Harappan 'non-texts'?

Tue Jul 4 16:21:08 UTC 2000

Steve Farmer has done it again...

And, again, he has not endeared himself to certain quarters on this list,
just as he has done on THE OTHER LIST (writing on spoked-wheel chariots,
which make the Rgveda clearly a 2nd mill. BCE phenomenon).

Again, we may notice how insulated and isolated some sections of
Indology/Indian Studies really are... and how often we extrapolate --
rotting manuscripts -- from what we (think we)  know (from modern medieval
and ancient India, as the case may be).

Comparison certainly opens our eyes, even if certain areas of the world may
have unique peculiarities.

In the present case, Indus Script/Writing, S. Farmer's comparisons  would
make the Indus  script and civilization unique in any 3rd mill. BCE  global
perspective... This has, indeed, been argued for certain aspects of its
culture (from the point of archaoelogy only!),  -- such as 'lack of cult of
personality'  (The "priest -king" and similar items seem to be very late
and Bactria-Margiana-influenced/imports!), the lack of palaces, etc.

This kind of uniqueness cannot be maintained for invention the script, as
it indicates similar "building blocks' as those that  have been used, more
or less independently,  in the earliest writing of Egypt, Mesopatamia,
China (and, as Indus script even seems to have some
similarities/connections with proto-Elamite writing).  However, as he
underlines, the USE and development of the Indus script are quite another
matter (see below), and the Indus civilization stands out here, again.

Yet, I maintain: in spite of what has been written about the script (see
Greg. Possehl's 1996 great index of "failed attempts" of decipherments!),
we have to start from scratch, and that includes the most recent attempts,
not yet recorded in Possehl. For the simple reason that all previous
attempts are based on the wrong premises and  materials.

We have discussed that at the 2 previous interdisc. Harvard Round Tables on
Central and South Asian Ethnogenesis in May 1999 and May 2000, and I draw,
subsequently,  especially on the contributions and the book by
Brian Wells, An Introduction to Indus writing. Second edition. Early Sites
Research Society (West). Monograph Series. Number 2. 1999   <ESRS West, Box
4175, Independece MO 64050>

Wells has studied both the Maya and the Indus scripts in great detail (as
well as Near Eastern contemporary developments) and therefore brings in a
lot of comparative expertise.

Briefly, from my perspective:

1) the number of characters (signs, mostly logographs, it seems) is not
close to 400, as most researchers maintain, but rather close to 600 (or
even 900 in his latest analysis).  This is not based on simple scribal
variants but on context:  e.g., on the  2-sign groups ('collocations'  of 2
graphemes)  that  one character (including its scribal variants)  forms
with another character (again including its variants).   - [For the fans of
an Indus alphabet:  what is the sign, occuring just once, of the "duck in
pond' ??]

2) the existing sign lists, therefore, are misleading, and

3) the concordances, therefore, are misleading as well. They list, say
(hypothetical case) the  sign *213  together with *465  and *505 as ONE
sign, as scribal variations, while the study of  collocations indicates
that they are used differently, and are, in fact, 3 separate signs. Nobody
can use the concordances without being mislead in the "decipherment":
Rta ain't artha or an-Rta, and  a-dharma ain't dharma  or gharma either.

4) there is some  historical development, disregarded so far

(5) and there is quite some is regional variation, e.g., between Harappa
and Mohenjo-daro, also disregarded so far.

In short : On what are all the analyses and 'decipherments', and all those
strong opinions  based???    Back to square one.

Details:

ad (5) Regional variation
To begin with the last, an always overlooked item: For ex., Harappa uses
the simple V-like sign, probably/but unproved meaning  "a measure of",  in
small tablets in combination with obvious numbers:   III  V,    IIII V
for example  "give me/here are 3/4 measures of.." (No one can tell that
this is the correct  interpretation, of course!) But this sign  is hardly
used at Mohenjo Daro (c. 8% only, by trade?).
Simple V occurs, barring the most recent excvations at Harappa etc., 194
times, out of which 175 x at Har., 16 x at Mohenjo-daro, and 3x elsewhere.
What does that teach? See for example at
http://www.harappa.com/indus2/index.html no. 149.

Note that this sign is different from the most common one (855x), a sort of
[genitive(??)] suffix:
=  =
  V             i.e. famous vessel with rims.

There are other signs with the opposite distribution, e.g. the simple 'man'
 sign   looking like
    |
  /|\
    |
   /\
It occurs 36 times at MD, only 7 x at Har., and 4 x elsewhere. Obviously
these characters are not uniformly used, to say the least. They could also
represent local traditions, different concepts, dialects, or even different
words of a different language altogether (if these 2 Indus languages  had
the same/very similar syntax! Ex.:  Chinese char. as used in Korean and
Japanese).

This regional variation again speaks for S. Farmer's concept of signs as
indicating personalities etc.

ad (4) Development:
So far, we could say very little about the development of the script: early
excavations did not always record the exact find spots & levels(!)
properly. Recent excavations at Harappa have reached "rock bottom", virgin
soil, and have  established a sequence for the script from the 4th mill.
onwards. We now see that the earliest Indus signs appear already at 3300
BCE and that they are different from  the (also used) (simple) potters'
marks (such as a cross, + , etc.). Some of the characters  look similar,
though not exactly the same, as in those of the much later Har. period,
2600-1900. Others disappear (for example, the star * symbol!
So much for the fish sign = "star"? and, as maintained recently by some,
an alphabet in 3300 BCE?? Another first!)

For recently found examples of early Indus writing AND potter's marks, see
harappa.com :
http://www.harappa.com/indus2/index.html  esp. photo 124 sqq. and cf. the
later, no. 131
http://news2.thdo.bbc.co.uk/hi/english/sci/tech/newsid_334000/334517.stm
http://sarasvati.simplenet.com/meadow/meadow.htm
= BBC, with some mistakes:  I have posted R. Meadow's corrections earlier
on INDOLOGY...

S.Farmer:
>1. Harappan script ...<according to specialists, so far>   shows unusually
>little internal development from c. 2600 BCE to c. 1900 BCE...
>This is *sharply* at odds with what is known of the
>development of logo-syllabic systems in societies with fully developed
>literate traditions

This observation still holds to a large degree. Note that some signs appear
already in 3300 BCE, and have been used throughout.  For example
http://www.harappa.com/indus2/index.html no.132, from a slightly later
phase than 33300 BCE, but still pre-Har.Civ., seems to represent the most
common sign, the rimmed vessel, mentioned above.
So what does that tell us?  Word or suffix? (which it could be here,  but
it's a fragment). Suppose it was a suffix, then: did the language not
change from c. 3000 to 1900 BCE? --   Another First!
(NB: Chinese characters did change several times, and not just in their
individual shapes. That can be found in any Chin.101 book. Most glaring,
the simplicifation of the "radicals",  from, as I recall, c. 500-odd to
just 214!). The Chinese script of 300 BCE is not that of 300 CE! All of
this quite part from the fact that N/S Chinese differ even in the
'collocations' of characters that they use, because their languages differ
more than the Romance languegs from each other. -- Same in the Indus ??)

S. Farmer:
> ... 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

NB: the "mytholocial tablets" tell whole stories, IN PICTURES, (for whome?
the literate scribes??)  but these use the same old characters as the seals
and tokens... i.e. (most of) the signs will not refer to stories...
Vale,  written literature...

                                        ****

The discussion has so far neglected the actual use of inscribed materials.
I underline the present Har. excavator's, R. Meadow's, observation on the
life of 'seal's (whatever their use, they were for the most part not used
to 'stamp' documents, but were, maybe, more of an identification tag, as S.
Farmer thinks: many are found, lost 'on the street', with the bosses on
their back torn off by the action of the string they were carried on!).

Meadow underlines that most of the inscribed  items had a 'life' of their
own. At a certain moment they were discarded in rubbish, --whole sets of
(the same) tablet/seal sometimes-- which indicates that they were no longer
'valid' (owner had died, changed status, business had faltered -- you go on
imagining). No one throws away a royal/business seal in China/Near East...
Again, not 'copious writings', but rather use of script for some sort of
identifiction/token...

(ad 1) number and use of characters

This  agrees with S.Farmer's point:
>4. Moreover, recent evidence
>suggests that there were probably more independent symbols around than
>Parpola thought..... 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.

Indeed,  4/5 of all Indus signs occur just 10 times in more than 4000
inscriptions.  And nearly HALF of them just ONCE!
(Note : numbers of signs established on using collocation analysis!)  What
kind of script is this? Extremely logographic, certainly not a syllabary,
and by NO means an alphabet. Show me a language with 400, 500 or 800
sounds ...

Rather,  we have to agree with S.Farmer:

>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!);

Again, who, and for what literary purpose, would invent a special 'duck in
pond' sign? Used once.

                                        ****

All of the above only underlines how how little we know.

In short, we need, as always, a geographical and a temporal study of the
script, before speaking about an Indus language and any interpretation or
'decipherment' of the unknown Indus  language(s).

For the Greater Panjab, however, we have some,  in fact, the ONLY available
evidence of the local language(s) spoken in c. 1500 BCE,  in my  EJVS paper
http://www1.shore.net/~india/ejvs/issues.html   ( Sept. 1999):
we get some 300-odd  loan words in the Rgveda  which are neither Dravidian,
nor Elamite or Sumerian, nor it seems now, taken from the BMAC language.
But Real Indian, resident before Indo-Aryan, Drav., Munda, Burusho etc.
make their appearance in the Panjab. Cf. the Nahali, Kusunda languages
mentioned in my recent post about South Indian river names (hydronomy of
Tamil...).

What we need, urgently, to sort out things further is an etymological
dictionary of Panjabi and Sindhi. None is in sight.

-----------

========================================================
Michael Witzel
Department of Sanskrit & Indian Studies, Harvard University
2 Divinity Avenue, Cambridge MA 02138, USA

ph. 1- 617-496 2990 (also messages)
home page:  http://www.fas.harvard.edu/~witzel/mwpage.htm

Elect. Journ. of Vedic Studies:  http://www1.shore.net/~india/ejvs
-------------- next part --------------
A non-text attachment was scrubbed...
Name: attachment.bin
Type: text/enriched
Size: 12897 bytes
Desc: not available
URL: <https://list.indology.info/pipermail/indology/attachments/20000704/9597105f/attachment.bin>