Fortunatov's Law and tolkAppiyar's rules

[Vidhyanath Rao]

I have finally located the book I had in mind. It is V. Labov
``Principles of linguistic change; vol. 1, internal factors'',
Blackwell, Oxford UK & Cambridge USA, 1994. The general principle
(Herzog's P.) that mergers tend to expand at the expense of
distinctions is discussed in chapter 11, pp. 313--331.

I switched the page references to CDP, due to my sloppy notes. I
apologize for any confusion this may have caused.

--
Sudalaimuthu Palaniappan <Palaniappa at AOL.COM> wrote:

>Having been familiar with the scientific method, I am yet to see a clear
>explanation of why using Pittsburgh dialect situation to draw conclusions
>about first millennium BC Sanskrit-Dravidian convergence on nasals should be
>deemed qualitatively different from using Malayalam to draw conclusions about
>the fate of stops after l/L.

This is hardly the place to discuss the Uniformatarian Principle in
general or its use in linguistics since the Neogrammarians. Labov has a
good discussion of the whole thing, with further references, to which I
refer those not already familiar with the ideas.

>As for Zvelebil's views, they are wrong and outdated.

Both Zvelebil in his `Inroduction to Dravidian lingusitics' and
DEDR, written well after 1972, maintain that n and _n are allophones
for Dravidian as a whole. Knowing nothing about North and Central
Dravidian or non-literary languages of South, I defer to them.

>Kuiper derives Skt. kavaSa (straddle-legged) from Dravidian related to
>Ta. kavar, kavalai, and kavaTu (fork of legs)  in Aryans in Rig Veda,
>p.26.  While Kuiper notes the absence of the form kavaz in Dravidian, if
>one takes into account Stephen Levitt's findings, one can see how a
>kaval >*kavaz/*kavaL which can lead to kavaSa in Sanskrit and with a
>dental suffix, one can get kavaTu in Tamil.

There are too many assumptions here for my taste. In any case, kava.sa
does not mean straddle-legged. Perhaps kava.saa+uuruu does, but that is
not clear. kava.sya.h occurs as an adjective to dura.h in MS 3.11.1 and
3, KS 38.6 and 8, VS 20.40 etc, where it seems to mean open.

Can somebody with access to the new Mayrhofer tell us what he has about
this? [I have access only to KEWA.]

This weak evidence must be set against the following two facts: One, zh
does not seem to retroflex a following t in Dravidian. [it certainly
does not in Tamil.] Second, and more serious, is that in MIA, starting
with Pali, all sibilants merge into s while in syntax we see greater
convergence with Dravidian, and greater retroflexion in stops. I am yet
see any explanation of this (other than what I proposed).

---
>As for the Hindutva folks, their arguments about Indian pre-history are
>marked by appalling ignorance about the Dravidian side. Of course, they
>are a lot more sure of their positions than scholars like Deshpande.

I guess there is a modified version of Godwin's law for Indology: If any
discussion lasts long enough, somebody will mention Hindutva/Parivar etc.
I thought that we successfully avoided that. Sigh.