The Indus Civilization
Page 24
This lack of data is not due to poor digging, since Mackay excavated several sites in the Near East, and he found many true seal impressions there, at Kish, for example. If seal impressions had been made at Mohenjo-daro he would have found them there, too.
Figure 7.4 A copper tablet from Mohenjo-daro (after Mackay 1937—38)
Another important observation regarding the function of the square stamp seals has to do with wear. Most of the Indus seals (not all) still retain the crisp edge of the original carving. They are unworn and clearly were not used to make many impressions, especially since the steatite from which they are carved is the softest on the Mohs’ scale of hardness, soft enough to be marked by a fingernail. Even when baked and hardened, this family of minerals registers as only about four on the Mohs’ scale. In contrast, the cylinder seals in Mesopotamia were intensively used and were consequently carved from hard stones: agate, crystal, lapis lazuli. There are many examples of Mesopotamian seals that were heavily worn from the taking of many impressions, even some that have been recut in antiquity to rejuvenate the scene and the written message. This fits well with the large number of seal impressions that have been recovered in archaeological excavations there.
There is a terra-cotta “seal” from Allahdino, complete with a boss on the back. The elephant and the script have been incised on the face in the manner of graffiti on pottery. The incisions are so shallow that they would leave little or no imprint on clay.
Some Indus seals have rather heavily abraded outside edges, especially those directly adjacent to the face. Joshi and Parpola have some excellent color photographs that nicely illustrate this.13 These are worn only on the edges; the interior carving retains a fresh, crisp line. This pattern suggests that the seal was wrapped in cloth (or another material).
Thus, the Indus seals were not used to make huge numbers of impressions, which in turn suggests that they may have been used primarily for visual identification. The seal itself, not the impression, was shown. It identified, or gave some form of legitimacy, to the bearer.
It is emerging that the Indus peoples and Mesopotamians used their seals in significantly different ways. The kind of text and seal administration found in the Mesopotamia archaic state was not a part of the economic fabric of the Indus. With the growing consensus that the Indus Civilization was not organized as an archaic state,14 the notion that a seal-based administration was used in the Indus political economy also comes into doubt.
The usual typology includes many objects in the “seal” category because of their form, not their function. At the moment, it makes sense to allow them into the domain of the seal and let the category of function expand to accommodate them.
The Indus Unicorn: A Real or “Fantastic” Animal?
There has been some discussion of the one-horned “bull” on the seals, usually called a “unicorn.” The question centers on whether this is a real animal, with the single horn shown for purposes of artistic perspective, or is it fantastic, a creation of the Indus mind? Mahadevan’s concordance lists 1,159 seals with this device, so it must have been important. 15 C. Grigson is a recent proponent of the fantastic animal hypothesis, which is not without merit.16 But most others, including Marshall and Mackay, agree that “the artist intended to represent one horn behind the other.”17 This point of view is supported by the fact that there are a few seals with unicorn-type bulls with two horns.
On the other hand, the original reference to the Indian unicorn was made by Ctesias the Cnidian, a Greek physician resident for seventeen years (c. 415—398 B.C.) in the Achaemenid court under Darius II and Artaxerxes Memnon. 18 There are also terra-cotta figurines that look like bulls with one horn coming out of the middle of the forehead from Harappa, Chanhu-daro, and Lothal.19 In the end, I believe that the single horn on the unicorn seals was an artistic convention. This implies that the animal is probably a real bull and that the unicorn of India as reported by Ctesias is a separate matter.
A Typology of What Is Inscribed
Indus inscriptions are found on a variety of objects, none of them lengthy texts. Since the context of Indus writing is an important element in decipherment, a typology of inscribed objects is needed, as in table 7.3.
Indus Pictographs
The Indus script is usually thought of as pictographic. There is a “fish” sign , a “rake” T, a “house plot” , a “diamond” ◊, and others. The Mesopotamian writing system also passed through this stage. Since these pictographs are so simple, there is, not surprisingly, some correspondence between their pictographs and those of the Indus system (figure 7.5).
The Proto-Elamite, or Linear Elamite, script of the Iranian Plateau is also pictographic and roughly contemporaneous with the Indus script. It has been found in eastern Iran, as near as Tepe Yahya and probably Shahr-i Sokhta. There is even better correspondence between this script and that of the Indus Civilization than there is between Sumerian and Indus. Unfortunately, Proto-Elamite is also undeciphered, although the counting system on the tablets is understood. The correspondence between the pictographs of Proto-Elamite and Indus script is close enough that G. R. Hunter noted: “That the languages (Proto-Elamite and Indus) are unconnected is probable, and the phonetic value of the signs may well be different. But that they are unrelated in origin seems to be contradicted by the number of resemblances that seem to be too close to be explained by coincidence.”20 Fairservis once did a transliteration of a Linear Elamite tablet into the Indus writing system, further documenting the close correspondence between the two.21
Table 7.3 Typology of inscribed material
I. Seals
I.A. Stamp seals with an Harappan-type boss
I.A.1. Standard stamp seal with an Harappan boss
I.A.2. Stamp seals with an Harappan boss and geometric designs
I.A 3. Narrative stamp seals with an Harappan boss
I.B. Harappan bar seal
I.C. Cylinder seals
I.D. Round seals of Persian Gulf type
II. Seal impressions
II.A. Standard seal impressions on clay
II.B. Seal impressions on pottery
III. Sealings and moldings
III.A. Sealings and moldings with only writing
III.B. Narrative sealings and moldings
IV. Inscriptions that are not seals, sealings or moldings
IV.A. Miniature inscriptions
IV.B. Copper tablets
IV.C. Inscribed copper artifacts
IV.D. Inscribed ivory and bone objects
IV.E. Graffiti
V. Miscellaneous inscribed material
The Total Number of Signs
While there is substantial disagreement on the total number of Indus signs, all experts agree that it is in the range of a few hundred. This number indicates that the writing system was not like our twenty-six-letter alphabet or as expansive as Chinese and Japanese writing, with thousands of logographic characters. This strongly suggests that the writing system was based on syllables, or something akin to them, and was neither alphabetic nor logographic. Virtually all scholars who have studied the Indus script agree on this point and have used the term logosyllabic to describe it.
The exact number of Indus signs, however, is an unresolved question. This is principally due to the fact that what could be considered one sign is open to interpretation. No one is quite sure which are variants and which are signs unto themselves.
The number of pictographic signs is important for several reasons. It sets the limit on the number of signs to be deciphered as well as determines which signs might be related and therefore have similar meanings or functions in the writing system. It also plays a role of fundamental importance in an examination of sign frequency, which can be a source of many clues and productive lines for research for the would-be decipherer. The thinking of some scholars is reflected in the number of signs they assign to the total size of the script, as given in table 7.4.
The Length of the Inscriptions
I
ndus inscriptions are usually short; they vary from isolated signs to as many as twenty-six characters in length. The longest inscription comes from two molded terra-cotta “bars” from Mohenjo-daro, both from the DK Area (figure 7.6). Inscriptions on one line are the most common, but they range up to seven lines.22
Sign Frequency
Since the number of signs and their frequency is important to most decipherment efforts, an example of where experts disagree on this matter is in order. The “diamond” and the “rounded diamond” were treated as separate signs by Langdon and Fairservis23 but as variants of one sign by Mahadevan and Koskenniemi, Parpola, and Parpola.24
Of the 417 signs given by Mahadevan, 113 occur once, 47 occur twice, and 59 occur fewer than five times 25 The “jar” sign is the most frequently used sign, occurring 1,395 times. Frequencies, as given by the Finnish team, for suffix signs appear in table 7.5
Table 7.4 Number of signs in the Indus script as estimated by various scholars
Langdon(1931b) 288
Hunter (1932) 149
Von Meriggi (1934) 270
Dani (1963: 16) 537
Koskenniemi and Parpola (1982) 396
Mahadevan (1977) 417
Fairservis (1992: 6) 419
Figure 7.5 Comparison of Mesopotamian and Indus pictographs (after Smith and Gadd 1924)
The “jar” sign is the only one that occurs more than a thousand times. It also has a very broad distribution, occurring at Rojdi as a graffito in the Posturban Harappan Rojdi C (c. 2000-1700 B.C.) and possibly as a design on the painted pottery of Cemetery H (see figure 9.9a). The “marker” l l sign occurs 649 times and is the second most frequently used. But this could be a sort of diacritical mark, and not a full sign in and of itself—another sign definition/ counting problem. Sixty-seven signs account for over 80 percent of the use. Of Mahadevan’s 417 signs, about 200 could be said to have been in general use.26
Table 7.5 Sign frequency according to the Finnish team
Total frequency of a sign Total number of signs with this frequency or higher
Over 800 1
Over 200 6
Over 100 24
Over 50 46
Over 20 86
Over 10 100
Note: After Parpola et al. (1969: 9).
The Origins of the Indus Script
There is little conclusive evidence for the development of the writing system in the Greater Indus region; although there is new evidence emerging from the ongoing excavations at Harappa. It is reasonable to assume that it arose from the same set of general culture processes that brought urbanization to the plains of the Greater Indus Valley.27 The Early Harappan-Mature Harappan Transition would be the time period within which the writing system developed from isolated signs to a system of symbolism that involved multiple characters. B. B. Lal has attempted to develop a sense of correspondence between the signs used in the writing system and the potters’ marks and painted designs on pottery from the Early Harappan Phase.28 It could be expected that the inventors, or codifiers, of the writing system drew on parts of their own culture when they turned their attention to writing. Lal’s position is interesting, but falls short of being compelling. In spite of observations from Harappa, there is not yet a convincing bridge that allows us to see a progression from the use of single signs to sign combinations, sign standardization, and the development of a system of representation that has an institutional rather than an individual quality.
Figure 7.6 Longest inscription
Prototypes for the square Indus stamp seal have been found at the site of Kunal as well as Harappa, as published on the Internet.29 These “proto-seals” seem to date to the Early Harappan-Mature Harappan Transition, which is just where they should be in terms of cultural history.
The apparent short period of development for the Indus script is a contrast to the history of writing in Mesopotamia, where cuneiform writing begins in the sixth millennium with three-dimensional counters or tokens. These slowly developed into the primitive texts of the late fourth millennium B.C.30
Lack of Change in the Indus Writing System
At the moment, it appears that the Indus script “arrives” full-blown, with the dawning of the Indus Civilization. As we learn more about the Early Harappan-Mature Harappan Transition this will change, and we will be able to trace the development of Indus writing within the 100 to 200 years of this transitional phase. With one or two possible exceptions, no one has detected any change or development in the writing system over the 600 years of the Mature Harappan. This is almost certainly due to the general lack of control over the detailed history of the Indus Civilization. It makes little sense to suppose that such a system was invented and then used for hundreds of years throughout the entire Indus region without any modification.
Two hints of change, not in the writing system itself, but in the inscribed objects, already exist. The first comes from Harappa, where Vats observes:
“Stamp seals . . . diminish in both size and numbers from the IVth stratum downwards in Mound F; and so also do the terracotta and faience sealings. Their place is taken by a class of very small seals and sealings which are not represented at all at Mohenjo-daro presumably because the strata exposed on that site are posterior to the age to which these early seals belong.31
The second piece of evidence comes from Mohenjodaro, where the copper tablets “were especially characteristic of the ‘Late Period,’ ”32 something confirmed by Franke- Vogt.33
B. M. Pande has also studied the evolution of signs and has produced some interesting suggestions, especially concerning the development of a sign that he derives from a quadruped .34Pande’s list significantly expands the corpus of copper tablets over that presented by Mackay.35
If the matter is given sensitive attention, evidence of change and variation n the Indus script will almost surely emerge. This is but one of many tasks that can be done with this writing system that does not necessarily involve the decipherment directly.
The Direction of Indus Writing
It is generally conceded that Indus script should be read from right to left in the impression.36 C. J. Gadd was the first to set forth this argument in detail.37 He noted that the Indus signs seem to face both to the right and the left, so the rule followed in Egyptian hieroglyphics that one reads into the face of the symbol is of no help. Of more significance, Gadd pointed out that on some seals signs (e.g., H 5629) seem to be cramped, as the engraver ran out of room (see figure 7.3). There are a number of other seals with this characteristic.38
Evidence for boustrophedon (the way an ox turns in plowing) writing has been noted by several scholars.39 One of the seals published by Marshall is of this general type with the “jar” sign at the end: .40 Seal number M 247 is also written in this form .41
Other important observations on the direction of writing have been drawn from split-sign sequences and one seal with writing on three edges on the face. The manner in which the strokes forming a character overlap, especially on unbaked pottery, is also a useful guide. The arguments for the direction of writing are covered in Possehl.42 In the end, we know that there is an occasional inscription written from left to right (in the impression). The boustrophedon form is present, but most Indus writing was to be read from the right to left, in the impression, and this feature of the Indus writing system is one of its the most securely known.
What You See Is What You Get?
A curious feature of the Indus script is that the writing system does not seem to have been expressed in texts or economic documents. The small “formulae” of the seals and sealings are all perfectly fine examples of ancient writing, but there are no lists, ledgers, letters, or law codes. It may have been that the Indus scribes used a perishable material, such as cloth or bark, to compose longer documents; both materials served as paper in historical times.
There is always a chance that a large corpus of long inscriptions in the Indus script will emerge, perhaps on imperishable materials. Bilingual text
s might also appear. But after a lot of digging there is no evidence for them today, and this may suggest that we know the principal cultural contexts of Indus writing, which is quite unlike Mesopotamia and Dynastic Egypt. The context for early writing in these civilizations, or archaic states, was administration. In the Indus it seems to be personal identity.
Professor J. Baines, an Egyptologist at Oxford University, made an intriguing observation pertinent to Indus texts written on perishable materials.43 There are many unique and low-frequency signs in the corpus of inscriptions on nonperishable materials, and Baines noted that this may well be evidence that the inscribed materials that have been recovered are sufficiently specialized that the balanced expression of the Indus corpus of signs is not present. If this is true, then it implies that at least some of the unique and low-frequency signs that are recorded are simply a reflection of the deep bias on the corpus of texts and that if we had the whole range of texts (e.g., those on perishable materials), this statistical point would vanish. A real writing system with 113 unique signs, or one where over half of the pictograms occur less frequently than five times, is odd. The point that Baines makes is not a complete answer, but I think it is important and should be kept in mind by students of this ancient script.