>> >The Sanskrit system could be and inmho IS the original.
This was debated over many decades when IE linguistics got going 150 years
ago. The arguments that convinced linguists then are still as strong
today. It is just not possible the Sanskrit is the origin of all other IE
languages.
It's always fun, though, to rehearse the arguments. Here's a starter (and
I'm sure there are better arguments):
(a) Some changes are impossible (or extremely unlikely), e.g. :
In the usual analysis, Sanskrit has lost some distinctions (e.g. k/kW and
e/a/o). So seeing it as the original language means a large number of
phoneme splits - where a phoneme becomes one thing in some words, but
another in others, without any logic or rhyme or reason. Such things do not
happen in any of the languages we have been able to observe historically.
(b) Some changes are much easier to explain in the usual system, e.g. :
Sanskrit shows irregular patterns such as perfect reduplication tat- but
cak- (not kak-); dad-, but jag- (not gag-). The usual explanation is that
Sanskrit was originally regular (tet-, kek-, ded-, geg-), but velar
consonants were palatalised before the -e- ( a very common occurrence in
languages). If Sanskrit were original, then you have to find a way of
getting from dad-, cak- etc to the regular ded- kek- we find elsewhere.
Or a second example: causatives normally have a lengthend -a- in the
root vowel (eg ka:rayati) but some verbs do not (eg janayati). The usual
system suggests that most of these are verbs where there was a laryngeal in
the root, which meant the root syllable ws closed, not open (g'onHayati).
The laryngeal shows up in other IE languages in the various reflexes. Can
we really suggest that a long vowel + cponsonant turned into a short vowel +
consonant + laryngeal? That seems a much less likely phonetic change.
(c) Some patterns reconstructed to explain Sanskrit are actually found
elsewhere, e.g. :
Perfect reduplication as above, or laryngeal lengthening (-h- preserved in
Hittite), or laryngeals blocking the lengthening of an original -o-.
To me it seems blindingly obvious that Sanksrit perfect reduplication had to
be originally regular (tet-, kek-, ded-, geg-), but the moment you admit
that, you are saying that it is not Sanskrit that was the origianl language,
but some ancestor of Sanskrit, and you have conceded the entire case.
Peter