Inscriptions and Dravidian sound changes "y" > "c" and "y" >

[Palaniappa]

In a message dated 98-05-25 03:24:12 EDT, bhk at HD1.VSNL.NET.IN writes:

<<That is exactly the time, medieval period, for hyperstandard forms to arise.
The old change of s>*h>y (in other positions) >0 (initially) was known to
writers through loanwords like a:yiram, e:Ni, avai,etc. An underlying y
forms were written with s. There are many examples from Tamil also to show
that R > r led later to some r's being represented as R. Hypercorrect forms
are generally less systematic and give the impression of a reversal of sound
change.>>

Fine. But I have shown why we have to assume that the composer was not
hypercorrective but was using existing forms. So, according to Krishnamurti,
the persons in the Tamil society who used "cAn2ai" were hypercorrective.
Krishnamurti is not considering an important factor, i.e., the basic social
factor involved in  who was literate and knowledgable in Sanskrit and who was
not and who uses -c- forms in native Tamil words. The -c- forms are
predominantly used by illiterate/lower class persons with little or no
Sanskrit or linguistic knowledge.

The hyper-corrective behavior Krishnamurti mentions can be expected from a
person with literacy in Sanskrit and comparative/historical linguistic
knowledge. An illiterate person does not know that the word ENi he uses is a
loan word let alone that the original word began with s or z or zr. If you
look at the -c- forms like ucaram, it is a mark of illiteracy. It is not a
mark of upper caste language. Some upper caste dialects use �z� instead of �c�
even in words where PDr forms would have "c" because of their knowledge of the
phonology of Sanskrit loan words. (Example: Sollu/zollu instead of
collu/sollu) So the colloquial use of -c- instead of -y- predominantly by
illiterate persons shows the truth is diametrically opposite to what Dr.
Krishnamurti says.

As for Tamil r/R, the confusion is there even in CT. Consider oLiRu and oLir.
I do not know if that is to be attributed to hypercorrection.

<<
mataca:nai is clearly a hyperstandard writing. If it was a sound change it
should continue into later times.>>

Not necessarily. As we all know, formation of new words and loss of old words
is an ongoing process in any language. This has even been recognized by Tamil
grammarians. According to nan2n2Ul, "Old ones disappearing and new ones
entering due to variation in times are not  wrong". There are many words in
the Tamil Lexicon which are not used today. Survival of a form has nothing to
do with whether it was a result of "genuine" sound change or hypercorrection.
Just following Krishnamurti�s reasoning, we know that "ucaram", etc. still
survive. So can�t one argue it is due to sound change?

<<
>ucar < uyar (DEDR 646)
>vacakkal < vayakkal (DEDR 5258)
>vAcal < vAyil (DEDR 5352) -
>vicalUr < viyalUr = viyal (DEDR 5404) +Ur
>
>uyar - to rise                  puRanAn2URu 334.8
>vayakku - to tame           akanAn2URu 344.10
>vAyil - entrance              puRanAn2URu 350.6
>viyalUr - a place             akanAn2URu 97.13
>vayal - cultivated field     puRanAn2URu 354.4  (Kannada has both vayal and
>bayal(u))
The above with -c- are hypercorrect forms. If y became c why not vasal and
vasil. >>

The reason is elementary. Language as a social phenomenon does not work like
Microsoft Word. You cannot guarantee global changes like you do in word
processing.

<<
>The forms with -c- are found in inscriptions at least 600 years later than
the
>literary attestations of -y-. None of the -c- forms discussed above are found
>in Classical Tamil. Some of the -y- forms are used even today in formal
speech
>while some of the -c- forms are considered very substandard. For instance,
>literate persons will not use ucar- in colloquial speech. The preferred form
>is -y-. -c- and -y- are clearly distinguished. On the other hand, vAcal is
>acceptable and vAyil is found in very formal speech. Inscriptions show not
>only vacakku, but also, mayakku, and macakku reflecting all the variant
>dialectal forms. Todau tiruvicalUr is the name of the town.
>
What you say proves my point. >>

Absolutely not. See the discussion above.


>The placement of DEDR 5259 Ta. vayiRu belly with an implied *-y- seems to be

>correct even though Kannada has basiR. Colloquial Tamil has forms like
vakuRu,
>vavuRu, but no vacuRu.

<<No. -c- is older as I said earlier. Konda and other SCD languages preserve
the s form, Konda vasking(velar nasal)'entrails'. You seem to think that the
oldest forms are only foound in Tamil in every respect. What is your basis
for taking -y- in vayiRu as older?>>

No, I am definitely not thinking only Tamil has the oldest forms. Only in a
posting yesterday, I noted the palatalization of *k- before front vowels as in
Ta. cey, with the result that  Tamil forms are not the oldest. But the case of
y/c alternation in the root final position we are talking about goes back to
PDr. stage according to G. Sambasiva Rao (A Comparative Study of Dravidian
Noun Derivatives, p. 14). In this case, I do not see any reason to choose a
form attested chronologically later to be the proto-form.

Krishnamurti seems to be thinking Tamil forms are later because he begins with
an axiom that -c- forms are older. That cannot be taken as an axiom. It has to
be proven. Without that axiomatic reasoning, one has to prove that SCD forms
are "preserved ones" and colloquial Tamil forms spoken by illiterate/lower
class persons are "hyperstandard". In other words, one has to show why the SCD
forms cannot be considered to have undergone sound change or "hyperstandard
forms" as he puts it.

Earlier  when I mentioned that the change y>j in Kui accepted by Subrahmanyam
and Krishnamurti as going against the weakening hierarchy of Hock,
Krishnamurti said, "Kui kaju for *kay is a doubtful transcription; I have to
check how regular is *y = Kui j. I know Schulze a German wrote j for y."

Dr. Krishnamurti is contradicting himself.   In Telugu Verbal Bases, p.120, he
has stated:

"2. Ta. kai hand; Ma. kai, kayyi; Ka. kayi, keyyi; Kod. kay; Tu. kei; Te.
cEyi, also kai-, and kE(lu); Go. kai; Kui kaju (-j- < -y-);....."

Please note that he was very explicit in showing -y- > -j- in Kui.

In "A vocabulary of the Kui Language [Kui-english]" by Rev. W. W. Winfield,
the author says, "This vocabulary has grown out of my own need for such a list
of words and from the attempts made to supply that need for private use as a
missionary among the Konds. It owes a very great deal also to Mr. A. J.
Ollenbach, formerly of Phulbani and now Divisional Officer of Angul in Orissa,
who most generously placed at my disposal the whole of his valuable collection
of Kui words gathered together during more than twenty-five years of
administration among the people of the Khondmals." This book, first published
in 1929 (more than 30 years before TVB) and reprinted in 1985, does show Kui
kaju "the arm, hand".

Moreover, according to Hock, "k" and "g" are higher than "y" in the weakening
hierarchy. If this rule is to be strictly and universally applicable, "y"
cannot become "g" or "k". But this is exactly what Krishnamurti derives for
Telugu. In Telugu Verbal Bases, p. 33, he states:

"Te. -g- < PDr. *-y-"

To show that -y- passed through the stage of -k- before becoming -g-, in
Telugu Verbal Bases, p. 34, he states:

"..-g- proceeds from an older *-y-; that -y- passed through the stage of -k-
is evident from Ta. vici (*viki) corresponding to Te. Ka. bigi < *viki <
*viyi; another striking proof of this can be found in the following etymology:
Te. vAkili doorway; Tu. bAkil; Ka. bAgil; Ta. vAyil, vAcal; Ma. vAtil (Ma. -t-
< *-c-)"

One should note that here -y- > -k- or -g- is presented not as a "hyper-
correction". It is obvious that he implies it is a sound change. Also, note
that Krishnamurti himself gives vAyil and vAcal as alternate forms. But, in
the present discussion the change from Tamil vAyil to vAcal is somehow
"hyperstandard" and not a sound change!

What can we conclude from all this? The Kui vocabulary was not transcribed by
Schultze. Kui does show y>j, as accepted by DED and DEDR. This means the one-
directional weakening hierarchy is not universal. So a -y- > -c- seems to be
possible.

Regards
S. Palaniappan