--- In qalam@yahoogroups.com, "Peter T. Daniels" <grammatim@...> wrote:

> Richard Wordingham wrote:
> >
> > --- In qalam@yahoogroups.com, "Peter T. Daniels" <grammatim@...>
> > wrote:
> > > suzmccarth wrote:
> >
> > > > now generally accepted that an akshara is a syllable. So this is
> > > > the feature that defines the script type for its users.

> > > > The akshara connotes wholeness I believe. Since the syllable

can be

> > > > pronounced, it can be matched up to the visible syllable.

> > > An akshara comprises all the C's in up to a CCCCCV sequence, and
> > > all those C's are not part of the same syllable.

> > While this is true for Sanskrit, how true is it of the vocabulary
> > that has not been (re-)introduced from Sanskrit? In Pali for
> > example, the only medial clusters different to initial clusters that
> > I can think of in un-Sanskritised Pali are the geminates and those of
> > the form NC. The latter can be simplified by using niggahita (=
> > anusvara). I think I read that geminated were degeminated in Middle
> > Indic, so how bad is the mismatch in ordinary words in modern
> > languages?

> > Am I correct in thinking that Brahmi originally did not show
> > geminates?

> Brahmi was devised for Prakrits, not for Sanskrit, and closed syllables
> are not common there; the virama was a later invention because there
> were no consonant-final words.

I'll make do with accessible information. I found an analysis of
_Sanskrit_ clusters at http://www.tphta.ws/TPH_SSCS.HTM . After
applying the assimilation/reduction rules of Pali, the clusters reduce
to the following forms:

syllable onset
nasal + syllable onset
duplicate of next syllable + syllable onset (e.g. -tty-)
-nm- (apparently non-existent in Pali!)
-ñm- (apparently non-existent in Pali!)

The 'duplicate' of an aspirated consonant is the corresponding
unaspirated consonant.

There might be some doubt about the identifcation of initial syllable
onsets - for example, while kry- and ghn- occur word initially in
Sanskrit, gny- does not. However, the list derived looks entirely
reasonable. (I do realise that the syllable onsets in Pali are
simpler than in Sanskrit; however, I have little reason to expect
anomalous behaviour invalidating my reasoning.)

The assimilation/reduction rules I applied were:

1) /l/ or /r/ before another consonants becomes the same, e.g.
Sanskirt _dharma_ but Pali _dhamma_ and Sanskrit _silpa_ but Pali
_sippa_ 'art, craft'.

2) Plosives assimilate to a following different plosive (disregarding
aspiration), e.g. Sanskrit _sapta_ but Pali _satta_ '7'. (A search of
an on-line Pali dictionary -
http://dsal.uchicago.edu/dictionaries/pali/index.html - found no
examples of consecutive but different voiceless plosives.)

3) s.t. > t.t.h

The rare clusters in -st- are a little discomforting, for Pali lacks
initial st-. However, with this one exception, one can, by the
consistent use of anusvara for final nasals, ensure that all Pali
aksharas with vowels are possible *phonetic* syllables, or different
from a phonetic syllable only by starting with a geminate. In this
context, it is worth noting that:

a) Latin used to write double consonants as single consonants
b) Japanese kana can represent geminate initial consonants by means of
the appropriate diacritic. (That is the basis for writing <pa> as
<ha> plus a diacritic).

It thus appears that at one time the orthographic syllables *did*
correspond well enough to the phonetic syllables.

Richard.