The Belitung shipwreck

In 1998, Indonesian villagers accidentally discovered, in shallow waters off the island of Belitung, near Sumatra, the remains of a ship that had lain on the seabed for over a millennium.Footnote ⁴ Two seasons of underwater archaeology revealed its astonishing treasure: a tightly-packed cargo of some 60,000 Chinese ceramics (Figure 1); about ten tons of lead ingots, used as ballast and probably also as a commodity; metalwork, mirrors and other precious objects; aromatics; and some personal belongings. The ship, having been loaded with these objects in China, appears to have set sail for a port in the kingdom of Srivijaya, perhaps in Java or Sumatra, and sunk before it could head back to the Gulf. The time frame for this short-lived journey can be circumscribed, on the basis of dated and datable objects found on board, to the first half of the ninth century, most probably the 820s to 840s.Footnote ⁵

Figure 1. Dusun-type jar recovered in the Belitung wreck with its load of Changsha bowls. Source: Krahl et al. (eds.), Shipwrecked, Fig. 47.

The existence of sea contacts between China and the early Islamic world had long been known from texts and archaeological finds across the Indian Ocean,Footnote ⁶ but this was the first time that such a concentration of evidence came to light. The intended route of the Belitung ship was, in all likelihood, a variant on the ones described by writers of the period: in Chinese, Jia Dan (729–805), a high-ranking Tang official who seems to have obtained information from West Asians; and in Arabic, the Abbasid court official Ibn Khurdādhbih (also Khurradhābih, ca. 825–911) and the anonymous compiler of the Akhbār al-ṣīn wa’l-hind (ca. 851).Footnote ⁷ While some place names remain difficult to identify, these writers are in agreement over the main elements of the itinerary. In West Asia, major cities in Iraq, especially Baghdad from the time of its foundation, in 762, were the main markets of this trade. Baghdad was connected to the Gulf by the Tigris and Euphrates leading to Basra, the primary port of Iraq, and Ubulla (the ancient Apologos).Footnote ⁸

The waters at these two ports were too shallow for the largest sea-going ships, and Siraf, on the Iranian coast, emerged as the major storehouse at which goods were unloaded and brought onto smaller vessels or despatched overland.Footnote ⁹ From the Gulf, eastbound ships would sail to Suhar or Muscat, in Oman, then either follow the coast of Sindh or use the monsoon winds to sail across the Arabian sea to Khambhat, in Gujarat, followed by southern India and Sri Lanka (Mantai), before crossing the Gulf of Bengal to reach Thailand, the straits of Malacca and head north towards Guangzhou (Canton, Ar. Khānfū), the main port for this trade in China. Some ships pushed further north towards Hangzhou or Yangzhou, at the mouth of the Yangzi River, which was connected to Chang’an, the Tang capital, through a network of rivers and the Grand Canal.

According to the indications given in the Akhbār al-ṣīn wa’l-hind, by departing in October in order to take advantage of monsoon winds, the trip from Muscat to Guangzhou took around five months of sailing, not counting stops and the initial leg of the journey from Siraf or Basra.Footnote ¹⁰ This time scale has found confirmation in the five months it took the Jewel of Muscat, a modern ship built on the basis of the Belitung wreck using traditional methods, to sail from Oman to Singapore in 2010.Footnote ¹¹ If distances remained unchanged, the time it took to link Basra to Guangzhou had shrunk to less than a year, bringing people at both ends of the continent into ever closer contact.

These direct links did not preclude the continuation of more segmented trade along the sea routes. The presence of Iranian and Arab merchants in coastal cities of India and Sri Lanka is attested long before the rise of Islam; it persisted during and after the period under study.Footnote ¹² Throughout the first millennium, ports such as Kullam and Mantai served as emporia where goods from China and the Islamic world were traded by Indian and foreign middlemen.Footnote ¹³ West Asian traders who had settled in these regions, if they ever became involved in the China trade, would still have been recognized as dashi and bosi in Chinese sources. A further pattern of trade is reflected by the mention by al-Masʿūdī (ca. 893–956) of a Sogdian merchant boarding an Arab ship from Oman to Kalah before 878, then completing his journey to Guangzhou on what is explicitly called a “ship of the Chinese” (markab al-ṣīnīyyīn): his goods were presumably trans-shipped across the Kra isthmus, thereby avoiding the Malacca straits.Footnote ¹⁴ Chinese ships, in other words, may sometimes have assumed the final leg of the journey in the South China Sea. These nuances reflect the complexities of trading diasporas: the relative weight of each mode of exchange may hardly be known, beyond the fact that direct seafaring increased significantly in this period.

The physical structure of the Belitung ship provides a unique piece of material evidence in this respect. The ship was built using glue and ropes, but no nails: this technique, like the shape of the hull, points to an origin in South or West Asia, rather than East Asia.Footnote ¹⁵ Furthermore, the scientific analysis of timbers used in its construction has shown that several of these were African. Trees native to the Arab-Persian Gulf were inadequate for shipbuilding, which led to the regular import of timbers for this and other purposes, whereas India was home to numerous suitable species. In all likelihood, the ship was thus built in the Gulf or the Arabian peninsula from imported East African wood.

The bulk of the cargo consisted of Changsha wares of different forms, including some 55,000 bowls and 1,500 ewers carrying a variety of motifs in underglaze painting, from ships and monsters to calligraphy (Figure 2). These were mass-produced for export by the Changsha kilns of Hunan province.Footnote ¹⁶ Other ceramics found in the ship responded primarily to Islamic tastes: green earthenwares, some probably from Gongxian, with painted and moulded decoration, as well as incised motifs inspired by Islamic pottery;Footnote ¹⁷ and three blue-and-white stonewares from Baihe in Gongyi county (Henan) echoing Islamic blue-and-white ceramics, but with an unmistakably Chinese approach to the painted line.Footnote ¹⁸

Figure 2. Changsha bowl with poetic inscription, from the Belitung shipwreck. Diameter 15 cm. Source: Krahl et al. (eds.), Shipwrecked, Fig. 106.

Finer pieces of stoneware were also recovered, but in their hundreds rather than their thousands, somewhat expectedly given that their production in China was more limited. These include Yue wares produced near Hangzhou, with their typical green glaze;Footnote ¹⁹ high-fired white stoneware from the Xing and Ding kilns of Hebei;Footnote ²⁰ and imitations of these two types produced respectively in Guangdong in the south of China and Gongxian (Henan) in the north.Footnote ²¹ The appreciation of fine stoneware was just developing in China itself during the Tang era, alongside the art of tea drinking. The Chajing (Classic of Tea) by Lu Yu (730s-ca. 804), the first great classic on the subject, discusses high quality bowls and utensils alongside the beverage itself. Local kiln names, such as Yue, Xing, Ding and Gongxian, became akin to brand names recognised—or avoided—by connoisseurs.Footnote ²² Hence the contents of the ship suggest that alongside dominant wares of lesser workmanship, ceramics highly valued in the Chinese market were sent to West Asia in more limited but still significant quantities. This conclusion is echoed by archaeological finds in Iraq and the Gulf,Footnote ²³ where these goods must have changed hands at high prices.

Smaller numbers of luxury objects in other media were discovered on board the shipwreck. Among them are 28 bronze mirrors of round, polylobed or square shape, some finely ornate, others of lesser quality. These were probably made in Yangzhou, the most renowned centre for this craft, in the eighth and ninth centuries. The only dated specimen was made in that city in 759; its metal, asserts the Chinese inscription, was “smelted a hundred times”—a claim often made by foundries, and taken with a pinch of salt by some contemporary Chinese writers.Footnote ²⁴ One additional mirror found in the wreck was made in the Han era (206 bce -220 ce): astonishingly, it would have already been an antiquity of about a thousand years when it was taken on board the ship.

The cargo also contained over thirty fine gold and silver objects, including cups, dishes, bowls, a flask, a golden bracelet, and silver boxes for herbs, medicines and other substances. These are typical of Chinese luxury metalware of the early ninth century, when such objects were mainly produced around the mouth of the Yangzi River, in the modern provinces of Jiangsu and Zhejiang, and again in Yangzhou.Footnote ²⁵ The finest Yangzhou mirrors were regularly presented as tribute gifts to the imperial court; likewise, high-quality metalwork of the kind found on the ship was a highly prized status symbol amongst the Tang ruling élite, as notably revealed by its presence at burial sites.Footnote ²⁶

Some of the most precious cargo may have been intended as tribute. In the course of their travels, merchants had to engage in gift giving in order to retain the favour of local authorities.Footnote ²⁷ The remarkable quality of some luxury objects in the ship could alternatively suggest the presence of a Chinese embassy headed for an unknown destination to the south or west, or of a foreign delegation returning from Chang’an with imperial presents. “Rewards” were commonly given in return for “gifts” in the course of diplomatic exchange, in a tacit form of trade between ruling élites.Footnote ²⁸ The iconography of the gold and silver vessels, with its tied knots, paired animals, auspicious symbols and exotic beasts could arguably be related to such a context. Indeed, two of the green-painted wares found in the ship carry Chinese inscriptions that identify them as tribute to the Tang imperial treasury.Footnote ²⁹

At one level, the ship thus provides a vivid illustration of the goods carried in this trade. While their bulk consisted of mass-produced export ceramics, the range of luxury objects reflecting the highest standards of Chinese production might reflect the activities not only of merchants, but also of political élites. The range of cargo also attests to the extensive supply networks garnered by West Asian merchants. These reached deep into China, from Hebei in the north through Henan and Hunan in the centre to the mouth of the Yangzi river and Guangdong in the south. The goods were probably sourced through Chinese middlemen and gathered in port cities: the mirrors and metalwork, together with the ceramics from Hebei, Henan, Yue and possibly Hunan, point to Yangzhou or Hangzhou, whence they would have been shipped to Guangzhou. Whether or not the Belitung ship began its journey in that region, its cargo must have been consolidated at Guangzhou, as suggested by the presence of ceramics from Guangdong, and most of all by the position of that city as the main Chinese hub of this trade, at the south-western edge of the Chinese coast.

Unusually for this period, the goods recovered within the Belitung ship also offer a glimpse of everyday life below the level of court and beyond the strict interests of commerce: among them were cooking utensils; a bone dice and ivory pieces from a board game; a Chinese lantern and oil lamp; a tiny glass bottle from West Asia, with unidentified organic content; Chinese tweezers, a Chinese needle, and a cymbal.Footnote ³⁰ A few objects could arguably reflect the imprint of Chinese travellers: a ceramic tea grinder; Chinese spoons; and a Chinese inkstone with an engraved insect at one end (Figure 3).Footnote ³¹ These eclectic finds point to the obvious: people lived in close contact and spoke to each other on board these ships. The Belitung wreck thus exposes with unprecedented immediacy some key aspects of this trade, whilst raising questions about its nature. Who were the merchants involved? How and why did they come to source such a wide range of goods in China? What other social groups may have been seen on board their ships? What information can be retrieved, more than a millennium after the fact, about the human interactions behind the transportation of goods?

Figure 3. Chinese inkstone with engraved insect, from the Belitung shipwreck. Length 10 cm. Source: Krahl et al. (eds.), Shipwrecked, Cat. 14.

Baghdad, Chang’an, and the market for luxury goods

In this period, northern China and Iraq were amongst the richest regions in the world, their wealth driven by the combination of imperial revenue, commerce and agricultural expansion in both the Yangzi basin and the Sawād, the fertile area south of Baghdad, between the Tigris and Euphrates.Footnote ³² The tax income of these two immensely rich empires spurred the rapid growth of their capital cities, which soon outstripped this economic base. Based on Chinese imperial records, Chang’an must have reached at least a million inhabitants during the Tang period. This figure could almost be doubled if one counted the suburban counties, but the population also periodically dropped due to political unrest, notably the An Lushan and Huang Chao rebellions.Footnote ³³ One estimate also put the population of Luoyang, the “Eastern capital” of the Tang, at 500,000 inhabitants.Footnote ³⁴

Comparable levels of urban concentration were only reached, outside China, in Iraq. Basra probably had more than 250,000 inhabitants within a few decades of its foundation in 636.Footnote ³⁵ Baghdad, the new Abbasid capital founded by the caliph al-Manṣūr in 762, soon grew into a city of enormous proportions. In the absence of reliable data, its population is notoriously difficult to assess, with estimates ranging from 280,000 to 2,000,000 inhabitants, though most writers agree to place it above 500,000.Footnote ³⁶ By all accounts, it far outgrew the largest cities of the Mediterranean world, Constantinople and Cordoba. Samarra, the second Abbasid capital, founded in 836, must also have reached a population of several hundred thousands, though it remained largely dependent on imperial revenue and quickly declined after the return of the caliphal seat to Baghdad in 892.Footnote ³⁷

These figures take on their full weight when considered in relation to world population, which stood around 200 million at the time.Footnote ³⁸ Between them, the five cities mentioned above—Chang’an, Luoyang, Baghdad, Basra and Samarra—would therefore have been home to well over one percent of mankind, the equivalent of 100 million people in today's world. These numbers, being based on more or less reliable estimates, only provide a rough order of scale, but they suffice to underline the pulling power of these centres in the Old World's economy.

Baghdad and Chang’an were the two great cosmopolitan cities of the age. The constant flow of foreign embassies into the Tang capital brought into its walls people and luxury goods from the breadth of Asia.Footnote ³⁹ Chang’an housed sizeable communities of merchants, being located at the eastern end of the land routes linking China to Central and West Asia, with further connections to the South China Sea through the Grand Canal. The city housed several East Syrian (“Nestorian”) churches, as well as Zoroastrian and Manichaean temples.Footnote ⁴⁰ Given its geographical location, the foreign population of Chang’an was mostly Japanese, Korean, Turkish, Uighur and Sogdian, whereas cities to the south, especially Guangzhou, received larger numbers of West Asians, Indians and Southeast Asians.

Baghdad, as a new foundation, drew its initial population entirely from outside its walls. The original city plan was round, with the mosque-palace complex at its centre. There has been an ongoing debate, which this is not the place to revisit, as to whether this form was based on Central Asian and Sasanian models or stemmed from functional imperatives.Footnote ⁴¹ The Church of the East relocated its seat to Baghdad from Ctesiphon, the former Sasanian capital, under the patriarchate of Timothy I (780–823), as did the Manichaeans until increased persecutions drove the latter away in the early tenth century. The city also housed the heads of the two Babylonian yeshivot, the foremost centres of Jewish learning. Soldiers, political élites and merchants from the lands between North Africa and Central Asia were attracted to the possibilities offered by the Abbasid capital; other populations, notably slaves of African, Turkish and Slav origin, were relocated there by force.Footnote ⁴² The city's higher classes were consumers of foreign and exotic objects, such as Indian and East African sandals, sari-like garments,Footnote ⁴³ and Chinese goods.

Upon founding the city, al-Manṣūr would have declared: “Here is the Tigris with nothing between us and China, and on it arrives everything that the sea can bring”.Footnote ⁴⁴ Whether or not the saying is apocryphal, it does reflect an economic reality: a century later, al-Yaʿqūbī (d. 897) noted that the Sūq Khuḍayr at Baghdad was “the treasure-trove of Chinese rarities” (maʿdan ṭarāʾif al-ṣīn).Footnote ⁴⁵ In this period, Chinese goods, and especially ceramics, were esteemed amongst urban élites of the Islamic world, to such an extent that any fine bowl or plate came to be generically called ṣīnīyya (“Chinese,” used as a substantive) in Arabic. As noted by al-Thaʿālibī (961–1038):

This propensity makes it difficult to single out references to actual Chinese wares in Arabic sources, although in some instances they are more explicitly identified.Footnote ⁴⁷ The appreciation of things Chinese was noted emphatically by al-Sīrāfī (early tenth century):

The Arabic word translated here as “designing” (naqsh) can indicate, depending on the context, painting, sculpture or engraving. The presence in the Belitung wreck, among thousands of Changsha bowls, of a few examples with poems painted on their surface (Figure 2) could suggest that some in their intended clientèle, probably in Iraq, developed an aesthetic appreciation for Chinese calligraphy.Footnote ⁴⁹ But this idea remains speculative, and buyers at either end of the continent were animated by fundamentally different tastes: the purity of naked forms sought by the Chinese was to some extent echoed by an admiration for the material qualities of Chinese ceramics in the Islamic world; but a strong appetite for decorative motifs, especially geometrical, prevailed in these regions (as indeed in the Mediterranean world)—a demand partly met by Chinese export wares and, as we shall see, by new types of Iraqi wares. The growing taste for this particular range of goods must have contributed to the expansion of sea routes, for—unlike overland caravans—ships could hold large amounts of cargo packed in heavy jars, with less risk of breakage.Footnote ⁵⁰ The quantities of ceramics stacked into such jars and fitted within the relatively small space of the Belitung shipwreck (Figure 1) attest to this factor; by contrast, finds of imported ceramics are extremely rare in Central Asia prior to the fourteenth century.Footnote ⁵¹

Triggering factors in the Islamic world

These dynamics ought to be replaced within a larger framework, to which further elements will be added as the discussion progresses. The Muslim conquests of the seventh century led to the growth in Iraq of substantial urban populations with disposable monetised income, an unusual phenomenon in the pre-modern era.Footnote ⁵² The first Muslim armies were mostly drawn from the ranks of Arab nomadic pastoralists who, having abandoned their traditional means of subsistence, were paid pensions (Ar. ʿaṭāʾ), and eventually regular salaries, in money by the state, as well as allowances for their families. In the Abbasid period, the social constituency of the military gradually shifted towards Turks and other Central Asians. Throughout the period under study, these expenses mobilised the bulk of the Umayyad and Abbasid state budget—over 80 percent, according to two quantified examples in textual sources.Footnote ⁵³ New Islamic cities such as Basra, Baghdad and Samarra were initially sustained by this continuous influx of money, which prompted the establishment of merchant communities to supply them. It is probably not a coincidence that direct sea trade with China witnessed its first great surge just as the first of these cities were founded in Iraq, in the seventh century.

The imperial courts at Baghdad and Samarra may have indirectly contributed to the inception of these trends. The Persian historian Bayhaqī (995–1077), in an account based on unnamed written sources, mentions 20 pieces of imperial China-ware (chīnī faghfūrī) and 2,000 other Chinese ceramic pieces offered to the caliph Hārūn al-Rashīd (r. 786–809) by the governor of Khurasan, ʿAlī ibn ʿIsā ibn Māhān. These are said to have included platters, cups, beakers, wine jars, “and various other kinds”.Footnote ⁵⁴ The anonymous author of the Book of Gifts and Rarities (Kitāb al-hadāya wa’l-tuḥaf, eleventh century) also asserts that coloured Chinese pottery was presented by a Christian secretary to al-Mutawakkil (r. 847–861), and 20 Chinese saddles were offered to the same caliph by a concubine.Footnote ⁵⁵ These anecdotes pose the problem, common in sources of the period, of the time lag between the reported facts and their record in writing; the Kitāb al-hadāya, in addition, has legendary overtones in many of its other accounts.Footnote ⁵⁶ But the pattern of gift-giving that they suggest is plausible and finds a confirmation in the archaeological finds of Chinese ceramics at Samarra. Tang court annals also mention, between the mid-seventh and late ninth centuries, numerous dashi embassies stemming from caliphal circles that left Chang’an loaded with luxury presents.Footnote ⁵⁷ The presence of these objects at the court would, in turn, have triggered an emulative demand amongst élite circles, and thenceforth into broader segments of society.Footnote ⁵⁸

The unification under Islamic rule of the former Sasanian empire with a large part of the Roman empire also created an enormous and relatively secure market for the onward dispatch of goods, from the southwest Mediterranean to Sind and western Central Asia, with Baghdad and Samarra as major sources of artistic tastes and trends. Likewise in China, the Tang empire was expanding economically and demographically from its traditional centres in the North towards the Yangzi basin and frontier zones of the southern coast, including Guangdong.

The makeup of trading networks from the Arab-Persian Gulf

The main actors of this trade were merchants from the Arab-Persian Gulf who started settling on the south China coast early on in the Tang period. These merchants tend to be called “Persian” or simply “Muslim” in modern scholarship, yet they must have been of diverse origins. Iranian Zoroastrians, Christians and Muslims would have naturally been among their ranks, since early Islamic trade with China was built upon Sasanian foundations, though Sasanian sea trade may not have reached beyond the western half of the Indian Ocean, at least on a regular basis.Footnote ⁵⁹ The South and East of the Arabian Peninsula had also been engaged in commerce with India and Sri Lanka since Antiquity; both diasporas, Arab and Iranian, were established in Indian ports before Islam.Footnote ⁶⁰ They may have increasingly intermingled in the sea trade from the third century onwards, as a result of Sasanian dominion over East Arabia.

After the rise of Islam, the main western termini of the China trade were, first, Islamic foundations, notably Basra and Baghdad; second, cities owing their growth to this trade, like Siraf; and third, ports on the east Arabian coast, such as Suhar. Thus they would all have shared a predominantly Muslim population, with a major Arab Muslim component, at a time when the rest of West Asia remained overwhelmingly Christian, and to a lesser extent Zoroastrian. The main socio-linguistic groups in the Gulf were Arabic-speaking Muslims, some of them with an Aramaic or Persian linguistic background; Arabic-speaking Christians (especially in southern Iraq and East Arabia); Persian-speaking Zoroastrians, Muslims and Christians; and Aramaic-speaking Christians and Jews (the latter mostly in Iraq).

The subtleties of West Asian identity were of little interest to Chinese writers, who referred to the merchants in question as dashi and bosi, if not simply hu; and little more can be derived from Arabic sources. Ibn Khurdādhbih and Ibn al-Faqīh (wr. ca. 903) do explicitly mention the sea trade of Jewish merchants between Baghdad, Ubulla, India and China; their name, the Radhanites (al-rādhānīyya), probably reflects an origin in the Sawād of Iraq.Footnote ⁶¹ The Church of the East is also known, from inscribed copper plates issued in 849 by the governor of Kullam, in modern Kerala, under the rule of the Chola king Sthāṇu Ravi (r. ca. 844–885), to have been granted some rights of jurisdiction over its community, tax exemptions and permission to levy custom duties from merchant ships.Footnote ⁶² This commercial involvement in India could conceivably have extended, within different parameters, to China, although evidence of this is lacking so far.

In practice, members of these communities would often have been multilingual and have spoken different combinations of Arabic, Persian, Aramaic and Hebrew. The Sthāṇu Ravi plates, reflecting the activities of overlapping merchant diasporas, attest to such diversity: written in Tamil, they bear signatures of witnesses in Arabic, Pahlavi (Middle Persian) and Hebrew script.Footnote ⁶³ Most Muslims, regardless of their linguistic background, tended to identify with the shared culture of Arab Muslim civilisation. Thus the informants of the anonymous Akhbār al-ṣīn wa’l-hind (ca. 851) repeatedly refer to themselves as “Arab” and “Muslim”, even though Persian loan words appear in their vocabulary.Footnote ⁶⁴ These communities were, in sum, probably as variegated in China as in their regions of origin. For practical purposes, unless specific information is provided, they will be referred to as West Asian in what follows.

West Asian merchants on the South China coast in the Umayyad period

Guangzhou was the first Chinese terminal reached by West Asian ships, and the most important Chinese hub of this trade, “the meeting point of merchants,” as the Akhbār al-ṣīn wa’l-hind call it.Footnote ⁶⁵ As early as 671, at the beginning of the Umayyad period, the Chinese Buddhist pilgrim Yijing (635–713) had embarked on a bosi ship in this port to head south, probably to Sumatra.Footnote ⁶⁶ In 714, a Christian priest known in Chinese as Jilie came to the court of Chang’an with Guangzhou's superintendent of merchant shipping, which suggests that he reached China through that port. The literatus Liu Zongyuan (773–819) wrote that the holder of this office in Guangzhou, “every year, conducted the sending of tributes”, and for this purpose he may sometimes have been accompanied to the capital by recently-arrived foreigners.Footnote ⁶⁷

In the same years that brought Jilie to Chang’an, the Umayyad court poet al-Farazdaq, who lived in Basra, wrote in a panegyric to the caliph Sulaymān (r. 715–717):

The “disease” in question was al-Ḥajjāj ibn Yūsuf al-Thaqafī, the Umayyad governor of Iraq and the eastern provinces, who died in 714. The poet implies that distant exile was no longer necessary to escape his fearsome grip: in so doing, he alludes to Muslims reaching India and China, presumably in this case from Iraq. The Korean pilgrim Huichao (Kor. Hyecho) incidentally confirms this, writing from hearsay about the land of bosi, around 727:

The description could imply that certain bosi ships only went as far as Sri Lanka and the Malay peninsula, or that they stopped there as part of a longer journey; in either case, it explicitly states that some of them did sail all the way to Guangzhou. In 750, a disciple of the Buddhist master Jianzhen (Ganjin) also related about this city:

These sources converge to show that West Asian sea traders were active at Guangzhou throughout the Umayyad period (661–750), and that at least some of them sailed directly from the Arab-Persian Gulf. The arrival of Jilie through Guangzhou in 714 could also suggest that the sea trade had already gained practical advantages over land routes, since the Church of the East had an extensive presence in Central Asia.Footnote ⁷¹

The growth of West Asian communities at Guangzhou and Yangzhou

This initial momentum was amplified after the Abbasid revolution (750 ce), which marked a shift in the centre of gravity of the Islamic empire from Syria to Iraq. In 758, according to the Jiu Tangshu (“Old Book of Tang”, 945 ce) and Xin Tangshu (“New Book of Tang”, compiled 1044–1060 ce): “dashi and bosi together ravaged Guangzhou, attacking warehouses and burning storehouses, and fled away to sea. The governor Wei Lijian escaped”.Footnote ⁷² The explicit mention of both dashi and bosi points to a range of West Asians, possibly Arab and Iranian. The text implies that they had grown numerous enough, by that time, to openly defy Chinese authorities.

The attack may have been an act of piracy, since pirates based in Hainan regularly attacked foreign ships: upon visiting the island in 749, Jianzhen had found it filled with bosi slaves captured by Feng Ruofang, a local chieftain. In a conspicuous show of wealth, the latter would provide lighting to guests by burning frankincense, a luxury commodity imported from Arabia and East Africa.Footnote ⁷³ Yet the context of Guangzhou and the mention of foreign assailants alone make another motive possible: reprisal against official corruption, a problem endemic to the city. Already in 684, its governor Lu Yuanrui had been assassinated by a kunlun (black, i.e. probably Malay) merchant after abusing his position in office; the perpetrators had then fled to sea. The events of 758 might represent the repetition of a similar pattern on a larger scale. At any rate, the sea routes appear to have remained open, since Du You states in his Tongdian (“Comprehensive Institutions”, 801 ce) about his relative Du Huan in 762: “he returned [from Iraq] via Guangzhou by means of a merchant boat”.Footnote ⁷⁴

Merchants suffering tax predation, in a city like Guangzhou, were a potential threat to public order, but the Chinese treasury was also set to gain from their activities, if suitably protected. The wealth brought to the southern region by this activity and by a growing agriculture became proverbial. As the Jiu Tangshu put it in relation to the eighth century, “its produce, together with the wealth brought by the sea route, was mountainous”.Footnote ⁷⁵ The efforts of some governors to re-establish equitable trade conditions are recorded in Chinese sources.Footnote ⁷⁶ An imperial edict of 829 (or possibly 834) also stated:

The edict identifies different types of tax imposed on merchants: anchorage duties and tribute to local governors, alongside official purchases. The Akhbār al-ṣīn wa’l-hind mirror this statement when they record about Guangzhou:

The text gives the example of camphor (Ar. al-kāfūr, a resin harvested in Southeast Asia but traded by West Asians in China),Footnote ⁷⁹ which was purchased by the government at twice the market price. Thirty per cent of the incoming merchandise was thus set aside for the state; the Arabic formulation leaves some ambiguity as to whether this was levied as tax, with subsequent purchases being made at a fair price, or bought upfront by the authorities. Common sense might suggest the former reading,Footnote ⁸⁰ yet the latter finds support in the edict of 829, which mentions court purchases but no import duties.Footnote ⁸¹ If this interpretation is correct, the only direct “taxes” officially imposed on merchants would have been tribute and anchorage fees, thereby providing a strong incentive for this trade. In either case, officials could extort further income through forced gift-giving.Footnote ⁸² The regulated release of foreign goods into the market at a given point of the year may have been meant to maximize competition and lower prices.

The edict of 829 again confirms the presence of foreign merchants in Lingnan, the area “south of the mountains” which comprised Guangzhou and coastal regions extending into modern Vietnam, and adds the important mention of Fujian and Yangzhou. The region name “Fujian”, in this context, probably refers to Fuzhou or Quanzhou, trading ports about which relatively little is known for this period.Footnote ⁸³ Further to the north, Yangzhou, by virtue of its location near the mouth of the Yangzi river and Grand Canal, was the main access point for international sea trade to Chang’an.Footnote ⁸⁴ Other evidence does confirm the presence of West Asian merchants there. The biography of the rebel commander Tian Shengong (d. 776) in the Jiu Tangshu and Xin Tangshu mentions that “several thousands of bosi and dashi merchants were killed” in 760 at Yangzhou, which suggests the existence of a sizeable community.Footnote ⁸⁵ Excavations at Yangzhou have revealed building remains with ceramic sherds and glass fragments that probably reflect the activities of their shops.Footnote ⁸⁶ Ibn Khurdādhbih also reports:

Figure 4. The Huaisheng mosque and minaret, Guangzhou (Public domain image, author unknown).

Figure 5. Islamic glass dish from a crypt of Famen temple sealed in 874 ce. Famensi Museum. Diameter 20.2 cm, height 2.2–2.6 cm. Source: Rastelli (ed.), China at the Court of the Emperors, Cat. 69.

“Qānṣū” could be Hangzhou or Yangzhou: the text is too vague to allow the distinction.Footnote ⁸⁸ It suggests, in either case, an awareness of the region surrounding the mouth of the Yangzi and of Korea, which was united until 892 under the kingdom of Silla. The latter was indeed famed as a source of raw gold and finished gold objects.Footnote ⁸⁹ Ibn Khurdādhbih suggests that Muslims had settled there, a plausible but otherwise unconfirmed assertion. The Akhbār al-ṣīn wa’l-hind do mention that “there lie beyond the sea the islands of al-Sīlā”, but add that “none of our companions (aṣḥāb) have reached there to tell us about them”; while al-Nadīm (wr. 987) simply lists al-Sīlā as a region of China, “one of the most pleasant and greatest lands, and one of the richest in gold”.Footnote ⁹⁰

West Asian merchants, whilst most significantly established at Guangzhou, were thus also present at Yangzhou, and probably in small numbers in some other ports of the South China Sea. To the first major factor contributing to the growth of this trade—the proximity of the Gulf to the major markets of Basra and Baghdad—a second may now be added: the official organisation of Chinese ports, which must have created a sufficient level of certainty in order for West Asians merchants to engage in the venture. Despite natural hazards and the risks of piracy, the sea routes appear to have been safer and faster than overland routes, which tended to remain more segmented.Footnote ⁹¹ Between the 760s and 840s, Tibetan expansion into northwest China would have further complicated and lengthened land travel from Chang’an.Footnote ⁹² In a sign of the times, several Arabic and Persian sources mention Sogdian merchants who sailed to China on ships from the Gulf in the eighth to ninth centuries. This account by al-Masʿūdī, for example, is about the period before 878:

This seems to reflect a world turned upside down, in which the patrons of age-old overland routes board ships thousands of kilometres south of their homeland, the traditional centre of this trade, to reach China. The importance of Iraq as a market, again underlined in the text, must also hold part of the explanation for this long détour.

The reconfiguration of sea trade in the late ninth century

The favourable conditions that prevailed in China for most of the eighth and ninth centuries were dealt a severe blow by the rebellion of Huang Chao (Ar. Bānshuwā), which massacred foreign merchants after besieging Guangzhou in 878. Abū Zayd al-Sīrāfī, who wrote a Supplement to the Akhbār al-ṣīn wa’l-hind in the early tenth century, noted:

The date and events mentioned here are accurate, but the figure of 120,000 Muslims, Jews, Christians and Zoroastrians should be treated with caution, despite the assertion that it is based on Chinese official records. Abū Zayd is a second-hand informer who, unlike the compiler of the Akhbār, relays a varied mix of reliable and less reliable information on China. Since some members of the above four communities presumably survived, the account would imply that the West Asian population of Guangzhou was well over half that of the entire Basra. This passage is best taken, then, as conveying large-scale implantation.

Regardless of exact numbers, the events reported do seem to have inaugurated a phase of decline at Guangzhou, which was superseded by Quanzhou (Ar. Zaytūn) as an international trading port from the tenth century onwards.Footnote ⁹⁶ In West Asia, Basra was likewise weakened by major sacks during the revolt of the Zanj, the African slaves, in the late ninth century, and again at the hands of the Qarmatian religious movement in the tenth century. The fortunes of Siraf also began to shift in the same period, in part due to less stable political conditions.Footnote ⁹⁷ Middlemen in India and Sri Lanka concurrently appear to have regained some of their former importance. Prevailing patterns of trade thus transformed at the close of the Tang era, as they had done at its beginning, although it remains to be seen whether direct sea trade between China and the central Islamic lands ceased altogether; but this later phase of evolution falls beyond the scope of the present study.

The Akhbār al-ṣīn wa’l-hind and Muslim perceptions of China

Among extant textual sources, the Akhbār al-ṣīn wa’l-hind (“Accounts of China and India”) stand out as a unique repository of information about Arab Muslim perceptions of China in this period. They were composed by an anonymous author around 851, which makes them the earliest extant Arabic work devoted to China, or indeed to any foreign land.Footnote ⁹⁸ A focus on matters of trade, business and administration shows that the contents were harvested directly from several merchants, whose vocabulary suggests an origin in Iraq or the Arab-Persian Gulf; at one point, the text explicitly calls them “Iraqi merchants”.Footnote ⁹⁹ These informants unequivocally regarded themselves as “Arabs” from the “Arab lands” who spoke “the language of the Arabs”; their remarks on ritual cleanliness, contract and Chinese religion make it clear that they were Muslim.Footnote ¹⁰⁰

Most other Arabic writings of the period that deal with the subject belong either to the genres of geography and topography, as with Ibn Khurdādhbih, or to literature and storytelling.Footnote ¹⁰¹ The relatively unpolished style of the Akhbār, which ignores established historiographical rules of transmission (isnād), would have been looked down upon by more established writers, yet it is a rare testimony stemming directly from the primary actors of this trade. As has already become apparent, a perception of fairness towards Chinese administration dominates among the informants of the Akhbār, who notably assert:

A certain respect and esteem for the Chinese social system transpire, as when the informants report that “poor and rich, large and small, all Chinese people learn calligraphy and writing (al-khaṭṭ wa’l-kitāba)”, and that public schooling is funded by the treasury. These remarks reflect the educational system instituted by the Tang in the seventh century and expanded in 738 by an edict of Xuanzong.Footnote ¹⁰³ Both the Akhbār and al-Sīrāfī's Supplement assert, after duly acknowledging that the Muslim ruler is foremost among the rulers of the earth, that the Chinese emperor is only second to him: as al-Sīrāfī puts it, “no king has more firmly established the foundations of peace and better maintained order” than him.Footnote ¹⁰⁴ Reading between the lines, this conveys the notion that China is better governed than the Islamic world.

The informants are familiar with the system of government, justice, taxation and commerce in China. They report about manners of being granted audiences at the court of Tang governors and about the protocol observed there. They show a close acquaintance with Chinese social life, as when they discuss funerary and wedding customs, as well as the culinary habits of the Chinese and their predilection for rice, out of which they make wine, vinegar, sweets and “other such things”.Footnote ¹⁰⁵ The Chinese, they add, “are not clean: they do not wash themselves with water when they have been, but wipe it with Chinese paper (al-qarāṭīs al-ṣīnīyya)”.Footnote ¹⁰⁶ Toilet paper had indeed appeared in China at least three centuries earlier, when the scholar Yan Zhitui (531–591) remarked: “Paper on which there are quotations or commentaries from the Five Classics or the names of sages, I dare not use for toilet purposes”.Footnote ¹⁰⁷ Muslim ritual purity, by contrast, required the use of water. We also learn in the Akhbār:

This is probably the first mention of tea recorded this far west—and what is more, by people who had tasted it.Footnote ¹⁰⁹ The observation must date to the 830s, when the state briefly attempted to impose a monopoly on tea production.Footnote ¹¹⁰

The tone of the Akhbār is, on the whole, overwhelmingly appreciative of China. The country, our Arabic informants assert, is more pleasant and beautiful than India: it is healthier, less plagued with disease, and has better air; it has fewer blind and handicapped people, the likes of whom are very numerous in India; there are impressive rivers, and the land is fertile. The Chinese are more beautiful than the Indians, and more like the Arabs from the perspective of dress, mount and beasts of burden.Footnote ¹¹¹ Their women do not cover their heads, but their men do so with hats—which must have seemed incongruous from a Muslim perspective. Chinese women put combs in their hair, or let it float.Footnote ¹¹² Here we see the premises of a fascination with things Chinese that would eventually grow into the stuff of legend in Arabic literature, as epitomised by the Chinese cities, kings and princesses of the Thousand and One Nights.

The “quarter of tributaries” (fanfang) and the outside world

The Akhbār al-ṣīn wa’l-hind also betray limitations of this acquaintance with China. By contrast with more down-to-earth matters, the understanding of Chinese religious beliefs shown by these merchants remains somewhat superficial, and again based on the normative framework of Islam: “The Chinese worship statues to which they pray and which they implore; they have religious books”.Footnote ¹¹³ They do show a distant awareness of Buddhism and its links with India:

These contrasting levels of knowledge may be related to the natural interests of merchant communities, but also to their social organisation. As the only named informant of the Akhbār, the “merchant Sulaymān” (sulaymān al-tājir), reports, Chinese authorities had granted the Muslim community at Guangzhou permission to arbitrate legal disputes under the authority of a Muslim judge. This amounted to a measure of extra-territoriality, perhaps for purposes of expediency, which echoes the situation of the Church of the East at Kullam.Footnote ¹¹⁵ Chinese sources indicate that foreign merchants, in Guangzhou and other cities, lived in a special “quarter of tributaries” (fanfang).Footnote ¹¹⁶ They might have been officially confined to it, since we read in the Akhbār:

This passport system for the internal travel of foreigners is known from the Tang Code, completed in 653 ce; the description given here also reflects an emphasis on genealogy and ethnic origin, the building blocks of the Tang social order.Footnote ¹¹⁹ Such rules restricted the movement and activities of merchants, but along with port regulations, they contributed to establish a relatively predictable framework for their trade within China, a benefit emphasised in the Akhbār.

Some West Asians who resided in China, whether for a period of months or years, must have learned at least the rudiments of the spoken language. This is the limited level of linguistic competency suggested by the Akhbār: their reports presuppose some social communication with Chinese people, and they contain the transliteration of half a dozen Chinese words, notably in relation to government hierarchies.Footnote ¹²⁰ Writing in the late eighth to early ninth century, Liu Zongyuan also noted that the superintendent of merchant shipping in Lingnan “dealt with the translation of languages”,Footnote ¹²¹ which hints at the existence of interpreters for official dealings, presumably with further linguistic proficiency.

The institution of the fanfang formalised the establishment of self-contained foreign communities living within a designated urban space. This, together with the restrictions on their inland travel, suggests official attempts to circumscribe the activities of these merchants. But these efforts were not necessarily successful. In the Jiu Tangshu, one thus reads about complaints from Guangzhou's superintendent of merchant shipping about their increasing intermarriage with local women, which the authorities were trying to curb.Footnote ¹²² This social reality is again indirectly reflected in the Tang Code, which states that marriage with foreigners is punishable by life in exile, but also adds:

As the years went by, such unions, whether officially sanctioned or not, would have produced children of mixed race, naturally conversant in Chinese and one or more West Asian languages.

According to the above-cited account of the merchant Sulaymān, the Muslim judge of Khānfū used to lead prayer during Muslim festivals and to pronounce the khuṭba (Friday sermon) in the name of Muslim authorities (sulṭān al-muslimīn).Footnote ¹²⁴ The khuṭba would have been delivered at congregational prayer, whether in a rudimentary public space or a permanent mosque building. One architectural monument, the Huaisheng mosque in Guangzhou, resonates with this context (Figure 4). Its prayer hall has been repeatedly rebuilt over the centuries, most recently in 1935. The minaret, articulated on two levels, is the oldest extant part of the building. Its date has not been precisely established, with estimates ranging from the tenth to the thirteenth century.Footnote ¹²⁵ It has the particularity of being built of plastered brick, and in a tapering cylindrical form that is unusual for China, but naturally brings to mind models in the Islamic world.Footnote ¹²⁶ A more precise dating will have to await a full documentation of its decoration, type of brickwork, and phases of construction. Local tradition ascribes the foundation of this mosque to Saʿd ibn Abī Waqqās, the companion of the Prophet said in later Chinese Muslim lore to have brought Islam to China.Footnote ¹²⁷ Without going this far back, a mosque could conceivably have stood on this site in the Tang era, as asserted by Fang Xinru in his description of the building (ca. 1206).Footnote ¹²⁸

Likewise the Shengyou mosque in Quanzhou was built in 1009/10 and restored in 1310 by a patron probably from Shiraz, in southern Iran.Footnote ¹²⁹ Other mosques in Hangzhou and Chang’an are said in local tradition to have originally been built in the Tang period, but the extant structures are later than the thirteenth century. This group of monuments remains to be fully investigated, but its very existence points to the continuous presence of Muslim communities in these port cities since at least the Song era, and probably under the Tang, when the socio-economic conditions for their existence were already in place.

Chinese ceramics and Iraqi potters

By creating regular direct links with Iraq and establishing a small-scale social infrastructure of their own in China, West Asian merchants opened the way for intensified cultural exchange, from the imitation of goods spawned by local demand, through transfers of technologies largely based on oral communication, to translations and the written record of ideas. These will now be studied in a sequence that broadly reflects this gradation of knowledge, starting with the most extensively documented items of this trade: ceramics.

As they reached the markets of Iraq, Chinese ceramics provided a crucial impetus for the growth of a distinctive Islamic pottery industry. As al-Jāḥiẓ, the famous littérateur (ca. 776–869) from Basra, noted with a characteristic hint of irony:

Up until the Abbasid era, the ceramics produced in Syria, Iraq and Iran were broadly akin to those of pre-Islamic times: they mainly consisted of unglazed vessels, some with appliqué decoration inspired by metalwork. The growing appeal of Chinese monochrome ceramics triggered creative responses from local potters, notably in Basra and Siraf.Footnote ¹³¹ The imitation of Tang white wares, and particularly of the widely imported Gongxian types, presented obstacles: the clays available in Iraq were different to those of China, and tended to have a buff yellow colour. The firing temperatures that could be achieved at local kilns were also lower than in China, particularly for white stonewares.Footnote ¹³² This challenge was addressed through a technical innovation: the addition of an opaque tin oxide glaze diffused light on the surface of the object, which produced an impression of whiteness.Footnote ¹³³ Vessel shapes were also imitated. The occurrence of Arabic signatures implies that the finished products were not necessarily sold as forgeries.

Tin oxide was not mined in West Asia, and had to be imported from Burma or the Malay Peninsula. The standard manner of creating bowls in China involves throwing the clay on the wheel, then moulding the interior and trimming the exterior to the desired thickness. The same method has been identified by scientific analysis of Basran production of this period. The agency of merchants might be invoked in both cases: their role extended beyond the transport of goods to the supply of raw materials and probably to the transmission of technologies, whether by bringing Chinese potters to Iraq or observing them at work in China. Merchants thus stand out not only as the backbone of the physical links between China and the Islamic world, but also as agents facilitating innovations from which they were ideally placed to benefit, and for which they might also have provided capital.Footnote ¹³⁴

Muslim potters, having created these new white wares, treated their surface as a blank canvas onto which they started adding painted motifs that responded to local taste for decoration. Islamic ceramic production had been thrown on a course of evolution that would be pursued into the modern era. The new types included semi-abstract foliate designs arranged symmetrically, in the typical manner of Islamic surface ornament of the period, and executed in cobalt blue. They elicited further responses in Chinese export wares. For example, three Chinese blue-and-white bowls discovered on the Belitung wreck appear to imitate corresponding West Asian wares. Their small number could suggest a test sample being dispatched to patrons in the Arab-Persian Gulf before starting production in China.Footnote ¹³⁵ Furthermore, the cobalt blue used in this category of Chinese wares may have been imported from West Asia, although this issue remains debated.Footnote ¹³⁶ In either case, the colour seems to have been applied to the wet glaze using the same method as in Abbasid Iraq, thereby pointing to a reverse transfer of techniques.Footnote ¹³⁷ Even in this most admired of Chinese crafts, the pattern of exchange was not entirely one-directional.

Glass, Textiles and the Diffusion of Styles

A range of other goods were also carried in this trade, a fact reflected by the metalwork, mirrors and aromatics found in the Belitung shipwreck and, for the route into China, by official dispositions for their storage and release at Guangzhou. Glass, being relatively heavy and fragile, was inherently suited to transport by sea rather than by land, like ceramics. The technology of glassmaking was more fully mastered in West Asia than in China, and this led to an eastwards movement of objects evidenced by the fragments found at Yangzhou, along with ceramic sherds, in what were probably West Asian shops.Footnote ¹³⁸ A few complete objects have also been preserved in stupas and burial sites from Tang and Liao China, as well as in Japan, so that today, these two countries hold some of the best-preserved specimens of early Islamic glass in existence.

The Shosoin repository in Nara (Japan), a building initially attached to the Tōdai-ji Buddhist temple complex, was endowed with the treasury of emperor Shōmu (701–756) by his widow Kōmyo (701–760) after his death. Amongst its contents are several examples of glasswork, from goblets through beakers and ewers to dishes. Their attributions are disputed, and sometimes range from West through Central to East Asia for the same objects.Footnote ¹³⁹ A crypt built to house a Buddhist relic at Famen temple, 120 km north of Chang’an, and sealed in 874 has also yielded, among other precious objects donated by emperor Xizong, an Islamic glass vase and drinking cup, as well as several glass plates with incised geometric and semi-abstract vegetal ornament typical of the Abbasid era (Figure 5).Footnote ¹⁴⁰ The imperial level of patronage, in both repositories, suggests that these were prized luxury commodities in China and Japan. Glass objects and fragments resembling Sasanian or early Islamic production have also been unearthed at Silla burial sites and temples in Korea, though again some of them may have been local products based on West Asian prototypes.Footnote ¹⁴¹

Textiles were another important commodity primarily exported from China, and this long before the rise of Islam. Starting in the fifth and sixth centuries, decorative motifs of West and Central Asian origin were making their way onto Chinese silks, presumably through land routes (likewise, Sasanian and Central Asian metalwork were sources of inspiration for Chinese silverwares and moulded-relief ceramics).Footnote ¹⁴² Silk had also started being produced in West Asia in the sixth century, and isolated specimens of this production appear to have reached China. Thus in 716, an embassy of the Umayyad caliph Sulaymān brought to the Tang court “a robe of golden thread”: this must have been an imperial embroidered fabric (Ar. ṭirāz), typically made of silk with Arabic inscriptions woven in gold thread.Footnote ¹⁴³ In broader terms, as noted earlier in this study, diplomatic activity was consistently accompanied by the circulation on a small scale, through gifting, of luxury objects.

The main movement of textile products, however, was from China towards West Asian markets. Their lightness and suppleness made them prone to transport by land, yet it is clear that by the 720s, they were also being carried on merchant ships, as attested by Huichao's mention of bosi sailing directly to Guangzhou and acquiring silk there.Footnote ¹⁴⁴ Silk is also the first in a list of Chinese products given by Ibn Khurdādhbih, whose focus is on the sea trade. Likewise in the Akhbār al-ṣīn wa’l-hind, one informant notes that the Chinese have gold, silver, pearls, silk brocade (dībāj) and silk (ḥarīr) in abundance, and use them as commodities (rather than currency).Footnote ¹⁴⁵ The Tang Code includes a statute forbidding the export of silk products and precious metals through the western and northern frontiers, which suggests an effort to channel their export to the sea routes.Footnote ¹⁴⁶

Survivals of textiles over such long periods of time are rare. Some silks from the Shosoin display motifs and compositions with strong Iranian or West Asian resonances, but as with glass, it is often unclear whether they were made in West, Central or East Asia, let alone how they were transported. In the example shown here (Figure 6), the symmetrical composition of the date tree with its flanking lions and attendants is typical of Sasanian and early Islamic designs; yet the treatment of the tree base, claws, mane and cloud scrolls seem to betray East Asian craftsmanship. To complicate matters further, the weaving technique is unconventional, and thus not easily localised.Footnote ¹⁴⁷ In this realm where goods, techniques and motifs had begun to be exchanged before Islam, it is particularly difficult to assert whether any specific objects, not to mention broader evolutions, may be tied to sea trade between Tang China and the early Islamic world.

Figure 6. Detail of textile stored at the Shosoin imperial repository in the mid-eighth century. Nara, Shosoin, South Section. Total dimensions 99 × 52.2 cm.

By their very uncertainty, the debates about the origins of glasswork and textiles from this period, together with the better-understood dynamics at play in ceramics, also underline the spread of shared artistic vocabularies across West, Central and East Asia – what one might call international Asian styles – for certain media and designs, partly diffused through sea trade, but also through other channels.

Plants, materia medica and alchemy

Foodstuffs and medicinal substances, which require either cultivation or proper administration as medicines, bring us one step closer to the transmission of ideas. Their material remains are unsurprisingly rare: the Belitung shipwreck is a remarkable exception, having yielded small quantities of spice, amber, nuts and aromatic resin as well as whole jars of star anise from Southeast Asia or southern China.Footnote ¹⁴⁸ Numerous fragments of Chinese utilitarian jars have also been excavated at Siraf and in the Gulf; these were probably used to carry consumable goods, like the turquoise jars from the Islamic world discovered at several sites along the South China coast, including hundreds of sherds and one complete vessel at Yangzhou.Footnote ¹⁴⁹

Textual sources also suggest that a wide range of plant products travelled between East and West Asia, although their interpretation requires caution. Most Arabic treatises on materia medica from the period, such as those by al-Kindī (ninth century) and Ibn Waḥshīyya (wr. 904), commonly list “Chinese” ingredients in their recipes: in the latter two works, for example, Chinese ginger (zanjabīl ṣīnī), as distinct from ginger (zanjabīl) and wild ginger (āsārūn), as well as Chinese rhubarb (rāwand ṣīnī).Footnote ¹⁵⁰ It is difficult to assert, on the basis of these texts, whether the plants were indeed Chinese: they may have been brought from China, from another part of the Indian Ocean, or have been acclimated in the Islamic world. Chinese cinnamon (dār ṣīnī) stands as a relative exception, since it is explicitly distinguished from common cinnamon (qirfa) by Ibn Waḥshīyya, is listed as an import from China by Ibn Khurdādhbih, and may be identified as cassia (Lat. cinnamomun aromaticum), the bark of a plant native to southern China.Footnote ¹⁵¹

Conversely, Chinese sources such as the Youyang zazu (“Miscellaneous Offerings of Youyang”) of Duan Chengshi (ca. 800–863) list numerous herbs, spices and aromatics from West Asia; these include dates, frankincense, oak galls, ambergris, and myrrh, to name but a few.Footnote ¹⁵² The date tree was the archetypal plant associated with Arabia and West Asia in China. It thus occurs on textile designs from the Shosoin illustrated above (Figure 6), in a representational mode of Sasanian or early Islamic origin. The informants of the Akhbār al-ṣīn wa’l-hind and its Supplement assert that the date tree did not exist in China, but Liu Xun (ninth century) provides an accurate description of specimens he had seen in the suburbs of Guangzhou, where they may have been acclimated of recent date. The date fruit must, in any case, have been imported from West Asia, since Liu asserts: “the foreign as well as the domestic kind is consumed in our country”.Footnote ¹⁵³

Alongside the plants themselves, knowledge about their medicinal uses made its way into Chinese pharmacological treatises of the early Tang period onwards; a few books were even devoted entirely to the subject.Footnote ¹⁵⁴ One of these, the Haiyao bencao (“Overseas Materia Medica”, early tenth century) by Li Xun, has been largely preserved through citations in later texts. Its author was born into a bosi family in China, possibly at Guangzhou, before making a career in the official administration of Sichuan and becoming a Chinese poet of note. According to Huang Xiufu (fl. early eleventh century), Li Xun's younger brother Li Xuan “made a living by selling fragrant herbs”, which suggests direct links with merchant communities. The very title Haiyao bencao implies transport by sea, an idea confirmed by the mention in the text of five maritime products, among dozens of foreign herbs, trees, fruits, animals, insects and jades.Footnote ¹⁵⁵

Alchemy is another field in which hints at these patterns of exchange can be gleaned. The Arabic alchemical corpus attributed to Jābir ibn Ḥayyān represents the production of a school active between the eighth and tenth centuries, probably in Iraq. In the Book of Properties (Kitāb al-khawāṣṣ), chapters 28 to 31 list recipes attributed to China and India, which explain in detail the fabrication of various Chinese oils (duhn ṣīnī) for belts, clothes, metals, swords, glass and marble; black Chinese glue; Chinese saddles, inks, and a Chinese cream for polishing mirrors.Footnote ¹⁵⁶ As with medicinal products, these beg the question of how Chinese the techniques really were. In order to start assessing this question, let us briefly examine a recipe from an unpublished manuscript in the British Library (Ms. Or. 4041, ff. 56r-v). The soot of pine wood rich in resin, says the recipe, was consummated in a stove (kānūn) covered with an [inverted] recipient (inā). The soot was then mixed with fish oil in a vessel (hāwūn), beaten for an hour, flattened with water, and rubbed until it thickened, becoming soft as a cream (marham). This substance was made into loaves that were hung on a string and left to dry.

This procedure corresponds to Chinese ink-making techniques, as recorded in manuals of the Song period onwards. These were based on pinewood that had to be rich in resin (after the Tang period, oil soots were gradually introduced instead as pine forests were being depleted). The soot was collected on a hood covering the fire and mixed with glue that could be made, among other ingredients, from fish, then beaten and spread by hand, before drying; the latter phase could be prolonged for years in order to remove any moisture from the ink bricks, which were given different shapes and moulded decorations.Footnote ¹⁵⁷

This recipe attests to a technological transfer occurring outside the realm of ceramics. The “Chinese” oils and glues in the Jābirian corpus also bring to mind the maintenance of ships in the Indian Ocean trade; whilst the polishing cream resonates with the mirrors recovered on the Belitung wreck; and the saddles echo the gift to al-Mutawakkil mentioned in the Kitāb al-hadāya wa’l-tuḥaf.Footnote ¹⁵⁸ These possible connections remain to be investigated in light of this text, which would deserve to be fully edited.

At the other end of Asia, Li Xuan, the above-cited plant trader of Iranian origin, was also an alchemist, and his brother Li Xun repeatedly cites alchemical remedies with Daoist affinities in Haiyao bencao.Footnote ¹⁵⁹ The (distorted) image of such figures emerges amongst the hu, bosi and dashi merchants of Tang-era tales: these characters, usually depicted as grotesque, are associated with alchemical and pharmaceutical pursuits, and sometimes turned into magicians with wonderworking precious stones.Footnote ¹⁶⁰ Although further study is still required, it seems that the milieu and mindset of alchemy and materia medica—semi-popular rather than courtly or religious, lying at the crossroads of craft and science, and requiring the acquisition of substances—were ideally placed to channel an exchange of ideas. Once again, West Asians emerge as its main driving force, along with Chinese people of West Asian origin, who were innately conversant with both cultures.

Chinese craftsmen, books and objects in Abbasid Iraq

Scattered fragments of evidence do also suggest that Chinese people, some of them skilled at a craft or literate, resided in Iraq during those years. This is most clearly shown by the case of Du Huan, an officer probably taken prisoner in the aftermath of the Talas battle (751), which saw Muslim forces led by Ziyād ibn Ṣāliḥ al-Khuzāʿī inflict a resounding defeat on the Tang armies of general Gao Xianzhi and his Turkish allies, thereby establishing a firm Muslim presence at the heart of Central Asia.Footnote ¹⁶¹ Du Huan spent the next decade in Iraq before returning to China on a merchant ship around 762. His original text has not survived, but his relative Du You recorded some sections in Tongdian, notably:

This direct testimony attests to the presence, in the 750s in Iraq, of weavers, metalworkers and painters, some of them from Chang’an and its broad region. The passage begins by stating: “Dashi is also called Ajuluo; its Arab king is styled mu-men [Ar. amīr al-muʾminīn, the main Arabic title of the caliph]. This [Ajuluo] is the place of their capital”.Footnote ¹⁶³ As suggested by modern commentators, “Ajuluo” might be a phonetic rendering of ʿAqūla, the East Syriac name of Kufa (or the more cognate West Syriac pronunciation, ʿAqūlo).Footnote ¹⁶⁴ But the text continues with a generic description of Iraq, making it unclear whether the craftsmen were active in this city or in another part of the land: Baghdad during its construction, which began between 758 and 762, comes to mind as another possibility.Footnote ¹⁶⁵ Whether this episode was isolated or part of a broader circulation of craftsmen remains an open question, given the lack of evidence.

In the Fihrist, written in 987, al-Nadīm also mentions Chinese objects he had seen in Baghdad. He writes:

This allusion calls to mind the so-called “butterfly” codices consisting of single or double folios glued together at the fold and kept together between the two halves of a wooden stick: their shape would indeed have made them resemble fans of the kind used in the Islamic world.Footnote ¹⁶⁷ The production of these codices may have begun in China as early as the eighth century, but it peaked at the time of al-Nadīm, under the early Song (tenth century).

“The Chinese”, al-Nadīm notes, “write with hairs which they fit into the top of stems (ruʾūs al-anābīb), like painters (al-muṣawwirūn) do”, thereby drawing an implicit contrast between Chinese brushes and the reed pens of Arabic scribes.Footnote ¹⁶⁸ He also writes:

This description reflects the reality of Chinese inks, sold as solid loaves onto which designs were stamped;Footnote ¹⁷⁰ the figure construed by al-Nadīm as the emperor may simply have been a man in Chinese dress. His use of the word midād accurately refers to soot ink, as opposed to the gallic ink (ḥibr) employed in Qur’ans of the same period. Interestingly, these observations represent the kind of second-hand information that could be gleaned by an inhabitant of Baghdad with no particular connection to China (al-Nadīm was a bookseller), be it in houses, shops or markets.

The range of Chinese goods that al-Nadīm saw at Baghdad raises the possibility that literate Chinese people lived there, without providing solid evidence that they did. The presence of books suggests that there were people to read them, yet one cannot rule out that they were sold as curiosities; and so it is with Chinese inks and brushes, which could readily be used by local buyers (indeed, some of them may have been produced locally, as suggested by Jābir's ink recipe). Likewise the inkstone from the Belitung wreck (Figure 3) may have belonged to a literate Chinese person headed for West Asia, but the owner could equally have been travelling elsewhere in the Indian Ocean, or the object may have been intended as a gift or a rare commodity. The footprint of Chinese people in Iraq, direct or indirect, is ultimately faint: it evokes an anecdotal presence based on individual trajectories, rather than an established and socially structured settlement.

Al-Rāzī and the Chinese scholar

A story recorded by al-Nadīm about the philosopher, medical writer and alchemist al-Rāzī (known in Latin as Rhazes, 854–925 or 935) adds a final piece of evidence to this puzzle, and is thus worth quoting at length:

This anecdote about the feats of a Chinese scholar may, at first, be received with a hint of scepticism. It does contain a description of shorthand that echoes Chinese “grass script” (caoshu).Footnote ¹⁷² The Arabic name, majmūʿ (literally “assembled”), is meaningless in this context, and may arguably be a phonetic transposition of caoshu. At any rate, the Chinese technique, which involved abbreviated and connecting strokes, had no known equivalent in the Islamic world. The central element of the story is therefore authentic, showing at least an awareness of this aspect of Chinese writing in tenth-century Baghdad. It is possible that the broader narrative was woven around this factual core by an oral tradition, but closer scrutiny tends to lend it some credibility.

Al-Rāzī was active at Rayy and Baghdad, two cities located at the western end of both the sea and land routes to China, and which might thus have been reached by Chinese travellers (Rayy is explicitly cited by Ibn al-Faqīh as an extension of Radhanite sea trade to China).Footnote ¹⁷³ Al-Rāzī was among the foremost medical writers and practitioners of his time; like several of his predecessors, he included Chinese drugs in his materia medica, as preserved through al-Bīrūnī's citations.Footnote ¹⁷⁴ The Sixteen Books mentioned in the anecdote were an abridged selection of Galen's medical works compiled at Alexandria between the sixth and seventh centuries, and which came to form the basis of Islamic medical education. They are cited repeatedly in Rāzī's own medical summa, the Kitāb al-ḥāwī.Footnote ¹⁷⁵ The Sixteen Books would thus have been extensively taught to students in his circle, whom al-Rāzī advised to “collect books on medicine” and to “compile books for [themselves]” to assist their practice.Footnote ¹⁷⁶

The translation procedure described in the text differs from what is known about Arabic and Syriac techniques, which involved a single translator carrying out his work in writing, though collation was sometimes carried out orally with a second person.Footnote ¹⁷⁷ It does, on the other hand, resonate with Chinese techniques originally established for the transmission of Buddhist texts: a committee of scholars would assemble, with some having at least spoken knowledge of both the source and target language—like al-Rāzī's Chinese student. The text was read out in the source language (usually Sanskrit), translated orally and simultaneously written down in Chinese. This raw output was checked for accuracy by scholars familiar with the source language, then for linguistic soundness by Chinese literati. The technique is brought to life by the colophon dated 286 ce of a Chinese version of the Lotus Sutra:

Here again we find a gathering of scholars—in this case, Chinese, Indian, Tokharian and Yuezhi—carrying out the translation orally and noting it down, then collating it, a technique which had been expanded and refined by the Tang period. Thus a colophon of 703 records an instance in which the Buddhist Sutra of Golden Light was translated into Chinese by a committee of 16 Chinese and Indian scholars. The Indians read the text in Sanskrit to Yijing, the pilgrim already mentioned above, who translated it orally into Chinese. This version was again “received with the brush” by two scholars before other members of the committee controlled its contents.Footnote ¹⁷⁹

A Chinese scholar steeped in this translation culture might naturally have replicated it in a foreign context, which could represent the background to the al-Rāzī anecdote. In comparison with Islamic techniques, this collegial procedure relying on orality had the notable advantage of speed. Seen in this light, the month it took the Chinese scholar to copy the Galenic compendium is not unrealistic: the colophon of the Lotus Sutra translation of 286 ce. states that it was completed in three weeks.Footnote ¹⁸⁰ The Buddhist work is of the same order of length as the Sixteen Books, which are extant across several fragmentary Arabic manuscripts.Footnote ¹⁸¹

One can only speculate as to whether the end result was a finished translation or simply a personal record of the Sixteen Books: the important point is that the anecdote is likely to be authentic. The story of this lone scholar, like that of Du Huan, does also comfort the idea of Chinese travellers reaching the Islamic world as a thin trickle, through individual circumstances – but nevertheless adding significant points of contact to the overall pattern of exchange.

Character index

Ajuluo 亜倶羅

An Lushan 安禄山

bishou 筆受

Bosi 波斯

Caoshu 草書

Cefu yuangui 冊府元龜

Chajing 茶經

Dashi 大食

Duan Chengshi 段成式

Du Huan 杜環

Du You 杜佑

fanfang 蕃房

Fang Xinru 方信孺

Fan Shu 樊淑

Feng Ruofang 馮若芳

Fujian 福建

Gao Xianzhi 高仙芝

Gutang guo骨唐國

Haiyao bencao 海藥本草

Han 漢

Hu 胡

Huaisheng si 懷聖寺

Huang Chao 黄巢

Huichao 慧超 (also 惠超, Korean reading)

Jia Dan 賈耽

Jianzhen 鑒真

Jilie 及烈

Jiu Tangshu 舊唐書

Kunlun 崑崙

Lingnan 嶺南

Liu Ci 劉泚

Liu Xun 劉恂

Liu Zongyuan 柳宗元

Li Xun 李珣

Li Xuan 李玹

Lu Yu 陸羽

Lu Yuanrui 路元叡

Poluomen 婆羅門

Quan Tangwen 全唐文

shizi 獅子

Tang dazhaoling ii 唐大詔令集

Tian Shengong 田神功

Tongdian 通典

Wei Lijian 韋利見

Xin Tangshu 新唐書

Xizong 僖宗

Xuanzong 宣宗

Yan Zhitui 顏之推

Yijing 義浄

Youyang zazu 酉陽雜俎

Yue Ke 岳珂

Zheng Fahuajing 正法華經