On Thu, 30 Jan 2003 21:52:12 -0000, "deshpandem <
mmdesh@...>"
<
mmdesh@...> wrote:
> I am wondering if there are signs of change in the nature of
>Sanskrit accent in the following way. Alternations like asti versus
>santi, and naumi versus nuva.h are linked to shifts of accent from
>the root to the suffix. If these sorts of shifts indicate that an
>unstressed vowel gets contracted, or at worst deleted, can one infer
>that the accent of Sanskrit in its formative stages was stress
>accent, rather than pitch accent as it get represented in Vedic
>traditions? Would pitch accent cause the same sort of contractions
>of vowels?
Not to my knowledge. Shortening, reduction or deletion of unaccented
vowels is surely the result of stress differences between unaccented
and accented syllables. That being said, stress accent and pitch
accent are not mutually exclusive: besides the possibility that PIE
went through a stress-accent stage first, then through a pitch-accent
stage, and then back to stress accent in many of its branches, we can
also imagine a scenario where PIE had had *both* types of accent
initially, and that some languages (e.g. Sanskrit) lost the stress,
while others (e.g. Latin) lost the pitch.
By the way, the alternation ásti ~ sánti (reconstructed PIE *h1és-ti,
*h1s-énti) can be accounted for by stress alone (**h1és-t(i) ~
*h1es-ént(i)), but in the case of náumi ~ nuváh. we need another
factor to account for the length (vr.ddhi) in the stressed member.
I'm not sure what it is: is the present tense of /nu/ a Narten
present?
>My second question is this. While the correlation of
>vowel contractions with accent shifts is visible in verb forms and
>certain nominal paradigms, why is it that there is no similar effect
>left in the formation of Sanskrit compounds? Consider the accent
>difference between a Tatpuru.sa versus a Bahuvriihi. There are no
>vowel alternations similar to naumi versus nuva.h between Tatpuru.sa
>and Bahuvriihi. Is it likely that the nature of accent changed from
>the stage when forms like naumi/nuva.h originated to the stage when
>compounds emerged?
Certainly. By the time these compounds formed, Sanskrit was already a
pitch-accent language, so the stress difference (tatpurus.a's accented
on the final member, bahuvri:hi's usually on the first) did not have a
great impact on the phonetic shape of the unaccented part, except
perhaps for the loss of -n in bahuvri:his with n-stems as the last
element (vis'vá-karma or vis'vá-karman), the reduction of *-o:(y) to
-a in compounds made with sakha: (kava:-sakhá [note the accent,
however]), and -a > -i in dhu:má-gandhi (gandha-) and a some others.
=======================
Miguel Carrasquer Vidal
mcv@...