VMs: Chinese thoughts [ was: languages etc]

["Barbara Barrett" <barbarabarrett@xxxxxxxxxxxxxxx>]

This eventually gets around to the vms; trust me ;-) BB

> Bruce Babbled;
> I'm not sure how much analyzing romanized Chinese text will tell you.
> For example, it seems like analyzing the same text in different
> romanization schemes (say pinyin and Wade-Giles) would give you
> radically different results.

Barbara Blithers;
All romanization scheames for chinese use different tactics to cope with
that language's use of aspirated and unaspirated devoiced consonants.
Pinyin, for example, uses europe's voiced consonants and surplus letters
keeping a one-grapheme-per-phoneme ratio with the exceptions of ch and zh,
but Wade-Giles uses digraphs and even trigraphs (eg;  _ch'_  for _q_) also
Wade-Giles digraphs changed value according to the following vowel (eg; WG's
"ch" = P's ch, j, or zh). The Yale system is actually the most accurate
system, but ugly because of its use of superscript numbers.

> Also, romanization would obscure the fact that Chinese has only a small
> number of syllables (about 500 I think) into each of which many
> different characters are mapped.

Under 500, about 450 last time I saw a chart. But don't forget the five
tones (so there's really over 2,000 syllables) and the huge number of
homophones! You'd think that a syllable would have only 5 meanings but
homophones increase this dramatically; MA for example, has over a dozen
semantic values!  In romanization the homophones are clarified by context
IF the tone is indicated (Pinyin uses diacritics, Yale uses numbers,
Wade-Giles has complex spelling rules).

Which brings us neatly to Voynichese, which in my opinion does not have
enough symbols to represent syllables+tones, or phonemes+tones. Just for the
basic vowels (ignoring the allophonic variants) and diphthongs there'd need
to be 12 symbols that had 5 variants with a further 30 symbols for the
consonants (IE 65 symbols, although this could be cut to 50 or so for
positional allophones).

Voynichese has possibly 3 basic symbols that have variants (e/b/s/y,
i/l/n/r, and ch + plume/arrow/teardrop diacritics) plus the gallows set
p/t/f/k. One could argue that  the i/ii/iii + r/l/m/n sets might be toned
vowels and while that would give us 7 vowels (digraphs of which *could*
provide the other 5) they're still a tone short,(I admit a vowel+consonant
digraph could act for the 5th tone) and these 7 vowels + 4 variants (28
graphemes) once subtracted from the character set would leave us hopelessly
short for rendering Chinese consonants without resorting to digraphs and
trigraphs. (and before someone says "chinese has only 4 tones; that's 4 plus
the "level" vowel - making 5 tones in all).

I know there's a great number of wierdos, but they do not appear often
enough to represent regularly used consonants and/or toned vowels.
Furthermore, if say one of the i/ii/iii sets were vowels we'd expect their
distribution to be fairly even within "paragraph/sentences" but they seem to
be mostly line-final.

So from the above I conclude that the voynich character set seems not to
have the complexity to render Chinese phonemically.

However, (Oh I love building up an argument then presenting its major flaw
<g>) this shortfall could be compensated by complex spelling rules (such as
in Mongolian script, where a grapheme can have up to 6 values depending upon
preceding/following graphemes and the gender of the word); in which case
every single voynich "word" would in fact be a single phoneme! Phonemes
would be run together only when the rules made the individual phonemes
non-confuseable. This could account for the high levels of repetition in
voynichese.

The complexity of the spelling rules of some of the world's writing systems
is truly astonishing, but they can be kept in ones head, so if the vms used
complex spelling rules then it is possible it could be written and read
without keys or references by anyone who understood the language and the
orthography!

Personally, I'm not of the "voynichese is Chinese" camp, for many reasons,
but I think I've presented the only orthographic scheme in which that could
be workable.

Barbara





______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list