[Fwd: Gallows bit as a tone marker?]

[Brian Eric Farnell <bfarnell@xxxxxxx>]

See if this makes any sense, it ties the gallows characters
occurrences or lack thereof with the tables characters as well
as their physical proximity:

You would assume if the writer had created his own transcription
alphabet that he would have the freedom to create a character
for whatever vowel sound he needed, but in a Mandarin
transliteration there would be some places where it would be
logical to use two letters to represent a vowel sound far
example, these are the vowel sounds in Mandarin as they are
represented in PinYin:
a, ao
e, ei
i, ia, iu, iao
o, ou
u, ui, ua, ue, uo
u umlaut which is just e and u put together

All of these could be represented by one letter each, BUT, the
combinations that begin with i or u are special in that they add
an English 'y' or 'w' sound to the vowel compound as an accident
of phonating their basic sound in combination with other
vowels.  Anyone whose pronunciation was corrected while saying
one of these 'i' or 'u' compounds would become aware that the y
or w was in fact just a slurring of the i and u vowel sounds. 
It would be logical to him to keep all compounds beginning with
an 'i' or a 'u' as at least two letters, possibly representing
the compound in a special way like a tables character.  That
would seem to be consistent of the split of tokens with tables
or without.  It indicates that it is an important aspect, but
not necessary to make a word, if tables characters represented
this type of vowel blend, then it would indeed be very common
but not necessary. Since it's logical to mark tones around the
vowel, then on the model of the gallows as tone marks, you would
expect to find gallows and tables together where they both
occurred in the same token.  I know that now I am really
stretching here, with nothing more than a hunch but the model is
very consistent with the numbers you are generating.  Also for
whatever it's worth, in Mandarin at least, those 'i' and 'u'
combos don't occur at the beginning of words, they must be
preceded by an initial consonant (at least I can't think of
any).  This is not definitive because the 'i' compounds do show
up in a way, but the sound of the 'i' is so different that it is
written is a different letter in modern Romanization methods
when it occurs as an initial and is treated as a consonant
instead of a vowel.  This is probably a natural result of their
phonation being difficult without an initial and/or the possible
confusion of those sounds trying to bond with the final of the
last word which is a huge no-no in a sound poor language,
syllables are not blurred together like we do.  I would assume
sounds in other languages analogous to those i and u compounds
would probably follow the same phonotactic rules.  Any thoughts?
Regards,
Brian



Jacques Guy wrote:
> 
> Brian Eric Farnell wrote:
> >
> > I'm very interested in your thoughts on this data about the
> > constructs of tonal languages as it might relate to the VMS.
> 
> We're back into Jorge Stolfi's hypothesis.
> 
> > Anyway, point is that every syllable in
> > Mandarin has three choices, it can end in a vowel, an 'n' or an
> > 'ng'.
> 
> And in some dialects, final 'n' and 'ng' are no longer
> distinguished. And Shanghaiese has even lost final
> 'n' and 'ng'
> 
> >  Does
> > this sort of extreme phonotactic limitation remind you of the
> > aiin problem?
> 
> As if we needed being reminded *sigh*
> 
> Gallows as tones? Could be. And then you have tone
> sandhi to make things worse.
> 
> We've come around full circle again it seems.