食狮史" (shi shi shi shi shi in Modern Chinese) was pronounced something similar to [ɕia dʲi̯ěɡ dʲi̯ək ʂi̯ər sli̯əɡ].

In addition, we have Old Chinese evidence like the Shijing which were written before Classical Chinese was established as a standard. In fact, "Classical Chinese" arose from an imitation of Old Chinese during Middle Chinese-speaking periods.

It is quite clear that Classical Chinese was quite divergent from the vernacular during Middle Chinese periods. Middle Chinese vernacular grammar is reconstructed to be more similar to grammars of current conservative Chinese languages such as Hakka and Cantonese. This means that, say, Su Shi would not be writing essays the way he speaks at all.


What you or other present as evidence to support "On the idea that Classical Chinese may not be direct ancestor of modern Chinese languages" may happen in other language too, it is just the difference between literary and the vernacular (spoken).

By the comparative historical method, one can construct systemic correspondence between dialects and Mandarin, between dialects of Mandarin and archaic Chinese. How can we imagine an artificial language systemically relates to a spoken one?

Yes, we have not done statistical research, which may be more likely to refute what you say. But what we have observed may lead to negation of your conclusion.


There are several facts speaking against this theory:

  1. Phonetic corresponences between Tibetan and Chinese words which are not borrowings from either of the related languages.

  2. The way the Chinese loan words are pronounced in neighbouring languages of the area which are not related to Chinese (e.g. Korean, Japanese or Vietnamese). They demonstrate regular phonetic correspondences between lexical items of Chinese/accepting languages.

  3. The correspondence between the ancient Japanese scripts (e.g. that of Manyoushyuu) and the corresponding Chinese characters.

  4. Chinese vocabulary items which are spelled / written differently yet cover close semantic fields. The difference implies, either simultaneously or per criterium: initial / tone / medial / final. Some regular corresponences between criteria can be observed.

  5. Tonality of the language and tone sandhi.

  6. Phonetics (that is, the items having phonetic meaning as well) in the Chinese writing system.

  • 5
    I don't see how any of these points address the question. Nobody is denying that Classical and Modern Chinese are closely related, from the same source, and so having the same lexical roots. The suggestion is that Classical Chinese was a significantly modified form of the language (in particular, eliminating most polymorphemic words) relative to the spoken language which is presumed to be the ancestor of modern Chinese languages. So Classical Chinese is an aunt, so to speak rather than a parent.
    – Colin Fine
    Commented Jun 19, 2013 at 15:14
  • 1
    These points prove that there is a correspondence between written and spoken forms of Chinese, which was preserved through language contacts.
    – Manjusri
    Commented Jun 19, 2013 at 15:29
  • 1
    If you mean by "correspondence" that they used words from the same roots, then nobody disputes that. If you mean that they used the same (monomorphemic) words, then I don't see how the points support that at all.
    – Colin Fine
    Commented Jun 19, 2013 at 15:33
  • 2
    The correlation is both lexical and phonetic. Same words are rendered with reggular sound/character pairings in neighbouring linguistic areas. Tibetan words are monosylabic, even when they happen to have complex consonant clusters at beginnings. Morover, there is a correlation for sounds between Chinese dialects as well.
    – Manjusri
    Commented Jun 19, 2013 at 15:37
  • 2
    Once again: nobody is suggesting anything other than a complete phonetic correspondence, and both forms of the language derived from the same source, and so sharing all their relationships to other Tibeto-Burman languages. What they are suggesting is that the creation of the Classical language was partly by shortening and simplifying words by omitting morphemes from them. Data on borrowing might be able to confirm or disconfirm this, if it happened at the right time; but the rest of the points you bring up are irrelevant to the question.
    – Colin Fine
    Commented Jun 19, 2013 at 21:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.