tantra and paa, suutra and nuul

[thompson at jlc.net]

>4/20/97
>
>N. Ganesan wrote, and so I will try to reply:
>
[snip]
>
>If the sanskrit term 'tantra' is related
>to 'tantu' whose main meaning is 'warp', (I agree. It can
>generically refer to the whole art of weaving/spinning sometimes),
>there seems to be connection between 'paa' of dravidian
>and 'tantu/tantra' of sanskrit.

We need to determine *what kind* of connection there is between these
terms.  It seems to me that it is not valid to jump from the assumption of
a semantic connection [or similarity] to a genetic connection.  The
semantic similarity is obvious, and is not disputed.  But I would dispute
the assumption of a genetic connection, unless you can provide *specific*
evidence of such.
>
>Are there any IE cognates between words meaning 'warp'
>and 'songs/verses' as well? Sanskrit and Tamil have one
>word to denote both.

There are *many* such in IE languages, and I have cited them already.  To
cite just one again: see Avestan vaf-, = "to sing, to weave".  Bartholomae
cites the Modern Persian form bAfad = "he weaves", whereas the Old Avestan
forms of this root = "sing, song, etc."  Compare Skt. ubh-, vabh-, clearly
a root extension of u-, ve-, already cited.
>
>The metaphor of relating weaving to composing a text
>can arise in any culture. I am reminded of the American usage,
>"may be in Timbuktu". But the 'warp' and 'text' meanings
>in 'paa' or `tantra` do not seem like a mere coincidence.

Maybe, maybe not.  But how can we determine whether this semantic
association, attested elsewhere in IE, the Near East, and Africa, is NOT
coincidence?  Surely, there is equally strong evidence that the metaphor is
widespread in IE.  It seems natural to assume that the Vedic RSis thought
that speech was like weaving because *it is* like weaving, and because IE
poets had been thinking so for millennia.
>
>I like Palaniappan's explanation that 'nuval' originally
>meant 'fine web' of spiders. From Nature, Dravidians
>got the clue and used 'nuul/nuval' also for composing texts. That is why,
>`nuul/nuval' refers to both 'thread' and 'text'.
>Similarly, in Sanskrit also, suutra (related to 'sIv-', 'syU')
>means 'thread' as well as 'text'.

See earlier reference to UrNavAbhi = "spider". -vAbhi derives from the same
root as Avestan vaf-, cited above, meaning "to sing".  Clearly,
Indo-Iranians, without any necessary assistance from Dravidians, also "got
the clue from Nature".

>Are there any other languages where some term from weaving industry and
>text are identical?
>Again, Sanskrit and Tamil use a single word to denote thread' and 'text'.
>
I could cite [already have cited] Indo-European, Semitic, and Niger-Congo
language families, if anyone is interested.

As I have already said privately to S. Palaniappan:

"Let me say first of all that I have sympathy and respect for the industry
with which you attack this problem.  In general, I am persuaded [as most
Vedicists are, I think] that Dravidian has left its mark on Vedic, even at
the earliest attested stage [i.e., the RV]. So our disagreement does not
involve this general point."

Our disagreement arises when it comes to specifics.

Best wishes,
George Thompson