Sandhi in `real' Sanskrit vs `conversational' Sanskrit

[Vidhyanath Rao]

I am curious about the use of sandhi in `real' spoken Sanskrit:
In print, it is customary to apply sandhi except at sentence
junctures. Do people pronounce whole sentences in one breath?
If not, do sandhi rules still apply when they pause to take a
breath? What about long compounds with twenty or more syllables?

My gut feeling, but one which I cannot check objectively, is that
it was easier to resolve sandhi when I heard the text being read
by someone familiar with it, than when reading printed texts.
I wonder if others have had the feeling. What does this tell us
about aural clues to word breaks?

We all have read about the `errors of sandhi' in manuscripts.
What is the type of error they are talking about? Are the sandhi
results distorted by Prakritic phonology? If is it a case of
not applying sandhi? Has anyone looked for patterns in these
`errors'?

All this talk about rigorous application of sandhi rules reminds me
of two essays on this topic. The first is in the series editor's
introduction to one of the HSS volumes published in the 1920's.
[It might be the tantraakhyaana.] The other is V. L. Joshi's
``Paa.nini and Paa.niniiyas on sa.mhitaa'' in the proceedings of
the Delhi Congress of 1964. I would be curious to know how
members of this list react(ed) to these two articles. In particular,
does `real' spoken Sanskrit have pauses of varying lengths?
Or is it punctuated like `real' printed Sanskrit: only da.n.das,
with commas etc. for the weak-minded only? To give a specific example,
if I want to speak the following in `real' Sanskrit, am I supposed to
run through the whole thing in one breath, or can I pause? If I can
pause, where should I pause? Only where there is a space in the usual
Devanagari printing, or should I be guided by synctactic breaks? What
if syntactic breaks are obscured by sandhi, as in `satiidam' in this
example?
        tarhi mamaanena ka.n.tharaktena t.rptaa
        satiida.mta.taaka.mjalai.h paripuur.na.mkuru

To return to a point I raised earlier: Oral communication often
contains information that is hard to communicate in writing.
Western practice has been to try and convey this by other means.
In particular, indicating word divisions and use of punctuation
were introduced and developed for this purpose. In India, teaching
has been soley oral, with books serving as only aids to memory.
But in the modern world, written communication continues to be
important. Just listen to all the lamentations about the (presumed)
deterioration in the ability to punctuate, to write clearly etc.
If Sanskrit is to serve a medium of modern communication, how can
we insist that Sanskrit must continue to do without such aids?

Isn't it the case that in Pali and Prakrits, sandhi is applied only
to standing phrases or very closely connected words? Isn't the same
the case in living languages?

Sanskrit learners face a chicken-and-egg situation: they need am
extensive vocabulary to resolve sandhi; but to get such a vocabulary,
the best way would be to read lot, to do which they need to know how
to resolve sandhi. Why not break this cycle by ignoring sandhi
in printing (as in done in >all< languages except Sanskrit)
and leaving sandhi in spoken language to be applied within short
phrases, with the boundaries determined by the speaker? If that
makes Sanskrit sound more natural and easier to follow, why is it
so bad?

To return to my original question, who owns Sanskrit? That is, who
decides what is `real' Sanskrit? Why is it `real' Sanskrit even
when intonation patterns that peek out from A.s.taadhyaayi are not
followed [for example, ``ki.m kriyaapra"sne ...'' suggests that
it was ``kim, ga'cchasiii'' for ``Are you going'' (literally
``What, you are going?'') opposed to ``ki"n gacchasiii'' for
``Whither are you going?''], but it is no longer Sanskrit when
sandhi is not applied blindly paying no attention to natural
pauses? Why is it `real' Sanskrit when we write ``ki.m gacchasi''
even though the pronunciation is mostly ``ki"n gacchasi''?



-Nath