A Study of the Mistranslation of the Port Khānfou in China in the Tang Dynasty ()
1. Introduction
In 916, the anonymous Arab authors1 and travelers Abu Zaid Hassan (Abou zeyd) and Ibn Wahab wrote Ancient Accounts of India and China (أخبار الصين والهند., Aḥbār aṣ-Ṣīn wal-Hind) according to the experiences of Arab merchants in China. Consisting of two parts, the first part of the book was written by an explorer in India in (851) based on Sulayman’s and several unknown persons’ accounts of China; and the second part of the book was Abu Zaid Hassan’s account of China. Ibn Wahab’s Chang’an Youji 《长安游记》 (Chang’an Travel Notes) and miscellaneous accounts of India were included in the book as an appendix. The book is regarded as an authentic record of the Tang and Song dynasties of China by the Arabs, and “its account of China is highly valued by scholars” (Fang, 2008). Abu Zaid Hassan considered that Khānfou, the trade port from which the Arabs entered China, is Canton.
Since the peak of Sinology in Europe in the 18th century, Ancient Accounts of India and China was translated into many languages, with many scholars studying the translation of Khānfou and arguing whether it should be Guangzhou or Ganpu.
French scholar Abbé Eusèbe Renaudot translated the book into French and named it Anciennes relations des Indes et de la Chine de deux Voyageurs mahometans qui y allèrent dans le neuvième siècle in 1718. Due to the loss of the source text, others doubted the reliability of Renaudot’s work. “He was a learned and accurate scholar, and possessed an extensive acquaintance with the orientals” (Morrison, 1832). His examination and correction of the accounts of China by Arab authors provide a great reference value for subsequent systematic studies.
Louis-Mathieu Langlès (1763-1824) sent the original Arabic version of Ancient Accounts of India and China to press in 1811. In 1845, French scholar Joseph Toussaint Reinaud retranslated the book, using Relation des Voyages faitse par les Arabes et les persans dans l’Inde et la Chine dans le IXe S de l’ère chrétiènne as the name. He also published the original Arabic text sorted by Langlès and examined it at length. Reinaud’s translation made Renaudot’s version fall into oblivion (Mu, Wen, & Huang, 1983).
In 1922, G. Ferrand published a new French version, Voyage du Marchand Arabe Sulayman en Inde et en Chine. This version is better than the first two in explaining historical facts, but the translation is too plain to reproduce the Arab style (Mu, Wen, & Huang, 1983). In 1948, Jean Sauvaget’s translation Relation de la Chine et de l’Inde (rédigée en 851, texte etabli, traduit et commenté par J. Sauvaget) was published, based on the manuscript found in Aleppo, Syria and stored in National Library of France (Paris) later. The number of this manuscript is 2281, named Histoire (Mu, Wen, & Huang, 1983). Sauvaget’s translation and explanatory notes are better than any other version published before. In 1976, Japanese scholar Katsuji Fujimoto translated Sauvaget’s version into Japanese and annotated it.
In the early 20th century, Chinese scholar Zhang Xinglang quoted several parts from Ancient Accounts of India and China and translated them. In the 1930s, Liu Bannong, a Chinese litterateur, and his daughter Liu Xiaohui published a Chinese version, using Ferrand’s version as the original text. In 1983, Zhonghua Book Company published another Chinese version of Ancient Accounts of India and China translated by Mu Genlai, Wen Jiang, and Huang Zhuohan. The first part of this version was translated from vol. I of Sauvaget’s version and the second part from vol. II of Katsuji Fujimoto’s version.
In the 17th and 18th centuries, the Jesuits in China were attracted by Ancient Accounts of India and China and examined some historical facts in the book. In 1696, the French Jesuit Louis Le Comte (1655-1728) published Memoirs and Remarks on China (Nouveaux mémoires sur l’état présent de la Chine) based on his journey in China. The book was translated into several languages later. The book introduces geography, history, physics, nature, astronomy, mechanics, military, commerce, politics, and the Jesuit mission in China; a trade port, Khānfou, mentioned by Arab travelers is also discussed in this book. By the first half of the 19th century, Elijah Coleman Bridgman (1801-1861), a Christian missionary in China, translated Nouveaux mémoires sur l’état présent de la Chine into English and concluded that “Canfu is not Canton, as has been generally supposed” (Bridgman, 1832).
Robert Morrison (1814-1843), a Protestant missionary in China, finished translating Ancient Accounts of India and China, by two Mohammedan travelers who went to those parts in the 9th century. With some additional notes, he published his translation in the section Review of The Chinese Repository respectively in Vol. I, No. 1, May 1832 and Vol. I, No. 2, June 1832 (Morrison, 1832). He agreed with Renaudot’s opinion, that is, “our two authors are more ancient, and that the two dates they give, the one of the year 237 of the Hegira (850), which is that of the first traveler, and the other of the year of the same 264 (877), when a great revolution happened in China, are true and just” (Morrison, 1832). Both the two Mohammedan travelers who had been to India and China in the 9th century mentioned Canfu (Khānfou) (Morrison, 1832).
In July 1834, John Robert Morrison published an article in The Chinese Repository to discuss the location of Canfu (Khānfou) (Morrison, 1832). Besides, another article “Alphabetical list of the province, departments, and districts in China, with their latitude and longitude” by Morrison in The Chinese Repository is about the location, latitude, and longitude of 18 provinces, all departments, and districts in China. Morrison wrote that “Kánpú (Chekiáng, Háiyen hien), is supposed to be the Canfu spoken of in Renaudot’s account of two Arabian travelers, and which in the ninth century was a large port. It is now of little importance, having been surpassed by Chápú” (Morrison, 1844). And he believed that “we rather too hastily adopted the opinion of the learned translator, that the port of Canfu, which the travelers frequented, was the same as the modern Canton, called by the present Chinese Kwangchou foo” (Morrison, 1834).
Since the 20th century, Japanese scholar Jitsuzo Kuwabara strongly supported that Khānfou was Hangzhou (Ganpu). In several parts of his work Somatsu no Teikyoshihaku Saiiki-jin Ho Juko no Jiseki (A Foreign Superintendent of Shipping Trade in the Late Song Dynasty, Traces of Pu Shougeng) such as “the location of Khanfou”, “the explanation of Khanfou” and “related records about Guangfu”, he summarized major viewpoints on the location of Khanfou. Jitsuzo Kuwabara said that “Almost all relevant scholars believed that Khanfou is the transliteration of Guangfu 广府 and thus Khanfou refers to Guangzhou”. For example, Fujita Toyohachi (1869-1929) is a strong supporter of the view that Khanfou refers to Guangzhou. He examined some related historical materials and the translation of Khanfou in the Chinese translation of his work 《中国南海古代交通丛考》 (literally A Research on Ancient Trade History in the South China Sea). But Jitsuzo Kuwabara provided new perspectives and materials to further study this topic and held that Khanfou refers to Hangzhou.
Apart from Fujita Toyohachi, Chinese scholar Liang Qichao also regarded Khānfou as Guangfu, another name for Guangzhou. He wrote that “Cantonese people still call Guangzhou Guangfu because they know that Canfu is the transliteration of Guangfu” in his work The Research Methods of Chinese History 《中国历史研究法》 (Liang, 1998). Canfu and Khānfou are different in spelling, but they have something in common according to records, thus Canfu is Guangzhou. In many works, the authors translated “Canfu” (Khānfou) as Guangfu or Guangzhou. And some even further mistranslated it as 墈府 or 坎府 (Kanfu).
In 1985, “the First Part of Chinese Article of Chinese Repository” was published and the title “Remarks concerning the situation of Canfu, formerly the chief resort of the Arabians and other foreign merchants in China” was translated as “关于‘广府’——古代阿拉伯和其他外国商人经常提到的一个中国城市——的所在地考证” (Li, Huang, Yan, & Cai, 1985). That is to say, the translator accepted the view that Canfu or Khānfou refers to Guangzhou, or Guangfu (广府). In List of Articles and Subject Index of Chinese Repository, the same article was translated as 关于“墈府: ‘墈府’ (Canfu)——古代阿拉伯和其他外国商人常提到的一个中国贸易城——的地点考证” (Gu & Yang, 2008). Canfu was mistranslated as 墈府 (Kanfu).
Li (2020) supports that Khānfū refers to Guangzhou based on his analysis of historical materials about Banshoa, Khānfū, Ganpu, and Guangzhou. Banshoa is the mistranslation of Pang Xun, but the author still holds that Banshoa is the transliteration of Huang Chao, which led to his unreliable conclusion.
In Ancient Accounts of India and China, the port Khānfou or Canfu praised by the Arabs refers to Kánpú in Haiyan County, Zhejiang Province but not Guangzhou or Guangfu. Kánpú is adjacent to Chápú and is 25 miles away from Hangzhou. In the 9th century AD, Kánpú (Ganpu) became a port for foreign trade and its name has remained till now since the establishment of Maritime Trade Supervisorates (Shibosi 市舶司) in 1277. At the end of the Qing Dynasty, the role of Ganpu in foreign trade was surpassed by Zhapu 乍浦 (Morrison, 1844).
Based on the notes in different translations of Ancient Accounts of India and China and related materials, this article examines the reason why Khānfou was mistranslated as Guangzhou and takes a closer look at the rise and decline of trade ports along the South China Sea in the Tang Dynasty. The article analyzes the problem from three perspectives: the location of the “Gates of China” from which the Arabs entered China; the Bonshoa Rebellion, the Pang Xun Rebellion, and the Huang Chao Rebellion; the mistranslations of Khānfou as Kanpoo, Kwangfon, Kwangchow fon, Canfu, and Canton.
2. Disputations about the Location of Khānfou: Hangzhou or Guangzhou
First, at the beginning of Arab’s account of China, Khānfou (Ganpu/Kánpú) and the “Pang Xun Rebellion” (the Banshoa Rebellion) are mistaken for Guangfu (Canton/Guangzhou) and the “Huang Chao Rebellion” respectively, leading to disputation about whether Khānfou is Guangfu (Guangzhou) or Hangzhou (Ganpu).
In the Middle Ages, Mohammedan traveler Abu Abdullah Mahomed Ibn Batuta’s work includes much untrue information about China. Fang (2008) said that “according to Batuta’s account, it seems that he hadn’t been to Annan, modern-day Vietnam and northern China; his account of China was disjointed, even saying that a great river had its source near Beijing, flowing through Hangzhou and Quanzhou, and entering into the ocean in Guangzhou. Since value and mistakes coexist in this book, it is necessary to examine the book before quoting”. Batuta made a mistake in recording the flow direction of the Jing–Hang Grand Canal and the Pearl River, which shows that not all the Arab’s account on China is reliable.
Ibn Khoredādbeh (Abu’l-kasimubaid-Allah, 820/825-913) wrote the Book of Roads and Kingdoms (The Book of Roads and Provinces) to record the name of courier stations in different places and the distance between them when he was Director of Posts and Police for the Abbasid province of Jibal in modern-day Iran. Four courier stations in China, namely Loukin, Khanfou, Djanfou, and Kantou, the distance between them and their local customs and products were introduced in the book (Fang, 2008). Loukin, Khanfou, Djanfou, and Kantou are today’s Hanoi, Guangzhou, Quanzhou, and Yangzhou respectively (Fang, 2008). Khanfou and Kantou are similar in spelling and can be easily confused. Due to the similarity in spelling and pronunciation, it is possible that Arab authors or travelers regarded Canfu as Kánpú and identified Khānfou/Canfu as Guangzhou (Canton).
According to Kuwabara (1929), Khanfou (Khanfu) first appeared in the work Ancient Accounts of India and China and Ibn Khoredādbeh’s work on geography. Some regard it as Guangzhou and some as Hangzhou. There are so many viewpoints but none of them is convincing for all. He summarized some major ideas on this topic. Renaudot first translated Khānfou as Guangzhou and De Guignes agreed with him. Klaproth said that Khānfou refers to Gampou (Ganpu) in Marco Polo’s travels. Ganpu is located near the mouth of the Qiantang River and is about 120 or 130 li (1 li = 500 meters). To some extent, Hangzhou can represent Ganpu. Later, Yule, Richtofen, and Naka Michiyo all agreed with Klaproth. And the doctrine “Khanfou is Guangzhou” seemed not as convincing as before (Kuwabara, 1929). Later, Hield (1899…), Ishibashi Goro (1901…), Doctor Tsuboi (Shigaku Zasshi, 214-215), Pelliot (1904…), Doctor Fujita (1916…) all supported that Khānfou refers to Guangzhou rather than Hangzhou because Guangzhou was the greatest trade port for foreign business in southern China in the Tang Dynasty (618-907 AD). According to the Old Book of Tang and the New book of Tang, two classical historical works about the Tang Dynasty, Ganpu and Hangzhou were not important trade ports at that time because they first served as trade ports in the Song Dynasty (960-1279 AD). Therefore, Khanfou must be Guangzhou (Kuwabara, 1929). But the evidence provided by them is not strong enough to form a definite conclusion. Besides, they could not refute the argument made by Klaproth. His argument is based on analyzing the location of Khanfou from the time of its fall. analyzing the location of Khanfou from the time of its fall (Kuwabara, 2015). From Jitsuzo Kuwabara’s description of this topic, it can be referred that he is more likely to agree with the view that Khanfou refers to Hangzhou (Ganpu).
The translator of Somatsu no Teikyoshihaku Saiiki-jin Ho Juko no Jiseki Chen Yujing add explanatory notes in the translation, saying that the interpretation of the statement “Banshoa captured Khanfou in 878 AD” in Ancient Accounts of India and China is the key to solve the problem. The time of Khanfou’s fall during the Huang Chao Rebellion was 877 AD while Old Book of Tang, New Book of Tang, and Zizhi Tongjian say that Huang Chao captured Guangzhou in 878 AD. “If Khānfou refers to Guangzhou, its time of fall is inconsistent with the account in Ancient Accounts of India and China. But the fall of Hangzhou was a year earlier than Guangzhou, matching Hassan’s account. It is regarded as strong evidence by those who regard Khanfou as Hangzhou. Those who hold that Khanfou refers to Guangzhou guess that in Arab’s account 878 AD should be 879 AD, and thus Khanfou refers to Guangzhou. Both of the two parties could not convince each other” (Kuwabara, 1929). Hereditary House of Southern Han (chapter 65) in Historical Records of the Five Dynasties and Biography of Luxie in the Old Book of the Tang say that Guangzhou fell in 878 AD; 120,000 Muslims, Jews, Christians, and Zoroastrians were killed after Huang Chao’s capture of Khanfou. In the Tang Dynasty, only Guangzhou housed so many foreigners. And according to Ancient Accounts of India and China, residents in Khānfou built houses with wood or rattan which frequently caused the fire.
Jitsuzo Kuwabara listed three arguments to show that Khānfou refers to Hangzhou. Firstly, Marco Polo who came to China in the Yuan Dynasty said that Gampou, a port near Hangzhou (Quinsai), is a hub of foreign merchant ships and Gampou was the transliteration of Kan-phou located near the mouth of the Qiantang River. Secondly, Abulfeda recorded that Khansâ, the name of Hangzhou in the Southern Song Dynasty was Khanqou, the misspelling of Khanfou. Thirdly, according to Chinese historical materials, Hangzhou fell in 878 AD and Guangzhou fell in 879 AD during the Huang Chao Rebellion. Khanfou must be Hangzhou according to the fall of its time.
To figure out whether Khanfou is Guangzhou or Hangzhou, it is important to understand the relation between Khanfou and Guangfu and the time when Guangzhou was set as 州府 zhou and fu (both traditional administrative divisions). Jitsuzo Kuwabara thought that Ishibashi Goro was the first to make the point that Khanfou is the transliteration of Guangfu (Guangzhou), with his article published in Shigaku Zasshi in 1901. Three years later, Pelliot wrote an article to express the same idea (Kuwabara, 1929). Jitsuzo Kuwabara summarizes that the Old Book of Tang mentioned Guangzhou and used Guangfu to refer to it in related accounts. Besides, volume III of Tang Liudian 《唐六典》, Buddhist Pilgrim Monks of Tang Dynasty by Yijing monk and Todaiwajo Toseiden (Eastern Expedition of the Great Tang Monk) by OMI no Mifune include the expression Guangfu 广府.
In his work Somatsu no Teikyoshihaku Saiiki-jin Ho Juko no Jiseki (A Foreign Superintendent of Shipping Trade in the Late Song Dynasty, Traces of Pu Shougeng), Jitsuzō Kuwabara pointed out that Pelliot holds that Guangfu 广府 is the abbreviation of Prefecture de Kouang-cheou 广州府 (1904, Bulletin de l’École française d’Extrême-Orient, 205). Pelliot thought that the official title of Commanders-in-chief or military commissioner can be called Dafu 大府 or Huifu 会府 and the prefectural governor under their leadership can be called Zhoufu 州府 or Fu 府. Therefore, Zhou 州 can be called Fu 府 no matter there is an Area Command (Dudu Fu 都督府) or not (Kuwabara, 1929). Pelliot’s statement is not convincing enough because the combination of Zhou and Fu as an administrative region like Guangzhou Fu and Hangzhou Fu appeared after the Ming Dynasty. Thus, Guangfu couldn’t be the abbreviation of Guangzhou Fu. According to that Liu Julin was the commander-in-chief in Guangzhou Area Command in 741 AD, Doctor Ishibashi’s view is more acceptable since he considered Guangfu to be the abbreviation of Guangzhou Dudufu (Guangzhou Area Command) (Kuwabara, 1929).
Jitsuzo Kuwabara says that “In the Tang Dynasty, Guangzhou is the place of Dudu Fu and Jiedushi Fu and thus Guangzhou was called Guangfu then. But Dudu Fu and Jiedushi Fu are different” (Kuwabara, 2015). In the Ancient Accounts of India and China, the time the Arabs arrived in Khanfou was the Tang Dynasty when no Zhoufu was set in Guangzhou. The name Guangfu was formed after the Ming Dynasty. How the name Guangfu was formed should be figured out to reveal the relation between Guangfu and Khanfou.
3. A Study on Misreading of Khānfou and the Reasons
To conclude the disputation about the location of Khānfou, three aspects are helping us to research systematically. First, the Arabs first arrived in Ganpu rather than Guangzhou when they came to China. Second, Arab geographers and historians misread Khānfou and later generations were misled by their mistakes. Third, they also misread the Pangxun Rebellion as the Huangchao Rebellion happened at the end of the Tang Dynasty.
From 1848 to 1883, Joseph Toussaint Reinaud and S. Guyard translated Arab Abulfeda’s Taqwim al-Buldan (A Sketch of the Countries) into French. The book includes information on the Yuan Dynasty of China which is quite refreshing and the author himself could not help saying that he knew little about the Orient (Fang, 2008). In his work, he mixed up Canton, Khanfu, Khansã, and Ganpu. “The author wrote that there was Sikhu (the West Lake) in Hangzhou. Previous Arab travelers had not been to some places near there. But he mistakenly thought that West Lake was located in the north of Hangzhou. He regarded Quanzhou (Shanju, Shinju in the past or today’s Zaitun) and Guangzhou (Kwangchou foo) as the two largest ports in China. He appeared to know Sila, but took it for an island in the East China Sea” (Fang, 2008). It shows that Arab geography was in decline at the end of the Middles Ages but the unverified accounts of China were regarded as reliable sources of history by later generations.
3.1. Ganpu, “the Gates of China” from Which the Arabs Entered China by Sea
Robert Morrison’s (1782-1834) A View of China, for Philological Purposes recorded that “During the reign of Emperor Huan of Han (146-168 AD), ancient India, Daqin (Egypt or Arab) and other countries sent tribute to China through translation. And since then, merchants and foreigners gathered in Yangzhou and Guangdong” (Morrison, 1817). At the end of the Tang Dynasty, Arabs entered China by sea through the “Gates of China” which was exactly Khānfou, a place recorded in Ancient Accounts of India and China by Arab authors and travelers.
In 1685, the King of France Louis XIV sent Jesuit Louis Le Comte (1655-1728), a royal mathematician, and five other Jesuits to China. The six arrived in Siam and parted ways, then Father Le Comte and four other Jesuits gathered and arrived in Ningbo “by a small Chinese vessel” on June 17, 1687 (Bridgman, 1832). In the 1830s, according to the description of Louis Le Comte and his party docking at a Chinese port in Nouveaux mémoires sur l’état présent de la Chine, Memoirs and Remarks on China, Bridgman (1832) suggested that Khānfou was not Guangzhou (Canton) as rumored. In his view, Louis Le Comte’s account of the Chinese coast was similar to that described 144 years later in The Chinese Repository: “the rude junk, the narrow cabin, disorder among the sailors, idolatry”. “There is a notable difference, however, in two particulars: in the first case, no opium scenes were exhibited; and in the latter, no attempts were made to work miracles. Le Comte speaks of the ‘Typhon, than which nothing is more terrible in the seas of China and Japan’, and also of the frightful appearance of an ‘infinite number of rocks and desert islands, through which they were obliged to pass’; and of channels ‘so narrow, as not to exceed ten paces in breadth’, and of a ‘pretty wide bay, in which the Chinese observe a profound silence, for fear of disturbing a neighboring dragon’”. But Bridgman couldn’t find out the accurate places where Le Comte docked since the author didn’t record that in the book. Karl Friedrich August Gützlaff (1803-1851) was certain that the place where Le Comte and his partners docked was not Canton, but what the Arab travelers called the “Gates of China”. In this case, Khānfou in the Arab traveler’s accounts is not Guangzhou but Ganpu located between the coast of a narrow entrance between Shuiniu Island 水牛岛 and Zhitou 峙头 in Hangzhou. And Ganpu was replaced by Zhapu located at the same site.
In Vol. I, No. 2 “Review” section of The Chinese Repository published in June 1832, an introduction was given to Ta Tsing Wan-neën Yih-Tung King-wei Yu-too (A general geographical map, with degree of latitude and longitude, of the Empire of the Ta Tsing Dynasty, may it last forever) drawn by Le Mingchě (Le Tsinglae), a priest of the Taou sect and a native of Canton. The review pointed out that “The narrow strait which separates Hainan from the mainland is, probably, the place called by the Mohammedans of the 8th and 9th centuries, ‘the Gates of China’” (Le, 1832). Williams (1913) wrote that “South-west from Chapu lies the old town of Canfu (called Kanpu by the Chinese), which was once the port of Hangchau, but now deserted since the stream on which it is situated has become choked with sand. This place is mentioned in the voyages of two Arabian travelers in the ninth century, as the chief port of China, where all shipping centered. The narrow entrance between Buffalo Island and Kitto Point is probably the Gates of China mentioned by them”.
Morrison (1844) wrote that “Chápú (located in Kiáhing fú, Chekiáng) is a seaport town of note in Pinghú district on the north side of Hangchau bay”. He also quoted the account given of Canfu by the Mohammedan travelers. “Canfu is the port for all the ships and goods of the Arabs who trade in China. When a ship has got through the Gates of China, she, with a tide of flood goes into a freshwater gulf, and drops anchor in the chief port of China, which is that of Canfu; and here they have fresh water both from springs and rivers, as they have also in most of the other ports of China. The city is adorned with large squares, and supplied with all the necessities of defense against an enemy, and in most of the other provinces there are cities of strength fortified in the same manner” (Anonymous Author, 1834).
In the 1830s and 1840s, Kanpoo was surrounded by circumvallation, with its location a little different from its former address. It was famous for its salt industry, being one of the most prosperous cities on the east coast of Zhengjiang. After the waters receded, a large area of land was revealed here. And Chapoo 乍浦, a few miles to the east, was already the center of trade and commerce with Japan. After studying the location of Kanpoo, John Robert Morrison quoted Gützlaff’s description of Arabs’ arrival in Chapoo. “On the 8th, we steered for Chapoo, the emporium of the Japan trade. None of us had ever been there, nor were we in the least acquainted with the situation of the harbor. After having rounded the first bold head-land, a large trading place gradually opened, and we perceived a great number of junks at anchor. We could no longer be ignorant of the place of our destination… To prevent all trouble, we resolved not to go on shore, and strictly to abstain from all intercourse with the authorities, Chapoo, however, looked too invitingly. The city itself is built at the bottom of a bay. The anchorage is shallow, and the junks lie high and dry at low water. There are many fine shops in the suburb, but the streets are narrow and crowded. The principal part of the city is surrounded by a massive wall, which is now tumbling down, and has considerable breaks” (Anonymous Author, 1834). Besides, The Canton Register and “Journal of a voyage along the coast of China from the province of Canton to Leaoutung in Mantchou Tartary, 1832-1833” by Gützlaff (1834) also contain the same account of Arabs’ arrival in Chapoo.
As early as 300 AD, Kanpoo was a port for anchoring and mooring ships. In 717 AD, the imperial court of the Tang Dynasty established Kanpoo Town. In 720 AD, a position for the management of maritime affairs was established. From the 7th century to the 9th century, the imperial court established a business management position in Kanpoo.
Ganshui Choreography 《澉水志》 by Song Changtang says that Ganpu was set as Maritime Trade Supervisorate in 1246. Zuili Choreography 《檇李记》 by Wang Qiao says that Ganpu locates to the west of Haiyan and served as a place for foreign ships to anchor during the Song and the Yuan dynasties. Zhapu was adjacent to Ganpu, and both were foreign trade ports. Zhapu is part of Pinghu, with foreign ships gathered here in the Yuan Dynasty. But no Maritime Trade Supervisorate or Changwu 场务 (an organization supervising the production of iron and salt and taxation in ancient China) was set here (Fang, 2008).
During the Song and Yuan dynasties, Yangfa, Yangzi, Yangshu, and Yangying, members of the same family, stayed at the port of Ganpu, where they built a checkpoint to check merchants arriving at Ganpu by sea and the business they ran. In the Song Dynasty, the imperial court of Song established business relations with 50 or 60 countries, enabling foreign merchants to do business in China. In the Yuan Dynasty, Emperor Shizu of Yuan Kublai resumed trade activities shortly after the establishment of his reign.
At the beginning of the Yuan Dynasty, Marco Polo came to China and stayed in Hangzhou for a long time. His residence became a place for government officials to stay called Zhanchi 站赤 (courier station). In his travels, Marco Polo called Hangzhou as people in the Southern Song Dynasty did. Hangzhou was called Xingzai 行在 or Jingshi 京师 (Khinzai)2. In different versions of The Travels of Marco Polo, Xingzai or Jingshi was called Quinsai, Quiensay, Quisay, Chisai, Chesai, Cansay, Campsai, Chansay, Khanzai, and Khanza. Marco Polo said that Ganpu was a great port near Hangzhou, with merchant ships loaded with goods anchoring here (Bridgman, 1832). The river that flows through Hangzhou forms a harbor here, reaching out to the sea. Ships were carrying goods to be loaded or unloaded and exports were loaded on ships to be shipped to ports in India.
In 1277, the imperial court of the Yuan Dynasty set up the Maritime Trade Supervisorate to revive its trade connections with countries in Southeast Asia. Ganpu was one of the four oldest Maritime Trade Supervisorates established in the Yuan Dynasty. Volume 94 of the History of Yuan 《元史》 says that “in 1277, Maritime Trade Supervisorate was set in Quanzhou, with Mang Gudai 忙古䚟 being the supervisor; three Maritime Trade Supervisorates were set in Qingyuan, Shanghai and Ganpu respectively, supervised by Yangfa, Pacification Commissione 安抚使 of Fujian Province” (Song & Wang, 1976). Ten years later, Maritime Trade Supervisorates were set in Hangzhou, Guangdong, and Wenzhou. In 1293, “there are seven Maritime Trade Supervisorates in total. They are Quanzhou, Shanghai, Ganpu, Wenzhou, Guangdong, Hangzhou, and Qingyuan” (Song & Wang, 1976).
In 1298, the Maritime Trade Supervisorate in Ganpu ceased trade. Shihuozhi 食货志 (literally treaties on foods and goods) in The History of Yuan says that “in 1298, Ganpu and Shanghai were incorporated into Qingyuan to run one Maritime Trade Supervisorate, being led by the Zhongshu Sheng 中书省 (the Palace Secretariat)” (Song & Wang, 1976). Since then, the arrangement of Maritime Trade Supervisorates has undergone some changes. By 1322, Quanzhou, Qingyuan, and Guangzhou were set as Maritime Trade Supervisorate again (Fang, 2008). Yao Tongshou was a native of Ganpu in the Yuan Dynasty and his work Yue Jiao Si Yu 《乐郊私语》 also records the changes in the Maritime Trade Supervisorate in Ganpu.
3.2. Misreading of Kankhou or Canfu by Arab Geographers and Historians
The changes in Guangzhou’s name and Maritime Trade Supervisorate can be evidence of Arab geographers’ and historians’ misreading of Kankhou in Ancient Accounts of India and China. Due to the differences between Chinese and Arab pronunciation, Kankhou was misread as Kanpoo/Kwangfon/Kwangchow fon 广州府 or Canfu 广府. To figure out the misreading of Kankhou helps determine whether Khanfou is Guangzhou (Guangfu) or Hangzhou (Ganpu).
The development of Guangzhou can be dated back to the reign of King Nan of Zhou, king of Zhou from 314 to 256 BC, more than 2000 years ago. It was called Nanwucheng (literally Southern military city) at that time. On ancient maps of Guangzhou, the city was identified as “Kwangtung sǎngching” (literally the urban area in Guangdong Province). In ancient Chinese records, Guangzhou was called Keaouche, Keaouchow, Lingnan, Kwangchow, Nanhae, Nanyuě, Pwayu, Pihyuě, Yuě, and Yuětung. The latter names are usually used in classical writings and official documents (Anonymous Author, 1833). Other names like Yangcheng, Xiancheng, and Suicheng come from legends.
The name “Guangzhou” first appeared in 210 BC and covered a wider area than it does today. During the reign of Emperor Wu of Liang (502-549) in the Southern Qi Dynasty, Guangzhou split up and part of it became Guizhou 桂州, the present-day capital of Guangxi Province, Guilin. Since then, the name “Liangguang” (literally two Guangs) appeared, meaning “the vast eastern province” Guangdong and “the vast western province” Guangxi. The name “Guangdong Province” first appeared in the Song Dynasty in 1150 AD. At the end of the Song Dynasty, Guangdong was called Guangdong Lu 广东路 and was called Guangdong Dao 广东道 in the Yuan Dynasty. During the reign of the Hongwu Emperor (1368-1398) in the Ming Dynasty, the name “Guangdong Province” was first used and is still used today. “It was at about 1368 AD that Kwangchow, the principal district of the province, was first called a foo; previously it had been usually called Kwangchow loo” (Anonymous Author, 1833). According to Ancient Accounts of India and China, Arab travelers arrived in China long before the name Kwangchow foo was first used. Therefore, Arab travelers can’t use Canfu to refer to Kwangchow (Guangzhou).
John Robert Morrison wrote an article to examine why Canfu was misread as Guangzhou. In Abu Zaid’s account, Canfu was Kwangchow foo which was called Canton by Europeans in the middle of the 18th century (Morrison, 1832). “Canfu is the port for all the ships and goods of the Arabs who trade in China… They say that in the kingdom of China there are above two hundred cities that have jurisdiction over several others, and have each a prince or governor, and a eunuch or lieutenant. Canfu is one of these cities, being the port for all shipping and presiding over twenty towns” (Anonymous Author, 1834).
John Robert Morrison’s account was consistent with the feature of several ports in China at that time. All ports along the South China Sea were near great navigable rivers due to the tide (Fujita, 2015). Kwangchow, Fuhchow, and Hangchow are big cities with great rivers flowing through them. The rivers widen at the mouth and bays are formed. Therefore, only when it is at high tide can these cities be freshwater ports. The pronunciation of these cities is different from that of Khānfou. In Arabic, kh (ق) differs from f (ف) only by the addition of a single dot to differentiate between these two sounds. “Chow” in Kwangchow fon is omitted and what is left is Kwangfon which is similar to Canfu in pronunciation. Khānfou and Canfu sounds alike, misleading people to take Canfu for Cwangchow fon. But Canfu was not what people called Guangzhou in the Tang Dynasty or any time before Tang. No existing evidence or discussion can prove that Canfu is Guangzhou.
The place called Kanpoo, about 30 miles west of Hangzhou, is Kánpú. In Arabic, there is no /p/ sound, but /f/ sound is similar to /p/. Due to differences in pronunciation, Kánpú is read as Kanfoo or Canfu in Arabic. The translator of Somatsu no Teikyoshihaku Saiiki-jin Ho Juko no Jiseki by Jitsuzo Kuwabara, Chen Yujing cites Ishibashi Goro’s statement about Loukin as an annotation. Ishibashi Goro holds that “In Arabic, there is no /p/ sound but /f/ sound. And /f/sounds like /p/. Besides, F (·g) and K (:g) are similar to each other in spelling, causing confusion. As a result, Longbian 龙编 was originally spelled as Loupin or Lung-pien and was changed to Loukin by the Arabs” (Kuwabara, 1929). Chen Yujing strongly agreed with Ishibashi Goro. Jitsuzo Kuwabara also explained the change of pronunciation in his work A Study on Trade Ports in Tang and Song Dynasties. He said that “Gampou in Marco Polo’s travels is the transliteration of Ganpu (Kan-phou), a place near the Qiantang Estuary. In the Tang Dynasty, K and G sounded the same. Therefore, Kan-phou and Ganpu were pronounced in very similar ways. The Arabs did not have a /p/ sound in their language and had /f/ to replace it. And Kan-phou was changed to Khanfou by them. In conclusion, Khanfou must be Ganpu rather than Guangzhou.
Once a port, Ganpu was filled with sand afterward. Hangzhou was located farther from the mouth of the Yangtze River and was inaccessible to ships because of the blocked channel. For this reason, Ganpu became an anchorage and perhaps a gathering place for Arab traders who carried imported goods and received exports from here. Due to linguistic factors, Kanpoo was regarded as another city near it. However, Kanpoo could not be a small port or Guangzhou which was not a prosperous city for barbarians to live in at that time. Later, Europeans mistook Kanpoo for Guangzhou which became a large port over time. And it is reasonable to infer that Kanpoo was prosperous and great as recorded by Arab authors and travelers.
3.3. The Misreading of “Pang Xun Rebellion” as “Huang Chao Rebellion” in the Late Tang Dynasty
To figure out whether Kanfhou is Guangzhou (Canton) or Hangzhou (Ganpu), it is important to know whether the “Banshoa Rebellion” refers to “Pang Xun Rebellion” or “Huang Chao Rebellion” and the year of the fall of Kānkhou. Some Arab’s inaccurate accounts make it difficult to figure things out. “Famous historian and geographer Abu-I-Hassan ali-el-Mas’udi wrote about China and mentioned Huang Chao Rebellion in his work The Meadows of Gold and Mines of Gems” (Fang, 2008). But the “Huang Chao Rebellion” in Masudi’s work should be the “Pang Xun Rebellion”. Available information shows that it is Huang Chao Rebellion that interrupted Arabs’ connection with Guangzhou or Hangzhou.
Huang Chao Rebellion serves as an argument for the view that Khānfou refers to Hangzhou (Kuwabara, 1929). When Huang Chao captured Khanfou, many foreigners were killed. It was recorded by Arabs. Ancient Accounts of India and China says that Banshoa captured Khanfou in 878 AD while both the Old Book of Tang, New Book of Tang, and Zizhi Tongjian say that Guangzhou was captured in 879 AD. Hangzhou fell in 878 AD. If Khānfou refers to Guangzhou, there must be something wrong with Arab and Chinese accounts of this rebellion. Those who hold that Khanfou refers to Guangzhou guess that in Arab’s account 878 AD should be 879 AD, and thus Khanfou refers to Guangzhou. Both the two parties could not convince each other. There is a discussion about rebellions in the late Tang Dynasty in Shigaku Zasshi, saying that the accounts in historical records contradict. Though the New Book of Tang and Zizhi Tongjian say that Huang Chao captured Khanfou in 879 AD, the Hereditary House of Southern Han (chapter 65) in Historical Records of the Five Dynasties records that Khanfou fell in 878 AD. It seems that Arab’s account is more convincing than that of Old Book of Tang, New Book of Tang, and Zizhi Tongjian. If the above argument is reliable, then Khanfou is more likely to be Guangzhou. What’s more, 200,000 Muslims, Jews, Christians, and Zoroastrians were killed in the Huang Chao Rebellion. No other city in the Tang Dynasty housed so many foreigners as Guangzhou. Hangzhou was underdeveloped then and thus was impossible to house so many foreigners. Ancient Accounts of India and China also says that residents in Khānfou built houses with wood and rattan which frequently spark the fire. According to the New Book of Tang and Old Book of Tang, Song Jing, an official in Guangzhou at that time, advised residents to build houses with tile roofs to stop the fire from spreading because they covered roofs with bamboo or straw. All the residents felt grateful to Song Jing. The account in this book is consistent with that in Ancient Accounts of India and China, supposing that Khanfou refers to Guangzhou.
The fact about Khanfou’s fall is an important factor to figure out its location. Jitsuzo Kuwabara compared some historical materials in the Old Book of Tang, the New Book of Tang, Zizhi Tongjian, and Historical Records of the Five Dynasties in terms of their records of the time and facts of Guangzhou and Hangzhou during the Huang Chao Rebellion. He found that only the biographic sketch of Emperor Xizong of Tang in the Old Book of Tang says that Hangzhou fell during Huang Chao Rebellion. And the records of the year of Guangzhou’s fall contradict. The differences among these records lead to an inconclusive discussion about the location of Khanfou. However, the point is that the Banshoa Rebellion in Ancient Accounts of India and China refers to Pang Xun Rebellion rather than Huang Chao Rebellion. Banshoa is the transliteration of Pang Xun rather than Huang Chao. The time of Hangzhou’s fall during the rebellion is a better fit for historical records than that of Guangzhou. Mulberry orchards were destroyed in Khānfou during the rebellion, leading to the decline of the silk industry.
Fang holds that the Bonshoa Rebellion in Ancient Accounts of India and China should be Pang Xun Rebellion. “Klaproth was the first to interpret Banshoa as Huang Chao and this view became popular over time. The suppression of the Pang Xun Rebellion did not get help from other countries. But Hassan recorded that Taghazghaz helped suppress the rebellion. It is highly likely that Hassan mixed up the Pang Xun Rebellion and the An Lushan Rebellion. As for Hassan’s record of the year of Guangzhou’s fall, it was a year earlier than it actually happened. It is just a common mistake” (Fang, 2008). When talking about Nestorianism in China, Fang Hao says again that Bonshoa Rebellion refers to the Pang Xun Rebellion because he thinks that Pang Xun killed 120,000 Muslims, Jews, Christians, and Zoroastrians in Khanfou (Fang, 2008).
Apart from Huang Chao, some scholars think that Banshoa refers to Banxue 班雪. Examples are Huang Zhuohan’s translation of Ancient Accounts of India and China and Hu Yaofei’s translation of Arabic Historical materials about the Huang Chao Rebellion 《关于黄巢起义的阿拉伯文史料译注》. Both of them translated Banshoa as Banxue. And on this basis, Li (2020) concludes that Banxue refers to Huang Chao and the Pang Xun Rebellion is the same as the Huang Chao Rebellion.
Jitsuzo Kuwabara cited the biographic sketches of Emperor Xizong of Tang in the Old Book of Tang which says that “residents in zhou and fu in southeastern China suffered losses in mulberry industry and farming and couldn’t farm on time after the towns were invaded. He pointed out that after Khanfou’s fall, many foreign residents in the city left. The city was in disorder, agriculture and mulberry industry were damaged, commerce was depressed, and there were no foreign merchant ships. These events were so considerable that Arab authors and merchants couldn’t make mistakes when recording the time of these events.
In Arabs’ record, “the Chinese carefully cultivate the mulberry for the sake of its leaf, wherewith they subsist and propagate their silkworms. This devastation is the cause why silk has failed, and that the trade which used to be driven with it, in the countries under the Arabs, has stagnated” (Anonymous Author, 1834). The rebellion that caused the devastation of mulberry trees took place in 877 AD. Related account of this rebellion in The Chinese Repository perfectly matches that of The Chinese Annals, that is, the rioters besieged Kanpoo, plundering, killing, and destroying all mulberry trees.
Guangzhou was never famous for its silk industry; thus, destroying mulberry trees and damaging the silk industry was not likely to inflict heavy losses on Guangzhou’s foreign trade. But according to related records, mulberry trees and the silk industry mean a lot to the city’s economy and prosperity. And Hangzhou was famous for its silk industry. Therefore, it is really possible that Banshoa refers to Pang Xun. Guangzhou became a wealthy city thanks to the prosperous commerce 30 years after the Pang Xun Rebellion. The merchants suffered from the turmoil in Hangzhou and moved to Guangzhou to continue their business. Hangzhou declined after 877 AD and was rarely talked about afterward. Based on the above information, Khānfou is possible to be located somewhere adjacent to Hangzhou.
Samuel Wells Williams cited a part from The Chinese Repository in his book The Middle Kingdom. “Much of the statement made by Abu Zaid respecting the wealth, extent, and splendor of Canfu really refers to the city of Hangchau. The bore in the Tsientang River makes it impossible for ships to lie off that place, and this had its effect in developing Kanpu. The destruction of the capital in 877 contributed to direct part of the trade to Canton, which even then and long after was comparatively a small place, and the people of that part of the country but little removed from gross barbarism. In Marco Polo’s time, Ganpu was frequented by all the ships that bring merchandise from India” (Williams, 1913). And Williams concluded that Canfu was Ganpu.
4. The Historical Value of the Examination of the Mistranslation of Khānfou
From the 6th or 7th century AD to the end of the Middle Ages, the Arabs controlled the world’s sea and land transportation under their geographic advantages. The same religious and linguistic roots served as a link between the Arabs and other caliphate countries from India to the Atlantic Ocean. Some learned Arab scholars first recorded the arrival of Persian and Arab merchants and travelers in China during the Tang and the Song dynasties, many of whom had traveled to China themselves while some of them had not. As a result, there is some unreliable information in Arabs’ account of exchanges between Chinese and Arab in cities along the South China Sea coast and of the Maritime Trade Supervisorate. And these unreliable materials lead to an inconclusive discussion on some topics like the location of Khānfou. “So many things have happened in the late Tang Dynasty. Most of them are associated with each other and there are always some new questions when trying to understand the facts. In this case, it is difficult to figure things out thoroughly based on these materials” (Kuwabara, 2015).
Ancient Accounts of India and China was the first Arab book about things in China. It was published in many languages, and errors are common in transcription and translation when it comes to placename, personal name, or loanwords. Since the 18th century, whether Khānfou is Guangzhou or Hangzhou remained an unresolved question. Searching for new materials to help answer this question and figuring out why Khānfou was misread will provide a reference for further study on this topic and similar topics.
4.1. Providing Key Evidence to Give a Definite Answer to the Location of Khānfou
For a long time, historians and geographers focus on whether Banshoa refers to Huang Chao or Pang Xun and the time and details of Khānfou’s fall. The stringency of arguments provided by those who regarded Khānfou as Guangzhou or Hangzhou is also evaluated. But these efforts failed to end the debate.
In Ancient Accounts of India and China, the gate of China from which Arab travelers and traders entered China by sea at the end of the Tang Dynasty, is actually referred to Ganpu, not Guangzhou. Misspelling and unreliable records of Kankhou by Arab authors led later scholars to misinterpret Kankhou/Canfu as Guangzhou. Banshoa’s transliteration as Banxue in Chinese and the unreliable part of Ganpu’s history misled related researchers to mix up the Pang Xun Rebellion and the Huang Chao Rebellion. These are new aspects to analyze the location of Khānfou.
4.2. Helping Figureout Mistranslation of the Name of Trade Ports in the Tang and Song Dynasties
The debate about Khānfou’s location is mainly caused by mistranslation. In Ancient Accounts of India and China and other works about trade ports in ancient China by Arab authors, there are many mistranslated placenames. Here are some examples. Ibn Khoredādbeh’s Book of Roads and Kingdoms records four courier stations in China Longbian, Guangfu, Quanfu, and Jiangdu were transliterated as Loukin, Khanfou, Djanfou, and Kantou respectively. The spelling and translation of these placenames were distorted and Jitsuzo Kuwabara tried to figure out this problem and found out the real location of these places in the first chapter of his book Somatsu no Teikyoshihaku Saiiki-jin Ho Juko no Jiseki (Kuwabara, 1929). Phillips (1874, 1875, 1876, 1877) wrote a series of writings “Zaitun Researches” in The Chinese Recorder and Missionary Journal to study Zaitun and its location. Fujita Toyohachi also examined the name and location of Quanzhou and the Arab residents living here in his work A Research on Ancient Trade History in the South China Sea.
In Arabic, kh (ق) differs from f (ف) only by the addition of a single dot to differentiate between these two sounds. Besides, there is no /p/ sound but a /f/ sound which is similar to /p/ in Arabic. Therefore, Khānfou was misspelled as Kanpoo, Kwangfon, Kwangchow fon, and Canfu. From the perspective of pronunciation, it is reasonable that Kánpú (Ganpu) became Kanfoo or Canfu in Arabic.
Chen Yujing pointed out that Xingzai 行在 (today’s Hangzhou) is an example to study the change of pronunciation in the process of translation. Xingzai is spelled as Hang-tsai or Hing (Hsing)-tsai in Wade-Giles. In the Song Dynasty, whether the initial of 行 (Xing) was K, or H or kh has yet to be studied. In the Yuan Dynasty, the pronunciation of 行 was different in northern and southern China. In the northern part, the initial of 行 was not K or kh but h while in the southern part the initial of 行 was kh rather than h (Kuwabara, 1929). And Chen infers the changing pattern of pronunciation of some Chinese characters like han 汉, han 寒, han 汗, he 何, hu 呼, hu 胡, hu 湖, hu 虎, he 贺, he 鹤, hai 海, hei 黑 and so on; that is to say, the initial of these characters changed from K or G to kh and last to H. 行 has the same initial as the characters mentioned above. Therefore, it can be concluded that the pronunciation of 行 changed from Khang or Khing to Hang or Hing (Kuwabara, 1929).
In Arab and Persian works about the Tang Dynasty, a great port in China Zaitûn (also called Zaytun, Zayton, or Zeytoun) refers to Quanzhou, which was written as Djanfou in Ibn Khoredādbeh’s Book of Roads and Kingdoms. “At that time, Arabs and other foreigners called Quanzhou Zayton, Zaytun, Zeytoun, or other names with similar pronunciation. The name Zayton was from Erythrina variegate which was called 刺桐 Citong in Chinese. During the Five Dynasties and Ten Kingdoms (907-979 AD), Liu Congxiao had Quanzhou rebuilt and many Erythrina variegates were planted around the city. From then on, Quanzhou was called 刺桐城 Citong City or 桐城 Tong City. The Arabs thus translated the name of Quanzhou as Médnet Zeytoun (Médnet is the translation of city and Zeytoun is the transliteration of Citong). Those who only called Quanzhou Zeytoun omitted Médnet” (Kuwabara, 1929). There are five viewpoints on what Zaitûn exactly means: the transliteration of Citong桐城, the transliteration of Ruitong 瑞桐, the translation of olive tree, the transliteration of Haicheng, a city near the mouth of Zhangjiang, and the transcription of Caykong or Carchan (Kuwabara, 1929).
In the Tang and Song dynasties, many Arab travelers visited China and recorded things about China, including trade ports. But the mistakes in their account have driven later generations to do related research. And one of the topics is the transcription and translation of placenames, which can help correct mistranslation or errors in the transcription of the name of trade ports in the Tang and Song dynasties.
4.3. Serving as References for Studies on Ancient Trade Ports along the South China Sea
The systematic absorption of Greek and Indian knowledge from the middle of the 8th century AD onwards provided the Arabs with many advantages, helping them play a major role in the cultural exchanges between East and West and the Silk Road. Along the River Tigris, Islamic culture and the caliphate increasingly developed. Libraries and schools were built and talents were gathered here. They highly valued philosophy, astronomy, and medicine but neglected geography at that time (Morrison, 1832). Even though geography was not valued by the Arabs then, their accounts of trade ports along the South China Sea are still valuable and are important sources of historical materials for scholars to study related topics on these ports.
The historiographical value of Ancient Accounts of India and China is unparalleled by any other work. Finished four and a half centuries before The Travels of Marco Polo, the work is the oldest account of China by foreigners. Comparative studies on the translation and location of Khānfou will help find out more details about intercourse between China and the West during the Tang and Song dynasties and provide a reference for the study of the trade ports along South China Sea and the history of India and Southeast Asian countries.
NOTES
1Some hold that the anonymous Arab author is Sulayman/Suleìman.
2Khinzai: In the Middle Ages, western countries called Hangzhou Khinzai, Khansai or Khanzai. There are two views on the translation of Hangzhou. One is that Khinzai is the transliteration of Jingshi 京师 because the imperial court was set in Hangzhou in the Southern Song Dynasty and foreign merchants called Hangzhou Jingshi. Another view is that Khinzai is the transliteration of Xingzai 行在.