The northern part of room 4 contained a big clay jar lying on its side. Scattered near it were twenty-three willow sticks with writing on them, as if the sticks had fallen from the opening of the clay jar. These sticks contain the records of household expenditures compiled by the steward for his lord and kept in this compact local archive.71 He used willow sticks as a writing material—not paper or leather—because they were cheaper and more readily available.
Since the steward recorded the dates on which different amounts of wine and wheat were consumed in hosting various visitors, his ledgers offer a sketch of the local economy. On several occasions people from nearby villages brought cartloads of grain to the fortress, which they gave to the lord, perhaps as a tax paid in kind, and the steward’s accounts report that they received grain from him as well. Herding animals was a major economic activity: people ate sheep and goats, and they made suits from animal skins, sometimes from as many as fifty, but usually from a smaller number. One such document (A17) lists various expenditures: 200 dirhams for a horse; 100 dirhams to build a roof; 50 dirhams to the Zoroastrian priest; 15 dirhams for both a doctor and a wine pourer; 11 for a cow to be eaten at a New Year’s dinner; 8 dirhams for a document drafter; 8 dirhams for paper, silk, and butter; and 5 dirhams for an executioner. Although scholars are not certain what kind of coin circulated in Samarkand, the dirham was the main unit of silver currency at the time throughout the Arabic-speaking world and had superseded Sasanian silver coins. Almost all the goods appearing on the willow-branch accounts, with the important exceptions of paper and silk, both from China, were produced locally, giving the impression that the local Sogdian economy, at least during these years of conflict, functioned largely through barter.
In addition to the willow sticks, the site produced nearly sixty documents on paper and leather that had originally been stored on the second story and were scattered among the remains of the collapsed ceilings of the first and second stories of rooms 2 and 3.72 A third findspot for leather documents was the basket the shepherd boys had unearthed.
Among these ninety-seven documents were three legal contracts written on trapezoidal pieces of leather. They reveal a sophisticated legal apparatus for that time. Although leather may seem like a cumbersome writing material, it was used throughout the entire Arabic-speaking world (Europeans were using parchment, also made from treated sheepskin, in the same centuries), and experienced scribes could record detailed agreements on it. By far the lengthiest, and so the most informative, text from Mount Mugh is the marriage contract and the accompanying document, labeled “the bride’s script,” or copy, in which the husband restates his obligations to the bride’s family. Both were found in the basket Puloti handed over to the authorities.73
The marriage contract and the bride’s script are dated to the tenth year of King Tarkhun’s reign, or 710. Together, the two documents run a full ninety lines on two pieces of leather, respectively 8 inches (21 cm) and 6 inches (15.5 cm) long. These documents specify the terms of the exchange of the Sogdian woman Chat from her guardian, Cher, the ruler of Navikat (a Sogdian city in the Semirech’e region of modern Kazakhstan), to her new husband Ot-tegin, whose name clearly marks him as a Turk. Because her father’s name is mentioned even though he plays no role in the arrangements, it seems that Chat was the ward of Cher.
This pre-Islamic contract is striking for what it reveals about the strict reciprocity of obligations in this society: just as a husband can end the marriage under certain circumstances, so too can a wife end the marriage under those same circumstances. The Sogdian contract invokes the legal term for a special kind of marriage that granted husband and wife equal rights in a number of respects.74 The agreement begins by explaining that the husband is obliged to provide “food, garments, and ornaments,” as his wife is “a lady possessing authority in his own house, the way a noble man treats a noble woman, his wife.” She, in turn, “must always conform to his well-being and obey his orders as befits a wife, the way a noble woman treats a noble man, her husband.”75
Then, much like a modern prenuptial agreement, the contract outlines what will happen should things go wrong. If the husband should take “another wife or concubine, or keep another woman that does not please [his wife] Chat,” he promises to pay her a fine of “thirty good, pure dirhams of [the type] Den” and send the woman away. If he decides to end the marriage, then he may do so, but he must provide his wife with food and return her dowry and all the presents that she gave to him when they were married. Neither husband nor wife will owe the other any compensation. The husband will then be free to remarry. Notably, the wife, too, is entitled to end the marriage, but only after returning her husband’s gifts; she is to retain her own property as well as a payment from him. Should the marriage end, neither party will be responsible for the other’s crimes, and only the criminal party shall be obliged to pay any penalties.
The contract confirms the fluidity of Sogdian social divisions. Should either the husband or the wife become the “slave, hostage, prisoner, or dependant,” of someone else, his/her former spouse will not be responsible. Clearly some people in this society are better off than others, and those signing contracts with 30-dirham penalties are among the better off, but they, just like the less privileged, faced the real prospect that, if their fortunes changed, they, too, could be reduced to slavery.
The statement of the husband’s obligations in the wife’s script repeats much of this information but adds a few new clauses. Ot-tegin begins by saying, “And, Sir, by god Mithra! I shall neither sell her nor pawn her.”76 Mithra, the guardian of truth and contracts, was one of the three most important gods of the Zoroastrians and ranked just below the supreme deity, Ahura Mazda, to whom the word “God” would ordinarily refer. Ot-tegin promises to return Chat to her guardian in the event that the marriage ends, at either his or her instigation. In addition, if “someone, from my side or from the enemies’ side takes her or detains her,” he will obtain her immediate release. He promises, too, that he will pay a fine of one hundred dirhams if the marriage ends but he fails to return her unharmed to Cher’s family. If he does not make the payment promptly, he will pay a 20 percent late penalty on the unpaid balance. Much of this document spells out the procedures by which the guardian can obtain payment; for instance, it names a guarantor whom the guardian can seek out. The entire population is enjoined to monitor the agreement, which was signed in a “Foundation Hall” in front of witnesses.
The two other contracts from Mount Mugh, one for the rental of mills (B-4), and the other for the sale of a burial plot (B-8), share the same overall structure as the marriage contract, though they are much shorter. Both give the date (the year of the king’s reign, the month, and the day), the names of the parties involved, the item to be transferred, the conditions of transfer, and the names of the witnesses and of the scribe.
The contract for the lease of three mills to a man by Devashtich specifies an annual rental of 460 units of flour.77 Like the willow-branch documents, this contract requires payment in kind, here in flour. But this contract goes beyond a simple statement of rent. Forty-two lines long, it is a sophisticated legal instrument specifying the time within which the lessee must pay the ruler and the consequences should he fail to make complete payment.
The third contract is for the rent of a burial place for 25 dirhams.78 The contract sets the terms by which two sons rent a mud-built “eskase” location for burial from two brothers. It may mark a truce between two feuding families—the family of the brothers renting the plot and that of their enemies whom they fear may disturb their mourning. Zoroastrians disposed of their dead first by placing them in a structure outside, called a Tower of Silence by modern Zoroastrians, where animals of prey could eat the flesh, and then placing the cleaned bones in a well, called an eskase in this contract.79 Yet, because no such burial wells have been found yet in the region of Sogdiana, others suggest the word may refer to a naus structure for the remains of the dead like those found in Pa
njikent.80
The Mount Mugh contracts help us understand that the fortress contained more than just the personal archive of the ruler Devashtich. Certain documents, like the agreement specifying the amount to be paid to him as rent for his mills, clearly belonged to him. But why would he have kept a copy of the contract giving the complicated terms of the marriage between a Turk and his Sogdian bride? Or the document leasing the burial place?
Quite possibly the residents of Mount Mugh, including the bride Chat, brought all their important legal documents with them for safekeeping, perhaps during the final siege of the fortress. They may have hoped to recover their documents after the Arab threat had been eliminated. But the contracts remained in the fortress of Mount Mugh—untouched—until the shepherd boys found them in 1932. If this is what happened, it explains why the Mount Mugh documents included the correspondence of not just the ruler Devashtich but also the letters of several other lower-ranking lords who also took refuge in the citadel.
By combining information in al-Tabari’s detailed chronicle with that from the Mount Mugh documents, we can reconstruct the events leading to the fall of the Mount Mugh fortress.81 The chronicles record that a new Arab governor, nicknamed “the Lady,” fought the Sogdians between the autumn of 720 and the spring of 722. The Sogdians were allied with the Turgesh, a people who had originally been the subjects of the Western Turks but who, between 715 and 740, gained control of some of Western Turks’ territory.82 In 721, Devashtich, who had been the local ruler of Panjikent for some fourteen years, was officially crowned as “king of Sughd, lord of Samarkand.”83
Devashtich claimed to be the successor of Tarxun, the last ruler of Samarkand. Tarxun had surrendered to Qutayba in 709, but following a local revolt he either committed suicide or was executed in 710. A man named Ghurak took over as his successor. Qutayba, claiming to avenge Tarxun’s death attacked the city again and gained control in 712. When Ghurak surrendered, he signed a treaty promising a onetime payment of 2 million dirhams and then 200,000 dirhams each year thereafter.84 Qutayba and some local lords accepted Ghurak’s claim to succeed Tarxun, but others who lived to the southwest of Samarkand supported Devashtich. For a decade the two rivals coexisted, but not much is known about this period.
In 719 Devashtich wrote to the Arab governor of Khurasan deferentially as though he was his subordinate, but by the summer of 721 he was optimistic about his chances of defeating the Arabs. At this time, he wrote a letter (document V-17) to Afshun, a lord of the town of Khakhsar, 8–10 miles (12–16 km) southwest of Samarkand, in which he described a “large army come down, both of Turks and Chinese.” Apparently the Turgesh, some Chinese, and the king of Ferghana to the east had formed an alliance against the Islamic forces. The Mount Mugh letters provide the only evidence of Chinese involvement in these events, and another letter (document V-18) mentions a “Chinese” page (the word for “page” is uncertain). The word “Chinese” may refer to someone ethnically Chinese from the Western Regions—not necessarily forces sent by the central government in Chang’an.85
The documents reveal that a year later, most likely in 722, the circumstances have completely changed. A messenger reports that the “Turks” are nowhere to be found, while another, possibly a postmaster, describes the fall of Khujand in Ferghana to the Muslim forces and the surrender of 14,000 people.86 The chronicler al-Tabari reports that the Sogdians divided into two groups. The larger, consisting of at least 5,000, went to Ferghana, where they were denied entry, and a Muslim army slaughtered them.87 A much smaller group, of perhaps one hundred families, sided with Devashtich and fled to the fortress of Mount Mugh.88
During the final massacre by the Arab forces, only the merchants in this larger group could afford to pay a ransom to the invading forces in exchange for their safety. Taxation was a major issue for the newly conquered peoples of Central Asia, who hoped to avoid heavier taxes by converting to Islam and thereby becoming eligible for the preferential tax rates for Muslims. During the eighth century, however, the caliphate desperately needed all revenues for its war effort, and individual governors did not always grant their new converts the preferential rates. Many Sogdians fled to the Turkish lands or to China as a result.
Devashtich and his followers, who numbered perhaps only a hundred men and their families, moved to the Mount Mugh fortress (called Abghar by al-Tabari).89 They sent a small force outside the fortress to fight the Muslim forces, who drove them back to the fortress and defeated them after laying siege. Following his defeat, Devashtich requested a safe conduct from Said al-Harashi, who initially granted his request. The one hundred families remaining in the citadel surrendered the contents of the fortress in exchange for their freedom. The Arab army commander, al-Tabari reports, then auctioned off the contents of the fortress, leaving a fifth for the state treasury, as Islamic law prescribes. That is why the fortress was nearly empty when the Soviet archeologists excavated the site in 1933. Everything of any value had already been removed, and the paper and leather documents must have escaped notice.
Although the Arab commander had promised Devashtich safe conduct, he reneged on his word. Al-Tabari describes his grisly end. The commander “slew al-Diwashini [Devashtich], crucifying him on a [Zoroastrian] burial building [naus]. He imposed upon the people of Rabinjan the obligation to pay one hundred [dinars] if the body were removed from its place.... He sent al-Diwashini’s head to Iraq and his left hand to Sulayman b. Abi al-Sari in Tukharistan.”90 The method of killing suggests that Devashtich was an important figure. Because Devashtich represented Sogdian resistance, the Arab commander chose an extreme method to dispose of his body.91 (He was subsequently dismissed for having used such a gruesome punishment.)
Devashtich’s death was just one chapter, and a minor one at that, in the Islamic conquest of Samarkand. Within a few decades Muslim troops had secured the region, and over time Persian displaced Sogdian and Islam replaced Zoroastrianism. At the 751 battle of Talas, which took place at what is now Kazakhstan, Muslim troops defeated a Chinese army largely because the forces of a nomadic people, the Qarluqs, defected to the Islamic side. Four years later General An Lushan rose up against the Tang dynasty, and the dynasty was forced to withdraw its troops from Central Asia to suppress the mutiny. These two events in rapid succession meant that, after the mid-eighth century, the city of Samarkand and the surrounding region of Sogdiana no longer looked east to China. The Islamicization of Sogdiana prompted many of the Sogdians living in China to settle there.
The Mount Mugh documents predate both the Islamicization of Central Asia and the spread of paper-making technology to the region. The different materials used to write the Mount Mugh documents show that local rulers were willing to pay for Chinese paper because of its durability and convenience, but the residents of Central Asia continued to use leather for important documents, like the single letter in Arabic that Kratchkovsky deciphered, and reserved willow sticks for less important matters like household accounts.
A rare indicator of long-distance trade at Mount Mugh is the Chinese paper found there. Three Chinese-language documents were pieced together from eight fragments, all on recycled paper originating in China. No one at Mount Mugh actually wrote Chinese. One of the documents originated as an official document written in Wuwei, Gansu, a thriving city on the Chinese silk route to the east of Dunhuang. After it was discarded and sold as paper (the reverse side was blank and could still be used), Silk Road traders brought it west some 2,000 miles (3,600 kilometers) to Mount Mugh.92
In the eighth and ninth centuries, Chinese paper reached far into Central Asia, as far as the Caucasus Mountains, to the site of Moshchevaia Balka, literally “ravine of the mummies/relics.” The site, consisting of burials either on limestone terraces or dug out of the hillside, near the northeast edge of the Black Sea is the farthest location from China where Chinese paper has been found to date. In the early twentieth century, excavators unearthed a few scraps of paper with Chinese writing on it. The most complete document
, 6 inches (15 cm) by 3 inches (8 cm), contained a few lines in running hand that give the date and then the amount of various expenditures (2,000 coins, 800 coins). Although extremely fragmentary, it appears to be a ledger of expenses.93 The site produced other items, all clearly of Chinese origin: a piece of silk with a design showing a Buddhist deity on it as well as a man riding a horse (the prince Siddhartha before he left the palace?), a scrap of paper with a Buddhist text, and a fragment of an envelope from some paper mache object. These items clearly demonstrate that at least Chinese paper and silk paintings—and possibly even Chinese merchants—had reached the Caucasus region sometime between the eighth and ninth centuries.94
During the eighth century people living in Central Asia learned how to manufacture paper. An Arab account reports that at the 751 Battle of Talas the Abba-sid Caliphate won a resounding victory over the Chinese and brought back prisoners of war to their capital at Baghdad. Some of the POWs taught their captors how to make paper.95
As with other legends about technological transfer, this one is not necessarily credible.96 The techniques of making paper—beating a mixture of organic material and rags to a pulp and drying it on a screen—were not difficult to replicate. The technology steadily spread out from central China, reaching Central Asia by the eighth century. After the year 800, paper gradually replaced leather as the primary writing material of the Islamic world. Paper had many advantages over leather: low in cost and quickly produced, it was far more convenient than leather and much more available than papyrus, which grew only in Egypt. Paper entered Christian Europe through its Islamic portals at Spain and Sicily, in the late eleventh and early twelfth centuries.
The Silk Road: A New History Page 17