Origin of retroflexion in IA

[Vidhyanath Rao]

George Thompson <thompson at JLC.NET> wrote:
> But I am not attempting to figure out the origin of retroflexion in all of
> IA. I am trying to figure out the origin of retroflexion in a specific
> text, the RV.

I am afraid that I don't understand this. Sound changes must affect
dialects/langauges, not texts, right?

> There is none in Avestan, there is some in RV, there is more
> in later Vedic, and still more in Classical Sanskrit.

But in RV, retroflexion occurs in phrases, across word boundaries.
Why does this not become more common.

> I assume that Dravidian is a factor in this development, but I do not know
> how to characterize its role [convergence? substratum? an open question].
> Remember, the RV is an orally composed and orally transmitted text.
> Therefore I think it is reasonable to take Deshpande's theory seriously.

Only if the phenomena I mentioned can be explained: Preservation of
aspiration, loss of retroflex sibilants in MIA that seems to go with the
lack of retroflex sibilants in Dravidian, presence of alveolar vs
retroflex contrast for stops and nasals in proto-Dravidian.


My problem is that ``substratum influence'' has become a deux ex machina (sp?)
in Indology. If we see any change, then the immediate reaction is to
say ``substratum influence'', without looking for possibility of
internal evolution.

---

BTW, as long as I limit myself to words of IE origin that have undergone
only the >regular< sound changes to Sanskrit, I find it very hard to come
up with examples of contrasting pairs that exhibit dentals and retroflexes
as different phonemes. Of course, if we go to MIA, it is easy. Or if we
take unexplained retroflexion, as in .sa.s/.sa.t. But otherwise,
only laryngeals and voiced sibilants become zero in Sanskrit. I don't
see how to create contrasting pairs just with those. The difficulty of
finding contrasting pairs in RV may be due to the relative rarity of
backformations/borrowing from MIA, and loan words with retroflexes.

----

Regards
-Nath