The Boundless Sea

by David Abulafia

  For Nanna, the chief son of Enlil, his lord, Ur-Nammu, the mighty male, the king of Ur, the king of Sumer and Akkad, the king who built the temple of Nanna, caused the former state of affairs to reappear, at the edge of the sea in the customs house trade was [gap in the inscriptions] … Ur-Nammu restored the Magan trade [literally: boat] into Nanna’s hands.23

  These expeditions into the still uncertain waters of the Persian Gulf were, then, dedicated to the gods, whose protection was sought and whose temples benefited from the copper and luxury goods brought from Dilmun, Magan and beyond.

  The best evidence for the reality of Dilmun lies in what would be regarded as quite humdrum documents about merchants and their import–export businesses, were they not so exceptionally old. Lu-Enlilla, for instance, was a garaša-abba or seafaring merchant from Ur, one of the greatest Sumerian cities, who traded on behalf of the Temple of Nanna and was commissioned by its administrator, a certain Daia, to take fine cloths and wool on a trading expedition; he was to exchange these goods for copper of Magan. Copper was the great desideratum: Sumer had emerged at a time when copper and, subsequently, copper smelted with tin to make bronze, was required in ever vaster quantities not just to forge strong weapons and tools but to create beautiful objects – figurines, panels, bowls. Sumer was rich in the produce of the soil and of its flocks of sheep and cattle, but poor in metals, robust timber and good quality stone. The copper-bearing mountains of Oman were the place to which to turn for metal, and there is no doubt that Magan corresponds to what is now the Omani peninsula (now partly under the rule of the sultan of Oman and partly under the rule of the sheikhs of the UAE).24 Proof that the copper came from there lies in the small natural nickel content of Omani copper, which is closer to the nickel content found in copper objects from Sumer than it is to the copper of lands to the north. Copper from the mountains to the north of Mesopotamia was also more expensive and, given the vast quantities that were traded, was more difficult to transport than the seaborne copper of Magan. Oddly, Magan bought barley from Ur but provided Lu-Enlilla with onions, which were plentiful enough in Mesopotamia, so maybe the sailors on board a ship bound from there had overstocked their larder with them and Lu-Enlilla simply had to put up with that.25 Meanwhile life in Magan improved; settlements became more permanent, stone towers were built, as well as monumental tombs. This was still a dispersed society, and nothing remotely comparable to the great cities of Sumer emerged, but merchants in search of copper (of which fragments survive in local graves) had given a stimulus to an area that in earlier centuries had been a deserted backwater.26

  The sea was becoming more important as Ur and its neighbours became greater and greater centres of consumption. Difficult routes across the mountains of Afghanistan could be avoided by taking the sea route to India, where access could be gained to the increasingly powerful cities of the Indus Valley. A gift to the goddess Ningal at Ur included two shekels’ weight of lapis lazuli, carnelian, other prized stones and ‘fish-eyes’ (pearls); these goods had arrived from Dilmun, ‘the persons having gone there by themselves from the month Nissannu till the month Addaru’. The names of these months were eventually passed all the way down to the Hebrew calendar as Nisan and Adar, and, bearing in mind the astronomical sophistication of the peoples of ancient Mesopotamia, we can be sure that this means they were absent for eleven months. The gifts had arrived from Dilmun, but they had originated in a scattering of places – the presence of ivory among items listed on some of these tablets suggests links to India, a land of elephants, and its presence was not incidental, for ivory objects were lovingly carved in Ur or were imported ready-carved, like some painted ivory figures of birds, brought from Meluḫḫa, as was much of the carnelian that was so cherished in Sumer. Sometimes, it appears, natives of Dilmun came with these goods; and sometimes men of Ur, such as Lu-Enlilla, set out for Dilmun and carried on their trade there. Some Ur merchants did so as agents of a temple; but, increasingly, others worked on their own.27 Loans at interest, business partnerships, trading contracts assigning risk, and other indications of a commercial economy with many of the attributes of mercantile capitalism abound, for the first capitalists on record were Sumerian merchants long ago in the third millennium BC:

  Lu-Mešlamtaë and Nigsisanabsa have borrowed from Ur-Nimmar 2 minas of silver, 5 kur of sesame oil, 30 garments, for an expedition to Dilmun to buy copper there. On the safe return of the expedition, the creditor will not make a claim for any commercial losses. The debtors have mutually agreed to satisfy Ur-Nimmar with 4 minas of copper for each shekel of silver as a just price; this they have sworn before the king.28

  But for the use of weighed silver in place of coin (not such a great difference, all told), and the names, this could almost be a commercial document from Barcelona over thirty centuries later.

  The contract just cited forms part of the business correspondence of Ea-nasir, a wealthy merchant of Ur, whose house was identified by Sir Leonard Woolley during his triumphant excavation of Ur of the Chaldees in the 1920s and 1930s; it was not a particularly large house, and consisted of five rooms around a main courtyard, though a couple of rooms had been ceded to a neighbour. He lived around 1800 BC, at the end of the Sumerian ascendancy and at a time, as will become clear, when trade to India had contracted. But he was still a wealthy man. His speciality was copper, which was delivered in ingots, and he apparently supplied the royal palace. He was surely one of the most prominent businessmen of his day, maybe a little unscrupulous, but looking at his wealth it is impossible not to be impressed: one of his shipments weighed eighteen and a half metric tons, of which nearly one third belonged to him.29 The character of trade had changed somewhat in the century since Lu-Enlilla had lived, and the temples of Ur were no longer (so far as we know) heavily involved in expeditions down the Persian Gulf; this was business conducted by private merchants, and they preferred to pay for their goods in silver, weighed by the shekel, rather than in textiles, as had been the case with Lu-Enlilla. Probably many of the textiles Lu-Enlilla had sent to Dilmun and beyond were woven in temple workshops by female slaves attached to the temple itself. Silver, on the other hand, suited the needs of mobile merchants who constantly and energetically bought and sold, and traded on the open market in Ur.

  Far from being a dry enumeration of imports and exports, Ea-nasir’s private archive conjures up the passionate disputes that were bound to arise about the quality of goods and the obligations to fulfil a contract:

  Speak to Ea-nasir; thus says Nanni: now when you had come you spoke saying thus: ‘I will give good ingots to Gimil-Sin’; this you said to me when you had come, but you have not done it; you have offered bad ingots to my messenger, saying: ‘If you will take it, take it, if you will not take it, go away.’ Who am I that you are treating me in this manner – treating me with such contempt? And between gentlemen such as we are!’ … Who is there among the Dilmun traders who has acted against me in this way?30

  The word translated as ‘gentleman’ is a technical term for a citizen of a very respectable social position, and ‘gentlemen’ were bound by a code of honour to which they swore their adherence in the temple of the sun god, Shamash; it was this divinely guaranteed contract that he was being accused of ignoring.31 This is only part of a longer complaint, and just one of several that Ea-nasir filed away. Even though many of his partners were perfectly satisfied with his conduct, he has been described as a difficult, perhaps sleazy, businessman. This may be unfair; he is far more likely to have preserved documents about transactions out of which claims could arise than others, perhaps the vast majority, that could safely be discarded once everyone was quit of their profit.

  IV

  Magan, then, lay close to the exit from the Persian Gulf, reaching up to the tip of the Musandam peninsula, where Oman almost brushes Iran. For many sailors, Magan must have meant the island of Umm an-Nar near Abu Dhabi, a significant settlement where plenty of Sumerian pottery has been found and where copper from the Omani mines
reached the Persian Gulf; Umm an-Nar was a sort of storehouse for goods being despatched towards Iraq.32 The massive reconstructed tomb at Mleiha in the Emirate of Sharjah (UAE), a product of the Umm an-Nar culture, is 13.85 metres across.33 However, the term ‘Magan’ was also applied to the Oman peninsula. As Harriet Crawford has observed, ‘the ancient scribes seem to have had a rather elastic and foggy concept of location.’34 But where was Meluḫḫa? Everything suggests that it was a wealthy and desirable trading partner for Sumer. The creation of a sea link between Sumer and another centre of high civilization has special importance in the maritime history of humanity, as one of the first moments, perhaps the first moment, when civilizations that had developed independently to a comparable cultural level entered into dialogue with one another across the sea. Once it has been located it will be possible to return to the question of where Dilmun was to be found, whether it was a specific place or a broader region. Sumerian documents often listed Dilmun, Magan and Meluḫḫa together, because they obviously lay on the same sea route, with Meluḫḫa at the end. Since ivory was one of the most precious exports from Meluḫḫa, the choice is narrowed down to the coast of east Africa or that of India, the two areas from which we could expect elephant ivory to be exported; and it has been seen that Indian goods did reach Sumer.

  In much later centuries, when the Assyrians dominated Mesopotamia during the early first millennium BC, the name of Meluḫḫa became attached to parts of east Africa; but that certainly does not mean it was always identified with the area. In the first place, the route out of the Persian Gulf tends eastwards, to lands rich in carnelian as well as ivory; there is a short, clear run from the Strait of Hormuz to the coast of Pakistan. Indeed, the relationship between that coast and Oman has been so close that from the eighteenth century to the middle of the twentieth century the sultans of Oman possessed an outpost on the coast of Pakistan, at Gwadar, 240 miles from Oman proper. Moreover, if ships that left the Persian Gulf then turned south and west, coasting along the shores of Yemen past Aden and maybe as far as east Africa, we should expect evidence of Sumerian contacts in Yemen too, but there is none. Nor is there evidence that the inhabitants of what are now Yemen, Somalia and neighbouring regions were able to launch their own trading fleets, whereas this was certainly true of the Meluḫḫans. Ships from Meluḫḫa and Magan are known to have reached Sumer in the days of King Sargon around 2300 BC and to have docked at his capital, Akkad, where lived ‘Su-ilisu the Meluḫḫa interpreter’. There were enough Meluḫḫans around Lagash in the century before 2000 BC to create a ‘village of Meluḫḫa’, and they had a garden and fields producing barley, so Meluḫḫan migrants were familiar in Mesopotamia at this time.35 Looking eastwards, the journey along the shores of Iran and Baluchistan to the mouth of the Indus was far less challenging than that to Africa, and perfectly manageable under a captain who understood the monsoon seasons.36

  It is not beyond credibility that ‘Magan’ actually meant ‘copper’ (rather as Cyprus, Kupros, did in Greek) and that ‘Meluḫḫa’ meant ‘ivory’, in the language of proto-historic India. Or Meluḫḫa may originally have meant ‘the place across the sea’, rather like the medieval term Outremer that Europeans eventually attached to the crusader kingdom of Jerusalem, but could have meant anywhere across the sea before becoming attached to a particular place; this might be deduced from the Arabic term milaḥa, possibly derived from ‘Meluḫḫa’, which was used in the early Middle Ages to mean navigation, seafaring or seamanship.37 Afghan lapis lazuli could be obtained from Meluḫḫa; this would have been carried down the Indus Valley to ports where Sumerian traders were making their purchases. The area also produced fine woods, including a ‘black wood’ that must be ebony, and sometimes wooden objects decorated with gold were brought from there, another sign that Meluḫḫa was no backwater. Finally, and conclusively, inscribed Indian seals have occasionally been found on Sumerian sites and in reasonable quantity in the Persian Gulf, so there can be no doubt that contact existed between Sumer and the Indus Valley. Moreover, Indus pottery has been found at Abu Dhabi.38

  The Indus Valley civilization remains the least known of the great Bronze Age civilizations, for such evidence as there is often remains impenetrable – there are inscriptions in a script that cannot be read and a language that cannot be guessed at, and there is little that can be said about the social and political organization of a culture whose impressive cities stare blankly at the excavator. For the Indus Valley appears to have been dominated in the second half of the third millennium BC by two massive and tightly planned cities, very similar in layout and construction, Harappā and Mohenjo-daro, though these are only their modern names. They lie a full 350 miles apart, and the more southerly city, Mohenjo-daro, stands very roughly 200 miles up the Indus River.39 These were not, then, cities that had immediate access to the sea, though from Mohenjo-daro it is easy to imagine craft reaching the Indian Ocean, and the Indus civilization possessed many dozens of towns and ports along the ocean shore, in the general area of Karachi, ruling over as much as 800 miles or 1,300 kilometres of coastline, far beyond the Indus estuary. One of the most important harbours lay at Lothal, in the Gulf of Cambay (north-west India), which gave access both to the local river system and to the open sea, and which offered the facilities ships would require at the end of the long journey from the Persian Gulf. Lothal possessed a substantial dockyard and several anchors have been found there. It traded in several directions, for there were contacts with the Neolithic peoples living further south along the west coast of India as well as with the Persian Gulf.40

  So much attention has concentrated on the mystery of how the two highly organized mega-cities came to be built along the Indus river system that rather little interest has been shown in other places, and even the assumption that these were the twin capitals of a single empire is pleasant speculation. Rigid central control was in place, for, as the archaeologist Stuart Piggott observed, the size of bricks used to build the two great cities, the highly standardized pottery, and the weights and measures show ‘absolute uniformity’: ‘there is a terrible efficiency about the Harappā civilization which recalls all the worst of Rome’, while he also observed ‘an isolation and stagnation hard to parallel in any known civilization of the Old World’. Over several hundred years during the second half of the third millennium BC not much changed.41 Although it has proved hard to identify any grand palaces or temples, this was a highly stratified society in which labour gangs were set to work pounding grain into flour. It is assumed that the major rural activity, apart from food production, was the growing of cotton, which is a convenient argument, as this is a product that leaves little residue archaeologically; there is no obvious mention of cotton in Sumerian documents that mention Meluḫḫa, and the Sumerians sent their own textiles to Meluḫḫa, whose main attraction was the luxury items, semi-precious stones, ivory and fine woods, mentioned already. Very occasionally one of the distinctive Sumerian cylinder seals, carved to be rolled across clay, reached Mohenjo-daro, though Sumerian items rarely appear on Indus sites, even though Indus products such as carnelian beads appear quite often in Sumer.42 A piece from an Omani vase also reached Mohenjo-daro, no doubt by sea and river. The best place to look for signs of contact is the port of Lothal, and, to be sure, gold pendants of Sumerian origin and (possibly) Mesopotamian pottery have been found in a merchant’s house there, as has a clay model of a boat. A circular seal found at Lothal, showing goats or gazelles and a dragon, bears close resemblance to circular Sumerian seals.43

  Seals, however, provide the best evidence for contact the other way, from the Indus Valley towards Sumer. It is not just that seals, being made of stone, survive well; they were also used by government officials, priests, merchants and anyone else who wished to set his seal, literally, on property, and that could include goods sent by ship to foreign ports; they were functional, but they were also declarations of identity, and provided vehicles for some of our earliest written texts. The Indus
seals are quite distinctive. Rather than being rolled on clay they were used as stamps, so they are flat and square; they generally portray local animals – tigers, humped oxen (zebu), elephants – and they often carry inscriptions in a distinctive linear script very different to Sumerian cuneiform.44 So if these seals turn up in the Persian Gulf in reasonable quantity, we have evidence of visits by Indian travellers to the area, evidence, in other words, of merchants passing between Meluḫḫa and Sumer. And they have indeed been found in the ruins of major cities such as Lagash and Ur, sometimes showing the sorts of animals familiar from Indus seals, sometimes also containing a few letters in the Indus script. One seal thought to have been found in Iraq portrays a rhinoceros, which never appears in Sumerian art, for the animal was unknown in Mesopotamia. This seal has some Sumerian characteristics as well, such as its shape, and it may provide evidence for Indians settled in the Sumerian lands, who have already been encountered at Lagash; but maybe a better explanation, more in accord with what we know from the Persian Gulf, is that it originated among the mixed settlements to which it is now time to turn.45

  V

  Having identified Magan and Meluḫḫa with reasonable confidence, we are left with the location of Dilmun. This was another place that wandered around the Babylonian map, or rather acquired several identities: Dilmun the abode of the Blessed; Dilmun as a region; Dilmun as a specific place. The merchants Lu-Enlilla and Ea-nasir clearly knew what they meant when they talked of Dilmun: for them, it was a place where one could buy copper and other goods, and it had its own community of traders. What began as a general term for (most likely) the Persian Gulf roadsteads that ran all the way down the Arabian shores from Kuwait to the tip of Oman came to mean one place in particular; the term may originally have signified simply ‘lands to the south’.46 Archaeologists have identified where that was, and where its outstations on the route to Magan were as well, notably Umm an-Nar, which has been mentioned already. The discovery of Dilmun was the work of two scholars from a not particularly prominent Danish museum in Aarhus, Geoffrey Bibby and P. V. Glob, later renowned for his book The Bog People, dealing with the very different topic of the prehistoric victims of human sacrifice found almost perfectly preserved in Danish bogs. Indeed, as in Denmark, the issue was bodies, or rather the 100,000 burial mounds visible in Bahrain and assumed (though this was exaggeration) all to date to some prehistoric era. Glob and Bibby questioned the easy assumption that Bahrain had been created as a vast isle of the dead, a cemetery island, which might, after a fashion, fit with the idea of Dilmun being located there, as a holy island and the abode of the Blessed.47 According to one tablet King Sargon of Akkad had conquered Dilmun, which lay in the ‘Lower Sea’, that is, the Persian Gulf. About 1,600 years later, in the eighth century BC, a warlike king of Assyria, also named Sargon, sent his armies as far as Dilmun, this time said to lie on the ‘Bitter Sea’ and to include an island; its king was Upēri, ‘who lives like a fish’ out in the sea.48 Dilmun could not, then, be very far from Sumer, and the island of Bahrain was an obvious candidate for its location. The connection went back far in time: around 2520 BC King Ur-Nanše, ruler of the Sumerian city of Lagash, asserted that ‘ships of Dilmun, from the foreign lands, brought me wood as tribute’. This visit to the king of Lagash fits neatly with archaeological evidence that settlement on Bahrain became denser around the middle of the third millennium BC.49 The coast of Arabia was an unlikely source of wood for the king of Lagash, and the wood must have been brought to Dilmun from further afield – from Iran or India.50