Persians, Arabs and Muslims

China’s maritime contact with western Asia – the Western Regions (xiyu 西域), as they are often referred to in Chinese sources – long predated the coming of the Muslim merchants. Textual and archaeological evidence points to the existence of maritime trade ties connecting China with southeast Asia, southern India and Rome in the west as early as the first century C.E., a commerce in which Chinese silks, Roman glass, wine and specie, and pearls, ivory and peppers from various parts of maritime Asia were actively traded.Footnote ³ In the period following the early third-century fall of the Han empire, and especially during the fourth through sixth centuries when China was divided into northern and southern dynasties, Chinese ports hosted merchants from Kunlun 崑崙 (in Malaya) and southern India as well as Buddhist monks who had made their way from India.Footnote ⁴ Trade with the distant west, however, was the domain of Persian traders from the Sassanid empire, which ruled a vast swathe of western Asia from 224 to 651, and over the course of that period they extended their activities from the Indian Ocean east into China (Figure 1.1).

Figure 1.1 Arab merchants sailing to India

(from Hariri’s Maqamat. Bibliothèque Nationale, Paris Ms. Arabe 5847)

It is impossible to date the arrival of Persian merchants in China with any precision. The dynastic histories describe tribute envoys arriving from Persia in 455, 530, 533 and 535, but they almost certainly traveled by land.Footnote ⁵ Only in 671, half a century into the Tang, do we have definitive evidence of Persian seafarers in China, in the form of a travel account by the Chinese Buddhist pilgrim Yijing 義淨:

From a half-century later (717), we learn of an Indian Buddhist who sailed in a convoy of 35 Persian ships from Ceylon to Palembang (Srivijaya), eventually arriving in Guangzhou in 720, quite possibly being met my merchants such as those in Figure 1.2.Footnote ⁷ A Chinese account by a Chinese monk from 727 of Persian commercial activities asserts that they

The Chinese monk Ganjin (Jian Zhen 鋻真) describes in his travel diary being shipwrecked in southern Hainan in 748, where he encountered a local warlord who reportedly captured “two or three Persian ships” each year and enslaved their crews, a topic to which we will return.Footnote ⁹

Figure 1.2 Tang merchant figures

(Macao Museum)

These references to Persian maritime engagement in the China market are fully consonant with Edward Schafer’s portrait of Iranian merchants in Tang dynasty tales, in which he argues that the Persian merchant had become a common cultural figure in Tang China, stereotypically regarded as wealthy, generous and at times something of a magician, though Schafer also makes the point that many of these merchants are described as living in the north and presumably came to China by land.Footnote ¹⁰ Less clear, however, is how the Persian presence in Tang China relates to the Muslim Umayyad’s conquest of the Sassanid empire in 651 and the subsequent arrival in China of Arabs and Islam.

One cannot overstate the transformative impact that the founding of Islam by Muhammad (traditionally dated 622) and the subsequent rise of the Umayyad Caliphate (661–750) had upon western Asia, indeed upon the Eurasian and north African world. Their reverberations were certainly felt in China. In the 660s the Tang court entertained two embassies from Firuz (Pilusi 卑路斯), the son of the last Sasanian ruler Yezdegerd III, who had fled to Tokharestan (Tuhuoluo 土火羅) and was soliciting Chinese help in reviving the Sasanian cause. In the early 670s Firuz came to Chang’an himself and sought Chinese help in restoring his empire, and in response Emperor Gaozong sent him (or his son Narses; the sources differ on this) with a Chinese force that was already being sent to the west, but in fact the Chinese force never went beyond the Tarim Basin.Footnote ¹¹ More important than this interesting sideshow was the steady stream of Umayyad envoys who came to Chang’an, beginning in 651 and continuing to 750.Footnote ¹² Although many of these were traditional tribute missions, several in the early eighth century came demanding Tang submission, for the Muslim armies were engaged at this time in their dramatic expansion through central Asia.Footnote ¹³ Subsequent Umayyad weakness together with the expansive foreign policy of Emperor Xuanzong (r. 712–756) allowed for a return of Tang power in central Asia, but that ended abruptly in 751 when a Tang army under the Korean general Gao Xianzhi 高仙芝 was defeated at the Battle of Talas (near modern Tashkent) by an Arab army of the newly established Abbasid caliphate (750–1258).Footnote ¹⁴ The Abbasid caliphs sent no fewer than twenty embassies to the Tang between 751 and 798, while a recently discovered tomb stele indicates that the Tang sent at least one mission to the Abbasids.Footnote ¹⁵ We also have the revealing account by one Du Huan 杜環, a member of Gao Xianzhi’s army who was captured by the Arabs at Talas and returned to China on a merchant ship to Guangzhou in 761. In his account of Kufa (the initial capital of the Abbasids) and Abbasid society he describes Chinese painters, silk weavers and gold and silver craftsmen living and working there.Footnote ¹⁶

The presence of Arab merchants and Muslims in Tang China is more difficult to document, even though there is no doubt that both were there. If one excludes Arabs in non-Chinese armies and, of course, those associated with the Abbasids, Tang references to Arabs are scarce.Footnote ¹⁷ There are a couple of mentions of Arab merchants from Tang stories, such as one where a young man sells a fabulous pearl to an Arab in the Persian Bazaar of Guangzhou, and another in which “A party of noble Arabs purchase a supposedly valueless gem from a Chinese temple. Their king had offered an emirate to its finder, for it had formerly belonged to the Arabs, who used it to bring forth water in the desert.”Footnote ¹⁸ The Tang histories have only two sets of references (in each case, from both the New and Old Tang Histories), albeit very important ones, to Arabs in southeastern China. The first documents the 758 raid and brief seizure of the city of Guangzhou by Arabs and Persians that was described at the beginning of this chapter.Footnote ¹⁹ The second recounts the massacre of “several thousand” Arab and Persian merchants in Yangzhou in 760 by rampaging government troops that had occupied and looted the city.Footnote ²⁰

We will revisit these important incidents below; here, I would note the coupling of Persians and Arabs by the Chinese authors. This may reflect some understandable confusion by Chinese writers as to who these foreigners were, for “bosi” and “dashi” were first and foremost the names that were given to distant countries and not the terms that they commonly applied to the foreigners living among them. In fact, the most common terms used to describe foreign merchants in Tang China were “hu” 胡, a term most commonly applied to Persians but also to Tibetans, Turks and other pastoral nomads, and “fan” 番 (alternately, 蕃 or 藩), typically used for foreigners or aliens and often found in port cities, and combinations such as “fan-guests” (fanke 番客) and “hu-merchants” (hushang 胡商).Footnote ²¹ Without a doubt these terms or ethnonyms all involved degrees of ethnic stereotyping, a subject to which we will return, but the point to be made here is that these were the terms used for most of the evidence relating to the west Asian merchants in the ports of China.Footnote ²²

There is also reason to believe that Persians and Arabs made common cause in China. During the century following the Abbasid defeat of the Sasanians, a process of conversion was underway that resulted in the vast majority of Persians converting to Islam by the middle of the ninth century.Footnote ²³ We also know that the Muslim Persian Samanid empire (819–999), a vassal state of the Abbasids in eastern Iran, was actively engaged in maritime commerce.Footnote ²⁴ Even at the beginning of the Abbasid period, we have evidence from Du Huan about the commingling of Arabs and Persians. He writes that, in Dashi (the Abbasid caliphate), “Arabs and Persians are mixed and live together” (dashi bosi canza juzhi 大食波斯參雜居止).Footnote ²⁵ So it is reasonable to assume that as they began making their way to China by sea, Arab merchants accompanied their Persian counterparts and traveled on Persian ships. Thereafter, the number of Arabs undoubtedly increased, but, given the commingling of Persian and Arab merchants, it might be best to consider their presence in China as that of an Arab-Persian community.

If Tang references to Arabs are rare, those for Muslims are virtually nonexistent, even though there is no doubt about their having been present in Tang China. We must first dismiss the intriguing but legendary accounts of Sa’d ibn Abi Waqqas (Sahaba Saadi Gangesi 撒哈八撒阿的乾葛思), who, according to Ming and Qing accounts, made three trips to China, the first as an envoy from the Prophet in 629, and finally to Guangzhou where he built two mosques and was eventually buried. Although an important part of Chinese Muslim lore, there is no support for this story from Tang sources, and it is moreover highly implausible that an associate of Muhammad would have made his way to China that early.Footnote ²⁶ The first Chinese description of Islam of which I am aware comes from Du Huan, though it not named as a religion but rather presented as the religious practices of the Arabs:

There are also claims, mainly in stelae dating from the fourteenth century and beyond, of Tang origins for China’s most ancient mosques, namely those in Guangzhou, Hangzhou, Yangzhou, Quanzhou and Xi’an (Tang Chang’an), and also for the Lingshan Holy Tomb (Lingshan sheng mu 靈山聖墓) in Quanzhou. In none of these cases is there Tang evidence for these early dates, and the scholarly consensus is that none of them predates the Song.Footnote ²⁸

There are, finally, Persian and Arabic sources claiming a Muslim presence in Tang China, and here we are more fortunate. The Arab physician and geographer Sharaf al-Zaman al-Marzawi (d. 1120) described a group of Shi’a Muslims fleeing Sunni persecution in Khurasan during the late Umayyad (c. 740s) who came to China and settled on an island in a river across from a large port (a port that, Schafer speculates, was Guangzhou) and continued in existence there for some time.Footnote ²⁹ While plausible – the Shi’ites in Khurasan suffered persecution by the Umayyads and could well have fled to China – the lack of any corroboration from Chinese sources leaves the story’s veracity in doubt.

Very different is the account of Muslim merchants in Khanfu (Guangzhou) from the Akhbar al-Sin wa-‘l-Hind (Account of on China and India). This work is actually a collection of three documents with separate authors that was compiled by Abu Zayd al-Sirafi in 916. The account comes from a section written in 851 by an anonymous merchant who had been to India and quotes a merchant named Sulayman about China:

This quotation, which is found in a collection of observations about China, is very like descriptions of Muslim merchant communities elsewhere in Asia, and is widely accepted as authentic.Footnote ³¹ It is, to my knowledge, the first reliable description of the practice of Islam in China, but describes it as a practice limited to the foreign merchant community. As we shall see, that insularity of religious practice characterized Islam in the port cities of China throughout the period covered by the book.

Abu Zayd al-Sirafi is also responsible for our only account of an Arab in Tang China. Ibn Wahb al-Qurashi was a native of Basra and a member of the family of Muhammad who, after the sacking of Basra by the Zanj in 871, went to Siraf. There he came across a ship departing for China, and on a whim embarked. On arrival in Guangzhou or Khanfu, as it was known to the Arabs, he further decided to proceed to the capital in hope of an imperial audience. Arriving after a journey of two months, he submitted petitions announcing himself as a descendent of the “prophet of the Arabs.” The emperor, in response to his petitions, ordered the governor of Khanfu “to make investigations and inquiries among the Arab merchants about Ibn Wahb’s alleged kinship with the prophet of the Arabs.” After receiving a positive report, the emperor granted an audience – described in detail and involving back-and-forth exchanges via the interpreter – which dealt with Islam and its prophets, the states of western Asia, the age of the world and Ibn Wahb’s reasons for coming to China. The emperor then plied him with gifts, ordered the use of post horses for his return to Khanfu, and instructed the governor there to treat him with honor until his departure. Ibn Wahb further gave Abu Zayd a description of Chang’an that included such realistic details as the east/west division between official households and merchants and commoners.Footnote ³²

This is a curious story. That an elderly man – described as being of advanced years but with his senses intact – who was neither an emissary nor a merchant, but an individual whose claim to fame was his religious lineage, would travel to China and then succeed spectacularly is implausible, and there are elements of the story that particularly defy belief. Notably, it is unthinkable that the Tang emperor would say, as Ibn Wahb reports, that he esteems only five kings: first and foremost the king of Iraq, who is “at the center of the world,” with “the other kings … ranged about him.”Footnote ³³ However, other parts of the account have the ring of veracity, such as a remarkably accurate description of Chang’an, which Ibn Wahb provides to Sulayman.Footnote ³⁴ I would accept the basic outline of the story, but with the understanding that it was creatively shaped for its Arab audience. But in terms of our concern about Arabs living in China, it is noteworthy that the merchants of Khanfu are represented as an established group who, when consulted about Ibn Wahb, are able to vouch for his identity.

The Way to China and Its Trade

It is remarkable that the Tang–Abbasid trade existed at all. The sea route from Basra to Guangzhou was over 6000 miles in length, complex and treacherous (see Map 1). That a direct link not only existed but flourished during the Abbasid period is attributable to three factors. The first was the existence of a ship that was capable of making the voyage on a regular basis, namely the Arab dhow – known in southeast Asia as the Kunlun ship – characterized by sewn rather than nailed planking and, until the eleventh century, the only sea ship capable of such journeys.Footnote ³⁵ That such ships actually made their way to China has been demonstrated by two discoveries of shipwrecked dhows of likely west-Asian origin. The Belitung shipwreck, whose remarkable cargo of ceramics was described earlier, was discovered in 1998 off the coast of the island of Belitung, which lies between Sumatra and Borneo (see Map 2). It has been dated to after 826 and, given its overwhelmingly Chinese cargo, had clearly come from China.Footnote ³⁶ Then, in 2013, a remarkably well-preserved dhow was discovered in the Thai province of Samut Sakhon at the northern edge of the Gulf of Thailand.Footnote ³⁷ Known as the Phanom Surin shipwreck and preserved in a mangrove swamp that preserved timbers, ropes and wadding materials, it has been dated to the late eighth century, and its cargo – while not large – was revealing. It included ceramics from Guangdong, the Mon-speaking areas of Thailand, and the Persian Gulf, and, most remarkably, an inscription on a Persian jar in the Persian Pahlavi script. The inscription, which reads “Yazd-bozed”– a proper name, presumably for the merchant aboard the ship or perhaps the producer of the jars – is the earliest Pahlavi inscription to be found in south, southeast or east Asia, and points to the role of Persian merchants in the trade between western Asia and China.Footnote ³⁸ The two heads of west Asians – one incised onto a brick and the other terracotta – both discovered in Thailand and dating to the eighth century (Figures 1.3 and 1.4), provide yet further evidence for a Persian presence in southeast Asia.

Figure 1.3 Graffiti-caricature of a west Asian merchant on a brick, early eighth century

(courtesy of Fine Arts Department and John Guy)

Figure 1.4 West Asian merchant head, eighth century

(courtesy of Fine Arts Department and John Guy)

The second factor was the Asian monsoon, an annual weather pattern that both facilitated and conditioned long-distance travel in Asian waters. Specifically, the prevalence of southwest-to-northeast winds in the summer months and northeast-to-southwest winds in the winter months did not simply facilitate west-to-east and east-to-west travel, respectively, but also made possible the traversal of vast stretches of sea in the Indian Ocean by significantly shortening travel times.Footnote ³⁹ Third was the trade itself, which was based upon the demands by the rulers and ruling classes of two great and prosperous empires for precious goods from the other end of Asia. We shall return to this trade, which constituted the lifeblood of the maritime merchants. Suffice it to say that both textual and archaeological evidence bear witness to a vital and flourishing commerce.

It should be stressed that west Asian merchants were not alone in their commercial endeavors. From the Han into the early Tang, China’s most important sea trade was with the states of southeast Asia, and in the port cities the Kunlun merchants of that region predominated. According to Wang Gungwu, by the mid-eighth century a transition was underway in which the Kunlun merchants were giving way to Persians and Arabs with their long-distance trade, a change that became fully apparent in the ninth century. It should be stressed, however, that there was always an active trade with southeast Asia, most particularly with Srivijaya, the maritime power centered in eastern Sumatra, either in the course of travel between China and west Asia, as in the case of the two shipwrecks, or exclusively between China and southeast Asia.Footnote ⁴⁰

By the ninth century, the knowledge of how to accomplish this lengthy voyage was sufficiently widespread to result in descriptions of the route in both Chinese and Arabic. In his “Route to Foreign Countries across the Sea from Guangzhou” (Guangzhou tong haiyi dao 廣州通海夷道) from 801, the statesman and geographer Jia Dan 賈耽 (729–805) provided a highly accurate sailing itinerary from Guangzhou to Baghdad, not only with the primary route past Sumatra and Ceylon and on to the Persian Gulf, Basra and Baghdad, but also providing alternate routes through southeast Asian waters, and a further route skirting the Arabian Peninsula and going down to the northeastern coast of Africa.Footnote ⁴¹ This itinerary, which was excerpted from his now-lost 40-chapter geography of the world, was clearly based on the reports of mariners who had come to China, who had made their way to the capital, for Jia himself was not a traveler and had not even served in Guangzhou.Footnote ⁴²

As valuable as Jia Dan’s itinerary is, the Arab accounts are more useful for our purposes, drawing directly as they do on the accumulated knowledge of the west Asian mariners. Both the anonymous traveler writing in 851 (in An Account of China and India), whose description of the Muslims in Guangzhou was cited above, and the slightly later Book of Routes and Realms (Kitab al-Masalik wa-l-mamalik) of Ibn Khurradadbih (d. 885) offer detailed descriptions of the route from the Persian Gulf to Khanfu (Guangzhou).Footnote ⁴³ The geographical information offered by the two is similar, and while The Book of Routes and Realms is more scholarly and authoritative, we will use the Account of India and China, since, as an example of rihla or travelogue literature, it is more likely to reflect the information actually used by Arab and Persian mariners.Footnote ⁴⁴ After describing the primary western terminus Siraf (where goods from al-Basra and al-Ubullah were transshipped) and the sometimes-dangerous journey (owing to pirates and reefs) through the Persian Gulf, the ships cut across the ocean to the port of Kollam Malay on the southwestern coast of India,Footnote ⁴⁵ where large China-bound ships were assessed a toll of 1000 dirhams (in contrast to other ships, which were assessed only 10 or 20 dirhams) (see Map 1). From there, the China-bound ship skirted the southern coast of Ceylon, made for the Nicobar Islands in the Bay of Bengal to replenish food and water, stopped at Kalah Bar in Malaya, passed through the Malacca Straits, made additional stops at the island of Tiyumah, Sanf in Champa and the nearby island of Sanf Fulau, and finally headed to Khanfu. The account further notes that the south China coast had a reputation for dangerous reefs and storms. The author provided a general timetable for the whole trip: roughly a lunar month (29–30 days) for each of the four legs of the trip, marked by Kollam Malay, Kalah Bar, Sanf and Khanfu. With stops, the whole trip would take around six months.Footnote ⁴⁶ The most striking feature of this account is how unremarkable it is. The route it describes was lengthy and complex but also well known and frequently traveled.

Khanfu was not the end of the road for many of the Arab and Persian merchants who made their way to China. In its account of the route to China, The Book of Routes and Realms has merchants stopping first at Luqin (Annan or Hanoi), then at Khanfu, then Khanju (Quanzhou), and then Qantu (Yangzhou) at the beginning of the Grand Canal.Footnote ⁴⁷ Yangzhou was a major emporium for inter-Asian trade, with substantial populations of Arabs and Persians, which will be discussed later. There is evidence, moreover, that Persian merchants were active not only in the ports but in many Tang cities,Footnote ⁴⁸ this in marked contrast to the Song period, when foreign merchants were restricted to designated port cities.

It was, of course, the wealth of exotic and much sought-after goods that caused this travel. Consider Ibn Khurradadbih’s catalog of the goods to be had from across maritime Asia in The Book of Routes and Realms:

Among this plethora of goods, two stand out. Ever since Roman times, silks from China had been highly sought after throughout the Eurasian world, and the fact that Ibn Khurradadbih begins his list with three varieties of Chinese silk bears witness to the demand for it in Abbasid society. Second is porcelain, which was described by the author of Account on India and China: “They have excellent cohesive green clay, out of which they manufacture goblets as thin as flasks, through which sparkle one sees the sparkle of water can be seen.”Footnote ⁵⁰ However, if the Belitung shipwreck is any guide, porcelain constituted only a small portion of the ceramic goods that made their way west. Among the 60,000 artifacts excavated from this shipwreck, 400 were porcelain, which Regina Krahl identifies as referring to the translucent Xing ware from Hebei,Footnote ⁵¹ but there were also green-splashed wares from Henan, celadon from Zhejiang, and, most importantly, stoneware from Changsha in Hunan, of which there were 57,500 objects, many of these clearly intended for west Asian consumption.Footnote ⁵² These include white ware cups and saucers of a style widely imitated in west Asia; blue-and-white wares from the Gongxian kilns in Henan, whose cobalt-blue was produced using cobalt presumably imported by Arab or Persian merchants; and Guangdong jars (Dusun jars) – large vessels in which were stored smaller ceramic pieces, lead ingots and fruit star anise – which have been found across maritime Asia, including one excavated from the floor of the Friday Mosque in Siraf and dating to 841 CE.Footnote ⁵³ Some of the bowls from Changsha also have what appears to be roughly copied Arabic script. According to Chen Dasheng 陳達生, Tang kilns in the Hunanese city of Changsha were using imported Muslim ceramics as prototypes for the mass production of ceramics, including some with Arabic inscriptions, “expressly for export to foreign Muslim markets through the port of Yangzhou.”Footnote ⁵⁴

As for imports into China, Han Yu 韓愈 (768–824) is evocative: “The commodities of the outer nations arrive daily: pearls and aromatics, rhinoceros and elephant [horn and ivory], tortoise shells and curious objects – these overflow in the Middle Kingdom beyond the possibility of use.”Footnote ⁵⁵ We must note that while demand was great for certain commodities such as frankincense and myrrh, which came exclusively from western Asia, the majority of those luxury goods for which the demand was seemingly insatiable had a variety of sources across the Southern Seas and occasionally in China as well.Footnote ⁵⁶ These included rhinoceros horn, ivory, kingfisher feathers and, in fact, many of the goods that Ibn Khurradadbih attributes to India and southeast Asia. Yet increasingly during the Tang it was the Arab and Persian merchants who arrived with them at Chinese ports.

Special mention should be made of pearls, for which there was a ready market in both east and west. Since pearl beds were to found throughout Asian waters, no region had a monopoly on them. However, given the great esteem in which pearls – especially large and lustrous pearls – had across Asia, and their portability, it is not surprising that they played a significant role in international commerce. Indeed, as Edward Schafer has shown, Persian merchants in Tang China were typically regarded as very wealthy and bearers of (or seekers after) valuable pearls, not infrequently pearls with magical qualities ascribed to them.Footnote ⁵⁷

We have no way of even estimating the quantity or value of the trade that flowed between Chinese ports and the Persian Gulf during the Tang. During his tenure as prefect of Guangzhou, which began in 769, the refusal by Li Mian 李勉 (715–786) to extort bribes was credited with increasing the number of ships arriving from the Western Regions from four to five per year to over forty.Footnote ⁵⁸ This development occurred when Guangzhou was still suffering from the aftereffects of the Arab-Persia raid of 758. But the corruption to which it alludes was ongoing; Jitsuzō Kuwabara has amply documented the reputation of a posting in Guangzhou for allowing the accumulation of fabulous wealth, and that might be seen as another measure of the great value of the trade.Footnote ⁵⁹

From the west Asian side, we can also cite the Kitab ‘Aja’ib al-Hind (“Book of the Wonders of India,” c. 950), a travel book by the sea captain (nakhuda) Buzurg ibn Shariyar (c. 952) that tells of Ishaq bin Yahuda, a Jewish merchant from Siraf (Oman) who visited China between 882 and 912, and who upon his return had transformed his initial capital of 200 dinars into “a shipload of musk, silk, porcelain, jewels and other precious-stones and other wonderful Chinese merchandise. The musk, silk and porcelain alone were reported to have been worth 3 million dinars.”Footnote ⁶⁰ Buzurg also relates an account of an audience that Ishaq had with the ruler of Lubin (a Chinese province), at which the ruler asked him if he had seen such wealth as was evident in his court and also addressed Ishaq as ya ‘arabi (Arab).Footnote ⁶¹ Apart from the evidence that Ishaq’s story offers for the involvement of Jewish merchants in the China trade – something confirmed by accounts of a massacre in 879 that will be discussed later – a story like this, with its aura of wealth in both China and Siraf, surely helped establish the promise of the China market as a place where fabulous profits could be gained.

Tang Supervision of Maritime Trade

Throughout the medieval maritime world, local rulers and governments had a natural interest in the merchant ships that arrived on their shores, and their policies included patronage of the traders, taxation, forced purchase and free trade. In many cases the foreign communities themselves acted on behalf of the local rulers. Although he is writing about a later period, André Wink’s analysis is pertinent: “More often than not typical diaspora communities like the Badija Naidus, the Sayyids of Golconda or the Mappillas of Malabar appear to have been rooted in the revenue collection and even to have been able to obtain access to court politics.”Footnote ⁶² Even in western Asia, with its highly developed political institutions, governmental interest in maritime trade was confined primarily to a concern for revenue and the demand for specific goods, and was little involved in the encouragement of trade.Footnote ⁶³

In Tang China, by contrast, the role of the government was far more central. According to Wang Zhenping, the Tang central government was theoretically not involved in foreign trade. Rather, trade was overseen by the governors of Guangzhou (for the south seas trade; trade with Korea and Japan was largely channeled through Yangzhou), and on a more ad hoc basis by “commissioners for trading with foreign ships” (shibo shi 市舶使), who were typically eunuchs.Footnote ⁶⁴ The latter represented the compelling interests of the imperial palace and imperial family in the luxuries provided by the maritime trade, and, as Edward Schafer has noted, these eunuch officials were notorious for their exactions and corruption.Footnote ⁶⁵ Indeed, so prominent was the role that they played that the author of the Account of India and China, writing in 851, described eunuchs governing Guangzhou alongside the civil governors.Footnote ⁶⁶

That said, to foreign eyes the Chinese approach to imports seemed highly organized and even generous, as can be seen in the description of the Chinese procedures in Account of India and China:

Contemporary Chinese accounts are considerably more critical, with an emphasis on the abuses of local officials and eunuchs. To quote an imperial edict from 834, unusual in addressing the issue of maritime trade:

Whether such imperial attitudes had much impact is questionable, for Chinese sources suggest that the actions of the early ninth-century prefect and Lingnan military governor Wang E 王鍔 were more representative:

However different their perspectives, Arab and Chinese authors are agreed on the major role played by officials in the treatment of maritime trade, and that was a fact that stood in sharp contrast to other ports of Asia and informed the lives of the merchant communities residing in the ports of China.

Merchant Life in China

As the edict of 834 clearly demonstrates, a number of cities served as the termini for ships arriving from the Nanhai (South Seas) and hosted foreign communities, though information about most of those communities is frustratingly sparse. We have already encountered Yangzhou, with its strategic location at the entrance to the Grand Canal, as the site of the 760 massacre of Persians and Arabs as well as the port from which Chinese-manufactured Muslim ceramics were exported. We also know that Jiaozhou (near modern Hanoi but then the southernmost port of the Tang empire) was an important port of call for Arab and Persian ships coming to China – Ibn Khurradadbih described Luqin, presumably Jiaozhou or its port, as having “Chinese stones, Chinese silk, Chinese porcelains of good quality, and rice”Footnote ⁷⁰ and also that it prospered in the years following the Persian and Arab raid of Guangzhou in 758.Footnote ⁷¹

For a glimpse of what life was like for the maritime merchants in China, we must turn to the emporium of Khanfu (Guangzhou). It was, in the words of Wang Gungwu, a large trading settlement or frontier settlement, inhabited by merchant-adventurers, foreign traders, and non-Han peoples of Guangdong, a city in which Han Chinese were a distinct minority.Footnote ⁷² Writing in the tenth century, al-Masʿudi described the remarkable geographical spread of the merchant community in Khanfu in the mid-ninth century. Within the city “there were buildings [with occupants] from Basra, Siraf, Oman, the cities of India, the islands of Zabedj (Java) and Sinf (?), and other realms, and they were stocked with their merchandise and cargoes.”Footnote ⁷³ The city was substantial enough to impress the Chinese monk Ganjin (Jianzhen 鋻真), who visited Guangzhou in 750 and marveled at the “enormous variety of races” there and offered this description: “The city has triple fortifications. The governor general commands six banners, each constituting an army, and their dignity is no different than that of the Son of Heaven [i.e., the emperor]. The city is filled with purple and crimson and it is surrounded by the press of the suburbs.”Footnote ⁷⁴ Such a respectful impression of this frontier outpost might well not have been shared by Chinese officials coming from the great cities of the north, but from the perspective of Heian cities or the port cities across the expanse of maritime Asia, Guangzhou could well have looked magisterial.

Within Guangzhou, the foreigners – and the west Asians particularly – resided primarily in the “foreign quarter” (fanfang 蕃坊), under the authority of a foreign headman. Discrete residential quarters for foreign merchants were a common feature of ports across maritime Asia, so Arab readers would not have been surprised by Sulayman’s description of Khanfu’s Muslim community with its judge that was cited earlier. Chinese sources provide confirmation of this. Li Zhao, writing in the early ninth century, mentions a foreign headman (fanzhang 蕃長) who presided over the foreign traders and cooperated with the authorities in drawing up the manifests for the arriving ships.Footnote ⁷⁵ Liu Xun 劉恂 of the late Tang described this encounter with the Guangzhou headman: “At the house of the Fan-ch’iu [Fanqiu] 番酋 (lit. foreign chief), I once ate the Persian dates brought over from his own country. The fruit had sugar-like colour, soft skin and flesh and tasted as if it was first baked and then boiled in water.”Footnote ⁷⁶ Since the Chinese sources almost always use the ambiguous “fan” for “foreigner,” we generally cannot determine the ethnic identity of the headman, or even determine whether there was only one headman at a time, but in the case of Liu Xun’s account the Persian dates strongly suggest that the “chief” was in fact west Asian.

The adjudicatory functions that Sulayman ascribed to the Muslim judge in Guangzhou are supported by an important Tang text. According to the sixth chapter of the Tanglü shuyi 唐律疏議 of 635,

The commentary elaborates:

Whether such fine distinctions as those that were applied to Koreans in the Chinese application of extraterritoriality were also made of west Asians we cannot say for sure, but the evidence from Sulayman suggests that the Chinese authorities were content to recognize religious rather than geographic identity in their case.

That Guangzhou had a foreign quarter does not mean that foreigners – and their families – were content to live in it. In the biography of the infamous Wang E, whose commandeering of imported goods for his private gain was detailed above, the statement is made that “The Cantonese and foreigners (yi ren 夷人 or ‘eastern barbarians’) lived amongst each other [in the foreign quarter]. Because the land was undesirable, they sought to live in the river market [area].”Footnote ⁷⁸ Far more informative is the remarkable account in the biography (or biographies; the versions in the Old and New Tang Histories vary slightly) of Lu Jun 盧鈞, who came to Guangzhou as prefect and military governor in 836, a generation after Wang E had been there. After describing how Lu had reversed the corrupt policies of his predecessors and governed honestly, thereby relieving the vexations of the foreign merchants, it describes his response to conditions in Guangzhou that he found unacceptable. Foreigners were living together and intermarrying with the Chinese, and many foreigners had bought rice fields and built houses. If the local authorities tried to interfere with them, they combined and rose in revolt. In response, Lu Jun enacted laws forcing the foreigners to live in a separate quarter (yi chu 異處) and forbade them from marrying with Chinese or acquiring land and houses.Footnote ⁷⁹

To some extent this account reflects the social fluidity of a frontier city, a fluidity that Lu Jun, good Confucian official that he was, was attempting to counter. In the same biography we are told that Guangzhou had become a place of exile where the children of disgraced officials who had been sent there found themselves stranded, unable to return even after pardons had been secured. Jun arranged for assistance for their medical and marital needs, in all helping several hundred families. After his three-year term, “several thousand Chinese and foreigners” requested that a shrine be built to honor Jun.Footnote ⁸⁰ For our purposes, the critical question is, who were the foreigners who were intermarrying and settling with the local population? Were they tribal peoples from Lingnan, merchants from southeast Asia or west Asians? The Old Tang History uses the term man liao 蠻獠 (both terms for southern peoples) to describe those who lived together with local inhabitants (turen 土人), and that could be taken to mean the local tribal peoples. However, it also talks about the Man ships (manbo 蠻舶) of the South Seas arriving in Guangzhou, while the New Tang History states that it was fan liao 蕃獠 who lived together with the Chinese (Huaren 華人), thus employing the character commonly used for west Asians.Footnote ⁸¹ From this evidence we can conclude that the maritime merchants were certainly among those who were intermixing with the local population in Guangzhou, and while we cannot say conclusively that these included Persians and Arabs, there is no reason why they should have been excluded. More broadly, the passage indicates that at least portions of the foreign merchant community had put down roots in Guangzhou and assumed settler rather than sojourner status. This was an important development, and foreshadowed the Muslim merchant communities of later centuries.

One limitation of the Tang–Abbasid sources for the merchant communities in China is that they give us no almost information about individual merchants or about their internal functioning. Apart from the invaluable account of the Guangzhou community with its Muslim judge in the Account of China and India, which has been discussed above, the remainder of the ninth-century portion of this work containing the accounts of Sulayman and other anonymous sources contains nothing about individual merchants or their lives. What is does present, however, is a wealth of descriptive material concerning China (and India, though only the Chinese parts concern us here). If we view this material as constituting the Muslim maritime community’s collective knowledge of China, then an analysis of these texts can tell us a great deal about the community’s social positioning and the kinds of information to which they had access.

The descriptions of China with which we are concerned come from the ninth-century portion of the Account (later in the work, Abu Zayd has some additional information from the tenth century) and cover a wide range of topics presented in 72 numbered items, most of which are short entries in the Akhbar al-Sin wa-l-Hind translation.Footnote ⁸² Some of the entries are devoted to maritime travel, Asian port cities and India, and some compare India and China. However, most concern China, and among them certain subjects stand out by virtue of the frequency of their occurrence or the detail of their coverage.

The subject of commerce is an instance of the latter. Although treated in only three entries, these are among the longest entries in the work. They detail how officials processed and taxed the cargoes of arriving ships (#34, quoted earlier), the mechanisms for borrowing, lending and handling defaults (#44), and the serious consequences of bankruptcy (#45), all matters of the utmost importance for merchants. The credit practices, we might note, were based upon written agreements and were backed up by the force of the law for those who defaulted. Related tangentially to commerce are the accounts of Chinese buildings, which are described as having been built with wood (#60, 72) and therefore a cause of the fires common to Guangzhou. The result, we are told, was to increase the rarity of merchandise in the Sino-Arab trade, since it would burn in the warehouses.Footnote ⁸³

By far the two most frequently occurring subjects are government and the personal lives of the Chinese, and the former are largely skewed towards local government. Concerning the empire as a whole, we are told that the king of China has over 200 urban metropolises, each with its prince and eunuch (#33), and the king himself is described in only the vaguest of terms: as one of four kings of the world, beneath the Arab king but above the Roman king (#24), as lacking designated heirs (#54), and as secluding himself two months of the year in order to inspire awe among his subjects (#39). Concerning the functions of government at the local level, the entries are far more knowledgeable, reflecting the first-hand observations of Arab merchants. They describe local officials in some detail (#37, 38),Footnote ⁸⁴ as well as taxation (incorrectly stating that the Chinese had no land taxes, only head taxes) (#40, 47), legal proceedings (#38, 58, 67), coinage (#34), schools (#48) and the documents required for travel around the empire (#43). One intriguing entry describes a public bell – to be found in every locality – that anyone who has suffered an injustice can ring, and then present his grievance to the “prince.” From Song Chinese sources we know that the Chinese indeed had such a public grievance system, though using drums rather than bells.Footnote ⁸⁵ Finally, armies and warfare are mentioned only twice, and then briefly (#56, 72), a reflection of the largely pacific character of this world of commerce in the mid-ninth century.

The entries describing the lives of the Chinese present abroad and fascinating array of observations. Concerning appearances, we are told that the Chinese “are handsome and large,” with skin “of a white hue and a tint of red” and very black hair, and also that women leave their hair uncovered, in contrast to the men, who cover theirs (#49), that all Chinese dress in silk, using multiple layers in the winter and a single layer in summer (#21) and that Chinese almost never have beards (#65). Food and the preparation of food is another common subject; the Chinese staples of wheat and rice, fruit trees, and the butchering of meat are all briefly described (#62, 72, 71). Most informative is #22, with its details of cooking and foods of all sorts:

Other topics include marriage (#57, 61), illness and medicine (#46, 72), death and burial (#35), writing by all Chinese, “poor or rich, small or great” (#36), and the love of music (#55). Remarkably specific information is also given on toilet practices (#23, 71) and the males’ lack of circumcision (#63), and the charge is made that Chinese “surrender themselves to sodomy with young slaves” (#59).

There are, finally, several references to Chinese religion: to their worship of statues (#64), the role of priests speaking for the statues (#70), and their practice of Buddhism, with its belief in the reincarnation of souls (#72). The most judgmental entry (#23) likens them to the Zoroastrians: “They eat carrion and other similar things, just as the Magians do; in fact, their religion resembles that of the Magians.”Footnote ⁸⁷ There is nothing, however, that would hint at any concern with proselytizing the Chinese.

Taken as a whole, these ninth-century descriptions of China and the Chinese reveal a remarkable breadth of knowledge, but knowledge with definite limitations. Against vague and at times fanciful ideas about the empire and monarchy, we have concrete and detailed accounts of those elements of government, law and products that one would expect from merchant observers. I would suggest that the accounts of the customs and activities of the people reflect a level of social and even personal intimacy that came from extended residence in Guangzhou, a residence which, as we have seen, included living among the Chinese and even intermarrying with them. At the same time, nothing in these accounts suggests any significant interaction with the local elite, with the exception of the “prince,” probably the provincial governor, whose habits are described, perhaps as a result of official interactions, including banquets, held for the merchants.

About the interactions of the maritime merchants themselves we are almost entirely ignorant. We know from the Account that the Guangzhou Muslims had a chief who led them in prayers and delivered the weekly sermon (khotba). Whether the merchants had additional corporate practices we do not know as in southern and southeastern Asia, but Edward Schafer provides an intriguing glimpse of what he calls mutual benefit associations among Persian merchants.Footnote ⁸⁸ In the three examples that he found in Tang tales, merchants gathered to socialize, compare their treasures, and in one case even to pool their resources to purchase a valuable pearl. In the most revealing story,

Beyond this, we can speculate that the mechanisms for Islamic trade in west Asia and the Mediterranean were employed by the Muslim merchants in Tang Guangzhou. These included the universal and limited investment partnerships (mufawada and ‘inan), which offered a degree of financial security in the pooling of funds, and commenda contracts, in which an agent-manager was entrusted with capital or merchandise.Footnote ⁹⁰ There were also corollary practices that “rendered possible the delegation of power and authority to associates, colleagues, and even strangers, as economic circumstances required,” for example, by allowing an investor or merchant to entrust his goods to another, who would act for him in disposing of the goods and provide him with the proceeds at no charge.Footnote ⁹¹ Although this description is largely based upon eleventh-century sources from western Asia, it is highly likely that the Guangzhou Muslims made use of these practices – or some much like them – that relied upon trust and honor more than written contracts of the sort ascribed to the Chinese. As an example of the importance of ethics in commerce, we might cite the case of Abu’Ubayda ‘Abdallah ibn al-Qasim, known as “al-Saghir” (the small), from a small Omani market town, who was involved in the China trade and traveled there, probably before 758. On one occasion, when he discovered that his partners in the aloes wood trade had disparaged a shipment of wood – probably from China – to drive down the price, and had then praised the same goods after purchasing them so as to drive up the price, he broke off the partnership.Footnote ⁹²

The Arab-Persian merchants in China were not simply isolated individuals pursuing wealth on their own, but part of a diaspora creating the most effective and integrated long-distance trade network that maritime Asia had ever seen, and their success was to a large extent dependent upon shared values and mutual trust. At the same time, neither the trade nor the trading diaspora was a fixed entity. As we will see below, both underwent dramatic developments in the eighth and ninth centuries that had profound consequences for their subsequent histories in China.

The Waxing and Waning of the Settlements

Over the 700-year course of the Muslim communities that we are exploring in this book, the period of the late Tang was perhaps the most dramatic, marked by three traumatic events: the Arab-Persian raid on Guangzhou in 758, the massacre of Persians and Arabs in Yangzhou in 760 and the larger massacre of Muslims, Christians and Zoroastrians in Guangzhou in 879. These have all been mentioned earlier; here our attention will be focused upon how these events shaped the history of the west Asian communities in China.

Although there is no discernable connection between the events of 758 and 760, both occurred during the Rebellion of An Lushan 安祿山, which wracked the Tang from 755 to 763 and for a time threatened to topple the dynasty. The rebellion was fought out mainly in the north – most famously with the rebel capture of Chang’an in the sixth month of 755 and the flight of the emperor Xuanzong 玄宗 and his retinue to Sichuan – but the entire empire was profoundly affected, initially by the chaotic conditions that prevailed for the better part of a decade, since the initial rebellion spawned many local rebellions (especially after An’s assassination in 757), and, in the longer run, by the multi-faceted weakening of the power of the central government.Footnote ⁹³

The Yangzhou massacre was in fact a product of this season of rebellion, for it occurred when so-called government troops, sent to put down the local rebellion of Liu Zhan 劉展, entered the city of Yangzhou, and in their looting and killing made the prominent Persian and Arab merchant community a particular target.Footnote ⁹⁴ There is an interesting discrepancy among the four accounts of the massacre that sheds some light on the nature of that community. They are found in the biographies of the two Tang generals responsible for putting down a local rebellion. In the Old and New Tang History biographies of Zheng Jingshan 鄭景山, both state that “several thousand Arab and Persian merchants were killed.”Footnote ⁹⁵ By contrast, the biographies of Ma Shengong 馬神功 both mention only Persian merchants as having been killed.Footnote ⁹⁶ Why were the Arabs omitted? I would argue that this reflects the long-standing nature of the Persian community in Yangzhou, to which Arabs were a recent addition. We know nothing about that community in the aftermath of the massacre, so we can only speculate as to length and extent of the damage visited upon it and the South Seas trade with which it was involved. There is one bit of evidence from the ninth century indicating that a discrete Persian community continued. The Japanese Buddhist monk Ennin 圓仁, who recounted a lengthy trip that he made to China (in a diary that remains an important source for information on late Tang China), described how, during his stay in Yangzhou in 839, an official solicitation for funds to repair a balcony at a local Buddhist temple resulted in a donation of 1,000 strings of cash (out of 10,000 needed for the repairs) from the “Persian state” (Bosiguo 波斯國).Footnote ⁹⁷ The use of guo is curious, since it typically refers to a state. However, because of the local nature of the restoration project, it seems most likely that Bosiguo referred either to the Persians collectively (perhaps including Arabs) or else to a Persian headman who spoke for the community. In any case, Persians had clearly survived the 760 massacre.

In contrast to Yangzhou, Guangzhou in 758 was far from the scenes of rebel activity and the government was preoccupied with its campaign to take back the capital (accomplished with the help of Tibetan and Uighur soldiers), so this raid by Arabs and Persians in a frontier port seems to have elicited no reaction. The raid, an account of which began this chapter, is recounted in four places, two each in the Old and New Tang Histories. The least informative, from the New Tang History annals, simply says that Arabs and Persians “plundered” (kou 寇) Guangzhou.Footnote ⁹⁸ In their treatises on Persia (and Persians), the two histories are largely in accord, describing how the two groups plundered the city, burning its warehouses and storehouses and then leaving by sea, though the New Tang History states that the Persians followed the Arabs in the “raid” (xi 襲), presumably coming from the sea, capturing the city of Guangzhou, and burning its storehouses.Footnote ⁹⁹ The annals of the Old Tang History provide a very different account: “[Officials from] Guangzhou memorialized, [reporting] that soldiers from the countries of Arabia and Persia besieged the city, and the prefect, Wei Lijian 韋利見, abandoned the city and went into hiding.”Footnote ¹⁰⁰ Whatever their differences, all four accounts agree that this was indeed a raid and not a takeover of the city.

To return to the question with which we began this chapter, who were these raiders and from where did they come? Two suggestions – that they were the product of increased trade activity following the establishment of Baghdad as the Abbasid capital in 750,Footnote ¹⁰¹ or that they were disgruntled Arab troops sent by the Caliph to help Guo Ziyi (the loser in the Battle of Talas) to quell an insurrection – both seem highly unlikely, since neither theory explains how these groups would have made their way to the coast of China. That it was the work of unhappy traders also seems unlikely: they might have engaged in an urban riot, but the raid as described suggests plunder rather than simple destruction. Rather, the most likely explanation is that they were followers of the piratical strongman of southern Hainan, Feng Ruofang 馮若芳.Footnote ¹⁰²

In 749, the Chinese monk Ganjin, whose description of Guangzhou was quoted earlier, made unexpected landfall on Hainan Island, when the ship on which he was traveling almost sank in a typhoon.Footnote ¹⁰³ Once there, he was escorted by the inspector general to Wan’an prefecture 萬安州 (modern Lingshui) in the far south, where he was entertained for three days by the prefectural chief, Feng Ruofang. According to Ganjin,

This curious tale, dated just ten years before the Guangzhou raid, is supported by an entry in the early Song literary compendium, the Taiping guangji 太平廣記 of Li Fang 李昉, which recounts a Tang tale of one Chen Wuzhen 陳武振, whose mansion in Zhenzhou (modern Yaxian in southwestern Hainan) was filled with gold, rhinoceros horns, elephant’s ivories and hawksbill turtles. The source of this wealth came from “merchants from the west” whose ships had foundered on the coast. His success in doing so was attributed to his moude fa 牟得法 (method of capture), which involved reciting incantations from a mountain when a merchant ship appeared so as to call up wind and waves and trap the ship on the coast.Footnote ¹⁰⁵ This account is from the “magic” (huanshu 幻術) section of the compendium, and so one might question its reliability. However, the striking parallels with Ganjin’s account make it likely that Feng Ruofang was the model for Chen Wuzhen, and it should be noted that the southern coast of Hainan lay right along the most common sea route from the south to Guangzhou. Perhaps most important, in recent years archaeologists have found conclusive proof of ancient Muslim communities dating to the Tang and Song periods in southern Hainan. Two abandoned Muslim cemeteries were discovered in coastal areas (one in Lingshui, the other in Yaxian) with numerous tombs and stelae with Arabic inscriptions. Although none of them provide dates, they may well be associated with an eleventh-century influx of Muslims into Hainan that will be discussed in the next chapter, some of them, stylistically at least, can be dated to the ninth century, and several have Persian titles, thus indicating a connection with the Tang community (see Figure 1.5).Footnote ¹⁰⁶

Figure 1.5 Tang Muslim tombstone from Hainan (Guangzhou Museum)

Returning to the raid of 758, were the Persians and Arabs shipwrecked merchants and seamen who had been captured by Feng Ruofang and perhaps operating under his command? Could they have been a group that had escaped from the clutches of Feng and were acting on their own? We can only speculate, but the very existence of these west Asians in Hainan, living outside of the normal bounds of Tang–Abbasid trade, makes them the likely candidates for those who undertook the pirate-like raid of Guangzhou.

Whatever the identity of the raiders, the raid itself marked the beginning of a period of great difficult for the port of Guangzhou. Li Mian’s praised tenure as governor of Lingnan, which, as we observed earlier, resulted in the increase of arriving ships from the Western Regions to over 40 per year, was the exception in the late eighth century. Earlier, in 763, the Commissioner of Maritime Trade (a eunuch) got rid of the military governor and allowed his men to ransack the city.Footnote ¹⁰⁷ Then, in 773, his successor was killed by a mutinous officer who held the city for three years. Order was restored in 775 when the general Lu Sigong 路嗣恭, with a force of 8000, took the city and killed 10,000 of his “fellow traitors.” He also acted against the merchant community, executing those merchant ship crew members (shangbo zhi tu 商舶之徒) who had served the rebels and confiscating the family property of the merchants, worth several million strings, keeping that for himself rather than sending it to the capital. This displeased the emperor, who did not reward Lu for his military success.Footnote ¹⁰⁸ Not surprisingly, it also soured the foreign merchants on Guangzhou as a port, with the result that Annan (Hanoi) became their favored port.

In 792, the court received a request from the governor of Lingnan to address the trade woes of Guangzhou by imperial fiat:

Although the emperor was inclined to grant the request, it was countered by the minister Lu Zhi 陸贄, who in one of the clearest statements of market forces to be found in any Tang document, submitted that:

Lu argued further that, since Annan and Guangzhou were both part of the empire, it was unfair to discriminate against one in favor of the other. It is not clear whether there was an attempt to close the Annan port (the entry implies that the request was denied), but the fact is that in the ninth century Guangzhou was able to regain its dominant position in the maritime trade, and the Arabic Account that were analyzed above are a reflection of this, for they do not make any mention of the eighth-century troubles. Then came the massacre of 879.

Like the events of 758 and 760, the massacre came at a time of national convulsion caused by the rebellions of Wang Xianzhi 王仙芝 and Huang Chao 黃巢, which transpired during the years 874–878 and 878–884, respectively. In the judgment of Robert Somers, this lengthy period of rebellion was “… the final stage of a long period of social dislocation and widespread militarization that had begun many decades before.”Footnote ¹¹⁰ Ironically, Guangzhou was an almost accidental victim, for when Huang Chao and his troops approached the city in the fifth month of 879, they were ending a long period of flight south from stronger government forces that had begun some nine months before in Henan. After Li Tiao 李迢, the military governor of Guangzhou, refused to surrender, Huang’s forces stormed the city and sacked it. Four medieval Arab authors have provided accounts of the plunder and killing, but those of Abu Zayd al-Sirafi and al-Masʿudi (896–956) are the most valuable. Abu Zayd’s account is the earliest (c. 914) and the most detailed.Footnote ¹¹¹ After describing Huang and the origins of the rebellion, Abu Zayd continues:

Al-Masʿudī’s version largely agrees with Abu Zayd’s, though it puts the number of those killed at 200,000, “a prodigious number of inhabitants.” It also states that Huang’s forces destroyed the mulberry plantations outside of Guangzhou, thereby striking at the silk trade, “as the destruction of mulberry stopped the export of Chinese silk to Muslim countries.”Footnote ¹¹³ That the two accounts agree on Huang Chao’s destruction of mulberry trees provides important testimony for the importance of silk as an export commodity during the Tang.

Although Abu Zayd’s figure of 120,000 has been accepted at face value by many historians, perhaps because of the claim that they came from the Chinese census, it and al-Masʿudi’s even larger figure are almost surely exaggerations. There is no suggestion in the Account from 851 that such a great number of westerners had congregated in Guangzhou, a number that would have exceeded the populations of many of the leading cities in the ninth-century world. If one considers, further, that the only Tang statistic we have for the annual shipping traffic into Tang Guangzhou is for 40 ships (in the early 770s), it should be clear that there was no way for such a number to have been employed in Guangzhou, even if they could have made their way there.

That said, there can be no doubt that a tragedy of major proportions occurred at the hands of Huang Chao and his followers. For Abu Zayd, moreover, the consequences for the Arabs’ maritime trade with China were profound. In his account, after describing how the Chinese appealed unsuccessfully to the Turkish King of Taghazghaz for help in putting down the rebellion, he writes of the trade:

Although we cannot corroborate the particulars of these Arab accounts with Chinese sources, what is important is the Arab sense of rupture, betrayal and loss (the loss of properties being no small matter for merchants). As we will see in the next chapter, they and (at least) many of their fellow merchants from across Asia departed China and moved their operations to southeast Asia. The massacre therefore marks the beginning of a period of transition that would result in the reconstitution of the Muslim communities in China, but with distinctly different parameters and practices, and quite possibly contributed to a transition to a more segmented long-distance trade, with fewer ships traversing the entire route from western Asia to China.

Reflecting on the Tang period as a whole, I would suggest that the west Asian merchants of Guangzhou occupied an anomalous position. As essential middlemen in a highly lucrative trade that fed a luxury market in China (especially at the court) and supported large-scale ceramic production well beyond southern China, they were the recipients of liberal government policies and were able to settle and flourish in Guangzhou (as well other cities like Yangzhou). Yet they were also at the mercy of extortionate demands by corrupt officials and, more seriously, they were the targets (and on occasion agents) of violence. This undoubtedly owed much to the frontier, colonial nature of Guangzhou in the Tang empire, where both eunuchs representing the imperial household, a primary consumer of imported luxuries, and the military were able to play outsized roles, the latter in both fomenting and suppressing uprisings. But the great wealth and racial and cultural foreignness of the west Asian merchants, when combined with their significant numbers, would also have made them easy targets. This element of violence seems to have been a characteristic of the period, for as we will see during the succeeding Song period (960–1279), large-scale violence involving the foreign merchants in the port cities of China was nonexistent.