Pramāṇa · Search · About · API

The Indus Script: Texts, Concordance and Tables c4.s8: 8. The criteria of arrangement and related conventions of the Corpus

The Indus Script: Texts, Concordance and Tables c4.s8

Language
English

Contents

8. The criteria of arrangement and related conventions of the Corpus

General considerations

c4.s8.t1.1

Theoretically, the Indus seals and inscriptions could be classified in several ways. For example, the inscriptions could be arranged according to the pictographic sequences they contain. However, this arrangement would only serve the needs of scholars interested in the script and is better left to the concordances of the script. If the concordances are keyed to the Corpus, cross-reference and verification will be easy, whatever the principles of arrangement.

c4.s8.t1.2

Ernest Mackay, in *Further Excavations at Mohenjo-daro*, arranged the objects coming from Mohenjo-daro according to the different areas of the site and the absolute depth of the finding place from the surface.⁶⁸ He wished to control the data from the point of view of archaeological distribution, looking for evolutionary and other trends. The result was chaotic: objects of different types and sizes were mixed with each other. Unless one knows the number of the object, it is impossible to locate it without scanning through the entire material. In the present Corpus, the archaeological context is taken into account in the arrangement of the objects when it is feasible and useful: thus the objects from the Late Harappan period from Lothal (graffiti only) and Chanhujo-daro (seals) are presented as a separate section at the end.

c4.s8.t1.3

⁶⁸ On Mackay's "stratigraphy", cf. Jansen 1986 (fn. 3), 52-54.

c4.s8.t1.4

The aim of the classification must be efficiency in placing and locating any given object within the whole. The type of the object, form, material, iconographic motif, size, style and state of preservation have been chosen as parameters in the Corpus, in this order. A solution of this kind, which makes a neat layout possible, was followed by Sir John Marshall in *Mohenjo-daro and the Indus Civilization* and, less successfully, by M.S. Vats in *Excavations at Harappa*.

The 1st criterion: the owners of the objects; and the overall publication plan

c4.s8.t2.1

Ideally, of course, one would like to see all the objects coming from a single site, for example Mohenjo-daro, neatly arranged into one single sequence. There are, however, other considerations and realities, which have made it impracticable to realize this ideal. Instead, the Corpus is divided into three volumes according to the first criterion of physical location and ownership of the original objects. In this the *Corpus of Indus Seals and Inscriptions* follows the example of the *Corpus of Minoan and Mycenean Seals*, for instance, which is divided into different volumes according to the museums in which seals are preserved. The first volume of the Corpus presents the collections housed in the museums of India, the second volume the collections in the museums of Pakistan.⁶⁹

c4.s8.t2.2

⁶⁹ We want to emphasize that the order of the volumes is due simply to the fact that the Indian material first reached the stage when publication could be begun and has no political implications. In fact, the possibility of leaving out the volume numbers altogether in order to avoid the issue was considered, but then dropped as impractical.

c4.s8.t2.3

The third volume will contain the relatively few objects known to exist in collections outside India and Pakistan and the large number of lost objects, which are not directly documentable but must be published as old photographs only. Besides addenda to the previous volumes, this third (and for the time being last) volume will also contain a detailed catalogue of all the objects of the Corpus, documenting (in addition to the excavation and museum numbers, which are given separately in the first two volumes as well) such matters as the archaeological context, measures, notes on the material, manufacture, text and iconography, and published references. Furthermore this information will be fully indexed.⁷⁰

c4.s8.t2.4

⁷⁰ This arrangement has practical reasons. The first two volumes are bulky because of a large number of photographs, while the third volume will contain an essentially smaller number of photographs. Therefore, it has more space to accommodate both the lengthy catalogue and the indexes, which naturally should be cumulative. Even the museum indexes, which would have been handy in the first two volumes, have to be published in the third volume for these reasons.

c4.s8.t2.5

This first volume, then, contains 1537 Indus seals and inscriptions physically existing in public collections in India. We have excluded the objects stolen from the Prince of Wales Museum of Western India, Bombay, although photographs of these objects are available; they will be published as lost objects in the third volume. As far as possible, we have tried not to publish old photographs, but to procure new better ones. When originals almost certainly existing in Indian collections could not be located, however, we have resorted to reproductions. If better pictures of such objects are obtained later, they will be published in the third volume.

c4.s8.t2.6

The 2nd criterion: the provenance of the objects; and their numbering system

c4.s8.t2.7

It is clearly undesirable to lose control over the sitewise distribution of the objects; a purely typological arrangement mixing objects from all sites would be inadvisable. The site from which the object comes has to be a primary parameter of the classification. Now that seals and inscriptions coming from one and the same site will be distributed in several volumes, a flexible new numbering system is required which will both allow additions at will and make it easy to place the object in its proper context.⁷¹ The *Corpus of Indus Seals and Inscriptions* employs a separate consecutive numbering for each site, prefixed by a letter code which is more easily remembered than a numerical code. The major sites have a short, one-letter code. These sites are, moreover, arranged in each volume according to the total number of seals and inscriptions found at them, in the descending order. The sites which are "smaller" (in respect to the number of seals and inscriptions found at them) have a two-, three- or four-lettered code corresponding to their standard archaeological abbreviations and they are arranged in alphabetical order for easy reference. (See the table of contents.) The letter prefix for the site is followed by a dash and the number of the object assigned to it by its place within the classification sequence. Thus the objects from Mohenjo-daro in this first volume are numbered M-1 to M-620, and they are followed by the objects from Harappa starting with H-1. The objects from Mohenjo-daro in the second volume will start with M-621. Any number of additions can be made.

c4.s8.t2.8

⁷¹ In the recent editions and concordances, the Indus inscriptions from Mohenjo-daro and Harappa were keyed to the published excavation reports: a number code was allotted to each of these reports and prefixed to the consecutive numbers used for the objects in the plates of the respective report. This basic reference number system was then extended to cover the other sites as well and also the unpublished objects from Mohenjo-daro and Harappa found in museums. In principle, one could recognize the site from which any given object came from the first one or two digits of its four-digit reference number. However, this system has its obvious drawbacks and limitations. Mohenjo-daro required three separate first numbers: 1 for Marshall's report, 2 for Mackay's report, and 0 (in the Finnish concordance) or 3 (in Mahadevan's concordance) for the unpublished objects. The small sites required at least two consecutive first numbers, difficult to remember. And not only had this system become a bit complex, but it also started to run out of numbers. Cf. S. Koskenniemi, A. Parpola and S. Parpola, *Materials for the study of the Indus script I: A concordance to the Indus inscriptions*, Annales Academiae Scientiarum Fennicae B 185, Helsinki 1973, xvi; and I. Mahadevan, *The Indus script: Texts, concordance and tables*, Memoirs of the Archaeological Survey of India 77, New Delhi 1977, 30.

c4.s8.t2.9

The 3rd, 4th and 5th criteria: the object type, form and material; and the symbols in the page captions

c4.s8.t2.10

After the site, the next criterion of organization of the Corpus is the type of the object.

c4.s8.t2.11

Table 1 lists in order and explains the simplified symbols for the typological subcategories used in the page captions of volume one. Because this table simultaneously gives a convenient overview of the typological classification of the seals and tablets, the captions over each page are explained first in this context.

c4.s8.t2.12

The caption lists in order (1) the full name of the site and the numbers of the objects coming from it that are illustrated on the page; (2) the principal object type spelled out in letters; (3) simplified symbol(s) specifying the form of the object(s); (4) material (if metal), iconographic motif(s) and size class(es) expressed with Roman numerals. Occasionally, exceptional scaling or archaeological period is mentioned. Only one-line captions are used, and information that cannot be accommodated is dropped, starting from the last categories. The captions have been reversed on even-numbered pages, in order to place the first and most needed subcategories closest to the page number on the right.⁷²

c4.s8.t2.13

⁷² The decision to reverse the captions on even-numbered pages was perhaps not quite so felicitous, because they have become somewhat difficult to interpret: each subcategory forms an entity to be read from left to right, so two or more symbols of form class (each with or without information on the iconographic motifs on the left) are to be read from left to right on even-numbered pages.

c4.s8.t2.14

We have tried to keep the typological classification as simple and unambiguous as possible. Four broad categories are distinguished: (1) seals & seal impressions, (2) tablets, (3) graffiti on pottery and (4) miscellaneous. These main groups, which are functionally different from each other, are subdivided further according to formal criteria. The material of the object is taken into account next, but only in the form of a broad division into non-metal and metal (mainly copper or bronze)⁷³ objects, which are placed at the end of each class.

c4.s8.t2.15

⁷³ Most of the Harappan metal objects are copper rather than bronze; cf. D.P. Agrawal, *The archaeology of India*, Scandinavian Institute of Asian Studies Monograph Series 46, London 1982, 151. It is hoped that all the objects can be properly analysed in the near future, so that the results can be published in the detailed documentation of the third volume.

c4.s8.t2.16

Seals are the most important category of Indus inscriptions in terms of frequency, so they are placed at the beginning. The most common basic form of the Indus seals is square, which is placed first, and the next frequent form, rectangular, is placed after it. Within both forms, subcategories are distinguished.

c4.s8.t2.17

The square seal normally has a perforated boss at the back, which apparently served both for hanging the seal by a cord and as an aid in making the impression. This type is presented first, with the rare example of a metal (silver) seal at the end, followed by the exceptional seals of this category: those that have been inscribed on more than one side and those having a case (probably for an amulet) inside them. Next follow the square seals where the boss is absent: first perforated seals with one side inscribed, then perforated seals with two or more sides inscribed, and then the unperforated seals similarly subdivided. These seals without a boss share similar inscriptions and iconographic motifs with the ordinary seals having a boss, so they have been placed after them, before the seals with nothing but a swastika or some other geometric motif, although the reverse of these last-mentioned seals does have a perforated boss (usually smaller than the normal seals and undivided, see M-332).

c4.s8.t2.18

Imported foreign objects are usually placed at the end of each category; thus a fine Iranian square seal with a perforated undivided boss (M-353) is the last of the square seals of Mohenjo-daro.

c4.s8.t2.19

The normal type of rectangular seals has a profile that is straight on the front side and convex on the back side with a hole for the suspension cord going through the middle (cf. M-354 C).74 Whenever the side profile is rectangular, or the back has a boss similar to that of the square seals, this is shown by publishing the relevant side(s) (cf. M-407ff.).

c4.s8.t2.20

Other forms of seals are rare, and in most cases these forms have been inspired by foreign models, if the seal itself is not a foreign import (see above, chapter 2). The round seals of the Indus Civilization have a perforated boss of the same type as the square seals and differ in this respect from the round "Dilmun" seal (L-123). In the round seals of the Late Harappan period, the suspension hole goes through the flat body of the seal (cf. C-45 to 50). With regard to the cylinder seals, which come next, before the stepped seals, it has to be pointed out that two small cylinders from Harappa have been classified as incised tablets (H-368 and H-369).75

c4.s8.t2.21

The ancient seal impressions stand for the seals they were once made with, so they are placed next to the actual seals. A distinction is made between impressions on pots, which come first, and impressions on clay tags. Uninscribed sides of clay tags that have been attached to bales of goods are illustrated, if they bear significant traces of the package material. The tags have been arranged according to the number of seal impressions they contain, those with single impressions being placed first, then according to the iconography and the inscriptions of the seal impressions.76

c4.s8.t2.22

There is a large group of objects which we have lumped together and called, tablets. A basic distinction is made between stamped or moulded tablets, whose texts and iconography are in bas-relief, and incised or engraved tablets, whose texts are depressed. The incised copper tablets (placed at the end), so far found at Mohenjo-daro alone, can be divided into three groups according to their shape: square, rectangular and oblong (or long rectangular).

c4.s8.t2.23

Round tablets in bas-relief often bear a square seal impression on one side and are flat on the other side. These round 'tablets' are placed at the beginning, because they might also be classed as seal impressions;77 they may have functioned as tokens of identification, or 'passports' of representatives of the seal owners. Since some of the other tablets in bas-relief, too, may have been produced with the help of seals, these round tablets have not been separated from the rest.

c4.s8.t2.24

In both of the main categories, the embossed and the engraved, the tablets are subdivided firstly according to their form (and material) and secondly according to their iconography, size, and condition of preservation. We have tried to avoid form-based classifications that will lead to ambiguous cases and practical difficulties: thus, the class of rectangular shape includes both thin and thick tablets and evenly flat tablets as well as tablets that are slightly thicker at the centre than at the edges. Finer classifications have been proposed, but they are difficult to carry through in practice

c4.s8.t2.25

74 The arch of the back is usually smooth (as in the case of M-354), sometimes edged (cf. M-374 C), but as this distinction is often a question of degree, it is not systematically noted in the Corpus.

c4.s8.t2.26

75 The fact that the inscription has the normal direction of writing, from right to left, in the original cylinder but is reversed in the impression, is in itself not a sufficient proof for such a cylinder not being a seal, because the direction of writing has not yet been fixed in the early layers of Harappa. But the inscription in H-369 C connects this object with the vast majority of the 'tablets'.

c4.s8.t2.27

76 The Lothal tags with multiple seal impressions have been arranged in accordance to the preliminary analysis presented by Asko Parpola, The Indus Script: A Challenging Puzzle, in: World Archaeology 17: 3, February 1986, 401 f. with fig. 1.

c4.s8.t2.28

77 Incised tablets with a round shape have a different place in the sequence.

c4.s8.t2.29

and would complicate locating a given object in the Corpus.

c4.s8.t2.30

The term *graffito* is understood here to mean inscriptions incised on pottery before or after firing and inscriptions painted on pottery.⁷⁸ An attempt has been made to place graffiti with similar signs together, and the better and clear inscriptions at the beginning. When the text is very fragmentary, it is often quite uncertain in which direction the potsherd should be read. The reader, therefore, must never take the solution offered in the Corpus for granted, but be prepared to turn the photograph around.

c4.s8.t2.31

⁷⁸ Not infrequently, it is difficult to distinguish between a painted inscription on pottery and a painted pot decoration. This applies especially to the Early Harappan 'pot-marks', many examples of which will be published in the second volume of the Corpus.

c4.s8.t2.32

*Miscellaneous objects* is a heterogenous category designed to accommodate the few odd objects that fall outside the other typological classes. Inscribed copper or bronze weapons and tools is the most important object type here, but in this volume the category also comprises an incised terracotta cone (M-619) and an incised shell ladle (M-620).

The 6th criterion: the iconographic motifs

c4.s8.t3.1

The classification of the iconographic motifs in the Corpus is based on the following oppositions:

c4.s8.t3.2

iconographic motif : no iconography animate : inanimate moving animate being : unmoving animate being object : design animal : anthropomorph real animal : imagined animal single : group joined animals : composite animal

c4.s8.t3.3

A detailed analysis of the iconography of the Indus seals and tablets is in preparation and will be published elsewhere. What we offer here is a broad classification of motifs sufficient for organizing the material into coherent classes: 'unicorn' / urus / bison / zebu / buffalo / markhor / goat / deer / rhinoceros / elephant / tiger / hare / snake / gharial / animal group / joined animals / composite animal / anthropomorph / tree / cult object (variously interpreted as a manger, incense burner or filter) / ship / swastika / other geometric design.

c4.s8.t3.4

The 'unicorn' motif is placed first because it is the most common one of the Indus seals. The style of representing this animal in profile, so that just one single horn is shown, has in all probability been borrowed from the art of the ancient Near East. Although this representation undoubtedly has had a mythological explanation and importance in the Harappan religion, the 'unicorn' is likely to be a real animal (probably the urus, or *Bos primigenius*) which actually had two horns. It is in fact sometimes depicted as having two horns, but for the sake of analysis and classification, these two-horned representations have been separated from the 'unicorns' under the immediately following heading of 'urus'. These two headings are followed by other bovids, these by caprids and other cloven-hoofed ruminants.

c4.s8.t3.5

An "animal group" consists of two or more natural animals appearing on one object, either separately or forming one scene like the two confronting bisons. "Joined animals" usually have more than one head (as do the three tigers joined into a rhomb in M-295 or the bison which, in addition to its own head, has the head of the 'unicorn' in M-298) or, while composed of two or more animals, may not be complete animals (for example, just the heads and necks of two 'unicorns' are joined with each other and a fig tree and a cult object in M-296).⁷⁹ The "composite animal", again, is a complete beast whose body parts belong to different animals.

c4.s8.t3.6

⁷⁹ For the iconography of C-26 and C-41, cf. A. Parpola, The Sumerian 'bull-harp' motif in late Indus seals from Chanhujo-daro (forthcoming). 80 M-303 represents a deviant type, with not only the horns but also the hump of the zebu and a less human face. 81 An exception is Paul Yule, Figuren, Schmuckformen und Täfelchen der Harappa-Kultur, Prähistorische Bronzefunde I: 6, München 1985, 32-34. Yule's analysis is somewhat different in detail. 82 The scene in M-439 to M-441 B is classified as an "animal group" although three anthropomorphs are seen in it. 83 This sketchy list is not exhaustive for the anthropomophic motifs nor is it meant to provide an adequate description of the scenes involved.

c4.s8.t3.7

Usually only one type of composite animal is represented in the seals. It has the horns of the zebu, the face of man, the tusks and the trunk of the elephant, the neck and front legs of the goat, the middle body of the 'unicorn', the hind legs of the tiger, and the snake for a tail (cf. M-299 to M-302).80 But in the incised copper tablets of Mohenjo-daro, one can distinguish several composite animals. The composite nature of most of the animals depicted on these copper tablets has rarely been recognized so far.81 The "mastiff" of the excavation reports, for example, is actually a composite animal put together of the zebu (horns), tiger (head and front part of the body) and rhinoceros (back part of the body). In this fashion, we distinguish the following composite animals on the copper tablets (given separate Roman numerals when occurring after one another): buffalo + man + deer (?) + snake (M-504 to 506) / markhor + unicorn (M-543 to 549) / two-headed zebu + tiger (?) + unicorn (M-550) / markhor + camel + buffalo (M-551 to 566) / zebu + tiger + buffalo (M-567 to 570) / zebu + elephant + rhinoceros + snake (M-571) / zebu + tiger + rhinoceros (M-572 to 574) / zebu + camel + rhinoceros + snake (M-575 to 581).

c4.s8.t3.8

The "anthropomorph" is another broad category, which lumps together almost82 all the scenes in which any man-like figure is seen. This motif is broadly arranged as follows:83

c4.s8.t3.9

sitting anthropomorphic deity / anthropomorphic deity inside a fig tree / "contest": hero fighting with two tigers / man sitting in a tree and a tiger beneath looking at him / tiger-bodied goddess / deity holding by the hand two men who carry uprooted trees / archer / men jumping over a buffalo / man spearing a buffalo / tree-worship / sexual intercourse / religious procession with carried cult objects.

The 7th, 8th, and 9th criteria: the size, style, and state of preservation

c4.s8.t4.1

The size criterion implies that, other things being equal, the larger object comes first. Only in two categories of objects has it seemed necessary to distinguish between several size groups according to their height, for both intrinsic and layout reasons.

c4.s8.t4.2

The rectangular seals without iconography have been divided into three classes: (I) 18.5 mm and more, (II) 13 to 18 mm, (III) 12.5 mm and less.

c4.s8.t4.3

The square 'unicorn' seals have been divided into six groups: (I) 43.5 mm and more, (II) 35 - 43 mm, (III) 29 - 34.5 mm, (IV) 23 - 28.5 mm, (V) 17.5 - 22.5 mm, (VI) 17 mm and less. Within each size group, the 'unicorn' seals have been further arranged according to stylistic criteria. We have adopted the basic scheme developed by Paul Rissman by placing first the unicorns with a "collar", then the unicorns with "hatched neck", and finally the unicorns with "hatched face". Each of these groups, which apparently have a chronological significance, is subdivided according to the details of the "cultic object" in front of the unicorn.⁸⁴

c4.s8.t4.4

⁸⁴ Cf. Paul Rissman, The organization of stamp seal production in the Harappan Civilization (unpublished draft manuscript of a forthcoming study, 1986).

c4.s8.t4.5

As a general principle, badly broken objects are placed after the better preserved specimens of their category. However, exception is taken to this rule in the class of rectangular seals, which are arranged according to their length, since broken seals have once been longer than the full seals of the same length.

Linked texts (4)