Fortunatov's Law and tolkAppiyar's rules

[Sudalaimuthu Palaniappan]

In a message dated 98-07-09 10:03:27 EDT, vidynath at MATH.OHIO-STATE.EDU writes:

<< > There is no need for doubling of _t or T all the time. In fact, I had
shown
 > examples where the doubling does not occur. They were:
 > iyal + tEr > iya_tEr,  nAL + tO_tum > nATO_tum
 >
 > Moreover, from the point of view of pronunciation, Sanskrit intervocalic
-T-
 > is closer to Tamil -TT- than Tamil intervocalic -T- which is closer to
 > Sanskrit -D-.

 I am unable to look at comprehensive lists of Sanskrit borrowings
 from Dravidian at the moment (because the sources are far away or
 checked out of the library). I would appreciate it if someone can
 tell me if this implication are justified by such borrowings. [Because
 Prakrits voice intervocalic stops, borrowing in reverse are not conclusive.
 Learned borrowings are affected by writing and so I don't want use
 them. to claim that Tamil _t_t/.t.t cannot be borrowed as just .t.]
>>

This is what Burrow says: "Intervocalically a single unvoiced consonant in
Sanskrit very often replaces a double consonant in Dravidian: cf. capeTA
(no.21), piTaka-, puTa-, (no.39), etc." (p. 265) in Some Dravidian Words in
Sanskrit, in Collected Papers on Dravidian Linguistics, 1968, p.236-284.

What is interesting is Sanskrit shows variant forms with single and geminate
Ts. For instance, in addition to Skt. paTa cloth, we also have paTTa . These
are of IE origin. But a parallel case can be shown for loans from Dravidian
also. Burrow says, " There is another word puTa- n. in Sanskrit ....It means
'anthill', and is found only in the cpd. pipIlakapuTa-, MBh. The Dravidian
word from which it is derived appears in the various languages as follows: Ta.
pu_t_tu anthill, Ka. puttu Te. puTTa, Kui pusi, Kuvi puci, Malt. pute id.  Of
these the Telugu form with -TT- (out of original affricate most closely
resembles the Sanskrit word. On the other hand, in another loanword from the
same Dravidian source, namely Skt. puttikA f. the white ant or termite, the
Sanskrit word shows the same development of the internal consonant as
Kanarese." (p.274) (This example also shows variations between alveolar,
dental and retroflex.) So I don't think the presence or absence of gemination
can be used to reject Dravidian as the source of IA retroflexion.

<<
> [...] But  there is an important difference
 > between these consonants and t. k, c, and p remain as they are. But t
changes
 > either into _t or T.

 This can be split into two rules: (1) l=>_t and L=>.t before stops and
 (2) retroflexes absorb dentals and alveolars, and alveolars absorb
 dentals. (2) is needed for nasals anyway: ka.n + niir => ka.n.niir.
 (2) applies to IA also, but (1) does not, except for the strange fact
 that lt/ld are missing except for one occurance. [t+l and n+l do
 occur in Sanskrit where they become ll and nasalized l+l.]
 It is this difference that needs to be explained.
  >>

I do not understand exactly what you mean here. But, in order for the
Dravidian origin of retroflexion to be accepted, if you require that Dravidian
should exhibit a tendency where L > T only before t and not before k, c, p
replicating the situation in Sanskrit, I think a case can be made as follows.

As I have noted in my earlier posting, we should look to the colloquial
Dravidian language for clues to the origin of retroflexion in IA. After all,
the split of *l into l and L also is based not on centamiz but colloquial
Tamil (and other central Dravidian languages). In many colloquial situations,
L does not become T before consonants like k, c, and p. Malayalam which has
severed its connections with centamiz offers evidence for the unique nature of
t vs k, c, and p. Consider the root kEL- to hear. From it are derived forms
like kELkkunnu, kELppin2, etc. Note L does not change into T. On the other
hand, with the suffix tt, it is kETTu and not kELttu. Similarly with the root
vaRaL - to dry up, we have vaRaLcca dryness, but vaRaTTuka to dry up, vaRaTi
barren woman.

So I do not see anything unique about the above situation in Sanskrit which
cannot be explained with a Dravidian origin of retroflexion.

Regards
S. Palaniappan