>> So you do find it possible that at a period there was only one phonemic
>> vowel in PIE?
>
>Yes. The result would be as in Sanskrit where /a/ has precisely that
>status.
I am sure you don't mean this the way it looks. Sanskrit /a/ cannot come
from a single phonemic vowel in PIE. We know that the vowel we reconstruct
as PIE */e/ caused palatalisation in Sanskrit, while the vowel we
reconstruct as PIE */o/ did not, and that the vowel we reconstruct as PIE
*/o/ is reflected as a long /a:/ in an open syllable, whereas the vowel we
reconstruct as PIE */e/ is not. This is an internal matter within
Indo-Iranian, and cannot be taken to PIE. Indo-Iranian inherited */e a o/
precisely where Greek and other IE languages demand we reconstruct these
different vowels.
What you must have meant is that a period with only one phonemic vowel was
followed by a later period of PIE where at least 5 vowels can be found.
Peter