Sarasvati (texts & arch.)

Michael Witzel witzel at FAS.HARVARD.EDU
Wed May 20 21:49:14 UTC 1998

Yes, the Sarasvati, Edwin:
That really is an old horse (Oldham 1887, Raverty 1893, Wilhelmy 1959,
Yash Pal 1984), and we should look into her mouth :

For example, On 13 May 1998, Sn. Subrahmanya wrote:

> It seems that people are unwilling or unable to recognize the importance of
> the Sarasvati and its implications to the dating of the Vedas.
> The dating of the Vedas will in turn affect the rest of the Indian
> chronology and all of Indo-European studies.

Really, all of IE? as if there were no early data from the Hittites and
Mycenean Greeks.... (1900 BC onwards)

Indeed, in the Rgveda, the Sarasvati is well known and highly praised in
as a great stream. And, indeed, once it is called a river flowing from
the mountains to the 'sea' samudra (RV 7.95.2, cf. 7.6.7), i.e. apparently
in the Ghagghar-Hakra-Nara channel (see Oldham 1887, Raverty 1893!!), ---

unless the whole passage is not a reminiscence of the  E. Iranian
Sarasvati (= Avestan: Haraxvaiti)  and the great Hamum lake (Kangsaoiia,
cf. also Avestan texts on Vourukasha), into which the E.Ir. rivers flow,
as *sam-udra.

NOTE that Sarasvati/Haraxvaiti mean "(river) having ponds", certainly not
a good name for a mighty stream; that this must at least refer to a slow
flowing river with many bends and u-shaped cut-off ponds, or it may be a
river disappearing in the Iranian/Indian desert with many ponds in its
lower course.

However, the Brahmana texts (JaimBr 2.297, PancavB 25.10.1)  clearly state
that the Sarasvati disappears or "dives under" in the desert at a place
called vinazana / upamajjana. (Later texts such as the Puranas theorize
that it flows underground from there up to the confluence of the Yamuna
and Ganga at Prayaga/Allahabad).

Again, the present dry bed of the Ghagghar-Hakra is indeed cluttered with
Harappan sites. But these settlements are *on* the actual flood plain of
the old Sarasvati/Ghagghar-Hakra, which again speaks against an enormous
river during the Harappan (or a supposed pre-Harappan Rgvedic) period. Or
does one want to live in place that is flooded each summer?

The estimates of archaeologists on the exact date of the drying up of much
of the Sarasvati differ considerably: Mughal (1995) argues that the Hakra
was a perennial river in the 4th and early 3rd millennium BC and that it
had dried up about the end of the second. Other dates range from 2500-1700

Apart from the older studies quoted above (-- I always drew the
Sarasvati down to the Indus even on my student time maps...reenforced by
data from Wilhelmy whom I accidentally met in 1982 or so --), there now
are satellite photos but they taken *alone* do not provide dates.
Painstaking exploration on the ground is necessary. -- Even then, the old
Sarasvati-Sutlej can never have been larger than the Indus during the
millennia in question.

While RV 7.95.2 (cf. 7.6.7) indeed speaks of the Sarasvati flowing to
the/an ocean, this is not unambiguous, due to the various meanings of
samudra "ocean" or "confluence of rivers" (debated for more than 100 years
when the whole immigration/emigration debate was not known in its present
form) see:  RV 6.72.3, and cf.  Klaus, Kosmographie,1989; --- RV style is
generally quite hyperbolic: the Soma juice dripping through the sieve into
a cup is not exactly "rivers" "rushing" to the "ocean"... What does a
"Sarasvati flowing into an ocean" *really* mean??

Be that as it may, however, a *neglected* contemporary piece of evidence
comes from RV 3.33, which, on internal grounds, describes a situation of
only a few years before RV 7.95.2 (see my forthc. paper).  RV 3.33 *must*
speak of an already smaller Sarasvati as it refers to the confluence of
the Beas and Sutlej (Vipaaz,Zutudrii).

That means: the Beas had already captured the Sutlej away from the
Sarasvati. The later RV (books 3, 7 and the late book 10) thus depict the
present day situation, with the Sarasvati having lost most of its waters
to the Beas and probably also to the Yamuna; it was no longer the larger
river it *may* have been in the earlier Rgvedic period (and remembered in
RV 7.95).

This INTERNAL evidence, and the otherwise accepted date of the Rgveda at
c.1500-1200 BC, also agrees remarkably well with the new evidence from
Bahawalpur/Cholistan (Mughal 1995, who discovered all the now famous
multitudes of Harappan/Indus sites on the Hakra). He indicates that the
area along the lower Hakra (Sarasvati) was abandoned after c. 1400 BC by
its (post-Indus civ.) people, who moved eastwards, due to lack of water,
not due to "invading Aryans" (as J.Shaffer also says). The area was not
settled again until the establishment of some Painted Gray Ware (PGW)
settlements there (800 BC), when we indeed hear of sparse settlements in
the west (AitBr 3.45). More data in Taitt.Br. and Pancavimsa Br...

This also agrees with the scenario developed recently (Witzel 1995): early
Indo-Aryan immigration (*maximally* starting at the end of the Indus civ.,
1900 BC, and down to c.1400 BC) of the Yadu-Turvaza, Anu-Druhyu into the
Panjab, and immigration of the Bharata tribe (from across the Indus, well
remembered at JaimBr 3.238) only after that, when the major part of the
Sarasvati waters had been captured by the Sutlej/Beas (and the Yamuna).
This scenario, consistent with the geological, archaeological and textual
evidence is in striking contrast to that of the Indian homeland theory.
(for example, of a pre-Indus civ. Rgveda).
(And I am not even talking about copper (ayas) in the RV which would be
*globally* out of date in a pre-Harappan Rgveda).

A few final points:

> It is also interesting that scholars who argue that Max Muller's works are
> outdated have no qualms in still accepting a chronology that was proposed by
> him.

Everybody knows, and he himself says, that it was a guess, but it was a
shrewd one, as the above data again indicate. Also, we must remember the
contemporary data of the Mitanni, at c.1380 BC -- which he did not yet

> It is a symptom of the massive retrofitting of data.

I think the above shows that the retrofitting goes the other way.

The rest does not need comments as it is based on the Sarasvati argument

Michael Witzel                       witzel at

phone: 1- 617 - 495 3295 (voice & messages), 496 8570, fax 617 - 496 8571
my direct line (also for messages) :  617- 496 2990

More information about the INDOLOGY mailing list