solution to the "kuyava' etymology

[Palaniappa]

Regarding the formulation of the i/e  and u/o sound change model proposed by
Bh. Krishnamurthy,  there seems to be some disagreements between Dr.
Krishnamurthy and Dr. P.S. Subrahmanyam as has been pointed out by Dr.
Krishnamurthy himself in his foreword to Subrahmanyam�s book, "Dravidian
Comparative Phonology". But Subrahmanyam does agree with Krishnamurthy
regarding Krishnamurthy�s basic view : Proto-Dravidian had all the four vowels
i,e,u,o in the root syllable of (C)VC type. In Proto-South Dravidian they
remained as such when high vowels or zero followed in the next syllable, i.e.
(C)VC-i/u/0; but when a low vowel -a followed i,u merged with e,o;
i.e.PD*(C)i/uC-a >PSD*(C)e/oC-a- (rule of vowel harmony or umlaut). Then,
Early Tamil (which included Malayalam) changed these sequences to i,u (rule of
dissimilation) throughout.

Krishnamurthy based his conclusion regarding the nature of the proto-South
Dravidian phonemic status (whether they were *i/*u or *e/*o) on the
contraction of the root vowel and derivative vowel which takesplace when the
intervening *-k-, *-y- (<*-c-) or *-v- is lost. In his book "Telugu Verbal
Bases", Krishnamurthy gives the following processes

I. (a). *iC-[I,u > I                      II (a). *uC-[I,u >U
   (b). *iC-[a > E                           (b). *uC-[a> O
   (c). *eC-[I,u,a > E                     (c). *oC-[i,u,a > O
Krishnamurthy says, " From the above we notice that in contracted forms PDr.
*e and *o do not change their qualities but are simply lengthened when any one
of the three admissible vowel occurs in the derivative suffix. On the
contrary, PDr. *i and *u in the sequences *iC[-a and *uC[-a behave in the same
manner as original *e and *o in producing E and O in contractions. In contrast
to this *i and *u preserve their quality  even in contractions when they are
not originally followed by a. In other words in the environment [C-a all the
SDr. languages show E and O in contracted forms irrespective of whether the
Pdr. quality is *i *e or *u *o respectively". There seems to be a fundamental
problem with this.

Let us consider some words in Classical Tamil texts with first vowel "u" as in
our present problem. Consider the following words:
"uvaRu" (kalittokai 136.2)
"URu" (akanAn2URu 178.7) . Both mean "to spring, flow (as water in a well),
etc.

There are more pairs like this. If Krisnamurthy�s rule II (b) were to be
valid, we should have ORu instead of URu. So, this example shows that the
model of Krishnamurthy is not valid for all cases. This means at least in some
dialects, PDr phonemes must have continued to exist right through proto-SDr to
Classical Tamil stage as seen here. I am surprised that in 40 years Dravidian
linguists have not realized this. (This underscores the importance of
Classical Tamil for an accurate historical analysis of Dravidian.)

F. B. J. Kuiper in his book "Aryans in the Rig Veda" gives evidence of i/e and
u/o variation in foreign words in the Vedic texts.
zI�pAla (RV)  - ze�vala (AV)
ogaNa (RV) - u�gaNa (YV)

He also gives evidence of loss of initial "z" in two loan words between Vedic
and post-Vedic materials. Examples are

zirimbitha- (RV)    -  Irimbithi- (sarvAnukramaNI)
zailUSa�-    (VS)    -  AilUSa- (AB)

All these charactereistics remind us of situations recorded in South Dravidian
several centuries later. While Kuiper does not say the origin of these loan
words, the fact that such alternation has existed at Vedic or post-Vedic
periods indicates its long history.

Coming to the present case, I am attempting to explore only the etymology of
�kuyava" in RV and VS. I  leave the �ku/UyavAc" to others. Among these two (RV
and VS occurrences), I do not know if Krishnamurthy has any disagreements with
the VS occurrence in light of the discussion above. As for the RV occurrence,
Krishnamurthy says <<The Tamil word for potter is reconstructable to*kucawa in
Proto-South Dravidian. Old Tamil had both kucavan and kuyavan.See DEDR (1984)
entry 1762 The -c- form was older as in the case of PD*picar, PSD *pecar, Old
Kannada pesaru 'name', Ta. peyar, pe:r, Telugu pe:ru; Pkt. sa:siram (< Skt.
sahasra-), Ta. a:yiram, etc. Assuming that all non-SD languages have lost the
potter form, at the time of borrowing from Dravidian during the Vedic period,
the form should have been *kucawa[kusawa].. and not *kuyawa.>>

As far as I know, "kucava" does not occur in Old Tamil texts. What we do find
is "kuyava". I would be interested to know when "kucava" is first attested. Be
that as it may, even if the reconstructed form is *kucava, there is a
possibility that the variant "kuyava" might have existed even during the Vedic
period. According to P. S. Subrahmanyam, "*-c- is weakened to -y- in many
cases even in Proto-Dravidian". So, one cannot rule out the possibility of
Vedic borrowing "kuyava" from a dialect where such a weakening has takenplace.
In a later posting, I shall address the cultural factors which make me think
that RV "kuyava" might refer to the potter/chieftain. I welcome comments.

Regards
S. Palaniappan