Retroflex sounds

Bh.Krishnamurti bhk at HD1.VSNL.NET.IN
Sun Jun 21 18:08:54 UTC 1998

At 11:04 21/06/98 -0400, you wrote:
>The Australian example has been mentioned. Well, unfortunately, their
>retroflexes are not that old either:
>A. Capell, The history of Australian languages: A first approach,  in:
>S.A. Wurm, Australian linguistic studies, Australian National University
>1979,  p. 435, 437 regards them as secondary, derived from  *r+ cons.
>(just as some presuppose for parallel developments in IA and Dravidian).
>NB could the Drav. scholars on the list please provide some recent
>information on this problem?
>Bh.Krishnamurti wrote (june 21):
>.>>>.. Dravidian had both *_t and *.t basically in
>a large no of morphmes. It is true that in sandhi also _t and .t develop
>from l+t, and .l+t. <<<
>You you please elaborate on this and the preceding question? What do Drav.
>specialists think, these days, about the age of this sandhi effect and
>about the age of Drav. retroflexes in general?
>Michael Witzel                       witzel at

I forgot to mention. Drav retroflexes are as old as PD, maybe 3000 BCE. The
PD phonemes are: p t _t .t c k, m n .n ~n, l .l, r .z, w y H (a laryngeal)=
17 consonants; i e a o u (long and short)= 10 vowels. Alveolars and
retroflexes do not occur word initially. This distributional characteristic
is partly reflected in IA also, e.g no initial .n, .l, and .s. Through sound
change, initial  *_t (=*_r) .t, .d, l, *.l, *.z occur word initially in
South Central Drav languages. PD has contrastive minimal pairs for:

Initial: t-
Medial: t : _t : .t
also:   tt : _t_t; .t.t
post-nasal : nt: n_t: .n.t
             ntt: n_t_t: .n.t.t
Finally : no obstruent occurs; only sonorants occur

Alveolar and retroflex stops have the same distribution as the other
obstruents except that they do not occur word-initially.

There is no evidence to derive the alveolars and retroflexes as secondary
H.No. 12-13-1233, "Bhaarati"
Street 9, Tarnaka
Hyderabad 500017

More information about the INDOLOGY mailing list