Re: [tied] Re: Indo-Iranian Vowel Collapse (was: IIr 2nd Palatalisa

On Thu, 17 Nov 2005 14:39:08 +0000, Rob
<magwich78@...> wrote:

>I know that Tamil has a 5-vowel system, but it seems that it could
>have rather transparently derived from an earlier 3-vowel system in a
>manner similar to Sanskrit: /ai/ > /e/, /au/ > /o/.

Sankrit <e> and <o> are long vowels. Tamil distinguishes
/e/ and /e:/, /o/ and /o:/. No diphthongs are reconstructed
for Proto-Dravidian, and most Dravidian languages still
don't have them.

>> A third point is that the merger of short /a/ and short /o/ is
>> pretty widespread. It not only occurs in a once-contiguous swathe
>> from Indo-Iranian to Germanic, but also in Hittite. The unusual
>> feature is the further merger of *e with *a and *o.
>
>Is there a way to map the merger of short /a/ and short /o/ as an
>isogloss?

Hard to say. Indo-Iranian doesn't really merge /a/ and /o/
(the distinction is maintained in open syllables), and there
is no merger in Proto-Anatolian (Lycian keeps */a/ and */o/
distinct).

>Furthermore, how can we determine whether the apparent
>merger of short /a/ and short /o/ was indeed a merger or whether the
>separation of /a/ and /o/ (Italic, Greek, Celtic?)

Armenian, Tocharian, Lycian, Indo-Iranian.

>was an innovation?

It was a merger because there are no conditions for a split.

=======================
Miguel Carrasquer Vidal
mcv@...