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

> Kharosthi was pretty much out of use by the time Sanskrit came to be
written with any sort of regularity,

Maybe, but Andrew Glass writes in his thesis, 'The sign for OIA
visarga consists of two dots placed above the character, e.g. taḥ
[glyph] (Rapson 81). It is found in the few manuscripts written in
Sanskritized Gāndhārī or Sanskrit, notably Niya documents nos. 523 and
661, the tablet from Endere, and the manuscript fragments in the
Pelliot collection and Schøyen scribe 2.' WWS p377 also mentions that
there are 'a few specimens of Sanskrit written in Kharoshthi'.

> and Tocharian was usually written with Brahmi.

And the other script used was?

> A quick glance through Salomon 1999 shows only clusters with r and
y, plus one kS.

WWS p376 also gives <tv>, <rv>, <st> and <sp>. Transliterated
clusters in <v> can also be seen in a transcription of the Gāndhārī
Dharmapada available at
<http://titus.uni-frankfurt.de/texte/etcs/ind/mind/gandhpkt/dhpgpkt/dhpgp.htm>.
The first verse (unnumbered) has the word 'postaka' in 'halfverse' c.
'Salomon 1986:6' (I can't interpret that reference in Glass)
reportedly also has an example of 'spa'. While there may have been
some debate about the interpretation of 'sta', such clusters survive
in the word for '8' in several Dardic languages, including Kashmiri.

> Pali isn't Prakrit, and I believe it's much later than Kharosthi.

Kharoshthi suvived at least into the 4th Century AD in Khotan. Pali
was already being written in Brahmi-derived scripts by then, even if
you exclude Ashoka's edicts.

I agree that Pali isn't normally considered Prakrit, but some do
consider it as such. How do you reckon the language of Ashoka's edicts?

Now it does seem that CVC units were rare in Kharoshthi. If they were
not an internal development, what were they based on? Are they an
infrequent but 'natural' development in Brahmic scripts? According to
WWS p479, Javanese has four marks for final consonants, basically
anusvara, visarga, the repha form of ra, and a specialised form of
subscript ya. (Balinese seems to lack the latter usage - can anyone
here comment on that apparent difference?)

Richard.