Linear B texts

[Steve Farmer]

George Hart writes:

> It seems to me there is a fallacy in Steve Farmer's contention that  
> no civilization with writing lacks longer texts.  The fact is, if  
> any civilization committed all its writing to perishable materials,  
> we would not know they had writing.  There may be 20 or 30 ancient  
> civilizations that had writing about which we have no clue.  Thus  
> one cannot argue that if the IV civ. had writing we would  
> necessarily have longer texts.  All we know is that they left their  
> symbols on specialized seals and on a couple of artifacts (the  
> famous sign).  Carving something on stone or making a seal is a lot  
> more difficult than writing on a palmyra leaf -- there is no reason  
> whatsoever why they couldn't have written longer things (like  
> accounts ) on bark or some other material that is perishable and  
> easier to write on.  We have no long examples of writing in Tamil  
> from Sangam times, yet we know that they had writing and wrote  
> longer documents on palmyra leaves.  Of course, Farmer et al. have  
> other arguments that need to be considered.

Dear George:

Ignoring all the purely speculative "what if's" in your post, just a  
reminder, although I hate to repeat myself: what we in fact argue is  
that no known civilization that wrote long texts on perishable  
materials didn't also leave texts of significant length behind on  
durable materials.

On your claims about the earliest Tamil inscriptions: I know this is  
your field, but your claims here are demonstrably wrong. See the  
photo and discussion below.

The fact is, the Indus left *thousands* of short symbol strings  
behind on all sorts of durable objects -- not just on "a couple of  
artifacts." Huge numbers of symbol strings on potsherds haven't even  
been cataloged, as you'd find if you read the excavation reports.  
None of them is of any length, unlike the situation in all known  
literate civilizations in antiquity.

Moreover, if we believe the dates given by the Harappa Archaeological  
Research Project (HARP), run by Richard Meadow and Mark Kenoyer,  
those finds extend over a millennium. As already noted, these durable  
materials include *exactly* the same kinds of materials that truly  
literate civilizations routinely wrote long texts on.

There are issues of pausibility at stake here, leaving aside totally  
untestable speculation that "There may be 20 or 30 ancient  
civilizations that had writing about which we have no clue." Well,  
there usually are clues -- and also clues when you're not looking at  
"writing" but other types of ancient symbols, of which there are in  
fact many "flavors," as we've discovered over the past decade.

Quickly: Why everyplace else in the world where we KNOW there was  
literacy would we find literate peoples leaving long texts behind on  
potsherds, pottery, metal plates, vessels, weapons, etc.? -- while in  
the Indus civilization *only* they *exclusively* left symbols behind  
on the same types of objects that were *never* more than a handful of  
symbols long?

I guess you could in principle argue that there was a taboo on  
leaving long texts behind but not thousands of short ones. :^) Well,  
one Indus researcher whose whole career has revolved around "Indus  
writing" has actually said something, obviously rather in  
desperation. ....

You write that "carving something on stone or making a seal is a lot  
more difficult than writing on a palmyra leaf." This is in fact not  
true, George. As you'll recall when you think about it, preparation  
of palm leaves for writing is a quite elaborate and time-consuming  
practice. It is also expensive (as is writing on materials like  
cloth, parchment, silk, papyrus; wood also takes special preparation).

See for one review, Anupam Sah, "Palm leaf manuscripts of the world:  
material, technology, and conservation," Reviews in Conservation  
(2002): 15-24.

We take perishable writing materials for granted. The ancients could  
not.  Scratching on a potsherd, on the other hand, is quick,  
extremely cost efficient, and very useful, especially for accounting  
records. (Think here of the hundreds of thousands of ostraca with  
accounting records we have from the ancient Mediterranean.) We have  
many hundreds of thousands of examples of potsherd fragments with  
real writing and quite long texts from all over the ancient world --  
including India, but of course not from Indus times. In any event,  
take a look at our already detailed discussion of this issue on  
around p. 23 in "Collapse" (<http://www.safarmer.com/fsw2.pdf>), and  
esp. Fig. 1.

One final and obviously critical point -- since it involves your own  
field. You write:

> We have no long examples of writing in Tamil from Sangam times, yet  
> we know that they had writing and wrote longer documents on palmyra  
> leaves.

I hate to point out that claims from someone who is a specialist (as  
you are) in this field are wrong, but.... :^) If you in fact look at  
the photos of the very earliest Tamil inscriptions, in Iravatham  
Mahadevan, _Early Tamil Epigraphy: From the Earliest Times to the  
Sixth Century A.D._, Harvard, 2003 (in the series edited by Michael),  
you'll see unambiguous evidence that the inscriptions, scarce as they  
are, were *regularly* longer than *any* of the thousands of Indus  
"inscriptions" from many centuries of Indus civilization.

Ouch! I always thought that was amusing, and when Mahadevan published  
his wonderful 2003 book I pointed out the irony to him when we both  
gave talks at a big Indus conference. In fact, the very FIRST  
inscription he gives in his book (pp. 314-15), which he dates to the  
2nd century BCE, is in fact well over three times longer than any  
Indus inscription gathered over the past 135 years from large numbers  
of sites dating from the late 4th millennium to early 2nd millennium!  
Take a look at this scan I just made of Mahadevan's transcription of  
that piece:

http://www.safarmer.com/Indo-Eurasian/early.tamil.inscription.jpg

If anyone came up with one such real (not fake) Indus inscription,  
Michael, Richard, and I would have to fork over $10,000. (It won't  
happen, but we'd *gleefully* turn over the $10 K -- not our money,  
but an anonymous donor's -- for reasons we explain in our 2004 paper.)

We've carefully looked at all these counter-arguments, George --  
trying to falsify our own model. That's what goes on in science, if  
it is real science: trying to "save" a busted model is a scholastic  
and not scientific exercise. Things in fact only get interesting when  
models get busted, so you try every way to find holes in your own  
arguments -- not burying those holes anyway you can. (That's what I  
learned from my years hanging out with theoretical high-energy  
physicists: I quickly learned that unfortunately that wasn't the norm  
in ancient studies.)

But the evidence wins in the long run, and then you move on to the  
new interesting places that takes you. That's what we plan to do in  
Kyoto later this month.

Hmm, why do I feel as if I'm in a tag-team wrestling match with  
Dravidianists? Is there a bigger story here that involves things  
thousands of kilometers (and years) from any Indus sites? :^)

> Of course, Farmer et al. have other arguments that need to be  
> considered.

Indeed, and those arguments are constantly being missummarized in the  
discussions. We are very careful in our arguments --- and we try to  
careful summarize the positions of those on the other side.  That's a  
requirement for honest scientific discussion, as we see it.

Best wishes,
S. Farmer