Re: [tied] Re: Nominative: A hybrid view

On Tue, 3 Jun 2003, Glen Gordon wrote:

>
> Jens:
> >That just is not fair: You are asking for a catalogue of thousands upon
> >thousands of forms to demonstrate that one can get through the wordforms of
> >Sanskrit without ever opposing two vowels to each other.
>
> What's not fair is explaining everything with "shadow consonants". What
> stops us from using the same arguement on English, or Farsi, or Hakka?
> Any language can be monovocalic then.

If you posit enough, yes. It can be reduced to a single element if you
like. In the case at hand, however, it keeps within so narrow limits as to
be of typological importance. It is mostly not noticed by IE-ists that
there are marginal cases of deviation from the basic syllabification
rules, such as 3pl *yungenti 'they combine' with /yun-/ taken over from sg
forms like 3sg *yunegti where it is regular before a vowel. Nor, I
believe, is it generally known that it only costs the very infrequent use
of a single silent consonant (only one!) to allow Sanskrit to present all
its actual wordforms unambiguously by a notation using only the one vowel
/a/. That in fact is already more than what many typologists say is
excluded.

> >
> >Actually that is all the extreme theory says, and it is in my view quite
> >close to being correct, at least so close as to create the need for an
> >explanation why the distribution of vowels in PIE roots is so drastically
> >uneven.
>
> I think something along the lines of my own explanation is reasonable.
> It seems that *a is largely the result of lowered *e (and if my labial
> theory is correct, from *o as well). So if *a can be deduced to be
> originally an allophone of the other two, we are left with *e and *o.
> Yet this is an unbalanced system. Therefore, without introducing
> other assumptive vowels, the system must derive from a centralized
> one of *@ (*>e) and *a (>*o). We know that *e and *o oppose
> each other in IE because of roots like *reg- "ruler" versus *pod-
> "foot" so this is why I don't understand the point the monovocalists
> are trying to make.

I do believe I understand them, I just don't think the analysis is
correct. It is hard to specify the underlying vowel of *H3re:g^-s since it
never alternates - if that is due to levelling or potential allomorphs'
failure to survive, there is no telling that this had a vocalism
significantly different from that of 'foot'. The roots concerned were
*H3reg^- and *ped-, but perhaps the paradigms of the root nouns differed
as having short and long e respectively (the latter of which could be
written /ee/ if one wanted to).


Jens