Harappan 'non-texts'?

[Steve Farmer]

Can I invoke once more the immunity of an outsider and ask a naive
question -- and make a far-reaching (but maybe easily overturned)
suggestion?

Can anyone point to evidence of *any* sort that suggests that extended
texts existed in Harappan culture, as a lot of Indologists assume?
That assumption isn't limited to OIT proponents or Indian nationalists
hoping to tie Harappan culture directly to Vedic sources. Even Asko
Parpola (1994: 54), pointing to the Aztecs, imagines a lost Harappan
literature that his Aztec parallels imply may have included extensive
religious, legal, historical, scientific, and even philosophical
texts. Parpola hypothesizes that 'The normal writing material is
likely to have been the perishable palm leaf, as it was in India until
this century, or cotton cloth....'  Despite the stunning nature of his
suggestion, the whole of Parpola's analysis consumes less than 1/2
page of his dense 374-page book; no counterarguments against his
suggestion are mentioned, let alone discussed.

Gregory Possehl (1996) admits the 'undenial logic' of the notion of
lost Harappan texts, but cautions us about the perils in archaeology
of relying simply on 'logic.' Iravatham Mahadevan
(http://www.harappa.com/script/mahadevantext.html) also grants the
possibility of lost Harappan texts once recorded on perishable
materials, but is again more circumspect than Parpola; Mahadevan
registers his deep disappointment that even the massive wooden wall
inscription from Dholavira isn't much more than a 'magnified version
of a seal.' And, indeed, ten characters don't constitute much of a
'text.'

Others, like Koenraad Elst, who collaborates closely with extreme OIT
proponents (see the acknowledgements in Talageri 2000) are
understandably far less cautious. In one online paper, entitled 'The
Vedic Harappans in Writing,' Elst writes that 'There is not the
slightest doubt that this harvest of Harappan writings [the 4200-odd
seals he notes] is but the tip of an iceberg, in this sense that the
Harappan culture must have produced much more copious writings, but
that most of them disappeared because the writing materials were not
resistant to the ravages of time, particularly in the Indian climate.'


Is there an 'undenial logic' in the notion of an extensive Harappan
literature, or even the certainty of 'copious writings'? I'd like to
offer four arguments from comparative history that suggest, on the
contrary, that *no* written Harappan literature ever existed. That
evidence doesn't depend on 'arguments from silence' and has nothing to
do with the warm and damp climate of the Indus Valley. If valid, the
arguments suggest that we can safe;u predict that we'll *never* find
even indirect evidence of extended Harappan texts -- and that no
Rosetta stone is ever likely to appear that will allow us to
confidently decipher the script. (I leave aside here claims by Jha and
Rajaram that the code has been broken, joining the legions of past
claims of this nature documented by Possehl; my copy of their new
book, reportedly airmailed to me from India back in mid May, has yet
to arrive.)

There may very well be flaws in my argument; hopefully, someone will
point them out. I claim no specialized knowledge at all about Harappan
script:

1. Harappan script (as emphasized by Parpola 1994: 54) shows unusually
little internal development from c. 2600 BCE to c. 1900 BCE (to use
Parpola's dates). This is *sharply* at odds with what is known of the
development of logo-syllabic systems in societies with fully developed
literate traditions. Indeed, we can observe striking evolutionary
developments in such scripts in a number of Mesopotamian scripts
emerging in *exactly* the same period that Parpola identifies with
mature Harappan culture. Whenever literary production was extensive --
as Parpola himself emphasizes in respect to other historical cases --
scribal pressures gradually pushed scripts in simplifying directions;
expected changes included simplification of shapes, movements towards
fully syllabic forms, and a concomitant drop in the number of symbols.
These simplifying tendencies can even be identified when evidence
suggests that religious-political authorities actively tried to retard
these innovations. Admittedly, studies of the evolution of Harappan
writing have been limited by the uncontrolled nature of early IVC
excavations, and arguments can be given that Parola overemphasized the
'fixity' of Harappan script. Nevertheless, agreement appears to be
widespread that Harappan writing was relatively 'frozen' over many
centuries, at least when compared with scripts that rapidly evolved in
Mesopotamia in the same period. This is difficult to imagine in any
fully literate society producing 'copious writings'; it is *not*
difficult to imagine if Harappan symbols had highly restricted uses,
if centralized control was tight, and if there was little scribal
pressure (due to radically restricted use of the writing system) for
script simplification. Indeed, innovations in Harappan script appear
to be as heavily controlled as those in weights and spatial measures.

Try to imagine a modern literate society -- e.g., my beloved Italy --
in which no radical shifts in writing took place from Dante to the
present! Any such example would vastly understate the problem, since
after the 15th century the evolution of script was actively inhibited
by printed documents.

2. If extended Harappan texts were written on perishable materials,
auxillary evidence should have survived that points to the use of such
materials. As Mahadevan points out, e.g., if the copper stylus was
used with palm leaves, the stylus should be found among known
artefacts; the same argument can be applied to instruments associated
with other types of lightweight writing materials. Comparative
evidence from the Shang dynasty here is extremely instructive. Glyphs
and other evidence from the Shang dynasty point to the existence of
bamboo-slip 'texts' some 800 years before any such documents show up
in tombs (c. 400 BCE). But even in the absence of such evidence, many
different types of artefacts attest that the Shang and early Zhou
dynasties were truly literate: cast bronze texts, etchings on knives
and other metallic objects or jade, and of course the famous bone and
tortoise shell divination texts. But while the Harappans too left
metal tablets behind, once again -- just as in the case of the giant
Dholavira wooden 'billboard' -- all we find on these inscriptions are
the same types of truncated 'texts' found on Harappan seals.

Citing still further evidence from comparative history: Only four
pre-Columbian Maya texts survive, but even if these didn't extant, no
one could doubt the existence of extended Maya; thousands of stelae,
engraved bricks, and other objects argue to the contrary. But all this
is again lacking in Harappan culture.

In Egypt, of course, if not a single scrap of papyrus survived, it
would be child's play to demonstrate from architectual remains that
the ancient Egyptians had a full writing system. The earliest evidence
of the use of papyrus comes, in fact, from wall paintings of papyrus
rolls; and writing instruments too, of course, have survived in profusion.

Finally, it can be pointed out that much evidence exists that the
Babylonians and Assyrians not only wrote on clay, but (and this is far
less well known), on animal skins as well. This is attested by
extensive iconographical evidence known since the first decades of
this century. Again, no evidence of this nature has ever emerged, so
far as I know, from Harappan sites.

In sum, comparative history offers a lot of evidence -- from many
sources -- that 'copious writings' of no sort ever existed in the
Harappan script.

3. The suggestion that it is unlikely that any urban society as
complex as that of the Harappans could exist without literate
traditions -- a view claimed by Parpola and others -- may seem
obvious, but is demonstrably wrong. Parpola's choice (1994: 54) of the
Aztecs to make his case was a disastrous choice, since it is well
known that the Aztecs (unlike the neighboring Maya) did *not* possess
anything approximating a full writing system. Aztec picture books
provided elaborate mnemonic aids, but they could not be 'read' in the
same way that texts were read in societies with fully literate
traditions. Even in the absence of a true writing system, the Aztecs
efficiently managed cities with far larger and much denser populations
than those of the Harappans; many of the *suburban* centers
surrounding Tenochtitlan contained hundreds of thousands of
inhabitants! Other urbanized cultures also functioned without full
writing systems -- or without writing systems at all. These included
the Harappan's close BMAC neighbors (who had seals but no writing at
all), the Incas (who had to make due with their quipu mnemonics), and
(apparently) the early kingdoms of the Ganges plains that helped build
the foundations of later empires in India.

4. Another piece of evidence (pointed out to me off-List today by one
well-known Vedicist) involves the surprisingly low frequency of many
known Harappan symbols; for evidence, see the chart in Parpola 1994:
78. Of the 417 Harappan symbols listed in Parpola's chart, no less
than 264 -- a whopping 63% of Parpola's total-- appear no more than 1
to 9 times in all known inscriptions! Moreover, recent evidence
suggests that there were probably more independent symbols around than
Parpola thought, making this trend even more striking. This is most
easily explained by the ad hoc introduction of logographic symbols for
limited use -- e.g., to describe professions; the majority of these
symbols were clearly *not* syllabic. In any fully literate tradition,
the number of symbols like this should have lessened over time; again,
no evidence of any sort suggests that in Harappan this simplification
took place.

So in sum, after so many decades of excavations, all we still have are
some 4000-odd seals, a large number of those duplicates, averaging a
couple of characters in length, in which the largest percentage of
symbols appear only a handful of times. This is not what one would
expect in a fully developed writing system that produced 'copious
writings'; the suggestion is that this isn't the 'tip' of things but
the whole iceberg. This is what we would expect of an extremely
limited script used for ID badges, probably carried under compulsion
by every citizen (cf. Herodotus 1.195 for the Babylonians!);
communications with divine forces through votive offerings; and
commercial tags for customs purposes. But comparative history argues
*strongly* that the script was not used for broader 'literary'
purposes.

Comparative historians are like jesters and fools -- allowed as
misfits and outsiders to shout out obvious points that insiders don't
usually utter. Most of the time, of course, what jesters and fools
shout out *is* foolish; only occasionally can they hope to hit on some
half truth. Have I overlooked something terribly obvious? Do the real
Indologists have evidence that overturns these four arguments?

My best,
Steve Farmer