The Boundless Sea

by David Abulafia

  The settlement expanded, becoming a pole of attraction for Indian and foreign merchants. No doubt it was at first simply a place beyond the normal range of Greco-Roman shipping that accumulated Mediterranean goods as they were passed on from hand to hand through the ports on the western flank of India, most often in Indian boats. With time Arikamedu drew these westerners to its harbour, and what some like to think of as a Roman settlement in the Bay of Bengal came into existence, during the reigns of Augustus and Tiberius, when the town had already been flourishing for a century or more. It seems to have remained a lively place until the mid- to late second century AD. Among the excellent facilities it offered was a warehouse close to the river, 150 feet long. There were areas given over to industry, easy to identify from the large number of beads, bangles and cheap gemstones that are said to ‘litter the area’, and there were what excavators identified as vats for dyeing cloth, where the inhabitants manufactured the fine muslins that are mentioned in the Periplous as favourite exports of south India. Oddly, Roman coins are absent from Arikamedu.78 They were either sucked into royal treasuries or melted down; yet this did not inhibit intensive trade. Arikamedu was convincingly identified by Sir Mortimer Wheeler (who helped excavate the site) as one of the emporia in south-eastern India mentioned in the Periplous, specifically Poduké. This name also appears in almost the same form in Ptolemy’s Geography, and is a corruption of the Tamil word Puduchchēri, which simply means ‘New Town’, so one can imagine that the Greeks may also have called it Neapolis, which means exactly the same. Then, over time, the Tamil name was corrupted by the French and British into Pondicherry.79

  The lure of the Ganges can already be detected in the Periplous. Once this area became known by repute, the temptation to sail there in search of silk, pearls and other luxuries became overwhelming, though numbers were probably much smaller than in the western Indian Ocean.80 Strabo talked of traders who sailed from Egypt to the Ganges; he described them as ‘private merchants’, suggesting that these were people who went under their own steam; and he did not believe everything they told him, so whether they had actually reached as far as the mouth of the Ganges he could not really say.81 Ptolemy knew a fair amount about the city of Patna, on the Ganges; and he was aware that a thin tongue of land stretched downwards from south-east Asia, which he called the Golden Chersonese. This area was explored early in the second century AD by a sea captain named Alexander who may well have rounded the southern tip of Malaya, by way of the Strait of Malacca, and have entered the South China Sea, arriving at a place called Kattigara, which will be revisited in the next chapter. Greeks and Romans were not terribly sure where China lay, but they knew it was a source of fine silk, and that was probably the motive for occasional forays towards the South China Sea; an embassy from the court of Emperor Marcus Aurelius may have reached the South China Sea in the late second century, for Chinese records describe the embassy of ‘Antun’ (Antoninus, the emperor’s cognomen, or additional name), though they dismissed it as of little consequence because the goods offered as gifts – or, as the Chinese would prefer to think, tribute – were regarded as commonplace. This came as a surprise, because the Han court was aware, though vaguely, that the Roman Empire was a vast polity comparable to their own great empire. Since the embassy had set out with rhinoceros horn, elephant ivory and tortoiseshell the diplomats had probably lost their goods en route.82 Generally, though, even Burma was a stage too far; a few Roman imperial subjects settled there, as court entertainers in search of a patron, but there was no port city with a mixed population and markets full of goods from each end of the ocean and from the hinterland, in other words another Puhar or Poduké.83 In reality, Greco-Roman navigators very rarely ventured beyond Malaya, so that Ptolemy’s misunderstanding about what lay at the bottom of the Golden Chersonese was perpetuated for over 1,300 years.

  V

  From the late second century onwards Bereniké went into decline, and one explanation may be the pandemic (whatever disease it was) that struck the Roman Empire in the year 166. By the middle of the third century, Bereniké had not exactly vanished from the map, but there is no evidence that it was still a major emporium linking East and West. Recovery came, however. By the fourth century Bereniké was benefiting from developments further south: the creation of vigorous kingdoms on either side of the Bab al-Mandeb Strait, in Himyar (Yemen) and Axum (Ethiopia/Eritrea), ancient sources of incense, ivory and ebony. Meanwhile, goods from the western Mediterranean ceased to arrive in this part of the Red Sea. One view is that this reflected a growing fracture in the Mediterranean between east and west, though maybe what this reveals is, rather, the economic vitality of eastern Mediterranean ports such as Rhodes, Laodicea (Lattakiah), Gaza and Alexandria, which found themselves more than able to supply the needs of the Roman outstations in the Red Sea. This was accompanied by new opportunities to do business in south Arabia and east Africa, which are suggested by the discovery of a coin from Axum and another from western India dating to AD 362. Trade with Ceylon around 400 has even been described as ‘brisk’, and the town of Bereniké revived physically too, as new temples dedicated to Isis, Sarapis and other Egyptian gods were built, as well as a church and several warehouses.84

  The best indication of the manner in which this port looked two ways, towards the Indian Ocean and towards the Mediterranean, can be found in the preserved fragments of wood found on this site, which included a small quantity of Indian bamboo and large amounts of south Asian teakwood, including a beam more than three metres in length from one of Bereniké’s shrines.85 Teak was also widely used in shipbuilding, for instance by early Arab mariners. Yet the material remains also included cedar beams brought from Lebanon or somewhere in that area, and found in the remains of the temple of Sarapis. Meanwhile, at Myos Hormos, wood from ships taken out of commission was re-employed in everyday building. Just as these ships were put together in what amounted to a giant three-dimensional jigsaw puzzle, they could be taken apart quite easily, and the planks, beams and masts reused in quite different ways; wood was precious along the dry edges of the Eastern Desert. Beams and planks discovered during the excavation of the town buildings of Myos Hormos bore traces of pitch, iron nails and barnacles.86

  We are still left with the question of why Bereniké was abandoned in the sixth century, and its collapse can probably be attributed to a rich combination of factors: bubonic plague in the Mediterranean and Middle East, the so-called plague of Justinian; local wars out of which Axum in east Africa and Himyar in south Arabia emerged as the dominant political forces, under Christian and Jewish kings who were keen rivals; the ascendancy of Axumite and Himyarite merchants, with bases at Adulis on the African side and Kané on the Arabian side. Bereniké did not come to a cataclysmic end. It declined throughout the sixth century, and when it was finally abandoned no one lived close enough to raid the site for wooden beams or building stone. The result was that the dust of the desert blew over the town, and the dry atmosphere preserved its sand-immersed remains.87 What continued, however, was traffic up and down the Red Sea; what changed was in whose hands control of this traffic lay. Over time, the location of the key ports that linked the Nile, and beyond that the Mediterranean, to the Red Sea changed, as Myos Hormos, once a minor competitor to Bereniké, emerged in the medieval period as an important link in the chain connecting the Red Sea to the Indian Ocean and the Mediterranean, under its Islamic name of Qusayr al-Qadim. The connection was broken only for short periods, and the links between the Indian Ocean and Egypt, and beyond that the Mediterranean, were not severed even when different people to the Greco-Roman merchants took charge of the carriage of cargoes from India and east Africa.

  7

  Brahmins, Buddhists and Businessmen

  I

  Looking at the Indian Ocean from the perspective of the Periplous and Bereniké presents one overwhelming difficulty. The illusion is created that its ports interacted when the Greco-Roman ships arrived with merchants on board who craved the
spices of the East. When Bereniké and Myos Hormos were in decline, it might then be assumed, the whole of this network crumbled away. Without the Romans, it is true, the Indian kings would not be able to accumulate so much treasure; but whether they actually put much of this gold and silver into circulation remains doubtful. Fragments of evidence suggest that Barbarikon and Barygaza, or at any rate ports in their vicinity, remained lively centres of trade and industry in the fifth and sixth centuries, and that many of the stopping places and links that would be described by late medieval travellers such as the Venetian Marco Polo and the Arab al-Mas‘udi were already in place.1 In a word, the question concerns continuity, and relates directly to the idea that the opening of the Indian Ocean routes, whether it was achieved by Romans, Indians or Malays, or all of them in collaboration, should be seen as the first step in the creation of global networks of trade, in which sea routes functioned as the major links. In that case the India trade of the Cairo Jews around 1100, which will be examined before long, or the irruption of the Portuguese in the Indian Ocean in 1497–8, were only further stages in the bonding together of the Indian Ocean with the Mediterranean and with the markets of Europe that lay beyond.

  When the evidence for Roman trade is so rich it is tempting to dismiss that for ‘native’ trade as mere disconnected fragments. But the fragments can be connected, and they tell a remarkable story in which the Romans no longer appear as the main actors. To make sense of this story, places very far apart will have to be examined – as far apart as Madagascar and Malaya, and beyond Malaya to the very edges of China. This will reveal the sheer expanse of the area that was tied together by navigators in the first half of the first millennium AD. It will also show how the links in a chain that stretched all the way from southern China to the Mediterranean were being forged and attached to one another, so that the spice trade that had already obsessed the inhabitants of imperial Rome extended far beyond India and Ceylon. In particular, the mariners of Indonesia and Malaya became the great intermediaries sailing regularly between China and India; it was they who knitted together the networks that had previously functioned apart from one another; and it was they who made traffic by sea, rather than the arduous overland route, first of all worthwhile and then supreme.

  The subject of this chapter is a dizzying expanse of sea, then, encompassing the entire Indian Ocean and, beyond that, the South China Sea, which is ringed by southern China, Vietnam, the Philippines, Indonesia and the Malay peninsula. But the place to begin is a relatively small island very far from there, in the north-west of the Indian Ocean, with an area of 3,800 km2 (which still makes it the second largest island in the western Indian Ocean, after Madagascar). Socotra lies about 240 miles south of Arabia and the same number of kilometres east of Africa.2 It was not, therefore, visited by coast-hugging vessels, but by those who had mastered the monsoons and were willing to range out of sight of land, which was worth their while, since it functioned as a trading hub linking east Africa, the Red Sea and the routes to India. Even so, the local currents were difficult to manage; in addition, it could not offer a decent harbour and ships had to anchor off the coast. Between May and September it was unreachable, because the south-west monsoon was blowing. It was often chosen as a pirate base, though the pirates must have been constrained by the same difficult conditions as the traders. Yet the traders came. The Periplous gives an impression of exact knowledge garnered by a merchant whose passion for tortoiseshell knew no bounds:

  In the open sea is an island called Dioskourides; though very large, it is barren and also damp, with rivers, crocodiles, a great many vipers, and huge lizards, so huge that people eat the flesh and melt down the fat to use in place of oil. The island bears no farm products, neither vines nor grain. The inhabitants, few in number, live on one side of the island, that to the north, the part facing the mainland; they are settlers, a mixture of Arabs and Indians and even some Greeks, who sail out of there to trade. The island yields tortoise-shell, the genuine, the land, and the light-coloured varieties, in great quantity, and also the oversize mountain variety with an extremely thick shell … The so-called Indian cinnabar is found there; it is collected as an exudation from the trees.3

  The Periplous also explains that the island was ruled by the king of the opposing shore of south Arabia (Hadhramawt); it has often fallen under the control of the rulers of south Arabia, and at present it forms part of the Republic of Yemen. It used to be visited regularly by sailors from southern India and Barygaza; ‘these would exchange rice, grain, cotton cloth, and female slaves’, against great quantities of poor-quality tortoiseshell. The author then states, mysteriously, that ‘the kings have leased out the island, and it is under guard’. There are strong hints here that this island is neither worth visiting for its tortoises, nor worth the trouble of having to negotiate with the current tenant, who seems to discourage trade.4 Reading between the lines, it sounds as if the island had once again become a pirate base, for which its position suited it very well, in seas that are still notorious for piracy. Over the centuries, though, this island was visited by large numbers of travellers, for they have left about 250 inscriptions and drawings incised on the walls of a cave discovered in 2000, all of which happily confirm the insistence of the Periplous that Socotra was visited by people of very diverse origins. These inscriptions are in a great variety of languages, Indian, Iranian, Ethiopic, south Arabian, even Greek, and date from the period between the first century BC and the sixth century AD.

  The Hoq cave is very long – two and a half kilometres – and at some points as much as a hundred metres wide and thirty metres high, so it must have appeared an awe-inspiring place, and a cool one on an island baked by the sun. It was apparently a cult centre for one god or more, and the graffiti on its walls are not so very different from those that can be found on tourist sites today: bhadra prapta, ‘Bhadra arrived’, or simply the name of the visitor, most often in Sanskrit characters.5 One can imagine that sailors and other travellers who had braved the seas to reach Socotra were keen to offer their thanks to the gods for a safe journey there, and their prayers for a safe journey home. Who these gods were is far from clear, but the fact that the vast majority of inscriptions are in Sanskrit suggests that they were Indian gods, linked to Hindu or Buddhist cults; quite possibly the cave contained an image of Buddha himself. For at this period in Indian history the links between religion and trade were close; the Buddhist monasteries, in particular, saw no evil in honest profit. They discarded the caste system that relegated merchants to a less respectable role in society than the priests and warriors who were acclaimed as the leaders of the Hindu communities. The spread of Buddhism is generally thought to have stimulated the Indian economy.6

  The evidence that this island was visited over several centuries by Indian merchants is overwhelming; the earliest inscription is thought to date from before AD 100, but after 400 the number declined and the visits, or the practice of recording them, gradually died out – more likely the latter, for, as will be seen, the coming of Christianity rendered the cave cult obsolete. Amid the names there are several that speak loudly of contact with India: ‘Samghadasa, son of Jayasena, inhabitant of Hastakavapra’ had reached Socotra from a town close to Barygaza on the Gulf of Cambay; the town is still there, under the name of Hathab, but it was already known to the author of the Periplous and to Ptolemy under the name Astakapra. The Periplous mentioned it in its description of the hazardous approach to Barygaza, and excavation there shows that it flourished from the fourth century BC to the sixth century AD.7 Even more noticeable among the inscriptions are direct references to Barygaza: ‘Śesasya Visnusena from Bharukaccha arrived’; or a graffito simply saying Bharukacchaka, ‘the Barygazan’; and best of all: ‘the sea-captain Visnudhara from Bharukaccha’.8 Evidently, ships bound from north-west India for the Horn of Africa often stopped in Socotra. At a village now known as Kosh, archaeologists discovered Indian artefacts but could produce little or no evidence of links between Socotra and t
he Mediterranean; Kosh lies on the northern side of Socotra, confirming the statement in the Periplous that this was where the inhabitants chose to settle. It is even possible to say something about the boats that reached Socotra, as the Hoq cave also contains several drawings, the clearest of which shows a boat with two rudders and maybe three masts. This compares closely to a sixth-century image of a boat in the famous Ajanta caves in India, where a triple-masted ship with two rudders can be seen.9