The Indus script as proto-writing

Steve Farmer saf at SAFARMER.COM
Thu Jul 14 16:16:52 UTC 2011

Dear List,

It is interesting to note at a minimum that Asko's views (cited in  
full below) have radically changed since we published our 2004 paper.  
As Asko has generously pointed out in conferences that we have both  
attended -- most recently in Kyoto in 2009 -- due to our work he at  
least no longer claims that the so-called "Indus script" is full  
writing, but only qualifies as a "proto-writing" system.

The difference is not trivial, given his much more spectacular claims  
otherwise in his classic book on the so-called script (Parpola 1994)  
and in dozens of earlier and later articles, going back to the late  
1960s, when (like Rao 40 years later) Asko too claimed that he had  
"cracked the Indus code" using computers. (Asko, unlike Rao, later  
retracted his claims.)

So there has been big movement in the field since we published our  
paper in 2004.

However, the "proto-writing" idea Asko has in recent years adopted as  
a fall-back position doesn't work either, for quite obvious reasons we  
already discussed in our 2004 paper. (Over 200,000 reprints of that  
paper have been downloaded from my server alone since it was first  
published, but sometimes I wonder if anyone has actually read it  
through, given some of the odd comments made about it.)

On page 33 of that study, after discussing evidence of a lack of  
significant phoneticism in Indus symbol strings known from  
stratigraphic evidence to be exclusively very late (on bar  
inscriptions without iconography) we comment that this evidence

> suggests that the Indus system was not even
> evolving in linguistic directions after at least 600 years of use.  
> Since we  know that Indus elites
> were in trade contact throughout those centuries with Mesopotamia,  
> if the Harappans really had a
> script, by this time we would expect it to have possessed  
> significant phoneticism, as always
> assumed. (The usual claim is that the system was a ‘mixed’ script  
> made up of sound signs, whole-
> word signs, and function signs, like the Luwian system, cuneiform,  
> or Egyptian hieroglyphs.)
> The implication is that the Indus system cannot even be comfortably  
> labeled as a ‘proto-script’,
> but apparently belonged to a different class of symbols: it is  
> hardly plausible to argue that a proto-
> script remained in a suspended state of development for six  
> centuries or more while its users were
> in regular contact with a high-literate civilization.

Asko below (and elsewhere) has tried to counter this argument by  
anachronistically citing Archaic Sumerian as a parallel example. That  
is a very strange claim: it seems quite odd to us to draw parallels  
between uniformly short Indus symbol strings from ca. 1900 BCE -- a  
very high literate period throughout the Middle East -- with "proto- 
writing" from the Sumerians as much as 1500 years earler! Again, to  
quote our original 2004 paper, to which no additional comment is needed:

> it is hardly plausible to argue that a proto-
> script remained in a suspended state of development for six  
> centuries or more while its users were
> in regular contact with a high-literate civilization.

And a period of high literacy it was indeed! We have hundreds of  
thousands of extraordinarily long texts from this period in the Middle  
East -- and at the same the Indus were supposedly still working out  
the principles of writing in their "proto-writing" system??

The same of course goes for Asko's comparisons to early Egyptian  
writing from the late 4th millennium BCE -- an odd comparison again to  
make surely with early 2nd millennium artifacts carrying  
excruciatingly short symbol strings from the Indus Valley.

Moreover, as we tried to point out to him in Kyoto in 2009, he  
repeatedly miscites Baines' work on this point . Asko writes, clearly  
thinking of (but not citing) our comments about the Indus system not  
evolving in phonetic directions for over 600 years (as discussed in  
evidence in the locus cited in our paper above):

> ... for about 600 years equalling the duration of the Indus  
> Civilization — the Egyptians used a language-based, phoneticized  
> writing system, but did not write full sentences, only very short  
> texts fully comparable to the surviving texts in the Indus script.  
> Early administrative documents are assumed to have existed but have  
> not survived (cf. Baines 1999: 884).

I have in front of me Baines' fullest collection of his articles on  
early writing in Egypt, _Visual & Written Culture in Ancient Egypt_  
(Oxford 2007), which contains a number of his early and later essays  
on the origins of writing in Egypt. It contains a variety of  
illustrations of early Egyptian texts -- going back to *before* 3000  
BCE -- which are indeed "true writing" and are many times longer than  
*any* of the  thousands of short symbol chains we find from the Indus  
civilization from as much as well over a thousand years later.

The "proto-writing" thesis simply isn't any more credible than the  
"full writing" thesis that we buried in 2004.

Supposedly the Harappans couldn't figure out how to move from "proto- 
writing" to "full writing" for 600 years when their trade neighbors in  
the Near East were producing hundreds of thousands of massive texts?  
That wouldn't say much for the "Indus wisemen."

Massive and redundant evidence -- much of which Asko himself has  
gathered for us -- demonstrates that Indus symbols were of a different  
non-linguistic type. Moreover, similar non-linguistic symbol systems  
were common in the ancient world at the same time -- existing side-by- 
side with literate systems -- as we also discuss at length in our paper.

It is possible as Michael Witzel, Richard Sproat, and I have argued at  
considerable length -- also Steven Weber and Dorian Fuller, in papers  
we've given as far back as 2005 (again in Kyoto) -- to learn a great  
deal about the Indus civilization by systematically studying the  
spatial and geographical distribution of the signs using a  
computerized data base of known "inscriptions" -- aided by a special  
visual interface designed specifically for this purpose.

  Ironically, as we again argued in our original paper -- in part of  
it that is rarely cited, since everyone is still caught up in the  
(quite obvious) "it isn't a script" issue -- the symbols are more  
valuable to us as historical evidence simply because they are NOT part  
of a writing system.

In Kyoto in 2009 we offered to finance (up to $100,000!)  through the  
Cultural Modeling Research Group the creation of an online data base  
of the signs optimized for such purposes. Asko, Richard Meadow, and  
Mark Kenoyer among others were invited to join us in the Project. Our  
view is that once the evidence that we have access to was made more  
accessible to the public -- now the evidence is buried in  
extraordinarily expensive volumes --  the *obviousness* of the non- 
linguistic nature of the signs would be evident to everyone.

Asko declined the offer.

We are happy however that Asko at least no longer views the Indus  
signs as part or a full writing system, capable (as he suggested in  
his book in 1994: 54) of encoding long narratives of the sort found  
throughout the Ancient Near East in the same time period./

But the "proto-writing" thesis is no more credible, for reasons  
quickly sketched above. It is unfortunate given all the wonderful new  
evidence that we have in the symbols to still be caught up in "Indus  
script" discussions, which asks all the wrong questions.

The Indus civilization is much more interesting for what is was, and  
not for what many -- some with nationalist political agendas -- have  
imagined it to be. That is the theme of a book that Michael Witzel and  
I have wanted to write for some years (_Indus Valley Fantasies_) but  
have not yet found the time to put together.

Sorry again for the very quick note: on tight deadlines.

Steve Farmer

On Jul 14, 2011, at 7:50 AM, Asko Parpola wrote:

> Quoting "George Hart" <glhart at BERKELEY.EDU>:
>> The following reply from Steve Farmer lays out his position in a  
>> short and lucid way.  I am posting it with his permission, as I  
>> think others will be interested.  I don't want this forum to become  
>> a locus of IV speculation, as there is no end to that, but I do  
>> believe it's worth summarizing some of the current thinking.   
>> Perhaps someone on the other side ("It is writing") could post a  
>> similarly clear and abridged statement of their position.  George  
>> Hart
> I take up George Hart's challenge with the following reply:
> The Indus script as proto-writing
> Asko Parpola
> It is widely agreed that the Archaic Sumerian script or "Proto- 
> Cuneiform" is the world's oldest writing system, used in the Late  
> Uruk Period (Uruk strata IV and III, c. 3400-3000 BCE).  It was used  
> as an administrative tool to record on clay tablets such matters as  
> grain distribution, land, animal and personnel management, and the  
> processing of fruits and cereals. "The script can be 'understood' in  
> some sense, but it cannot be fully read; although there has been  
> some doubt concerning the language that was the basis for this  
> written expression, there is clear evidence that it was  
> Sumerian" (Michalowski 1996: 33).  Archaic Sumerian was logosyllabic  
> writing because its signs stood for elements of a spoken language,  
> words and morphemes, with initially rare phonetization. It was not  
> from the beginning able to record everything: it took many centuries  
> of ever increasing phonetization for this "nuclear writing" to  
> develop into a "full writing" where all grammatical elements were  
> written. Yet it is considered "true writing", because it was a  
> language-based system of visual aigns.
> 	The Egyptian Hieroglyphic writing was certainly used in Pre- 
> Dynastic times. The royal tomb U-j at Umm el-Qa'ab near Abydos in  
> Upper Egypt, dated to c 3200 BCE, contained 150 inscribed bone tags  
> originally attached to grave goods recording the places of origin of  
> these goods, as well as pottery inscriptions and sealings. These  
> were excavated in 1988 and published ten years later (Dreyer 1998).  
> This earliest form of Egyptian script was already a well-formed  
> logophonic writing system, which can be partially understood on the  
> basis of later Egyptian writing. "By the early 1st Dynasty, almost  
> all the uniconsonantal signs are attested, as well as the use of  
> classifiers or determinatives, so that the writing system was in  
> essence fully formed even though a very limited range of material  
> was written." (Baines 1999: 882). "Many inscribed artifacts are  
> preserved from the first two Dynasties, the most numerous categories  
> being cylinder seals and sealings, cursive annotations on pottery,  
> and tags originally attached to tomb equipment, especially of the  
> 1st Dynasty kings. Continuous language was still not  
> recorded" (Baines 1999: 883). Thus until the beginning of the Old  
> Kingdom starting with the 3rd Dynasty in 2686 BCE — for about 600  
> years equalling the duration of the Indus Civilization — the  
> Egyptians used a language-based, phoneticized writing system, but  
> did not write full sentences, only very short texts fully comparable  
> to the surviving texts in the Indus script. Early administrative  
> documents are assumed to have existed but have not survived (cf.  
> Baines 1999: 884).
> 	When defining the Indus script as logosyllabic, I noted several  
> constraints to be observed in its analysis: "the linguistic elements  
> that are expected to correspond to the signs are morphemes rather  
> than phonemes. Secondly, all of the morphemes pronounced in the  
> spoken Indus language may not, and are not even likely to, have a  
> counterpart in its written form. In the third place, all preserved  
> Indus inscriptions are very short, appearing on objects like seals,  
> which are not so likely to contain even normal sentences, with such  
> basic constituents as a verbal predicate or an object, let alone  
> complex sentences." (Parpola 1994: 89).  This was before Damerow  
> (1999) suggested the term 'proto-writing' for the earliest,  
> linguistically incomplete notations (cf. Houston ed. 2004: 11); on  
> these earliest writing systems see especially Houston ed. 2004.
> 	In my opinion Farmer, Sproat and Witzel (2004: 19 and 33) err when  
> they suggest that "the Indus system cannot be categorized as  
> 'script' ... capable of systematically encoding speech", and that it  
> "cannot even be comfortably labeled as a 'proto-script', but  
> apparently belonged to a different class of symbols."  Their  
> principal arguments, the shortness of Indus texts, their restriction  
> to only a few text types, and the long duration (c 600 years) of  
> this stage of script evolution, are effectively annulled by what is  
> said above about the early Sumerian and Egyptian scripts. For their  
> other arguments I refer to an earlier paper of mine (Parpola 2008).
> 	George Hart wrote yesterday (13 July 2011): "None of this proves or  
> disproves that the fish symbol might have been pronounced [in  
> Dravidian]  mīṉ. Steve Farmer wrote in reply (13 July 2011):  
> Probably one of the silliest claims ever made about the symbols,  
> with no evidence whatsoever to back it. My reply: there is actually  
> a lot of evidence to back it (see Parpola 1994: 179-272; and new  
> evidence in Parpola 2009). Due to a complete lack of bilinguals, it  
> is very difficult to verify sign interpretations, but not altogether  
> impossible. Perhaps the most important test stone is supplied by the  
> nominal compounds actually existing in languages that are  
> historically likely to be related to the Harappan language: these  
> can be compared to Harappan sign sequences that can be pictorially  
> interpreted and perhaps deciphered with the help of linguistically  
> acceptable homophonies (used in all early scripts for phonetication:  
> the rebus puns). The accumulation of iconically acceptable,  
> systematic and interconnected interpretations can eliminate chance  
> coincidences in a process comparable to filling cross-word puzzles.
> References:
> Baines, John, 1999. Writing: invention and early development. Pp.  
> 882-885 in: Kathryn A. Bard (ed.), Encyclopedia of the archaeology  
> of ancient Egypt. London and New York: Routledge.
> Dreyer, Günter, 1998. Umm el-Qaab I: Das prädynastische Königsgrab  
> U-j und seine frühen Schriftzeugnisse. (Archäologische  
> Veröffentlichungen 86.) Mainz: Verlag Philipp von Zabern.
> Farmer, Steve, Richard Sproat and Michael Witzel, 2004. The collapse  
> of the Indus-scrpt thesis: The myth of a literate Harappan  
> civilization: Electronic Journal of Vedic Studies 11 (2): 19-57.
> Houston, Stephen (ed.), 2004. The first writing: Script invention as  
> history and process. Cambridge: Cambridge University Press.
> Michalowski, Piotr, 1996. Mesopotamian cuneiform: Origins. Pp. 33-36  
> in: Peter T. Danies & William Bright (eds.), The world's writing  
> systems. New York: Oxford University Press.
> Parpola, Asko, 1994. Deciphering the Indus script. Cambridge:  
> Cambridge University Press.
> Parpola, Asko, 2008. Is the Indus script indeed not a writing  
> system? Pp. 111-131 in: Airavati: Felicitation volume in honour of  
> Iravatham Mahadevan, Chennai: Downloadable from
> Parpola, Asko, 2009. 'Hind leg' + 'fish': Towards further  
> understanding of the Indus script. Scripta 1: 37-76. (Downloadable  
> at

More information about the INDOLOGY mailing list