Tuesday, November 18, 2014

Merchant of the City or Master who Collects Wind?

Detail from seal H-12 showing inscription over unicorn: CIRCLED VEE / BI-QUOTES /
ZEE / CROSSROADS EX / POT (the last 4 signs are the focus of this blog post).

Iravatham Mahadevan recently published an article purporting to demonstrate that the Indus script represents a Dravidian language (2014).  His argument focuses on the interpretation of a single “phrase” or four-symbol series which I term PRAWN / ZEE / CROSSROADS EX / POT.  According to the concordance of Koskenniemi and Parpola, this sequence occurs at least 23 times, appearing in inscriptions from Mohenjo daro, Harappa, and Kalibangan (1982: 53).  As Mahadevan notes, this sequence can form the whole inscription (M-857 is cited but it also occurs on M-455 and M-950).  More often, this is part of a longer inscription (M-377 is depicted but additional such inscriptions include M-38, 369, 425C, 525-6, 626, 671, 682, 1061 [?], 1156, 1206, 1474-81, 1548-9, 1667, 1691; H-12, 61, 358, 703B [?],834; K-10; KP 2424, 7064E, 687).  In addition, there are two inscriptions that may contain the sequence but the final sign is illegible or broken away (H-579 and KP 6058).

Indus seal H-61 with inscription: FIGURE EIGHT WITH LADDER / PINCH /
Mahadevan’s interpretation of the sequence is “merchant of the city” (2014: 1).  The method he used to arrive at this conclusion is (1) identify what each sign represents, as if it were an ideograph; (2) find a word for the literal meaning in the revised Dravidian Etymological Dictionary (referred to as DEDR; Burrow and Emeneau 1984); and (3) determine the intended meaning by comparing homophonous words in the same source on the rebus principle.  This is an abbreviated version of the method proposed and used by Fairservis (1992: 23-24):
1.       Identify a given sign and what it represents;
2.       Find an equivalent word in the extant Dravidian languages;
3.       Choose the most suitable word, one with wide usage due to historical inheritance (not a loan from another language family);
4.       Consider both the sign’s substantive value (i.e., the literal meaning) and the chosen word’s phonetic value (assuming the rebus principle was in use);
5.       Ensure that values chosen for the signs hold true in texts other than those first considered;
6.       Ensure that orthographic variation is accounted for.

Interestingly, Mahadevan’s method does not seem to take into account the last two steps of Fairservis’ technique.  Mahadevan explicitly states that he has not deciphered the Indus script as a whole (2014: 40).  I will return to this point later.  As for Fairservis’ step 6, presumably Mahadevan took variants into account in producing his earlier works on the Indus script, including his own concordance (1977).

Detail from Indus seal M-38 with inscription over unicorn: TWO POSTS / BLANKET / MAN WITH DEE-SLASH /
CROSSROADS EX / POT (note the asymmetrical attachments at the top of the PRAWN, which Mahadevan
interprets as a wolf's ear {left side} and braided or knotted hair {right side}).

As his first step, Mahadevan identifies the initial sign in the sequence (my PRAWN) as a wolf as seen from the back, with one ear covered by braided and knotted hair, a feature indicating “the anthropomorphic character of the sign” (2014: 3).  This is a surprising interpretation, given that Fairservis identifies the same sign as a prawn (Fairservis 1992: 158; C-1, “a prawn {shrimp},” Dravidian root irā {DED No. 440}, homophones meaning “food; master”).  Elsewhere, this symbol is seen as a scorpion (e.g., Marshall 1931 & 2004: 450; an interpretation also found widely among claimed decipherments, as in Richter-Ushanas 2001: 48 & 184). 

If PRAWN were indeed a scorpion, this would be a problem for those proposals relying on the DED/DEDR for Dravidian, since there does not seem to be a word for this critter that is common to most of this language family.  “Scorpion” appears in the DED as entry 2362 (Tamil ñaṇṭu and cognates in 12 out of 18 languages, in only one of which it means “scorpion” – the others mostly mean “crab”), 2409 (Kolami tiṭor and cognate in only one other language), 2855 (Tamil tēḷ and cognates in 8 other languages), and 3672 (Telugu poṭṭi and cognate in only one other language).  In contrast to the paucity of Dravidian cognates for “scorpion,” we may recall the often cited FISH.  According to the DED, the Tamil word n “fish” has cognates in 14 other languages, out of total of 18 languages referenced (entry 3999).  As Parpola notes, this is one identification accepted by a number of scholars, a rare occurrence in this contentious field (1994 & 2009: 277).  Even so, Fairservis did not agree with it, identifying the FISH instead as a twisted thread (my LOOP) plus an affix of two strokes (1992: 185).  As Parpola also observes, “Many of the signs of the Indus script are so simplified and schematic that it is very difficult to understand their pictorial meaning unambiguously and objectively” (1994 & 2009: 278).  One cannot, then, simply accept Mahadevan’s identification of PRAWN as a wolf seen from the back.

Indus seal M-626 with inscription: FAT EX / PINCH / PRAWN / ZEE / CROSSROADS EX / POT /FAT EX /

Mahadevan cites the earlier work of G.R. Hunter as evidence for his proposal, although Hunter saw the symbol as a jackal (2014: 3).  Mahadevan states, “The animal is more likely to be the wolf as it looks larger and its tail is curled up.”  I am uncertain what he means when he says it “looks larger,” as the signs of the Indus script are all roughly equivalent in size.  But if one focuses on the curl of the tail, one is as likely to conclude that the animal is a domestic dog as either a wolf or jackal. 
The real problem with Mahadevan’s identification, however, is that it is hard to find a parallel anywhere for an animal shown in perspective from the back.  Animals are far more often depicted in profile in ancient scripts and indeed in much ancient art.  In Egyptian hieroglyphs, for example, Gardiner’s list includes 34 different animal depictions.  All are profile views.  The Egyptians also used depictions of various parts of animal bodies.  These include glyph F4, the front end of a lion – head and attached front leg – found in the word ḥ3t “front”; and F22, the back end of a lion found in pḥwy “hindquarters” (Gardiner 1976: 462 & 464).  Among the glyphs there are also various animal heads, legs, an ear, a jawbone, a tail, a heart, and so on.  But there is no parallel to Mahadevan’s proposed view from the back.

Egyptian hieroglyphic text from the tomb of Nefertiabet, showing
Gardiner's glyph F4, the front part of a lion, in the upper register.
Note also the owl in the upper register and the chick in the lower,
both in profile, the standard depiction for people & most animals.

Proto-cuneiform often makes use of a sort of graphic abbreviation, with only the head of an animal depicted, as for example dara3, a wild goat or mountain goat, pirig, a lion, and šaḫ2, a pig (CDLI sign list online).  But where the whole beast is depicted, as for aquatic fauna and birds, profile views dominate (e.g., mušen, bird, and sumaš, a marine fish). 

Proto-cuneiform sign ZATU 710, of unknown meaning, depicting a quadruped in profile,
rotated 90 degrees (so that the legs are to the right) as is typical.

As a third example, Chinese may be considered.  Early Chinese writing, as found on so-called “oracle bones,” was largely pictographic.  Here, too, animals are typically depicted in profile, even where later writing seems to show them from some other viewpoint.  For example, the modern quan3, “dog,” might be described as a flat rug of a dog with the legs spread-eagled.  But in the ancient writing, it was a profile view; one of the modern strokes originates from the tail, not a limb (Keightley 1985: 218).  According to Wieger, though, the character for sheep, yang2, is the animal “seen from behind,” with diagonal strokes representing the horns at the top, and horizontal strokes beneath this for the limbs (1965: 253).  A better way of describing this depiction is to call it a bird’s eye view.  This type of depiction is quite common – see, for example, the bird’s eye view of a lizard in rock art from various different continents (Le Quellec 2004: 64, for Africa; Bernardini 2009: 32, for North America; Layton 2009: 157, for Australia).

Old Chinese characters for sui, "pig," as found on oracle bones (top)
and old seal writing (below).  The modern form is still more schematic.

Thus, I find Mahadevan’s identification of PRAWN as a wolf seen from behind to be most doubtful.  Unfortunately for his interpretation of the four-sign sequence, the literal meaning of this first sign is vital.  Because it is the back of the wolf, in his view, the literal meaning is “back” or “to turn back,” which he specifies as DEDR entrees 4761 (Tamil mari ‘to turn back, turn about’) and 4834 (Tamil ru ‘to become changed, exchanged, retreat {as showing one’s back}’).  Then, applying the rebus principle, he gives the intended meaning (or extended meaning) as DEDR entry 4834 (Tamil ru ‘exchange of goods, barter, sell’).  Extending the meaning a bit further, he argues that PRAWN means the person who barters or sells, i.e., ‘merchant’ (2014: 5).

American rock art from Texas, showing a probable lizard (lower left) shown
in bird's eye view, alongside a possible mountain goat in profile (after
Newcomb & Kirkland 1967 & 1996: 192, Pl. 142).

If PRAWN is not a wolf, it would not mean ‘merchant’; if it not a view from the back, it cannot have that meaning either.  But for the sake of argument, let us assume this is correct and continue with the second sign, my ZEE.  Mahadevan identifies this as a depiction of a hook, which yields DEDR entry 2151, koḷ- ‘hook,’ and through the rebus principle the sign takes the meaning of the homophonous koḷ - ‘take, receive, buy; one who takes, receives, buyer’ (2014: 6).  To convince the reader of his initial identification, Mahadevan includes a photograph of a copper fishhook from Khirsara (fig. 5, p. 5).  This, he states, “offers a close parallel.”  Since the actual hook has a smooth and curving S shape, while the symbol ZEE is quite angular, it is not actually close, if a parallel at all.  But again, let us assume this is correct for the time being.  Fairservis makes the same proposal, after all (1992: 169).
The third sign is my CROSSROADS EX.  The meaning of this sign is obvious to Mahadevan -- it depicts actual crossed roads (2014: 7).  He cites a similar Sumerian sign, the proto-cuneiform symbol that came to be kaskal, “expedition, caravan; road, course; journey” (for the symbol see CDLI, for the meaning in Sumerian, see Halloran 2006: 136).  Mahadevan also cites the Egyptian hieroglyph of an outlined ex in a circle which is Gardiner’s sign O49, depicting a village with crossroads (Gardiner 1976: 498).  This glyph originated as an ideograph in the word niwt, “village.”  Usually it serves as a determinative in names of villages, towns, and inhabited regions, as at the end of Kmt, literally “the black land,” i.e., Egypt.  Mahadevan interprets the Indus CROSSROADS EX as an ideograph for DEDR entry 4064 (Tamil pāṭi ‘town, city, hamlet, pastoral village’) or for DEDR 5297 (Tamil vali ‘way, path, road’) (2014: 8).  The intended meaning becomes ‘resident of a city,’ by extension of the basic meaning of DEDR 4064 or by the rebus principle for DEDR 5297 which sounds like DEDR 5372 (Tamil vāl ‘to live, flourish’) (2014: 9).

Proto-cuneiform KASKAL, "road, journey, etc."
Cf. Indus CROSSROADS EX in inscriptions above.

Although he does not state this explicitly, Mahadevan’s reasoning seems to be that, because the protocuneiform kaskal, which resembles the Indus sign graphically, came to mean “road,” and the Egyptian glyph niwt, also similar graphically, can mean “city,” then the Indus CROSSROADS EX must mean either “road” or “city” and, by extension, “resident.”  However, one is on very shaky ground when using the signs of one civilization to interpret the symbols of another, especially in the absence of evidence of a link between the two. 

Egyptian glyph niwt, "village."  Cf. Indus CIRCLED EX.

Now, there was some kind of contact between the Indus civilization and Mesopotamia, as Harappan seals and trade items such as carnelian beads and lapis lazuli have been found in Mesopotamian cities (Possehl 2002: 221-222).  But there is no evidence of direct contact between the Harappans and Egypt.  Further, although it is possible to find parallels between symbols of the Indus script and protocuneiform, or Indus script and Egyptian hieroglyphs – or, indeed, between the symbols of any two writing or protowriting systems – this in itself is not proof of contact.  As noted in previous posts, the circled cross or circled ex is extremely common around the world, but the meaning of this ubiquitous symbol varies from place to place and over time.  In ancient China, the circled cross represented a field; in Egyptian hieroglyphs a village; in protocuneiform a sheep.  Thus, the meaning of a given symbol in (proto)writing system A does not necessarily reveal the meaning of the same or a similar symbol in (proto)writing system B.

Egyptian glyphs Z10, crossed sticks for "break," and Z11, crossed planks for imi.
Compare the somewhat similar Indus sign CROSSROADS EX.

But suppose that, for the sake of argument, we accept the validity of Mahadevan’s premise, that the proto-cuneiform kaskal “road” or the Egyptian niwt “town” does reveal the meaning of the graphically similar Indus sign.  In such a case, it seems to me that we face a conundrum.  If the appearance of kaskal in proto-cuneiform demonstrates that the Indus CROSSROADS EX means “road,” as Mahadevan proposes, then the CIRCLED EX and CIRCLED FAT EX should also be interpreted by means of proto-cuneiform.  In that case, it must mean “sheep” rather than “city.”  If, instead, we ignore the proto-cuneiform “sheep” and turn to Egyptian niwt, we derive Mahadevan’s meaning of “city” for CIRCLED FAT EX, but then CROSSROADS EX should no longer represent roads.  Instead, the latter sign has two possible parallels in Egyptian: Gardiner’s Z10, an ex shape formed by two crossed sticks, or Z11, two planks joined, forming a cross.  Unfortunately for Mahadevan’s proposal, Z10 functions most often as a determinative for words having to do with breaking; Z11 is phonetic for imi, as in “not be” (1976: 538-539).  Neither suggests roads or cities.

If one can simply pick and choose among ancient writing systems to find parallels and thus meanings for Indus symbols, one could argue with equal justification that the Indus CIRCLED EX means “sheep” (based on proto-cuneiform) and CROSSROADS EX “break” (based on Egyptian).  For that matter, if there is no need to demonstrate contact between two cultures, what would prevent one from choosing yet another writing system for one’s source?  Luwian hieroglyphs include a parallel to the Indus CROSSROADS EX in sign 224, made with three parallel diagonals in each direction, which may represent the syllable ha or pa (Payne 2010: 177).  There is also a circled cross, perhaps a wheel (ROTA) among the logograms (Payne 2010: 181).  But, of course, the relatively late appearance of Luwian makes it a much less attractive source for inferring the meanings of Indus symbols than either Egyptian hieroglyphs or Mesopotamian proto-cuneiform.  Luwian hieroglyphs only appear in systematic usage on Hittite seals from the 14th century BCE, after the demise of the Indus script (Payne 2010: 2).

Indus tablet M-1475 with barely legible inscription (from right to left): PRAWN /

In any case, the final Indus symbol in the four-sign sequence is Mahadevan’s “jar,” which I term POT.  Because this most common of Indus signs often ends an inscription, Mahadevan (as well as a number of other researchers) considers it a grammatical suffix.  Specifically, he regards it as the pronominal masculine singular suffix (2014: 10).  Adding this interpretation to the previous ones yields, literally, “barterer + receiver + resident + he-of-the,” or “merchant of the city” (2014: 10).
Mahadevan’s article continues with various cultural items he associates with one or another of the Dravidian words that he connects with the Indus signs of this phrase.  At the end, he boldly states, “The quality and quantity of interlocked findings at the three levels described in the paper have transcended the level of mere evidence and attained the level of proof: the Dravidian proof of the Indus Script via the ŖgVēda!” (2014: 40; author’s emphasis).  I have noted enough problems with his analysis to show that he has not proved his case. 

Returning to Fairservis, who used much the same methodology, we could equally well interpret the same Indus sequence as “master (who) collects wind.”  PRAWN is a prawn for Fairservis, Dravidian irā, with irai as a near-homophone meaning “master” (1992: 158).  ZEE is still a hook, Dravidian koṭu, with a near-homophone kūṭu meaning “to collect (as with food)” (1992: 169).  EX is a cross or hook (Fairservis does not include CROSSROADS EX in his sign list), Dravidian gāḷa, with the near-homophone gāḷ meaning “wind” (1992: 162).  Finally, Fairservis proposes much the same interpretation for POT as Mahadevan, though terming it a third person singular honorific ending (1992: 173). 

So, how is PRAWN / ZEE / CROSSROADS EX / POT to be interpreted: “Merchant of the City”?  Or is it “Master (who) Collects Wind”?

1.       Bernardini, W. 2009. Hopi History in Stone: The Tutuveni Petroglyph Site. Arizona State Museum Archaeological Series 200. Arizona State Museum, the Univ. of Arizona: Tucson.
2.       Burrow, T. and M.B. Emeneau. 1998. A Dravidian Etymological Dictionary (originally published 1984 by Oxford University). Munshiram Manoharlal Publishers: New Delhi.
3.       Çambel, Halet. 1999. Corpus of Hieroglyphic Luwian Inscriptions. Vol. II. Karatepe-Aslantaş. The Inscriptions: Facsimile Edition. Walter de Gruyter: Berlin & New York.
4.       Cuneiform Digital Library Initiative (or CDLI) online at http://www.cdli.ucla.edu/tools/SignLists/protocuneiform/archsigns.html
5.       Gardiner, Sir A. 1976. Egyptian Grammar: Being an Introduction to the Study of Hieroglyphs (3rd ed.). (orig. pub. 1927) Griffith Institute and Ashmolean Museum: Oxford.
6.       Halloran, J.A. 2006. Sumerian Lexicon: A Dictionary Guide to the Ancient Sumerian Language. Logogram Publishing: Los Angeles.
7.       Keightley, D.N. 1978 & 1985. Sources of Shang History: The Oracle-Bone Inscriptions of Bronze Age China. University of California Press: Berkeley & Los Angeles.
8.       Koskenniemi, K. and A. Parpola. 1982. A Concordance to the Texts in the Indus Script. Department of Asian and African Studies Research Reports No. 3. University of Helsinki.
9.       Layton, R. 1992 & 2009. Australian Rock Art: A New Synthesis. Cambridge University Press: Cambridge, U.K.
10.   LeQuellec, J.-L. 2004. Rock Art in Africa: Mythology and Legend. Transl. Paul Bahn. Flammarion: Paris.
11.   Mahadevan, I. 1977. The Indus Script: Texts, Concordance and Tables. ASI, New Delhi.
12.   Mahadevan, I. 2014. “Dravidian Proof of the Indus Script via the Rig Veda: A Case Study” in Bulletin of the Indus Research Centre No. 4. Roja Muthiah Research Library: Chennai. Pp. 1-44.
13.   Marshall, J. 1931 & 2004. Mohenjo-Daro and the Indus Civilization: Being an Official Account of Archaeological Excavations at Mohenjo-Daro Carried out by the Government of India between the Years 1922 and 1927. (Orig. pub. 1931 in London) AES Reprints: New Delhi.
14.   Payne, A. 2010. Hieroglyphic Luwian: An Introduction with Original Texts (2nd rev. ed.). Harrassowitz Verlag: Wiesbaden.
15.   Possehl, G.L. 2002. The Indus Civilization: A Contemporary Perspective. AltaMira Press: Lanham, MD.
16.   Richter-Ushanas, Egbert. 2001. The Indus Script and the Ŗg-Veda. ( 2nd rev. ed.) Motilal Banarsidass Publishers: Delhi.
17.   Wieger, Dr. L. 1965. Chinese Characters: Their Origin, Etymology, History, Classification and Signification. (Orig. pub. 1915 by Catholic Mission Press). Transl. L. Davrout. Dover & Paragon: New York.