The Silk Road: A New History

by Valerie Hansen

  The fifteen intact contracts buried in the moneylender Zuo’s tomb record that he usually made smaller loans: between ten and forty silver coins or between three and thirty bolts of degummed silk. Government regulations specified that purchasers should use bolts of silk for large purchases, such as of slaves and livestock, and coins for less valuable items, probably because coins were often in short supply. In line with these regulations, the moneylender Zuo purchased a slave girl for six bolts of degummed silk in 661, and paid 450 silver coins for ninety bundles of hay in 668. Eight contracts record loans of silk or silver coins, while five contracts record the rental of fields, at least once to someone who had borrowed money from him. Unlike so many of the other Turfan documents, the contracts were placed intact in his tomb, probably because Zuo had not managed to collect on them during his life but hoped to do so after his death.42

  The contracts buried in his tomb consistently charge interest of between 10 and 15 percent each month. The rate seemed high even to contemporaries: the Tang Code limited interest to 6 percent each month.43 Ordinary people who fell into debt for a variety of reasons came to the moneylender to borrow money to tide them over. We do not always know what happened to these people, but it is certain that they never paid back the loans, because if they had, the moneylender would have destroyed his copy of the contract as was customary on receiving the final repayment.

  While people in Turfan were using silver coins, people in central China used the same bronze wuzhu coins that they had been using since the second century BCE. The distinct currency spheres, silver in Turfan and points west, bronze in China, persisted after the 640 conquest of the oasis. Only sometime around 700 did people in Turfan shift to bronze coins, which they often strung together in bunches of a thousand, called strings. The latest Astana document to mention silver coins, a tax receipt dated 692, specifies the equivalent value in bronze coins: two silver coins were worth sixty-four bronze coins.44

  The use of silver coins in Turfan during the sixth and seventh centuries reinforces the point that China’s main trading partner at the height of the Silk Road trade, when the Tang stationed massive armies in the northwest, was the Iranian world, not Rome. Recall that no coins minted by the Roman republic (507–27 BCE) or the subsequent principate (27 BCE-330 CE) have been found—so far—anywhere in China. The most thorough survey by Luo Feng, the leading archeologist in Ningxia Province, identifies the earliest Byzantine solidus coins found in China (there are two) as minted in the reign of Theodosius II (409–50) and buried sometime in the early sixth century, and the latest in the mid-eighth century.45

  The period of these Byzantine coins overlaps with that of the Sasanian silver coins, and the two were often found together. Far fewer gold coins have surfaced in China than silver coins: eleven in Xinjiang and thirty-seven in central China for a total of forty-eight (versus more than 1,300 silver coins).46 These coins all appear to be solidus coins. First minted by Constantine (reigned 306–37), and containing 1/72 of a Roman pound’s worth of gold, or 4.55 grams, they show the reigning Byzantine emperor on the face and have an image of the Cross or Christ on the back.47 When Muslim troops conquered large chunks of the Byzantine Empire after defeating the Sasanians, Islamic mints removed all Christian elements from the solidus coins, just as they had removed Zoroastrian elements from the silver coins.

  On close inspection, many of the Byzantine gold coins turn out to be fakes.48 Sometimes they weigh less than the standard weight of genuine coins, or the iconographical details of the emperor’s portrait are wrong, or the lettering on the inscription is incorrect.49 Many have holes punched in them, an indication that they were sewn onto clothing, most likely for use as protective talismans (one is shown in color plate 4A).

  The largest number of gold coins found together in China is five, and it is far more common to unearth a single coin.50 Archeologists have not uncovered anything comparable to the hoards of silver coins from Wuqia and Turfan, yet another indication that the Byzantine gold coins were used for ceremonial purposes and did not circulate as a genuine currency, either in Turfan or in central China.51 None of the Astana documents record a transaction using gold coins, and those that have been excavated from tombs were often used as talismans. The Wuqia hoard, with its 947 silver coins and thirteen bars of gold, confirms this basic pattern: it shows that silver circulated in the form of coins while gold was used as bars.

  As the widespread use of silver coins indicates, Turfan lay halfway between the Iranian and Chinese worlds. During the years of the Silk Road trade, Turfan absorbed many foreign immigrants, none more numerous than the Sogdians who came from Samarkand. Sogdians came to settle in Turfan during the fourth, fifth, and sixth centuries, and the pace of migration heightened considerably after the fall of the Sasanian Empire in 651 and the Islamic conquest of Samarkand in 712.

  Although the Sogdians were famous traders, the Sogdians living in Turfan pursued a wide variety of occupations including cultivating the land, serving as soldiers, running inns, painting, working leather, and selling iron goods.52 When local officials, either under the Gaochang Kingdom or the Tang dynasty, drew up household registers, they did not label which residents were Sogdian and which were not. As a result, modern scholars must identify Sogdians by analyzing their family names and given names. Although the Chinese generally referred to the Sogdians as the people of the “nine jeweled surnames,” most Sogdians adopted one of seven Chinese family names: Kang, used by those coming from Samarkand; An, from Bukhara; Cao, from Kabudhan, north of the Zerafshan River; He, from Kushaniyah between Samarkand and Bukhara; Mi, from either southeast of the Zerafshan River or Panjikent; Shi, from Kesh, modern-day Shahrisabz; and Shi (written with a different character), from Chach, modern-day Tashkent.53 In recent years two Japanese scholars of the Sogdian language, Yoshida Yutaka and Kageyama Etsuko, have reconstructed forty-five different Sogdian given names from their Chinese translations.54 Many of the original Sogdian immigrants who moved to Turfan used these names, while those who lived in China for generations tended to give their descendants traditional Chinese names—in much the same way that immigrants often choose American-sounding names for their children.

  In addition to naming, the Sogdians who migrated to Turfan gradually modified their burial practices, bringing them in line with Chinese practices.55 Because Zoroastrians believe that flesh pollutes the pure earth, they traditionally exposed the dead to scavengers and buried the cleaned bones in ossuaries. Two ossuaries have been found in Turfan.56 Zoroastrians sacrificed animals to the major Zoroastrian deities, including tree, rock, and mountain gods; the god of wind; and the supreme deity, Ahura Mazda.57 It is likely that the political and religious leader of the Sogdian community, a man who bore the title sabao, led these sacrifices.58

  Many of the Sogdians living in Turfan adopted Chinese burial practices, including burying wooden slips to represent servants to serve the dead in the next world.59 Recent excavations of a graveyard to the northeast of Gaochang City in Badamu Village have found more than eighty graves of Sogdians, as indicated by their family names written on Chinese-style epitaphs.60 These naming patterns make it possible to identify Sogdians who appear in different documents, whether household registers recording the names of each member of the household or other materials.61

  Sometime around 600 officials of the Gaochang Kingdom recorded the names of forty-eight merchants who paid a tax, called a scale fee, on the goods they sold to one another.62 After the goods were weighed, the tax was assessed in silver coins. This much-studied document survives as ten paper shoe soles cut from four different sections of the original register. Offering a series of snapshots of individual transactions over the course of a single year in the early seventh century, it is the single most informative document about the commodities exchanged in the Silk Road trade. These documents embody all the joys and the frustrations of the documents pieced together from the Astana graveyard: they give more information than any other materials available, but missing sect
ions of the documents—where they were cut to make shoe soles—mean that they are incomplete.

  Even so, these records highlight the dominant role played by Sogdians in the Silk Road trade. Of the forty-eight names mentioned either as purchasers or sellers of a given good, fully forty-one are Sogdian.63 The scale-fee records suggest a relatively low frequency of trade—a handful of transactions each week—with many weeks in which no tax was collected.64

  Officials recorded all the sales by each day and then twice a month tallied up the total number of coins they collected. The rate of taxation was two silver coins (weighing 8 grams) on two Chinese pounds (jin) of silver, less than one percent. Scholars do not know how much a jin weighed in the year 600: either 6 ounces (200 grams) in the older system or about 1 pound, 3 ounces (600 grams) in the newer one. The lower weight is more likely, but the accompanying chart uses the original units of jin and liang (a Chinese ounce, with sixteen to a jin) because of the uncertainty.65

  The scale-fee register lists thirty-seven transactions over the course of a single year. Brass, medicine, copper, turmeric, and raw sugar traded hands only once, while other goods appear more often: gold, silver, silk thread, aromatics (the term xiang refers broadly to spice, incense, or medicine), and ammonium chloride. The one unfamiliar item on the list, the chemical ammonium chloride, was used as an ingredient in dyes, to work leather, and as a flux to lower the temperature of metals. These documents list ammonium chloride six times, in amounts from a low of 11 Chinese pounds to a high of 251 Chinese pounds. Fragrance, similarly, was traded in both small and large amounts, with a low of 33 Chinese pounds and a high of 800 Chinese pounds—the largest single amount recorded on the list. Gold, as one would expect, appears in small amounts ranging from a quarter to more than half a Chinese pound, and the largest amount of silver did not exceed 8 Chinese pounds. Surprisingly, these documents do not mention bolts of silk cloth, but, since their value was determined by width and length, they would not have been subject to a tax by weight.66

  TABLE 3.1 SCALE-FEE TAX RECEIPTS FOR ONE YEAR AT ONE CHECKPOINT NEAR TURFAN, CA. 600 CE

  The scale-tax documents do not list all the inventory of a given merchant, just individual sales, but even the largest quantity mentioned—800 Chinese pounds—could have been carried by several pack animals.67 We glimpse the same low level of trade in a series of affidavits given during the legal dispute mentioned in the introduction between a Chinese merchant and the brother of his business partner.68 The plaintiff’s Chinese name was Cao Lushan, a clear indicator that he was Sogdian; Cao was one of the nine jeweled surnames, and Lushan was the transcription of Rokhshan, a Sogdian name meaning “bright,” “light,” or “luminous,” cognate with our Persian-derived name Roxanne.

  Appearing before a Chinese court sometime around 670, the younger brother sued a Chinese merchant for the return of an unpaid loan. The Chinese merchant had violated Tang-dynasty contract law, in effect since the 640 conquest of Turfan, the Sogdian contended. As his brother’s heir, he was entitled to 275 bolts of silk. He brought the suit in Turfan, which served as the headquarters of the Anxi Protectorate between 670 and 692.

  At the time of the brother’s death, both he and his Chinese partner, like many merchants of the time, maintained households in the Tang capital of Chang’an and traveled the long route to the Western Regions when business required it. The brother had met his Chinese partner in Gongyuecheng (modern-day Alma-ligh, Xinjiang, near the Chinese border with Kazakhstan) and lent him 275 bolts of silk, which could be carried on several pack animals. Since the two men did not have a common language, they spoke through an interpreter.

  As this case demonstrates, plain silk—undyed with a simple basket-weave pattern—served as a currency alongside bronze coins during the Tang dynasty. Silk had many advantages over bronze coins. The value of the coins fluctuated wildly while the price of silk was more stable. The dimensions of a bolt of silk held remarkably steady between the third and the tenth centuries at 1 Chinese foot, 8 inches (22 inches, or 56 cm) wide and 40 Chinese feet (39 feet, or 12 m) long.69 Also, silk was lighter than bronze coins; the standard unit of one thousand coins could weigh as much as nine pounds (4 kg).70

  After making the loan the Sogdian went south to Kucha, leading two camels, four cattle, and one donkey. These seven pack animals carried his goods, which included silk, bows and arrows, bowls, and saddles. The Sogdian never made it to Kucha; one witness in the trial speculated that he had been killed by bandits who stole his goods. Even though the surviving brother did not have a copy of the original loan agreement, he was able to present two witnesses, both Sogdians, to the signing of the contract. According to Tang law, their oral testimony constituted sufficient proof of the original agreement. The Chinese court ruled in favor of the living Sogdian brother and against the Chinese merchant, ordering him to repay his debt.

  The deceased Sogdian brother had traveled with seven pack animals carrying his goods. Other caravans were about the same size, as we learn from twelve surviving travel passes found at Turfan. Like similar documents from Niya and Kucha, these record the members of each party—both people and animals—proceeding together from one destination to another and all the places to which they were allowed to go. At the beginning of his trip, each traveler applied for a travel pass that listed his ultimate destination, a few intermediate points, and the people and animals traveling with him. In addition, each time he entered a new prefecture, he received a document verifying the people and animals traveling with him.

  At every guard station, both within and between prefectures, local officials checked that all the people—classified as relatives of the primary traveler, servants (zuoren), or slaves—and all the draft animals rightfully belonged to him. Tang-dynasty law prohibited the enslavement of people to repay debts; the only legal slaves were those who were born to slave parents or who had been purchased with a contract registered with the authorities and who had the proper market certificate to show for it.71 Tang law was equally strict about animals: a traveler could bring a donkey, horse, camel, or cow by a checkpoint only if he carried a market certificate for any purchased animals. Like officials at Kucha, Turfan officials did not record what cargo each caravan carried. Still, the travel passes do give the size of caravans, which usually included four or five people and about ten animals.72

  One merchant, named Shi Randian (in Sogdian, Zhemat-yan, “Favor of the god Zhemat”), appears in several documents, so it is possible to track his movements in the years 732 and 733 and to grasp the level of government supervision. With his household registered in Turfan, Shi carried a travel pass allowing him to proceed from Guazhou to Hami via Dunhuang and then all the way west to Kucha, on a route similar to Xuanzang’s. Surviving documents record the approval of four different officials on the leg from Guazhou to Dunhuang. The caravan was checked two times on the nineteenth day of the third month, once on the twentieth day, and again on the twenty-first day.73 On his first trip, Shi traveled with two servants, a slave, and ten horses, but on his way back he had purchased an additional horse (for eighteen rolls of silk) and a mule.74 Since he carried the necessary market certificate demonstrating he had purchased them legally, he was allowed to proceed. A small-scale trader, Shi carried his wares on ten horses, and bought and sold individual animals from time to time to augment his income.

  Officials did stop those caravans whose papers were not in order. In 733, one resident of Chang’an, Wang Fengxian, was returning from a trip supplying the army in Kucha. He applied for a new travel pass, because he had diverged from his permitted route to pursue a debtor who, he said, owed him three thousand bronze coins. Local officials apprehended him in a town to which he had not been given permission to travel. When he explained that he had fallen ill, and others verified his account, they allowed him to proceed.75 The Turfan travel passes, like those from Niya and Kucha, confirm that all travelers were subject to the intense scrutiny of the authorities and could not diverge from their itineraries without officia
l permission.

  MAP OF GAOCHANG CITY

  By carefully examining the roads and buildings in the ruins of Gaochang, archeologists have identified distinct neighborhoods in the city. The Tang authorities divided Gaochang City into different wards, just as in the cities of central China, and these wards continued to be used under the Uighurs. The commercial district in the southwest of the city housed workshops where craftsmen made handicrafts sold at the local market. The authorities divided vendors of different goods into groups, each with its own row of stalls at the market, and visited regularly to record prices.

  Once a caravan passed through the official checkpoint into a new town, its members could find innkeepers, who also stored their wares for them, doctors, who treated their illnesses, and prostitutes, who, as today, have left little documentary evidence of their activities.76 The caravans visited the markets in each town they stopped in. Tang-dynasty law required certain designated officials, known as market supervisors, to inspect markets every ten days and record three prices—the high, the low, and the medium—of all commodities on sale.77 Such a register, in 121 fragments, for the main market in Turfan survives and is dated 743; some sections are dated to the fourteenth day of a certain month, others to the twenty-eighth, an indication that officials collected the data on two separate occasions.78 Chinese markets were divided into rows, where merchants sold related goods; the Turfan register lists over 350 goods divided into over ten different rows.

  As informative as it is, the price list does not reveal everything about the market. Some of the high-medium-low pricing sequences—6/5/4—seem suspiciously regular, for example, and the register gives the same prices for livestock, regardless of age or overall health. Nor does the register reveal how much of any single item is for sale or how many different stalls offer that particular item.