Retroflex Consonants in P.I.E.

--- In cybalist@yahoogroups.com, "Patrick Ryan" <proto-language@...>
wrote:

> 3) I assumed that Old Indian had three classes of cerebrals:
>
> a) those derived from contact with Dravidian;
> b) those derived from RUKI;
> c) and those which might have been inherited from PIE.

There are at least two other classes:
d) /n./ from a preceding /r/ in the same word.
e) The result of /l.d/ > /ud./, e.g. g^Hl.du- > hud.u- 'ram'

> 1) That retroflexion has been retained in Old Indian where RUKI

predicts it is, not of itself, completely convincing. Let us bear in
mind that RUKI was formulated to account for actual retroflexion not
theoretical retroflexion. So we would expect to see retroflexion where
RUKI "predicts" (really "observes") it. Could you agree that there is
a certain circularity here.

I wouldn't agree, for the rule also works for Balto-Slavonic. The
RUKI environment is one that excludes Sanskrit /s/.

> a)) All that notwithstanding, if you can confidently assert

that there is no observable tendency to

> eliminate retroflex articulation where it is observed in

earlier stages of Old Indian, based on your greater

> familiarity with Indian matters generally, I am prepared

to give up the idea that retroflexion is perceived

> as a marker of low register, and that some attempts to

limit or suppress have been made. That was a

> purely theoretical speculation, not based on any solid data.

Actually, /s./ and /n./ were not particularly stable! Some have
attributed the loss of /n/ v. /n./ distinction to Dravidian, on the
basis that total confusion arose because Indic had a dental v.
retroflex contrast, while Dravidian had a dental v. alveolar v.
retroflex contrast. A Dravidian would hear Indic /n/ as /n_/
(alveolar), and an Indic-speaker would hear Dravidian /n_/ as retroflex!

/s'/, /s./ and /s/ merge in Pali.

Richard.