Dates of written Rgveda

Steve Farmer saf at SAFARMER.COM
Sat Mar 11 20:51:42 UTC 2000

Dear Dr. Vassilkov:

Thanks much for your thoughtful note. I'll limit my comments here
to a few points, since I've already addressed some of these
questions in my reply to Dr. Thompson:

You write:

> I am afraid, you will not succeed in your attempts to apply
> to the text of the Rgveda patterns and approaches developed
> on the basis of the early written (or oral poetic) texts of
> other cultures. The technique of composition and transmission
> of Vedic texts is absolutely unique.

I admit that the idea that much in any one culture is "absolutely
unique" is anathema to the comparativist in me. For parts of my
work, over the past two decades I've closely followed specialized
developments in neurobiology -- especially in respect to
prefrontal functions, where important parts of memory as well as
cultural models are encoded. The deep biological constraints on
memory processes are, despite much neural plasticity, culturally
invariant. The same "wetware" and the similar means of internal
and external encoding (in use of rhyming patterns, imagistic
mnemonics, writing, etc.) implies that there will be broad
structural similarities in cultural byproducts as well.

> The concept of stratification which implies any
> *redactions* or *revisions* of the earlier texts does not work
> here. The composition of a Vedic hymn was a sacred process, its
> text was produced as a precious artefact (a gift to a god), the
> text was often strewn with alliterations and anagramms,
> which hinted at the name of the god; nothing could be changed
> or omitted in it.

Identical claims are made for the sacred canons in every other
major premodern society that I have studied. These religious
claims can't stand as philological proof, however.

> Special mnemonic technique made it possible to preserve the text
> word for word, sound for sound.There was even a myth stressing
> the mortal danger of the mispronunciation of a single syllable
> in a Vedic text.

Ditto for my remarks above. Jewish Rabbinical traditions make
similar claims for the Torah, for example; same for the
Kabbalistic texts of the later middle ages. Many other examples
can be cited. A lot is known about mnemonic techniques in a wide
range of half-literate societies. Again, neurological constraints apply.

> The "stratification" as regards Vedic texts means
> only that some mandalas contain old hymns and other mandalas
> (or their parts) contain late hymns; old and late hymns differ
> in style and language and no late revision of the
> old hymns was possible.

There are many signs of internal commentary in Vedic hymns that
at some point was integrated into the received text, following
familiar patterns seen in many other half-oral/half-literate
traditions. This seems most evident in the presumptively youngest
materials in the Rgvedas found in books 1 and 10, so far as I can
tell. Again, this kind of internal development is common in
other, non-Indian, traditions. This goes far beyond the question
of so-called interpolations: There were many ways in which
premodern oral/literate texts could become stratified.

> There is a consensus between the majority of
> scholars: the earliest "strata" is dated approximately by 1500
> BC, the latest one - around 1000 BC. The proposed correction of
> the date of the Buddha can make the Vedic
> hymns younger by a century, let us say, and not more.

Dating of similar texts that have always been assumed to be of
exactly the same period in other premodern fields is currently
being pushed radically forward. In China, e.g., the date of the
Yijing (Classic of Changes), which also has traditionally been
dated to the second millennium BCE, is now placed by prominent
researchers in the Warring States group in the 4th century CE --
a shift of over 600 years!

The philological grounds on which ancient texts composed under
presumptively preliterate conditions have been traditionally
dated are anything but solid, despite claims to the contrary.
Assumptions about rates of linguistic change (e.g., by the old
glottochronologists) are particularly untrustworthy when applied
to sacred languages, in which great lengths were made in priestly
classes to artificially retard linguistic drift. Violent
arguments about dating the Shijing (Book of Odes) and Yijing in
China on philological grounds, about dating various strata of the
Homeric corpus, and about dating different layers in ancient
Persian, Hebrew, and even Mesoamerican texts makes me question
the assumption that all layers of the Rgvedas lie so neatly in
the date ranges that you claim.

Good scholarship doesn't depend on majority opinions.

> The first written records in
> India belong to the 3rd century BC.

I know the studies of Fussman, von Hinüber, and Falk, et al. that
make these claims -- summarized, e.g., by Richard Solomon at

But much depends in the works of these scholars on the argument
of silence, which isn't a strong one when you consider that
neither the climate nor the nature of the typical materials used
for writing in ancient India would favor the survival of early
texts. (This is dissimilar to the situation in China, where a
long series of new tomb-text discoveries are now shaking the
field.) Moreover, as you point out, some evidence pushes that
date back as far as the 6th century, which is early enough for my
purposes. (Please don't confuse my views with Jack Goody's.)

In any event, to clarify matters: I don't argue against the
original oral composition of the Rgveda, by any means. My
interest revolves around how the canon was fixed and transmitted
starting around the mid part of the first millennium BCE, when
there was a greatly increased use of lightweight writing
materials in most major Eurasian civilizations. My views on many
points are in accord with those put forward on this List in the
last few days by Vishal Argawal and even by George Thompson, who
admits that "Brahminical orality" in the latter half of the first
millennium BCE (the period of Manu) "may have been a type of
secondary orality [influenced by literacy]." I can live with that

I would strenuously argue, on the other hand, against Michael
Witzel's comment that "near-perfect oral transmission" of the
Rgveda went on for over two millennia *before* the text was
written down, which he claims was not until c. 1000 CE. If
*that's* the consensus view in the field, it should be reexamined.

Deep down, on dating, BTW, I think there are many similarities in
the question of dating the Rgvedas, the Shijing in China, the
Psalms in the Middle East, and similar compilations in Eurasia.
All, I would guess, were reduced to writing in some way --
although official canonization didn't come until later -- by the
4th century CE.

My best,
Steve Farmer

More information about the INDOLOGY mailing list