FOUR ARTS AND CRAFTS

Compared with its architecture, the arts and crafts of the Indus civilization are well understood. Archaeologists are familiar with its tools and metalwork, carved stone weights, jewellery, shell and stoneware bangles, pottery, sculpture and seals, not only from excavations but also from technical analysis and experimental imitation of the likely processes involved in manufacturing them. These objects demonstrate, perhaps more than the Indus buildings and water engineering, just how sophisticated this ancient civilization was.

Even so, there are unsolved mysteries. The most perplexing is the small size of all the surviving arts and crafts, without any exceptions. Indus art objects always emphasize craftsmanship and technical qualities over monumentality – as in the exquisitely carved seals. To recall Wheeler’s perceptive comment, the seals have ‘a monumental strength in one sense out of all proportion to their size and in another entirely related to it’.¹ In the entire Indus civilization, there is absolutely no evidence for wall paintings, architectural ornamentation or life-size sculpture, in stark contrast with ancient Mesopotamia and Egypt. The most celebrated Indus sculpture, the steatite statuette of the ‘priest-king’, is a mere 17.5 centimetres in height. Maybe larger sculptures were created in perishable wood, but if so one would expect their bases or some other fragments to have been excavated. The finest surviving Indus objects – technically and aesthetically – are the elaborately drilled stone beads of jewellery and the seals, both of which found their way to Mesopotamia. But why the artists and craftsmen chose to persist ‘in what appears to have been a preference for miniature and less than life-sized art is a puzzle’, Dales noted in the 1980s.2 It has yet to receive a satisfactory solution.

Many Indus technologies appear to have been invented by craftsmen working at or around Mehrgarh before the Mature period of the civilization. Some of the latest Mehrgarh pottery, known as Faiz Mohammad Grey Ware (after its place of discovery on the Bolan Pass), is the equal of any pottery subsequently produced in the Indus cities. Mehrgarh itself has yielded micro-drills and other drilling tools used to work beads and various ornaments of semi-precious stone; a lapis lazuli bead was discovered there with a copper rod inside it, which may have been a drill abandoned during manufacture.

The majority of the Indus tools were made of stone: chiefly a very high-quality brown-grey flint or chert from an ancient source in the Rohri Hills to the east of Mohenjo-daro on the other side of the Indus. Several techniques were used to knap and shape the flint core, perhaps the most efficient being inverse indirect percussion, a technique unique to the Indus valley and peninsular India that continues to be practised at the agate bead-making centre of Khambat in western India. Inverse indirect percussion involved the core being pushed at a particular angle against an antler- or metal-tipped wooden stake set firmly in the ground, and then struck with a wooden or antler hammer, so that the stake’s tip detached a long, parallel-sided blade from the core. In this way, a single flint core could rapidly yield several blades.

However, some tools were made of metal, that is, copper and bronze, though not iron, which was unknown anywhere in the world until the second millennium BC. The copper was probably obtained from various sources, including Oman in the Persian Gulf and the Aravalli ranges in India. Amazingly, Indus bronze saws must have been as hard as steel, since they were capable of cutting shell as efficiently as modern steel saws. Although no ancient saws have survived, examination of the saw strokes on ancient fragments of shell from Indus workshops has enabled Kenoyer to reconstruct the saw’s basic shape. ‘It had a very thin serrated edge that was long and curved, similar to the saws still used in shell bangle-making in modern Bengal’, he notes.3 But, interesting to observe, an equally high technical standard was not applied to Indus knives and spears; metal blades went unstrengthened by the midrib required for military effectiveness. Such defective workmanship is generally true of the mass-produced and standardized objects of the Indus civilization, which were ‘sometimes quite inefficient’, as Parpola remarks, with ‘little effort . . . made to improve them’.⁴ Why this was so is another mystery, given the competence of Indus craftsmen in more technically demanding areas of manufacture.

For obvious reasons, the standardized stone weights – cubes or truncated spheres made of banded chert, agate or coloured jaspers – do not conform to the last observation. The weights are precisely made, well polished and systematic (though unfortunately not inscribed with any Indus script characters, which would have helped scholars to decipher the script’s numeral system). Unique in the ancient world, the Indus weight system does not correspond with any of the weight systems used in Mesopotamia or Egypt. It has left a remarkable legacy in India. It provided the weight standards for the earliest Indian coins, issued in the seventh century BC. It was identical with the system used by the first Gangetic kingdoms around 300 BC (just prior to the reign of Asoka). And it still functions, in the third millennium AD, for weighing small quantities in traditional markets in both Pakistan and India.

The fundamental weight may have been a tiny black-and-red seed known as the ratti, which comes from the gunja creeper (Abrus precatorius) and is used by today’s jewellers in Pakistan and India. The ratti’s average weight has been found to be 0.109 grams; eight ratti would therefore equal the smallest known Indus weight, 0.871 grams. From this basic unit, the first seven Indus weights double in size according to the ratio 1 : 2 : 4 : 8 : 16 : 32 : 64. The commonest weight is 13.7 grams (that is, approximately 16 × 0.871 grams). Thereafter, the system changes from a binary to an essentially decimal one, in which the weights rise in the ratio 160 : 200 : 320 : 640, and after a jump, 1,600 : 3,200 : 6,400 : 8,000 : 12,800 (i.e., in multiples of ten of the binary ratio). ‘The largest weight found at the site of Mohenjo-daro weighs 10,865 grams (approximately 25 pounds) which is almost 100,000 times the weight of the gunja seed’, notes Kenoyer.5

Graduated rules, for measuring length, have also been discovered. Of the four known examples, one is made of terracotta and comes from Kalibangan, the second is of ivory and comes from Lothal, the third is of copper and comes from Harappa, while the fourth is of shell and comes from Mohenjo-daro. The length of their divisions is about 1.7 millimetres, with larger units also marked: 17 and 33.46 millimetres on the scale from Lothal, and 67.056 millimetres (the largest unit) on the scale from Mohenjo-daro. It may be significant that 17 millimetres is very close to a traditional unit of length, 17.7 millimetres, noted in the classical Indian ‘economics’ text, the Arthasastra, dating from the fourth century BC or later.

Rules were presumably useful to a wide range of craftsmen, including architects and carpenters. Wood must have been widely used in brick buildings, transportation and other objects, judging from the holes left in brick structures for beams, seal images of wooden boats, terracotta models of oxcarts with solid wooden wheels and metal tools such as axes, chisels and saws, essential for woodworking. But unfortunately no samples of ancient wood have survived. Most likely, cedar and tropical hardwoods were used for buildings and furniture. The cedar, known as deodar (Cedrus deodara), which grows in the Himalayas, may have reached the cities via the Indus river and its tributaries. The hardwood was probably Indian rosewood (Dalbergia sissoo), also known as sisu and sheesham, judging from cuneiform references concerning the Mesopotamian trade with Meluhha to imports of ‘mesu wood’, a term which appears to denote rosewood. Because the heartwood of Indian rosewood is termite-proof, it is still used for making doors, windows and furniture in the Punjab (where it is the state/provincial tree on both sides of the India–Pakistan border).

Wood must also have been used for making looms, judging from various sizes of grooved bricks and stones that were probably loom weights, although depictions of looms are lacking. Textiles were certainly widespread in the Indus civilization, even if only a solitary piece of cloth survives: woven from cotton, it is dyed red with madder. There are multiple instances of impressions left by different grades of fabric on other materials, such as the imprint of rough sacking on the obverse of clay sealings and of jute cloth on a ceramic shard, traces of threads wrapped around the handles of copper objects, the insides of faience vessels originally moulded on a sand-filled bag, and even a toy bed imprinted with some tightly woven cloth made from finely spun thread. Indirect evidence for carpet production comes from the ‘distinctive curved copper-bronze knives that are functionally very similar to the curved blades used today for cutting the knotted threads of pile carpets’, suggests Kenoyer.6

The beads in Indus jewellery were produced from a wide variety of semi-precious stones, as well as from metals – gold, silver and copper – and shell, ivory, steatite and terracotta. Surprisingly, lapis lazuli and turquoise, though valued and traded in this part of South Asia from early times, were not staples of Indus bead makers. They preferred to work with the harder stones – carnelian, agate, chalcedony and jasper – because these retained their high polish, unlike the softer lapis and turquoise. The redness of carnelian (due to the iron oxides, haematite and goethite) was a particular Indus favourite, made by heating yellowish chalcedony packed in a covered pot with cow or goat dung. However, the blue of lapis and turquoise and the red of carnelian could also be imitated, in the first case by colouring beads of faience blue, in the second by painting long terracotta beads with red pigment. These terracotta beads, far less labour-intensive (and no doubt far less expensive) than carnelian beads, make a delicate clinking similar to the sound created by genuine carnelian beads when they are worn as a belt or necklace. ‘In modern Pakistan and India, the sound of ornaments is often referred to in poetry to evoke sensual beauty, and the sound of beads and bangles clinking against one another may have been just as important as the visual symbolism evoked by Indus ornaments’, guesses Kenoyer.⁷

The hoard of jewellery found at Allahdino contained one of only three known intact belts or necklaces of genuine carnelian beads. It consists of 36 long, drilled, carnelian beads with bronze spacer beads. When discovered by Walter Fairservis in 1976, following heavy rain on the exposed layers of his excavation, the belt was folded tightly into the centre of an earthenware jar. Stuffed around it were two or three multi-stranded necklaces of silver beads, eight coils of silver wire, fifteen agate beads, a copper bead covered with gold foil and a collection of gold lumps and ornaments. Why had all this been buried in the ground? The collection might conceivably have been the stock of an ancient goldsmith, given that the hoard resembles other hoards found at Mohenjo-daro and Harappa. But this seems improbable, given the very small size of the settlement at Allahdino as compared with those two cities; Allahdino is scarcely likely to have sustained a goldsmith. According to Fairservis’s colleague Kenoyer, the hoard of jewellery was probably ‘the hereditary ornaments of a woman or her family that had been hidden away for safekeeping’ – and never reclaimed, for reasons on which we can but speculate.⁸

Complete ornaments from Mohenjo-daro, as reconstructed by John Marshall after excavation in the 1920s.

Detailed studies of drilled carnelian bead manufacture, both ancient and modern (at Khambat), have been undertaken by Kenoyer and collaborators. The key technological element, the cylindrical drill bit – many of which have survived at sites such as Mohenjo-daro, Harappa, Dholavira and Chanhu-daro – was made from a rare form of metamorphic rock, modified by heating to produce an artificial material composed mainly of quartz, sillimanite, mullite, haematite and titanium oxide phases. Though not as hard as diamond, it was the next best thing. It has been dubbed ‘Ernestite’ by Kenoyer in honour of the early excavator Ernest Mackay, who first discovered the material and appreciated its importance at Chanhu-daro in the 1930s.

Drilled beads made of carnelian, such as these from Mohenjo-daro, are up to 13 cm long. Each bead required many days of steady, delicate drilling.

Kenoyer describes the likely ancient manufacturing process:

The drilling was probably accomplished by using a hand-held bow drill, with the bead held firmly in a wooden vice. Because of the intense heat produced, the whole process of drilling may have been done under water or with water continuously dripping onto the drill hole. Drilling experiments indicate that ‘Ernestite’ drills could perforate carnelian at a rate of 2.5 millimetres per hour, which was more than twice as effective as jasper or copper drills used with hard corundum (ruby) powder. Even then, it would have taken more than 24 hours, or three eight-hour working days of steady drilling, to perforate a 6-centimetre-long bead. The beads on the belts from Allahdino and Mohenjo-daro, ranging from 6 to 13 centimetres in length, would have required between three and eight days of steady drilling to perforate. Most modern bead drillers in Gujarat take long breaks after every few hours of work due to the strenuous nature of the drilling process. Considerable time is also taken in the preparation and repair of drill bits.9

On this basis, Kenoyer estimates that 480 working days would have been required to make the belt of 36 drilled beads found at Allahdino, from the initial heating of the carnelian to its final polishing.

No wonder, after such an extraordinary investment of technology and time, that such long carnelian beads were part of the treasures found in the royal burials at Ur in Mesopotamia. Some of the Ur beads appear to have been made in the Indus valley, while other beads, judging from their different style, were probably produced in Mesopotamia itself using Indus drilling technology. Presumably, the artisans doing the drilling were migrants from Meluhha to Mesopotamia.

Jewellery was exported from the Indus civilization to Mesopotamia. These Indus-made beads, including five made of heliotrope, were found in the Royal Cemetery at Ur, which dates from 2150–2000 BC.

Bangles made of conch shell, from Harappa. Bangle-wearing is still a common custom in South Asia.

The wearing of carnelian beads must have been a sign of high status and wealth. Bangles, by contrast, were an ornament common to all sections of society and seem to have been important; they are often found in Indus burials. Very likely, this Indus tradition of wearing bangles was one that persisted in South Asia: Indian warriors, for example, wore bangles in battle to protect the wrist and the arm; Indian women wore them to protect their families and ensure long lives for their husbands. The latter custom continues today. From the Indus civilization bangles ranging from simple clay circlets to hollow bangles of hammered sheet gold have been found; white shell and bronze were also used, as was faience, sometimes in imitation of a shell bangle with a stylized ‘womb-shaped’ motif. But the most prestigious material for bangles appears to have been stoneware.

The technology of stoneware bangles was clearly kept secret and disappeared altogether along with Harappa and Mohenjo-daro, the sole sites of production of these bangles. It first came to the attention of archaeologists at Mohenjo-daro in the 1930s, when Mackay encountered what seemed to be the vitrified remains of an unsuccessful firing of pottery, clay and bangles. Efforts to analyse and imitate stoneware bangle production by Mohammad A. Halim, Vidale and Kenoyer in the 1980s met with only partial success.

The clay they used for the stoneware bangles was prepared as a very fine paste. It was then thrown on a wheel to make a thick, hollow cylinder which, after some drying, could be cut with a cord into bangle blanks, which were allowed to dry until hard. Shaping with a stone blade, burnishing with a smooth stone and polishing with a cloth followed. Finally the bangles were fired at a high temperature inside a clay-coated jar, resting on stacks of terracotta bangles in a kiln closed at the top with a massive clay cap secured with the impression of a ‘unicorn’ seal. If all went well, the finished bangle was a fine, mottled greyish-black throughout (as revealed by surviving broken bangles) – a result encouraged by researchers’ inclusion of organic material, such as goat dung, in the jar.

The most unusual aspect of the stoneware bangles is not, however, their manufacture, but rather their inscriptions and their unexpected size. Every stoneware bangle carries an inscription: a single sign or group of signs from the Indus script traced onto the wet clay before firing. It may be a sacred symbol, a title or even a person’s name – there is no way of knowing at present. The diameter of every bangle is 5.5 to 6 centimetres, which is too narrow for it to have been worn around the wrist or the ankle. Perhaps the bangles were worn as a pendant or sewn onto clothing. The headband of the ‘priest-king’ statuette, and his upper arm, carry a prominent circular symbol that may depict a stoneware bangle. Bearing in mind the characters, the size and the secrecy, ‘It therefore seems a distinct possibility that the stoneware bangles were worn as badges of office by leading members of the hierarchy’, suggests McIntosh.10

Pottery from the Indus cities – both plain and painted – survives in large quantities. The potter’s wheel, employed at Mehrgarh, was in use for well over 1,000 years before the Mature period. But the urban Indus pottery is not particularly distinguished. Indeed, the pottery from Mehrgarh, especially the Faiz Mohammad ware, is probably superior to the pottery found at Harappa, Mohenjo-daro and other cities, both in form and decoration. Rather than pottery, the most vital sculpting energies of the Indus craftsman in clay, stone and other materials went into three types of object: numerous animal and human figurines, masks and toys ranging from the frankly crude to the almost sophisticated; a small but significant number of higher-quality sculptures; and, most of all, the thousands of seals.

Many of the female and male figurines appear to have been fertility figurines, judging from their nudity. They were presumably made by potters for worshippers to use in domestic rituals under a sacred tree or in the courtyard of their home. Once the ceremony was over, the figurine was probably discarded or given to children as a toy, rather than being dedicated to a shrine or building (as was common in Mesopotamia), given the figurines’ findspots. The animal figurines, which are somewhat more naturalistic and frequently charming and humorous, look as if some were simply toys, while others were intended for worship. For example, a hollow ram’s head figurine with a body like a tea cosy mounted on two wheels was assuredly a toy, whereas a finely modelled and incised, hollow, three-headed animal figurine, depicting an elephant with a hollow trunk, with two horns of a water buffalo curving along the cheeks of the elephant and the bottom jaw of a feline with bared teeth at the back of the elephant’s head, was probably some kind of cult object.

The sculptures number perhaps ten or a dozen, depending on how one categorizes the finest of the figurines and on the disputed status of two statuettes. The best known of the sculptures are undoubtedly the ‘priest-king’ in steatite and the ‘dancing girl’ in bronze, both of which were found at Mohenjo-daro during the early excavations. Neither identification is at all secure. The ‘priest-king’ (see page 115) is so-called largely on the strength of his commanding face, with its half-closed eyes and closely manicured beard, and the elaborate draped garment he wears, leaving his right shoulder bare – a style that is still considered appropriate in India and the Buddhist world when approaching a shrine or holy person. In addition, the prominent trefoil designs on the garment, according to Parpola, parallel the trefoils on cloaks worn by gods and priest-kings in Mesopotamia. The vivacious, naked ‘dancing girl’, confronting the viewer with her left arm sheathed in bangles and her almost bare right arm held akimbo, possibly (but far from definitely) poised in a dance posture, seems still more doubtfully identified. Marshall’s comment on her is surely over-confident: ‘here we clearly have a dancer or nautch girl of aboriginal stock represented, and we may reasonably infer that girls of this class were accustomed to wear nothing more than their ornaments when dancing, though it would be rash to suppose that they ordinarily went naked.’¹¹

‘Dancing girl’ statuette made of bronze, from Mohenjo-daro. It is 10.5 cm high, 5 cm wide.

Yet more problematic, as Marshall was the first to admit, are two damaged statuettes from Harappa: a headless male figure with one leg raised, this time undoubtedly dancing; and a headless male torso. Both cannot help but strike the viewer as distinctly influenced by Greek art, which was a common feature of the art of northwest India in the centuries after Alexander. As a result Marshall, and many subsequent scholars, have questioned whether these two statuettes belong to the Mature Indus civilization, despite the undisputed early provenance of the male dancing figure according to its excavator. The question of their true date remains open, to the extent that the two statuettes are often omitted entirely from current discussions of Indus sculpture.

About the aesthetic quality of the Indus seals there is virtual unanimity. Wheeler’s admiration has already been mentioned. Marshall considered the best seals to be ‘distinguished by a breadth of treatment and a feeling for line and plastic form that has rarely been surpassed in glyptic art’.12 Piggott – no admirer of Indus aesthetics, as we know – nonetheless conceded that the seals were ‘frequently carved with a brilliant sureness of touch’.¹³ Independent India’s great artist Satyajit Ray – a graphic designer and illustrator who became a writer and celebrated film director – was so enchanted by the seals that he wrote a short story inspired by the ‘unicorn’. Speaking for myself, I first became interested in the Indus civilization after seeing its seals.

Soft steatite (soapstone) etched with a bladed burin like a modern scalpel and then fired in a kiln so that it became slightly harder (about 4 on the Mohs scale of mineral hardness) was the commonest material and process for making Indus seals. The shape was typically square, but it could also take the form of a rectangle, though seldom that of a cylinder, the commonest shape for a seal in Mesopotamia. However, seals were also chiselled, inlaid, painted, moulded and embossed on many other materials: terracotta and glazed ceramic, shell, bone and ivory, sandstone and gypsum, and, rarely, the metals copper, bronze, silver and gold. No doubt the script was also carved into wood, woven into fabrics and basketry, inscribed on palm leaves and perhaps even painted on to human skin. Only semi-precious stones, including carnelian and lapis lazuli, were not used for seal carving – most surprisingly, given that semi-precious stones were often used to make Mesopotamian seals.

On the reverse of a seal was a carved knob or boss, perforated so that it could hold a thick cord. This suggests that a seal could have been worn around the owner’s neck or hung from his waist. Larger seals were probably kept in a pouch. It is also common to find that the boss of an excavated seal is missing. It most probably broke off after receiving a knock, as a result of the weakness of steatite combined with flaws introduced in the carving and drilling of the boss, causing the seal itself to drop into the street and become lost. ‘Many a surprised merchant must have reached for [his] seal to find a boss with no seal attached’, remarks Kenoyer.14 The loss of a seal would have been a serious matter, but it is not known how the Indus people coped with it; in Mesopotamia a herald announced the loss of a seal, and there were severe punishments for its illegal use. Perhaps this manufacturing defect in the boss is why the seal design was improved in the later part of the Mature period.

We have said nothing about the most important part of an Indus seal: the motif and the script characters. Since these are likely to have been connected with each other, and of course with the purpose of the seal, it seems appropriate to discuss them in chapter Ten, devoted to the challenge of deciphering the Indus script. The graphic impact of Indus seal art may be transparent; not so its semantic import.

Seal stones with bosses on the back made of steatite, from Mohenjo-daro. A seal stone probably hung on a cord that passed through the perforated boss and was attached to the seal’s owner.