Re: USAGE: Chinese Romanization (was: USAGE: Help with Chinese phrase)

On Tuesday, September 7, 2004, at 02:33 , Tam·s RacskÛ wrote:

> On 7 Sep 2004 Ray Brown <ray.brown@FREE...> wrote:
>
>> It does, but the GT treatment of the dental, retroflex & palatal
>> series does not strike me as having much bearing on the Pinyin
>> system:
>>
>>                   IPA          GT                 Pinyin
>> dental         ts ts=CA=B0 s      ds ts s            z  c  s
>> retroflex      t=CAÇ t=CAÇ=CA=B0 =CAÇ  =CAê   dsch tsch sch j    zh ch
>> sh=
>  r
>> palatal        t=C9ï t=C9ï=CA=B0 =C9ï      dj tj hs           j  q  x
Ach - it's got a bit mangled.

>   You are right here, but the task of the Chinese designers was to
> create one-letter initials (plus possible accent characters for
> aspirate, retroflex etc. quality). Therefore every predecessor
> system should be -- imaginally -- re-designed with monographic
> onsets before thainking it into consideration. In German |ds|~[ts]
> pair can be easily replaced by |z] and |c|, but Frech and English
> system have no such alternatives.
Yes - one could argue it that way. But I must confess that I saw nothing
particularly Germanic in |c| = [ts_h] and [z] = [ts]. It may be because I
am familiar with too many languages. I think there are other ways the
designers of Pinying could have arrived at this (even via Volapük and
Esperanto). Unless any records were kept of the deliberations of those
involved in the design of Pinyin, we can only speculate.

[snip]
>> Arguably, as I said in a mail I sent last evening, the palatals
>> are allophones of the dental series. The RT is by no means alone in
>> treating them this way.
>
>   (Just a remark: IMHO they are not allophones, in fact, they are
> homophones, i.e. historically different syllables that are merged
> into prensent-day Peking dialect.)
I assume you mean that the palatized dental series became homophonous with
palatalized velar series - yes, I said that in an earlier email. That
presumably is why both Latinxua and the French system doesn't have
distinct symbols for the palatal series (Pinyin j, q, x)

> Here I see an important point:
> all the systems, except Russian, uses a _third_ variant for the two
> merged syllables.
No - Latinxua doesn't and according to my information the French system
doesn't.

> On the contrary, Russian extended one orginal
> variant in a "pars pro toto" way. If the design would have been
> based on Russian influence,
But it wasn't.

>> I do not claim that the RT had any _direct_ influence on Pinyin;
>> what influence it had came through Latinxua.
>
>   Let me use some loose terminology here and let me denote the
> common set of Pinyin and both their endogenous predecessors
> respectively with a common term as "Pinyin".
Why? Doesn't this confuse things?

> But I do not see the
> Russian elements even in Latinxua
I can't say any stares me in the face. All I know is that Russian sinologs
helped in the design of Latinxua; therefore, if there is any Russian
influence at all in Pinyin it would be via Latinxua.

>
>> All true - which makes me even more of the opinion that the Maoist
>> government were concerned basically with the earlier _Chinese_
>> transcriptions.
>
>   Then let me rephrase the question: What kind(s) of conventions
> influenced e.g. Latinxua?
Not being party to the design, I do not know.

>   (I do not continue the thread on politics, however I do not see
> it as simple as you. Mao had a strong opposition in the party and
> he was gradually pushed into the background. The Cultural
> Revolution was Mao's break-out with the help of rural redicalism.
> It was against also previous communist era.)
I am aware that the politics of the People's Republic was not (and is not
simple), but in the 1950s the ruling party, whatever the actual position
of Mao at the time, was strictly communist and not exactly pro-western.
That's why I am skeptical about any direct influence from any one western
source.

>
>>>   Hungarian uses, too, but this is due to the German cultural
>>> influence (the early universities of the area were in Germany).
>>
>> But that's going way back in history.
> [...]
>> I did not say it was! I merely pointed out that by 1950, there was
>> nothing particularly German about |c| =3D /ts/. Why, that is the
>> Esperanto convention!
>
>   The latter is an interesting question. I would be significant to
> know the ideological status of Esperanto in 1950 in the Soviet
> Union and in China. (I am not aware of this.) An "Esperantoist"
> designer could have been a mediator of |c|,
Could be - but it is not a position I would push. My point was that I do
not see |c| = [ts] as particularly German. I would more readily describe
it as Slav usage (Yes, I know it is also Hungarian and Albanian and was
once Old French before front vowels [as in German, where |c| = [ts] has
never become the universal rule]).

>   You may call it "back in history", but I, as an inhabitant of
> that area, still feel this cultural influence.
No doubt - but it was way back a few centuries. German influence may
persist in central Europe - but China?

> And let me mention
> that the significant part of the Soviet elite was of Jewish origin.
> Their native bias was German. (Even Zamenhof's: the /g/ ~ /h/ ~ /x/
> triplet of Esperanto reflects German phonemicity, not Slavonic.)
Didn't it occur in his native Polish?

[snip]
>> No alphabetic system AFAIK consistently uses monographs; some use
>> them more than others. (I do not consider the Zhuyin Fuhao, aka
>> Bopomofo, to be an alphabet)
>
>   I treated Russian {dz} in my argumentation as a monograph (i.e.
> equivalent of Cyrillic {z}), since it is only the Grazhdanka
> solution of Church Slavonic monograph letter "dze^lo". Even {dzh}
> could be interpreted as a functional monograph.
In that case we ought to treat English |ch| = [tS] as a monograph.
Personally, I don't and would treat Russian |d.zh| a monograph.

> These are native
> Slavonic monophonemic solutions. But RT chose suprising solutions
> here instead of native ones, like {ts.z} and {ch.zh}.
Yes, they are odd.

> They are not
> even monophonemic. IMHO, this extreme markedness of these graphemes
> should be reflected in BL or PY (or in other transciprions as well)
> if there would be significant Russian Slavonic contribution.
But there isn't AFAIK.

>
>> But |=FC| is used only after |l| and |n|. In all other instances PY
>> writes [y] simply as |u|.
>
>   It is true but it does not affect the merits of my argumentation:
[snip]

> back |u|) or simply maintain Latinxua solution. Umlaut here reveals
> that German notation/orthography was taken into consideration and
Sorry - I think is pushing the argument to far. By the 1950s the
convention of u-umlaut = [y] is found in too many places to make it
specifically German. Indeed, the u-umlaut was used in the "English"
Wade-Giles system!

> it could be preferred even to previous Chinese conventions such as
> Latinxua |y| (even despite of the political-ideological
> considerations).
Yes, they could have adopted |y|; but chose to use |y| for [j] and
u-umlaut for [y] - exactly as in Wade-Giles. But in actual fact it is bit
puzzling why the retain u-umlaut as it is used only after |l| and |n] and
Pinyin does allow the alternative spelling _lyu_ and _nyu_. One wonders
why they didn't simply adopt the alternative forms as standard.

>
>> By the 1950s the Chinese People's Republic could find |=C3=A5| =3D [y]
>> in many other places, including Turkey - but I am *not*
>> suggesting Turkish influence!
>
>   Turkish influence can be ruled out
Of course it can!

[snip]
>   (It is a further question whether there is a correlation between
> Turkish choice of |=FC| and |=F6|, and the orthographic traditions of
> the nearest nations with Latin script that have these sounds: i.e.
> Hungarian and German. In Hungarian they are definitely a German
> borrowing, thus I think Turkish |=FC| and |=F6| are also such.)
Yes - it was certainly German that introduced these conventions. But my
point is that by the middle of the 20th century they had ceased to be
specifically German.

[snip]
>> I think most (all?) agree that [E]is an allophone of the mid
>> central unrounded vowel [=C9ô] (CXS [@])
>
>   This |e^| =3D [E] interjection I mentioned is an interchangeable
> phonetic variant of interjections |ei| with various tonemes. In
> syllabical situation, you could be right here, however -- if I
> remember correctly -- [E] is _also_ an allophone for [a] e.g. in
> codas |-ian|, |-uan| (AFAIK the latter can be pronounced both as
> /yan/ and /yEn/. [[I use X-SAMPA]]).
According to my copy of the "Faculty of Peking [sic] University" course of
1971, the |a| in -uan is [&] (IPA ae ligature), i.e. [y&n], but it does
agree that -ian /iEn/.

>
> ----
>
>> It is interesting that the French system of Ecole fran=C3=A7ais
>> d'Extr=C3=AAme-Orient adopted the same sort of solution as Beila,
>> namely:
>> - ts, ts' and s represent palatals before [i] and [y] but
>> dentals elsewhere;
>> - k, k' and h represent palatals before [i] and [y] but velars
>> elsewhere.
>
>   According to my sources, the French system is the oldest one used
> recently. It reflects an older, dialectal pronunciation.
I could well do.

>
>> Indeed, it is even further complicated by the fact that the
>> retroflex series are not found either before [i] and [y], therefore
>> it could be argued that the palatals are allophones of the
>> retroflex affricates and fricative
>
>   Yes, however the situation is the same with alveolar non-plosives
> and velar non-plosives. Thus three possible underlying structures
> could have a common surface representation.
I agree.

> However alveolar and
> velar non-plosives form triples similarly to the palatal series,
> but there are four retroflex onsets. If we would bind palatal
> intials to retroflex ones, syllables with |r-| would be orphaned.
They would be so orphaned, but that didn't stop Gwoyeu Romatzyh doing just
that.

[snip]
>>> An interesting point is that Ricci IIRC used _b d g_ the same
>>> way PY does.
>>
>> That's very interesting. If you recall rightly, it makes the
>> German theory of the origin even less likely IMO. So the tradition
>> of using _b d g_ this way possibly goes back 4 centuries!
>
>   It is possible -- nor I excluded Italian influence on |z| erlier
> -- but I wonder whoever preserved this tradition for centuries.
I understood that the varying systems over the last 4 centuries more or
less overlapped one another. There seems to me nothing surprising in bits
from one system or another being preserved in successors.

> Chinese interest in romanization is fairly recent.
 From the 1890s in fact. A scheme by a Chinese scholar called W·ng Zh‡o
gained some official support in 1900.

> And it became
> extinct in Latin systems, except German.
Is there definite evidence that this was so? Are you saying both Gwoyeu
Romatzyh and Beifang Latinxua decided to adopt this scheme specifically
from the German system? The more I look into the question of Romanization,
  the more I am aware of schemes that existed of which I know practically
no details.
=================================================================

On Tuesday, September 7, 2004, at 04:07 , John Cowan wrote:

> Ray Brown scripsit:
>
>> But |ü| is used only after |l| and |n|. In all other instances PY writes
>> [y] simply as |u|.
>
> I think that's a straightforward simplification,
I agree.

> There is no [u]/[y] contrast except after [l] and [n], so no reason to
> use the diacritic and get involved with the annoyances of double
> diacritics.
Yep - and apparently Pinyin does allow _lyu_ and _nyu_ as alternative ways
of writing _lü_ and _nü_, so the umlaut could have been avoided entirely,
leaving only a single set diacritics to denote tone.

[snip]
>> It may be argued that historically all the other uses of |ü| = [y] are
>> ultimately derived from german practice - but that IMO is an entirely
>> different matter.
>
> I think the evidence is in fact clear that the Turkish vowels with
> diaeresis are derived directly, not merely indirectly, from German
> conventions.
That may well be - but they could, for example, have got the idea from
Hungarian. The point I was trying to make is that by the 20th century, the
use of the trema over back vowels to denote front rounded vowels was not
confined to German; it could be found in many other places.

I have always been struck by certainly resemblances between Turkish and
Volapük orthography (the use of the trema and the peculiar value of |c| =
[dZ] are obvious examples), that I had long wondered if Atatürk had known
Volapük. But it appears that these resemblances are just co-incidental.

> BTW, Turkish has an interesting convention that I didn't know about
> until recently: a circumflex accent on a back vowel means that the
> preceding consonant is palatalized (normally palatalization is
> a sub-phonemic concomitant of a following front vowel).
I had known that for some time; tho according to my ancient copy of
"Turkish Self-Taught and Grammar" (1940 edition - I bought it second-hand
in my late teens), the circumflex is used only over _a_, _i_ and _u_. The
book notes with î that it is really the dotted-i that is circumflex,
although the dot is not written when the circumflex is added. The book
says they denote long vowels and gives palatalization only after |k| and
|g|. It appears they occur (almost) exclusively in loan words from Arabic
and Persian.

But this is digressing a bit from Chinese Romanization.

>> The whole debate about the phonemic status of [E] and [o] is not settled
>> AFAIK. I think most (all?) agree that [E]is an allophone of the mid
>> central unrounded vowel [??] (CXS [@])
Oh - the IPA shwa has turned into two question marks! A curse on the
tyranny of ASCII.

> Except that there is a contrast in the case of these two interjections
> /E/ and /o/ vs. the ordinary words /ei/ and /uo/.   Interjections are
> often exempt from the ordinary rules of a language: consider the
> English interjections [t!-t!] and /S/.
They are indeed. If we include interjections, then English has clicks!

>> My wonder, indeed, is why the "apostrophe system" ever got into
>> Wade-Giles.
> Probably because it was felt to be more phonetically accurate
> (i.e. closer to the European uses of consonants).
Basically southern European use. As Philip Johnsson has pointed out, the
use in all Germanic langs (including English), with the exception of Dutch
& Afrikaans, would favor the _b d g_ ~ _p t k_ system. the same is true
BTW of the Celtic langs (tho AFAIK there has not been a Welsh or Irish
system of Chinese Romanization).

>> I always thought that the Gwoyeu Romatzyh idea of having tones built into
>> the spelling rather than denoted by diacritics was a good idea
>> (anglophone
>> Newspapers always ignore diacritics); but I confess I thought GR made it
>> too complicated. I would have done it more simply.
>
> Well, there's always the "append one to four v's to each syllable"
> tonal spelling.  :-)
There is and that was, as you know, the convention in the Wade-Giles
system. But what happened when Wade-Giles got printed in newspapers etc?
The apostrophes, trema and tone-numbers were all regularly dropped. In
fact the "curtailed WG" that was more commonly used was simply awful; for
example, _chu_ could denote Pinyin _chu_, _zhu_, _qu_ or _ju_ - each in
any pf the four tones.

I know the fault is with anglophone publications, but the fact is that
such publications habitually omit the tone diacritics when printing Pinyin
and, in view of their past treatment of WG, would be just as likely to
omit 'tone numbers' had Pinyin used those. In the GR system the tone
indication was built into the spelling.

> But after much playing around with it about three years ago, I found a
> very simple scheme devised by Lon Diehl and promulgated on the Net by
> Mark Bosley, which I just loooove:
>
> tone 1: tang
> tone 2: ttang
> tone 3: taang
> tone 4: tahng
Yep - much simpler than the GR system - but not quite the way I would have
done it   :-)

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com
===============================================
"They are evidently confusing science with technology."
UMBERTO ECO				September, 2004