"syllabicity"

[Jens Elmegaard Rasmussen]

On Tue, 18 May 1999, Rich Alderson wrote:

[On my suggested one-vowel system of Sanskrit:]

> One can just as easily argue that [y] is a realization of /i/, [w] realizes
> /u/, and so on, when the vocalic phoneme is collocated with another vocalic
> phoneme, if one wants to accept this sort of deep-level phonology. [...]

Sure, but the point is that such an alternative is not demanded by
anything, i.e. one can do it one way or the other. The really
noteworthy thing about Sanskrit is that, if one does it this way (namely
by deriving a maximum of syllabicity by rule), this exhausts the system,
except for one single element, /a/, which is always a vowel. This is only
marginally affected by the erratic behaviour of Sievers' "rule"
(especially since competing variants are lexically equivalent) and by
loanwords that have only been integrated in a late (Middle Indic) period
in occasional disagreement with the rules. Even including these fringes,
it gives _exactly_ the picture of a one-vowel language of the style
suggested for PIE by Benveniste and others. But, unlike PIE, the Sanskrit
1-vowel system is very nearly true, since Sanskrit simply has no other
elements than /a/ that cannot be called a consonant, while IE has /e/ and
/o/, and in my opinion also /a/. Also for PIE, the full predictability has
been compromised a bit by analogical levelling, one of the best cases
being the weak forms of nasal presents: From *yw-ne'-g-t, the 3sg
*yune'g-ti is fine, but in the 3pl one expects *yw-n-g-e'nt to
yield **iwNge'nti and not the actual *yunge'nti pointed to by Skt.
yunja'nti, Lat. jungunt; in this paradigm *yun- has simply been carried
through from the sg. where it was regular before the following vowel.

   This is only meant as a refutation of one particular objection always
raised against the "classical" (in my opinions erroneous) analysis of IE
as having only one necessarily syllabic (morpho)phoneme, viz. /e/, this
objection being that the world knows of no other languages of such a
structure. Well, for what it is worth, there IS a language of _exactly_
that structure, namely Sanskrit. I realize that many now don't want
Sanskrit to exist, or, failing that, not to be given this analysis, which
they are at liberty to avoid. But nobody would be right in saying
that it could not theoretically be done to Sanskrit to the same extent as
it was (and sometimes still is) attempted for PIE.

Jens

[ Moderator's response:
  All right, I see the point you were making, and concede that under the same
  kind of phonological theory that led to claims that PIE, or Kabardian, or
  Ubykh, were monovocalic, Sanskrit too is monovocalic.  I understood you to be
  asking that this analysis be accepted as *correct*, rather than that it be
  taken as a strawman against which to argue.  Thank you for the clarification.
  --rma ]