Re: NATLANG: Chinese parts of speech (or lack thereof)

On Sunday, August 8, 2004, at 03:05 , John Cowan wrote:

> Philippe Caquant scripsit:
>
>> I skimmed a little through my documentation. I have for example
>> a small notebook entitled "Writing Chinese - the 214 keys". It is
>> entirely in French, I mean, they just give the Chinese sign and its
>> meaning, without any reference to pronunciation. Clearly here, the word
>> "ideogram" is fully justified, since I see different single drawings,
>> and I learn that such or such six-stroke sign means, for example,
>> "insect", or "blood", or "boat", or "tongue".
>
> That's true, but in the case of these particular 214 characters
> (I will use this as a neutral term for the units of Chinese writing)
> they have a second role, which I will discuss below.
Yes, they are _determinatives_ as john explains below. That is they used
in building up compound characters, where one part gives an approximation
of the pronunciation and the other, the determinative, gives a fairly
vague idea of the meaning. The technique is found in ancient Egyptian,
Akkadian, the so-called 'Hittite Hieroglyphic' & Mayan scripts inter_alia.

>> Then I notice, in another book, that the sign meaning "man" can be
>> pronounced in completely different ways, depending on the dialect
>> (Peking Mandarin, Cantonese, Hakka, Suchow, Fuchow, Amoy, T'ang Min). I
>> can't really read the phonetic writing that's used here, but clearly,
>> several of these pronunciations are totally different and not mutually
>> understandable by Chinese speaking only their own dialect; but as it
>> seems, every Chinese able to read will recognize the written sign for
>> "man".
Every European can recognize and read |&|. I read it as [ænd] or [n=],
according to context; Philippe, I guess, reads it as [e], our German
friends as [Unt], Spaniards as [i], Welsh as [ag] or [a] according to
context, etc.

I repeat again, since it seems to have been ignored, what Yuen Ren Chao
wrote about the character "man":
"It is making a false dichotomy to say that Chinese writing represents
meaning and that syllabic and alphabetc writing represents sound. The
written symbol MAN* represents as much the spoken word _jén_  as the
meaning 'man', the written form _man_ represents as much the meaning
'human being' as the sound [mæn]. The important difference is that of the
size and variety of units."

But {shrug] Y.R. Chao was only a Chinese speaker who became Agassiz
Professor of Oriental Languages and Literature at the University of
California. So what does he know about it?

> Just so.  Whether to call these dialects or languages is a question:
> they are about as different from one another as the Romance languages
> are, but unlike the Romance languages, they do not have separate literary
> and cultural traditions.
Until the last century, the only written tradition was Classical Chinese.
It was rather as though the old Roman Empire had survived and _Classical_
Latin (not your medieval stuff) continued to be the only written medium,
while the ordinary people spoke their various Romance dialects. In this
situation we would not, of course, have the standardized forms of the
modern Romance langs, but a whole continuum of related dialects from Roman
Britain, through western & central Europe and along north African.
_Literate_ people (a minority) would, of course, read the literary
language according to their own pronunciations.

>> Then the Japanese also use Chinese signs (kanji), and very likely
>> pronounce them their own way, having probably very little to do with
>> any Chinese dialect.
How wrong can you be!

In fact, the Chinese characters were adapted in _two_ ways: one based on
meaning, and called the _kun_ reading; the other based on *sound, with no
regard to meaning,* and called the _on_ reading.

> In fact they use the same sign to represent both a Chinese borrowing and
> its Japanese equivalent, so there are multiple possible pronunciations.
> It is as if in writing English we were to use the same Greek letters
> to write both the Greek borrowing "sphygnomanometer" and its English
> equivalent "blood pressure meter".
Absolutely! And to add to the fun, the borrowing of Chinese characters
spanned almost a millennium, with the result that the same character could
be borrowed more than once! It means that a kanji may have _three_
different _on_ readings:
- go-on, reflecting the pronunciation of the 6th cent. CE;
- kan-on, reflecting the pronunciation of the 7th & 8th centuries;
- tō-on, reflecting the pronunciation of the 14th century speech of the
Hangchow region.

To confuse the matter further, as meanings in real natlangs are rarely
precisely delimited, there could have developed more than one _kun_
reading   :)

For example, the kanji, which corresponds to the modern standard Chinese
character written _xíng_ in Pinyin and meaning "to do", "to go", has the
following six possible readings and meanings:
  go-on:  gyō ("line")
  kan-on: kō  ("sequence")
  tō-on:  an  ("travel")
kun #1:  i   - in _iku_ ("to walk")
kun #2:  yu  - in _yuku_ ("to drive")
kun #3: okana - in _okonau_ ("to do")

(The final syllable of the Japanese examples is written in the hirigana
syllabary)

As Florian Coulmas, Professor of Sociolinguistics at Chuo University,
Tokyo, writes:
"Most _kanji_ are associated with multiple pronunciations, and many with
multiple meanings."

In order to determine the correct reading of a kanji, one has to relay on
contextual information. They are hardly examples of ideograms!

> The Koreans also (used to) use Chinese characters, but only for words
> borrowed from Chinese.
No - at one time Chinese characters, called _hanja_, were used both for
borrowings from Chinese & for writing native Korean words. they adapted
same techniques as the Japanese, using the same character both to denote
Korean word(s) of similar meaning to Chinese, and to denote native words
with a similar pronunciation to the contemporary Chinese word.

But in the case of Korean, King Sejong developed an alphabetic script,
called Hangul, in the 1440s. But the educated classes continued to use
_hanja_ for scholarly & official writing till the 20th century. Indeed,
under the Japanese occupation, Hangul was proscribed. It is only since
1945, that Hangul has become the national script of Korea. In the north,
writing is completely in Hangul; in the south, as John observes, Chinese
characters continue to be used for writing words borrowed from Chinese.

Vietnamese was also once written using Chinese characters in a system
known as Chūnôm. Once again we find the same _dual_ adaptation of the
characters, i.e. by meaning and by sound, with no regard to meaning. The
Vietnamese now, of course, write their language wholely in a variey of the
Roman alphabet.

>> But, in fact, it seems that the Chinese use many polysyllabic words
>> too, and these polysyllabic words are represented by several single
>> characters following each other. For ex (sorry that I use "normal"
>> English letters): daxuesheng = student, zhongxuesheng = pupil of a
>> lyceum. Clearly, in both these "words", the signs "xue" and "sheng"
>> are the same, only the first (word ?  character ? syllable ?) differs.
>
> Just so.  There is no exact English or French equivalent for this: in
> Chinese it is called a zi4, which is also applied to a written character.
> "Meaningful syllable" would be the nearest equivalent.
"logogram" or "logograph" seems ro be the commonly used terms. Strictly,
it implies that each character represents a word, which is wrong; but if
we understand it to mean a character representing a morpheme, it is
basically correct. It so happens that the vast majority of morphemes in
Chinese are monosyllabic, and that each syllabic nearly always corresponds
to a single morpheme.

I guess, 'morphosyllabogram' is better, but it's a tad longwinded and
AFAIK not used.

>> Alas, my manual doesn't say what "xue" and "sheng" respectively mean, or
>> maybe it says it, but I don't know where. I can't believe that these are
>> simply meaning-empty "syllables", just like "stu" and "dent" in English.
>
> They are not.  Except in foreign borrowings (of which there are some,
> both ancient and modern), Chinese hardly has any such syllables.
Correct. In _dàxuésheng_ the morphemes are: dà (big) + xué (learn,
study) --> dåxué = "university" + sheng. Some books list the last as a
morpheme meaning, 'be born, life.' I know only in the compounds,
_xuésheng_ scholar, pupil; _xiānsheng_ teacher, Sir, Mr.

>> I understand "daxuesheng" is a compound of three concepts, "da",
>> "xue" and "sheng", whatever they may mean.
>
> "Big", "knowledge", and something I don't recognize.
See above.

>
>> Are there single Chinese signs pronounced as two
>> or more syllables, this I don't know, but I haven't
>> found examples yet
>
> There are not.  There is one character that is pronounced -r, so it
> is less than a syllable.
Yep - there are quite a few monosyllabic bimorphemic words in modern
Chinese, especially in the Beijing region. They are written with two
characters, the second denoting just -r which BTW is not pronounced as a
separate consonant but causes a rhotic pronunciation of the
vowel/diphthong of the first morpheme.

But there are also polysyllabic monomorphemic words in Chinese. These,
though being only a single morpheme, are written with multiple characters:
  one for each syllable. They are all foreign borrowings. But there are two
distinct categories:
- the first group came into Chinese so long ago that the origin of many of
them is either hypothetical or unknown. Nearly all of them are disyllabic,
  e.g.
pípa (four-stringed intsrument)
pípa (loquat tree) - yes, it is homophonous with the above!
zhīzhū (spider)
qīngtíng (dragonfly)
húdié (butterfly)
gănlăn (olive)
pútao (grapes)
Púsà (Bodhisatva)
bōli (glass)
etc.

These are written with two 'empty syllables', as Philippe termed it. More
properly, perhaps, we can say they are written as two bound
'pseudo-morphemes', as though, e.g. we analyzed 'spider' as consisting of
two bound morphemes, spi- + -der, both meaning 'an arachnid of the order
Araneida'.

The second subgroup are relatively recent borrowings, mainly from European
languages, such as tānkè (tank [military vehicle]), kāfēi (coffee), bālĕi
(ballet), ShāshìbĬyà (Shakespeare [sorry about the capital i-breve; my
mailer objects to lower case i-breve), Bèiduōfēn (Beethoven) etc.  These
are written with existing characters, so that, in theory, they could also
have other meanings. But the characters used have become so stereotyped
that most modern Chinese automatically regard such compounds as 'foreign
words', e.g. tānkè is automatically understood to mean 'tank', or at least
some foreign word, and not 'to-be-flat + to-overcome'. But every now and
then a speaker of modern standard Chinese will point out that, say, Lìmă
(Lima - capital of Peru) really means "keen horse", to give a mild example
    :)

>> (BTW, I understand that some complex ideograms, using many strokes,
>> also are compounds of simpler ones).
>
> The great majority of all such characters are indeed compound, graphically
> speaking, but they still represent a single meaningful syllable.
Yes, and the result certainly not an ideogram as one part of the complex
character hints at the *sound*.

> Most
> are devised out of two simpler characters on the principle "it means
> something like X, but it sounds something like Y."
Quite so, with emphasis upon the 'something like' in both cases. Hardly,
as I've said, the requirements of an ideogram.

In a tradition going back about 2000 years, Chinese scholars have
classified the characters under six headings. One of these, called
_chuănchù_ is of obscure nature with ver few established cases of
characters clearly falling into the class. The other five are:
1. Pictographs - characters that originated from pictures of objects. In
their present form the 'pictures' are not clear, but in their earliest
forms of 3000 years ago the pictorial aspect is clear.
2. Ideograms - diagramatic indications of ideas, such as the characters
for 'up' and 'down' and the numeral 1, 2 and 3 (consisting of one, two &
three horizontal strokes respectively).
3. Compound ideograms - in which the meaning of the whole characters is a
combo of the meanings of the individual parts, e.g. 'to stop' + 'arms' -->
'miltary'; 'man' + 'word' --> honest; 'sun' + 'moon' --> 'bright'.
4. Loan characters - this does not mean borrowings from other languages,
but the use a character to represent a word of similar *sound*; e.g. 'to
come' is represented by a character pronounced _lái_ which originally was
a pictogram for a type of grain (also pronounced _lái_.
5. Phonetic compounds - in which the compound is, as John says, 'devised
out of two simpler characters on the principle "it means something like X,
  but it sounds something like Y." '

Y.R. Chao observes that: "Characters in the [first] three categories are
often taken as representative of Chinese writing, but actually form only a
small minority of characters, and it must be remembered they represent
words (or rather morphemes) and do not directly represent meanings. They
are therefore strictly not pictographs or ideographs......"

The 5th class - the phonetic compounds - form by far the most common type
of Chinese character.
[snip]

>> So when we talk about "prefixes" or "suffixes" in Chinese, I suppose
>> they are also separate written (and conceptual) signs placed before
>> or after a main sign.
Yep - just like prefixes and suffixes in any other language!  Duh!

> They are separate written signs and separate meaningful syllables.
>
>> To me, this is not the same as in English or French prefix or suffix,
>> where for ex "rewrite" cannot be separated into "re" and write",
>> because a word "re" doesn't exist on its own (though "write again" is
>> split).
For goodness sake! In Chinese as in any other natural language some
morphemes are bound & some are free. The re- in rewrite _is_ a morpheme
with a meaning; it happens to be bound. 'again' is a free morpheme.
CHINESE DOES HAVE BOUND MORPHEMES - do stop trying to twist verifiable
facts.

>>  Of course I might be wrong here, I'm just trying to understand
>> the system.  So what is a word in Chinese ? It looks yet more unclear
>> than in French or English. In French for ex, you have groups like
>> "clin d'oeil" (a wink) or "bon march?" (cheap) which act exactly as
>> they were one single word (*clindoeil, *bommarch?). In English there
>> are too, for sure. Is this similar to "da xue sheng" ?
>
> English and French are organized around words, but Chinese is organized
> around meaningful syllables; the concept "word" does exist, but it is
> a technical term of linguistics, just as "morpheme" is a technical term
> of English or French linguistics.
These are historical accidents.

> To the Chinese mind, the distinction
> between "rewrite" and "again write" would seem completely arbitrary.
Except that bound morphemes are bound in Chinese. The perfective morpheme
-le, e.g. is bound, and so are word formative morphemes like -tou (with a
very vague meaning! it seems to be essentially a noun-forming suffix).

> Indeed, the most difficult question about writing Chinese in the Latin
> alphabet is where to put the spaces.
Always a bit of a problem when the native script doesn;t use white space.
How should we write SENATVSPOPVLVSQVEROMANVS? We usually write "Senatus
Populusque Romanus" (3 'words'), but the Roman abbreviation SPQR suggests
that maybe "Senatus Populus que Romanus" (4 words) is how the Romans might
have done it if someone had introduced the concept of white space. Maybe,
we should write "Senatus Populus-que Romanus"

> Spaces separate words, but what
> a word is is often unclear.  For another example, "tou teng" can be
> understand as the sentence "[My] head aches" or as the noun "headache",
> depending on the surrounding context.
Maybe we should rely on context.

> If you want to know what the word "jiguang" ("laser") means in Chinese,
> you would ask "Ji [and] guang, these two meaningful-syllables, what
> do they mean [together]?"  Ji is "stimulated" and "guang" is "light",
> reflecting the etymology of "laser" (an English acronym for "light
> amplification by stimulated emission of radiation").
? what's the problem? It's clearly a compound formed on the analogy of
English 'radar'.

>> Anyway, at the moment, I can't see why ideograms should not be called
>> ideograms any more.
>
> Because they represent not ideas (which have no language) nor words of
> Chinese, but rather meaningful syllables of Chinese.
AMEN!

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com    (home)
raymond.brown@kingston-college.ac.uk   (work)
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760

Reply