Dates of written Rgveda

[Steve Farmer]

Michael Witzel writes;

> Dear Dr. Farmer,
>
> Most of the pertinent arguments of the 'consensus' have been provided by
> Profs. Thompson and Vassilkov, so I can be fairly brief, and will limit
> myself to those items not yet mentioned.
>
Thanks much for your detailed note and the references, which I
will study carefully. Most of my reservations to these views have
also already been put forward. I'll limit my response to an even
smaller subset of your comments.

You write:

> The one point which you don't seem able to accept is that the text was
> indeed so faithfully transmitted *orally* only, over these maximally 1000
> years, until the redaction [which Prof. Witzel places just before c. 500
> or 400 CE).

Correct. I also have doubts about the supposed transmission of
the text from this period forward - over the roughly 1500 years
until the 1000 CE date that you set for the supposedly *first*
written version of the text. (That is the claim that I find least
credible.) I find that later date unlikely in the light of
references to written versions of the text found in sources from
somewhere not long after the start of the common era. Apparently,
at the least, not every scholar on the List is convinced by the
1000 CE date.

> As pointed out by my colleagues, this is nothing special inside India,
> while it may be elsewhere. It has been the ethos of this culture(also in
> neighboring Iran: Avestan) from the first hymn of the RV onwards. Even
> early post-RV texts refer to it, explicitly, by a story: Indra came to cut
> off someone's head because he had mispronounced the tonal accents in a
> nominal compound, and thus (grammatically) had turned himself  from an
> "enemy of Indra" into someone Indra hates ('whose enemy is Indra'). Indra
> had no choice, since Mantras always work... The same story was still told
> to me by some Nepalese Pandits when they corrected the manuscript with the
> Mantras for the coronation of the present king, in 1975.

As I've pointed out in an earlier post, I've seen *exactly* the
same sorts of claims about totally faithful transmission in
Hebrew Rabbinical documents. They don't pan out. Same for similar
claims in Islamic traditions. The same dangers regarding
mispronunciation are also raised in Kabbalistic documents from
the later Middle Ages -- and, in fact, in Maya written traditions
as well. On closer examination, you discover that the
transmission was far from faithful when it depended solely on
oral means. The only exception comes when texts were used as aids
to memorization, which helped (even if surreptitiously) to fix
the text.

Many studies of the contemporaneous interaction of literate and
oral traditions exist that depart considerably from (say) Goody's
simple model of oral/literate transformations. A review of some
of the issues involved is found in Thérèse de Vet, "The Joint
Role of Orality and Literacy in the Composition, Transmission,
and Performance of the Homeric Texts: A Comparative View,"
Transactions of the American Philological Association 126 (1996)
43-76. de Vet looks at the traditional work on this issue in
Parry and Lord and compares with the results of her own field
work on oral recitation in Bali. (I don't agree with all of de
Vet's conclusions, but the paper has its points.)

As a personal aside, I might point out that many intellectuals in
the European Middle Ages and Renaissance supposedly had eidetic
memories for texts cultivated by long mnemonic training. This
training supposedly allowed them to recite long works, including
the Aristotelian corpus (far longer than the Rgvedas) verbatim.
Perhaps the most famous of these figures was Giovanni Pico della
Mirandola (1463-94), who (according to many contemporary
accounts) could, after a single hearing of a text, immediately
repeat it "backwards as well as forwards" without error. In 1486
Pico proposed a debate "de omni re scibili" (on all knowable
things) involving twenty-eight past traditions and subtraditions
-- the largest such debate for which we have a record in Western
thought. I know the work well, since I've edited, translated, and
commented upon the 900 theses that Pico drew up for his debate.
That editorial work allowed me to test the "backwards and
forwards" thesis at length, leaving me very skeptical about
extraordinary memory claims in other premodern societies. In sum,
Pico's "near-perfect oral transmission" turned out to be an
illusion, and his quotations from memory turned out to have many
telltale errors of those found in tests of memory processes by
neuropsychologists. None of these were spotted by his
contemporaries.

In any event, if *truly* "near-perfect oral transmission" was
involved in India, my inclination would be to suspect that
somewhere, behind the scenes, written texts were playing a role
in helping *fix* oral memory. (See infra.)

> And, one can of course learn by heart the c. 800 pp. of the RV, in small
> Roman characters, perfectly. Just go to the various corners of India, and
> you will still find men who can do it! Make them start anywhere inside the
> text, without a written text in front of them. I have seen it. So why not
> in 500 CE, 500 BCE and 1000 BCE?

The faithful memorization of a text at one point in history with
aid of a written canon in the background is one thing; the
faithful transmission of a text over two millennia in the
supposed total absence of a written text is another. Note that
even in this paragraph you assume a *written* text against which
to *test* your thesis.

There are also, of course, scholarly dissenters from your
position in your own field -- which makes me question the claim
for a consensus. Vishal Agarwal's previous comments, which I've
quoted before, speak eloquently to your last point:

> Despite the mnemonic devices and a host of texts facilitating
> the oral tradition (these texts are called Lakshana granthas), the recitors
> still use the written text as an aid in memorization, and for occasional
> cross checking. The orally transmitted text is considered more accurate, but
> is nevertheless supplemented by a written text (some recitors place it in
> front of them, some behind them, while others who are perfect, do not use it
> at all at least in public because of the stigma attached to it).

You write, in conclusion:

> In one word, India is different here from the rest of the world (minus, to
> some degree, the Zoroastrian texts of their closest relatives, the old
> Iranians), due to the stress, put from the very beginning on correct
> pronunciation and recitation, -- from tonal accents, to words, to
> sentences, all in the proper textual order. No changes, no substitutions.
> As my colleagues (or the Nepalese pandits, above) have said: any change in
> pronunciation will result in unwanted or dangerous consequences.

Again, with all due respect, I've seen such claims in many
cultures about canonical sources, and remain highly skeptical.
There is apparent evidence that texts of the Rgvedas may have
considerably predated 1000 CE. What objection do you have to the
hypothesis that, despite the extraordinary emphasis that
premodern Indian tradition placed on "perfect oral transmission"
-- which had obvious political/religious uses, I might add --
these texts may have helped fix the canon in ways that were more
reliable than orality alone. Images suggested by Vishal Agarwal's
post keep reverberating in my head: of religious/political
reciters placing the text "behind them" or hidden away "in public
because of the stigma attached to it"?

I raise these issues only as questions, which had their origins
when I read those pesky passages in Manu that clearly refer to
early written versions of the Veda. Apparently, judging from
dissenters on this List, other evidence of early versions of the
written work exist as well from well before your 1000 CE date.

I'll watch the Indology List in coming months for more evidence
on this question. Thanks in the meantime for your detailed and
helpful response.

My best,
Steve Farmer