Islam in Champa and the making of factitious history
In 1922, a remarkable article appeared in the Journal Asiatique. Written by Paul Ravaisse, a professor at the École nationale des langues orientales vivantes in Paris, it described, translated, and discussed two Arabic inscriptions which had apparently been found somewhere near the coast of what, at the time, was French Annam (now part of Vietnam).Footnote 1 These inscriptions were both dated to the early eleventh century. They were of enormous significance, as they were the earliest Islamic inscriptions found anywhere in Southeast Asia. They quickly achieved considerable renown, especially because Georges Maspéro referred to them in his seminal work, Le royaume de Champa.Footnote 2 They have frequently been adduced as evidence of an Islamic presence in Champa during the eleventh century.Footnote 3 In fact, for some ninety years, the evidence of these inscriptions has been a principal foundation of the supposed history of Islam in Champa, and has had significant implications for the early history of Islam throughout Southeast Asia, as well as in southern China.
In 2003, however, Ludvik Kalus raised the question of whether these inscriptions really originated from Champa at all. One of them, the only one of the two to bear a date, he connected with Kairouan (Qayrawan) in Tunisia. Although he was unable to suggest an origin for the other, he cast doubt on the date proposed for it by Ravaisse, considering that it was probably significantly later.Footnote 4 There the matter rested for almost a decade. Then, in 2012, another article, by Abdeljaouad Lotfi, delivered the coup de grâce, as far as any connection between these inscriptions and Southeast Asia was concerned. Abdeljaouad confirmed that the first inscription originated from Kairouan, even identifying the workshop in which it had been carved. He also identified the person responsible for the second inscription, who lived in Kairouan during the first half of the nineteenth century.Footnote 5 It became clear, then, that both these inscriptions originated from Tunisia and had no connection whatsoever with Champa, or with anywhere else in Southeast Asia.
All this may seem astonishing. How could these two inscriptions have been identified as coming from Southeast Asia if, in reality, they originate from North Africa? In order to understand this apparent enigma, it is necessary to look carefully at the story of their discovery and the manner in which they came to the attention of Paul Ravaisse. His own account is as follows:
Les deux inscriptions arabes qui font l'objet de cette étude ont été découvertes, il y a quinze ou vingt ans, «sur un point non éloigné de la côte annamite, par un officier de la marine française», qui en prit sur place les précieux estampages – sans compter un croquis – dont on trouvera ici la reproduction. . . . A son retour d'Indochine, cet officier, dont on ne sais plus le nom, crut bon de remettre sa petite moisson épigraphique au savant indianiste A. Barth, lequel jugea meilleur, dès le premier coup d'oeil, de s'en dessaisir en faveur de son confrère de l'Institut H. Derenbourg . . . . . Mais H. Derenbourg . . . me passa, peu de temps avant sa mort, les trois feuilles vagabondes, en ne me donnant pour tout renseignement, sur ces inscriptions rarissimes, sur le lieu et la date de leur découverte et sur leur inventeur, que la très sommaire indication qu'on vient de lire. Il est probable qu'il n'en avait appris davantage de la bouche de Barth.Footnote 6
This bizarre story immediately raises the question of whether the alleged provenance of these inscriptions is credible. Who was the anonymous “officier de la marine française (French naval officer)?” Where exactly were the inscriptions found? Why did it take “fifteen or twenty years” for them to be published? Why did Derenbourg give them to Ravaisse? This narrative was soon questioned by Louis Finot. In a review of Maspéro's Le royaume de Champa, he asserted that:
Il serait bien hasardeux d'accorder une entière confiance à des documents d'une origine aussi vague. Loin de pouvoir affirmer qu'ils ont été trouvés « dans le sud du Champa », comme l’écrit Maspéro, on est en droit de se demander s'ils proviennent même de l'Annam.Footnote 7
This was a harsh judgment. Finot went so far as to suggest that it was entirely possible that the inscriptions had no connection at all with (French) Annam.Footnote 8 It must be stated immediately that he was well placed to make such a judgment. Born in 1864, he had originally studied palaeography, and had become deputy librarian at the French National Library (Bibliothèque Nationale). In order to be able to deal with Sanskrit documents, he took classes in the language with Sylvain Lévy, an eminent scholar of the period. This led him into a career in Indian studies. In 1898, he was appointed director of the Mission archéologique d'Indo-Chine, which became the École française d'Extrême-Orient (EFEO) in 1900. This was originally based in Saigon, but moved to Hanoi in 1902. In connection with this post, he travelled throughout French Indo-China and also visited Java. In 1901, he published an article on the religion of the Chams.Footnote 9 It was Finot who, in 1900, obtained the appointment of Paul Pelliot to the EFEO. In 1904, his term as director of the EFEO came to an end, and he returned to France via Thailand, Myanmar and India. Back in Paris, he took up senior academic posts. Early in 1914, he returned to Hanoi as acting director of the EFEO. The new director was prevented from taking up his post by the outbreak of the First World War, so Finot remained in Indo-China until 1918. His return to France was brief, however, for in 1919 he was again appointed director of the EFEO and was back in Cambodia by Christmas of that year. He remained in Indo-China until early 1926. Once again, however, he was called upon to fill his old post as acting director when his successor, Léonard Aurousseau, became ill. Consequently, in 1928, he returned to Hanoi. Sadly, Aurousseau did not recover. His death left Finot to await the appointment of a replacement. His suggested nominee was George Cœdès, to whom he handed over the directorship late in 1929. Finot left Indo-China for the last time in 1930, breaking his voyage home in Sri Lanka. He then retired to a house near Toulon, where he died on 16 May 1935.Footnote 10 He had spent more than fifteen years in Southeast Asia, travelling widely, even in remote regions of Laos and the highlands of Vietnam.
Looking more closely at the provenance of the inscriptions, it soon becomes apparent that Finot's reservations were not without foundation. The first point that must be emphasized is that the originals of the rubbings were never located. They could not be found anywhere in French Indo-China. It seems strange that the naval officer possessed not only the interest, but also the materials and the skill, to take rubbings of the inscriptions, yet he apparently took no pains to record with any degree of precision the location of their originals. It must also be wondered why he waited until he was back in France to pass the rubbings to a scholar, when a number of French scholars, many of them attached to the EFEO, were active in French Indo-China at the period. Again, there is a serious question regarding why Derenbourg should have handed the rubbings on to Ravaisse, rather than dealing with them himself. Derenbourg was not afflicted by illness before his death. He died very suddenly, on 12 April 1908, at the comparatively early age of 63. He was still hard at work up to the moment of his death. According to one of his obituaries, “il meurt debout” (he died standing).Footnote 11 Perhaps he was just too busy with other matters to deal with the inscriptions himself, but it seems an odd coincidence that he should have passed them to Ravaisse shortly before he died. Even odder is the fact that Ravaisse did nothing with the inscriptions for so long. He apparently realised that they were important. Indeed, Arabic inscriptions from Annam were certainly not common. Perhaps it is significant that Auguste Barth, the other scholar supposedly involved in the chain of transmission, died in April 1916.Footnote 12 Thus, by the time Ravaisse began work on the inscriptions, there was no one living who could either confirm or deny his story of their provenance, except perhaps for the French naval officer, but his identity was unknown. It is of course entirely possible that there was some confusion somewhere; that, at some point, an honest mistake was made. In the circumstances, however, it must be suspected that Ravaisse fabricated the whole story of the provenance of the inscriptions.
It seems that this suspicion may have been shared by French scholars of the time. It would certainly appear that Louis Finot seriously doubted that the inscriptions had really been found in Annam. When researching this article, it was not at all difficult to find information about most of the scholars involved. Normally, obituaries were published in one or more scholarly journals shortly after their deaths. These are often very detailed, sometimes even including bibliographies of the deceased scholar's publications.Footnote 13 Yet it has been impossible to find any such obituary of Ravaisse. His death, on 31 August 1929, apparently went completely unnoticed by the French academic community. Information about his life and activities has had to be collected, in fragments, from a variety of sources, of which one of the most useful is a French genealogical web site.Footnote 14 Shortly after his death, some of his work was published posthumously. A short monograph on an enamelled glass lamp from Egypt appeared in 1931. It was published by his family: apparently none of his former colleagues or pupils, nor any educational institution or learned society, played any part in this publication. It received a rather critical reception, even from one of his former pupils.Footnote 15
These two inscriptions “from Champa” must therefore now be excised from the history of Southeast Asia. The place which they held in it was due either to error or, very possibly, to deliberate imposture. It is therefore necessary to reconsider whether there was really any Islamic presence in Champa as early as the eleventh century. Without these inscriptions, the evidence is considerably weakened, but other claims remain. Finot discussed one of them at the same time as he expressed his doubts about the provenance of the inscriptions. The description of ChampaFootnote 16 in the Song shi includes the following report:
There are also mountain buffaloes [wild buffaloes], which are not used for ploughing, but are killed as sacrifices to spirits. When they are about to be killed, the presiding shamanFootnote 17 pronounces the following invocation: “A luo he ji ba”,Footnote 18 which means “May it soon be reborn”.Footnote 19
It has been claimed that the invocation “A luo he ji ba” is a phonetic transcription of “Allahu akbar”. This claim appears to have originated with Édouard Huber in 1903.Footnote 20 Huber dismisses the translation given in the source as “fantaisiste”, though without any explanation of why he considered this to be so. This is certainly unacceptable. There is no reason at all to believe that the translation is any more or less reliable than the transcription. If one is held to be inaccurate, then both must be. Conversely, if the transcription is considered to be an accurate representation of the original phrase, then the translation must also be accepted as correct, or at least not completely inaccurate, as Huber would have it. Since “Allahu akbar” does not correspond in any way at all to the meaning given by the source, the claim that the transcription represents this phrase must certainly be rejected. Moreover, it is extremely unlikely that such a foreign phrase would have been introduced into what was apparently a Hindu sacrifice. As Finot pointed out:
It would be very strange if the Hindu Chams, in a sacrifice to the gods, made use of an Islamic formula: since the success of the sacrifice was dependant on the exact observation of the ritual fixed by tradition, the introduction into this ritual of a modern invocation borrowed from a foreign religion cannot possibly be imagined.Footnote 21
Finot also suggests a quite different reconstruction of the Chinese transcription, in the Cham language. In view of the nature of the sacrifice, it is probably not without significance that the pronunciation of the last two Chinese characters, “ji ba”,Footnote 22 bears some resemblance to the Cham word for ‘buffalo’: kabav.Footnote 23 At least it is clear that there are other possible ways to reconstruct the original phrase underlying the Chinese transcription. However, as Finot noted, any such reconstruction would be too hypothetical for any solid conclusions to be based upon it.Footnote 24 Manguin referred to Huber's “very hypothetical rendering”, which he considered “must be left aside”, and also noted that qi ba (sic) might represent the Cham word for ‘buffalo’.Footnote 25 Conversely, Geoff Wade considers A luo he ji ba to be “undoubtedly the Arabic phrase ‘Allahu akbar’”.Footnote 26 The perils of relying heavily on phonetic similarities are well known. To cite only one example, it was at one time claimed that the Sabaeans had a presence in Sri Lanka during the early fifth century, on the strength of a reference by FaxianFootnote 27 to sabao Footnote 28 in the island.Footnote 29 This was long ago shown to be incorrect. Sabao is not a phonetic transcription of “Sabaean”, but of a Sanskrit word sārthavāha, “chef de marchands”.Footnote 30
In his note of 1903, Huber draws attention to the fact that the Song shi also records that: “Their [the Chams’] customs and clothes resemble those of the Dashi”.Footnote 31 Huber translates Dashi as “Arabes”, but at this period such a translation is not entirely correct. Certainly, the original meaning of Dashi was ‘Arabs’, but the term very quickly came to be applied to the people of all the countries conquered by the Muslim Arabs. As Zhou Qufei noted in about 1178: “Dashi is a general name for many countries. There are more than a thousand countries. Those of which the names are known are only a few particular ones”.Footnote 32 Thus, exactly what the comparison made in the Song shi may be is not very clear. By the twelfth century, Lahore had become the capital of the Muslim Ghaznavids (themselves of Turkic origin),Footnote 33 so that Dashi might have included parts of what is now Pakistan and even north-west India. It is also the case that not all those living in the lands under Muslim rule were themselves Muslims. It is clear from the documents from the Cairo Geniza, for example, that Jewish merchants living in Islamic lands were active in trade in the Indian Ocean region.Footnote 34 Jews from such places as Aden usually spoke Arabic and would undoubtedly have been included within the general appellation Dashi by the Chinese. Also, Odoric of Pordenone records that: “The people thereof [Tana, modern Thane, near Mumbai] are idolaters, for they worship fire, and serpents, and trees also. The land is under the dominion of the Saracens, who have taken it by force of arms”.Footnote 35
There is another possibility, too, although one which still requires further elucidation. A number of puzzling references occur in Chinese works which seem to place Dashi somewhere in Southeast Asia. To cite one example, the Song shi states that JavaFootnote 36 was “five days across the sea from the Dashi country”.Footnote 37 It must at least be possible that Dashi was sometimes used by the Chinese as a transcription of a Southeast Asian word. One that immediately suggests itself is tasik, Old Javanese and Cham for ‘sea’.Footnote 38 There is a place in western Java called Tasikmalaya, often abbreviated to ‘Tasik’. This question certainly needs further research, but, for now, it is raised as one possible explanation of the statement in the Song shi quoted above.Footnote 39 It would also make good sense in the context of the story of Ximo Footnote 40 the female ruler of HelingFootnote 41 in Java, who inspired fear in the Dashi, so that they dared not attack her lands.Footnote 42 This story appears to have something of the nature of a fable, however, and should perhaps not be taken too literally.Footnote 43 Undoubtedly, there are cases of the same Chinese term being used with different meanings. One is that of Kunlun.Footnote 44 In early Chinese texts, this term referred to the Kunlun Mountains, where the Queen Mother of the West supposedly resided.Footnote 45 Later, in the context of Southeast Asia, it was both an ethnonym and a toponym. The Jiu Tang shu states that: “From Linyi southwards, [the people] all have curly hair and black bodies and are known by the general name of Kunlun”.Footnote 46 As a toponym, it was applied ‘un peu partout’, in the words of Pelliot.Footnote 47
It may also be significant that, after stating that: “Their customs and clothes resemble those of the Dashi”, the Song shi continues: “They have no silk or silkworms, and wrap white cotton cloth around their chests, which hangs down to their feet, with an upper garment with narrow sleeves”.Footnote 48 Perhaps, then, the comparison with the Dashi, as far as clothing is concerned, was merely that, unlike the Chinese of the period, they wore cotton, not silk.Footnote 49 The style of clothing described here appears to be Southeast Asian rather than Middle Eastern. As for their customs, it does not seem at all likely that there could have been any great similarity with those of the Arabs. In the early fourteenth century, Odoric of Pordenone relates that: “When a married man dies in this country [Zampa = Champa] his body is burned, and his living wife along with it”.Footnote 50 A few decades earlier, Marco Polo noted that the people of Champa were “idolaters”.Footnote 51 His record is all the more significant because he probably visited Champa more than once, and because, shortly after this, he provides one of the earliest records of Islam in Southeast Asia, in Ferlec (Perlak) in northern Sumatra.Footnote 52 Had he seen Muslims in Champa, he would probably have said so. It may be noted that the earliest Islamic coinage issued anywhere in Southeast Asia was minted in Samudra-Pasai (northern Sumatra) during the early fourteenth century, which tends to confirm Marco's observation.Footnote 53 It was in northern Sumatra that Islam first became established in Southeast Asia. Taking all this together, there is little reason to believe that there was any Islamic presence in Champa even as late as the early 1300s.
It is also necessary here to deal with the other frequently-asserted claim relating to Islam in Champa (and elsewhere); that Chinese Pu,Footnote 54 a character often used in transcriptions of foreign names, was always, or almost always, derived from Arabic ‘Abū’. This claim can be traced back to Friedrich Hirth,Footnote 55 and then to Hirth and Rockhill,Footnote 56 but was later enlarged upon by Kuwabara Jitsuzō.Footnote 57 It has since been taken up by a number of scholars. To give just one example, Wade uses the fact that many people from Champa had the Chinese character Pu as the first element of their name (as written in Chinese) as evidence that: “some hundreds of Muslims removed in the 980s to Hai-nan and Guang-zhou to escape political turmoil in Champa”.Footnote 58 His source for this is poor, derived from a “seventeenth-century (sic) Chinese encyclopaedia Gu-jin tu-shu ji-cheng”Footnote 59 which is not necessarily reliable for a period several centuries before its compilation. In any case, the source says that this movement of alleged Muslims from Champa occurred “during the Song and Yuan dynasties”,Footnote 60 not necessarily “during the 980s”. Other sources refer to people from Champa moving to southern China at this time, but do not report that they were Muslims.Footnote 61 In fact, since no Chinese term meaning ‘Muslim’ (or ‘Islam’) existed during the Song period, it would scarcely be possible for any contemporary Chinese text to contain such a report.Footnote 62
Wade has also claimed: “That the envoys to China from the polities of Zābaj and Champa during this period [up to the eleventh century] were Muslim is undoubted, and verified by their names”.Footnote 63 By “Zābaj”, Wade presumably means Java. No such name as “Zabaj” occurs in the Chinese records to which he refers here. There is, however, no question of any Islamic presence in either Java or Champa at such an early period.Footnote 64 The supposed “evidence” of the names is entirely illusory. Ferrand pointed out long ago that when Pu occurs as the first element of a Southeast Asian name in Chinese transcription, it is very likely to be derived from the common and widespread honorific mPu, Pu, or Pō.Footnote 65 It is scarcely surprising that many Southeast Asians had names beginning with Pu in Chinese, as this honorific occurs, in similar forms, in more or less every Southeast Asian language. Even Chinese Bo,Footnote 66 “elder, lord, clan head, a feudal title”, probably belongs to this series, which may ultimately be Austro-Asiatic.Footnote 67 This is a much more likely explanation than the claim that Pu is always a transcription of Arabic ‘Abū’, and that all these Southeast Asians were therefore Muslims. An idea of how frequently Pō was prefixed to the names of Cham divinities can be gained by reading Cabaton's monograph on the Chams. Even Allah became Pō Ovlaḥ.Footnote 68 Anyone of high rank in Cham society was no doubt accorded the same mark of distinction.Footnote 69 It is therefore scarcely surprising that many Cham (and other Southeast Asian) names in Chinese transcription begin with Pu. In fact, it would be astonishing if they did not. There is certainly no need to postulate that the names are foreign (that is, Arabic) rather than Cham. In fact, during the early Song period (mid-tenth to mid-eleventh centuries), the pronunciation of Pu was probably not very similar to ‘[A]bū’.Footnote 70
The character Pu was commonly used in transcriptions of foreign words. It is found in the names of several countries, such as Puluozhong guo,Footnote 71Pulin guo,Footnote 72Puhuang guo,Footnote 73Pugan guo,Footnote 74Puduan guo,Footnote 75 and Pupo guo.Footnote 76 There are also cases where it occurs at the beginning of personal names, but clearly cannot be a transcription of ‘Abū’. For instance, it is the first character of the Jurchen name Pucha Yuannu.Footnote 77 There is also a clear example involving the names of Cham royalty:
In the fourth year [of the KaibaoFootnote 78 reign-period] (971–72), Xili Duopan,Footnote 79 the deputy king Li Nou,Footnote 80 the royal consort Madame Guo,Footnote 81 and the [king's] son PulujiboluoFootnote 82 together sent ambassadors to present tribute.Footnote 83
The king is clearly Hindu, his wife has a name which looks Chinese, and the son's name begins with Pu, which allegedly ought to mean that he was a Muslim. Obviously, this makes no sense at all. Perhaps Madame Guo was of Chinese origin, but there can be no certainty of that. Likewise, it cannot be claimed with any degree of credibility that Pu in this case is a transcription of ‘Abū’. The fact that a name in Chinese transcription can be plausibly reconstructed as an Arabic name does not mean that it must be. It is pure fallacy to think otherwise. In the case of this son of the King of Champa, it makes good sense to take Pu as a rendering of the Cham honorific Pō. It is also possible, however, that Pulu might be a Chinese rendering of Cham Braḥ, ‘royal’.Footnote 84
The attempt has also been made to reconstruct Cham names beginning with the common Chinese family name Li Footnote 85 as Arabic names, taking Li to represent ‘Ali’. This is even less convincing than the claim that Pu always represents ‘Abū’. For one thing, the name Li was common in Vietnam,Footnote 86 as Lý. Wade has claimed that the personal names following Li in the names of Cham envoys recorded in Chinese sources “are not Chinese or Vietnamese names, as the characters which follow the ‘Li’ do not represent Chinese given names but are phonetic representations of other names”.Footnote 87 Perhaps so, but this certainly does not mean that the “other names” must be Arabic. In fact, however, most of the Cham names beginning with Li which Wade citesFootnote 88 could be Chinese or Vietnamese. In the great majority of cases, Li is followed by two Chinese characters, exactly as would usually be expected in the case of Chinese or Vietnamese personal names. To mention but one instance, I can see absolutely no good reason to try to reconstruct Li Bozhu,Footnote 89 which certainly might be a Chinese name, as “Ali Bashir”.Footnote 90 The phonetic resemblance is scarcely close.Footnote 91 It makes much better sense to think that people from Champa with the name Li were descendants of Vietnamese or Chinese settlers, rather than Arabs. If their personal names were “foreign”, this is most probably because they were Cham (or perhaps Vietnamese or Chinese in Cham pronunciation). There is no reason whatsoever to think that they were Arabic. A serious issue here is that little is known about Cham names of the period. It is therefore impossible to be sure that these Chinese transcriptions do not represent Cham names.
There is in fact no sound evidence of any significant Islamic presence anywhere in Southeast Asia before the second half of the thirteenth century. Marco Polo noted that there were Muslims in Sumatra, but the process of conversion was quite obviously at an early stage:
in this kingdom of Ferlec all the people used to worship idols, but by reason of the many Saracen merchants who frequent there with their ships, they have converted them all to the . . . law of Mahomet. And these are those of the city only . . ., but those of the mountains . . .are such as beasts.Footnote 92
This is Marco's only mention of Muslims anywhere in Southeast Asia. In “the kingdom of Sumatra”, for example, where he stayed for five months, the people were “savage idolaters”.Footnote 93 It must be remembered that Marco was travelling with envoys who had journeyed to the court of Qubilai Qa'an from the Ilkhanate. No doubt some of them, or some of their attendants, were Muslims, who would quickly have taken notice of any coreligionists met in the course of the voyage. Marco's testimony is very valuable and highly likely to be accurate.
Confirmation of Marco's observations is provided by the fifteenth-century accounts of Champa in records of Zheng He'sFootnote 94 famous voyages. As Manguin has stated, neither Ma HuanFootnote 95 (himself a Muslim) nor Gong ZhenFootnote 96 say anything about Muslims in Champa. The Sejarah Melayu also seems to indicate that Chams were not Muslims. In the early sixteenth century, Tomé Pires states flatly that there were no “Moors”, that is Muslims, in the kingdom of Champa.Footnote 97 Again, Manguin emphasized that the Portuguese paid great attention to the question of religion, and particularly to the presence of Islam.Footnote 98 It would seem, then, that all the Muslims who allegedly were present in Champa during the periods of the Tang and Song dynasties had mysteriously disappeared by the late thirteenth century, and were still missing during the fifteenth and sixteenth centuries.
It is probably also of significance that there is a clear distinction between the Muslims of China and those of Southeast Asia, one which is very rarely taken into consideration. The great majority of Chinese Muslims are Sunni of the Ḥanafī school, while Southeast Asian Muslims, although also Sunni, are of the Shāfi’ī school. It is this latter school of Islam which predominates in almost all areas around the littoral of the Indian Ocean.Footnote 99 It is therefore very unlikely that Islam could have been introduced into China from Southeast Asia, or by travellers using the maritime routes to southern Chinese ports. There is no clear connection between Islam in China and Islam in Southeast Asia, which very likely had separate origins.
As far as the assumption that Arabs were regularly arriving in Chinese ports via the maritime routes across the Indian Ocean and the South China Sea during the Tang and Song dynasties is concerned, there are some very serious objections to be raised. Bielenstein calculated that only about five of 30 “Arab” (that is, Dashi) embassies to China during the Tang dynasty could have arrived by sea. His calculations are incorrect, however. Bielenstein believed that: “Arab ships left the Middle East with the beginning of the summer monsoon and reached China in 130 to 140 days”.Footnote 100 This is impossible. He says that ships would have sailed about the middle of April and arrived in China about the middle of August. However, the southwest monsoon wind of summer does not begin to blow at the same time throughout the entire area of the Indian Ocean. At Mumbai, it does not normally set in until the middle of May. Further north, it commences even later.Footnote 101 Ships therefore could not have departed from the Middle East before late May or early June, and could not have reached the northern tip of Sumatra much before early August. It took more than a month to complete the voyage from Sumatra to the China coast, and the southwest monsoon in the South China Sea begins to fail at about the end of August. September is a month of changeable winds.Footnote 102 By the beginning of August, therefore, it was too late to set sail across the South China Sea for any Chinese port. It must be borne in mind that long ocean voyages at this period frequently encountered problems at some stage, so that allowance always had to be made for the possibility of delay. A ship might be blown off course by a storm, becalmed, need to stop to make repairs, run short of provisions, or, even more commonly, of fresh drinking water (which was difficult to keep in drinkable condition for a long time in containers aboard ship), or encounter some other misfortune. Ships did not set out on a voyage of more than a month with only about a month when favourable winds were likely to prevail along the route. Time had to be allowed for possible delays.
It is also now often alleged that ships could have set out from the Persian Gulf in October.Footnote 103 This may be possible with modern sailing vessels, but was not possible in earlier times. Ships did not sail eastwards when the northeast monsoon wind was blowing. As Zhou Qufei noted, “fundamentally, the foreign ships sail with the wind”.Footnote 104 They sailed eastwards when the southwest wind blew in summer. Chinese sources are quite specific about the times of year when ships arrived in Chinese ports and departed from them. Zhou Qufei states that foreign ships arrived “after the summer solstice”,Footnote 105 that is, during late June and July. Those making for Arabia (Dashi) set sail on their return journeys “from the second month of winter on”,Footnote 106 that is, during the eleventh month of the Chinese year (normally from early to mid-December into January: in 1178, when Zhou completed his book, the eleventh month began on 11 December, and ended on 9 January 1179). Departing from Guangzhou in December or January, ships reached Lamuri (also written Lambri)Footnote 107 in northern Sumatra in about 40 days. There, however, they had to stop, as they could not be sure of completing the onward voyage within the remaining period of the northeast monsoon. So, “they stay there until the next winter, before again sailing with the northeast wind. With favourable winds, they can arrive [in Arabia]Footnote 108 after sixty days”.Footnote 109 Ships from Arabia could not complete a round trip to China and back within two years.Footnote 110 In fact, the voyage in one direction took at least a year, and that was assuming that no problems were encountered at any stage.Footnote 111 What Zhou says is confirmed, at least in part, by Zhao Rukuo, who also relates that ships sailing to Arabia from China had to make a long stopover in Sumatra.Footnote 112
Other sources provide details of departure times from ports in the Arabian region. In about 1169 or 1170, David Maimonides wrote to his elder brother Moses from the Red Sea port of ‘Aydhāb, where he was waiting to take ship across the Indian Ocean. He expected the letter to arrive “when I, God willing, shall have already made most of the way”, so his departure must have been imminent. The letter is dated 22 Iyyar, which was certainly sometime in May.Footnote 113 It would have taken several days to exit the Red Sea from ‘Aydhāb, so the voyage across the Indian Ocean is unlikely to have begun much before the beginning of June. Another letter from ‘Aydhāb, which is believed to date from 16 May 1141, indicates that ships were about to set sail from there towards India.Footnote 114 It is therefore reasonably clear that the normal time to sail eastwards from Arabia was late May or June.
A Chinese source referring to one of the early fifteenth-century voyages of Zheng He records that the ships set sail from China (the coast of Fujian province) in the sixth year of the XuandeFootnote 115 reign-period, on the ninth day of the twelfth month (12 January 1432), arriving at Champa two weeks later, on the 24th day. On the sixth day of the second month of the following Chinese year (7 March 1432) they arrived in Java. They then passed through the Strait of Melaka, reaching the port of Sumatra (Samudra) on the eighteenth day of the eighth month (12 September). Sailing from there on the tenth day of the tenth month (2 November), they arrived at Sri Lanka on the sixth of the eleventh (28 November). On the tenth day (2 December) they departed, reaching Calicut (Kozhikode) on the eighteenth (9 December). By the 26th day of the twelfth month (17 January 1433), just over a year since they had left the coast of Fujian, they arrived at Hormuz. The return voyage began on the eighteenth of the second month of the following Chinese year (9 March 1233). During the early sixth month they were in sight of the coast of China. The ships were back in Nanjing, then capital of the Ming Empire, on the sixth day of the seventh month (22 July 1433).Footnote 116 The ships set sail on the return journey before May or June, but they did not sail in October. It is likely that these fifteenth-century Chinese ships were considerably superior to vessels built in other countries at the time. In particular, their sails allowed them to sail very close to the wind: “Of all the eight quarters whence the wind may blow, there is only one, the dead ahead quarter, which cannot be used to make the ship sail”.Footnote 117 Despite its advantages, the Chinese type of sail was very rarely copied in other countries, with the exceptions of Japan and the Maldives.Footnote 118 Also, the route taken by the ships in 1432–33, through the Strait of Melaka, was probably different from, and shorter than, that used before about 1300.Footnote 119 These factors allowed the Chinese ships to complete their return journey in a single season, but this would not have been normal for ships of other countries at an earlier period. It should be noted that a principal reason why mariners preferred to sail with a following wind was that a greater area of sail could safely be raised than when the wind was abeam. This, of course, meant greater speed. Navigation was also easier, as a side wind tends to push the whole ship sideways to some degree. It may seem to sail straight through the water, but will in fact drift off the steered course, away from the wind: it will make leeway.Footnote 120
According to my evaluation of Bielenstein's data, it is possible that as many as 16 of the embassies from the Dashi which arrived in China during the Tang dynasty might have come by sea. This includes all the embassies for which no month of arrival is given, so it is very much a maximum figure.Footnote 121 This means that it is very likely that half of the embassies travelled overland, possibly considerably more than half. It is probably unlikely that any of the earlier embassies, prior to about 750, travelled by sea, although I have counted some of them as possibly having done so. Before 750, the Umayyad Caliphate had its capital at Damascus, far from the Persian Gulf. Then the Umayyads were defeated by the ‘Abbasids, and the centre of the Caliphate shifted from Syria to ‘Irāq (eventually to Baghdad).Footnote 122 This move away from the Mediterranean region towards the Persian Gulf certainly gave the Caliphate easier access to the maritime routes of the Indian Ocean. However, the last recorded embassy to the Tang court arrived in October/November 798,Footnote 123 less than 50 years after the Umayyad defeat, so there was not a very long period during which the sea routes were likely to have been preferred. It is worth noting that, of the total of 30 recorded embassies, no less than twelve arrived in China during the period from 744 to 762/3. It must be likely that this was because of the conflict between the Umayyads and ‘Abbasids, both sides seeking the recognition (or perhaps even the support) of Tang China.Footnote 124 At this time, travel overland is at least as likely as travel by sea.
Bielenstein believed that “only about 10 of the 47 missions” which arrived at the Song court could have travelled overland.Footnote 125 Again, I am not at all sure that his calculations are correct. In reality, it is highly unlikely that more than a few “Arab” (Dashi) embassies travelled to Song China by sea. The Song shi contains the following report:
When they [the Dashi] came to present tribute in the first year of the TianshengFootnote 126 reign-period (1023–24), it was feared that they might be plundered by the people of the West. It was therefore commanded that, from then on, they should take the maritime route via Guangzhou to the capital.Footnote 127
“The people of the West” here refers to the Tangut state,Footnote 128 which was often hostile to the Song Empire. Dashi embassies had been passing through Shazhou,Footnote 129 which was conquered by the Tanguts not long after this time. The essential point in the present context is that envoys from the Dashi had clearly been travelling exclusively overland during the first several decades of the Song period. The Chinese actually had to tell them to travel by sea! Li Tao confirms that “whenever the country of the Dashi presented tribute, their route was from the western boundary of Shazhou, arriving at Qinting”.Footnote 130 He does not mention an embassy in this year. His version of events indicates that it was suggested that Dashi embassies might travel by sea, to avoid possible problems on the land route, and that the Emperor decreed that this was permissible.Footnote 131 All this would be quite extraordinary, if the frequently-repeated claims that maritime trade with China was dominated by Arabs were true.Footnote 132 Obviously, something is very wrong with such a view. Although the Song huiyao indicates that the Dashi envoys obeyed the Imperial command,Footnote 133 it appears to be very likely that, in reality, they merely indicated acceptance, but continued to travel overland.Footnote 134 This would explain why so few Dashi embassies reached the Song court after the Song Empire had lost northern China to the Jurchens. After 1125, only five embassies are recorded, compared with a total of 47 for the entire period of the Song dynasty. Two of these five are reported to have arrived by sea, but in the other three cases the route is unclear. One which arrived on 28 March 1129 most probably travelled overland.Footnote 135 If Arabs were not regularly sailing to ports on the southern coast of China, then they were not passing Champa, so that Islamic influence there is unlikely.
However, if it seems very probable that Arab embassies rarely travelled to China by sea, this does not preclude the possibility that merchants from the Islamic world arrived by ship more frequently. Indeed, there are a few references to “Arab” (that is, Dashi) shipmasters apparently acting as envoys during the Tang and Song periods. Such references are very few, however, amounting to only seven out of the total of 77 recorded embassies. It is not clear if they were genuine envoys or merely claimed to be.Footnote 136 Overall, the impression is that Arabs were not frequent visitors to Chinese ports, at least during the Song period. They may have arrived more frequently during the Tang period, but the sack of Guangzhou by Huang ChaoFootnote 137 in 879 apparently ended such trade contacts: “the trading voyages to China were abandoned, and the country itself was ruined . . . . . The reason . . . for the end of the China trading voyages from Sīrāf, was an uprising led by a rebel . . .”.Footnote 138
There are several important Chinese sources which give accounts of China's maritime trade during the Song dynasty. They all suggest that Dashi merchants were infrequent visitors to China. Writing during the early twelfth century, Zhu Yu recorded that frankincense,Footnote 139 a typical Arabian product, was carried to China in Srivijayan ships.Footnote 140 He also states that: “This country (Srivijaya) is directly to the south in the sea, westwards to Dashi is still a long way. ChineseFootnote 141 going to Dashi repair their ships in Srivijaya and exchange goods”.Footnote 142 These two mentions of Dashi are the only ones which occur in Zhu Yu's work. SrivijayaFootnote 143 is mentioned five times.Footnote 144 It is perhaps significant that Zhu Yu refers to Chinese sailing to Dashi, but says nothing about people from Dashi sailing to China.
Nevertheless, as I have said above, it is likely that some Dashi merchants travelled to China during the Song period. Zhou Qufei reports that:
Coming from the Dashi country, [merchants] travel southwards in small ships. When they reach the country of Kollam,Footnote 145 they change to large ships and travel eastwards. When they get to Srivijaya, they also come to China like the Srivijayans [that is, by the same sea-route, and perhaps also in Srivijayan ships].Footnote 146
Clearly, then, those Dashi merchants who ventured as far as China usually did not travel all the way in their own ships. Zhao Rukuo relates that:
Dashi is extremely far to the northwest of Quanzhou. It is difficult for the foreign ships to arrive there directly. Setting out from Quanzhou, ships sail for more than forty days to Lamuri,Footnote 147 where it is most convenient to spend the winter. The next year, they set out again. With a favourable wind, they reach this country only after more than sixty days. The products of this country are mostly carried to Srivijaya and traded there. Merchants then resell them onwards to China.Footnote 148
The Chinese terminology used by Zhao Rukuo here indicates that goods from Dashi were sold onwards to China by different (presumably Srivijayan) merchants. This is confirmed by the documents from the Cairo Geniza, none of which makes any mention of merchants travelling as far as China: “not a single Geniza letter referring to direct contacts with China has been discovered thus far”.Footnote 149 These documents do, however, contain mention of merchants from Arabia travelling to Sumatra,Footnote 150 which is in complete agreement with what Zhao Rukuo says. During the Song period, then, the available evidence suggests that merchants from Dashi travelled to China only occasionally, and that official embassies from Dashi usually reached China overland.
As already noted, the best evidence of Arab contacts with Southeast Asia and China certainly relates to the period of the Tang dynasty. Tibbetts stated: “By the end of the tenth century, Arab knowledge of South-East Asia had passed its zenith. All the more original sources are earlier than this date, and the works which follow consist mainly of extracts and quotations from earlier authors”.Footnote 151 The most informative texts are the al-Masālik w'al-Mamālik by Ibn Khurdādhbih, the work edited and augmented by Abū Zayd al-Sīrāfī usually known as the Akhbār al-Ṣin w'al-Hind or Accounts of China and India, and the partly fictional ‘Ajā’ib al-Hind attributed to Buzurg ibn Shahriyār.Footnote 152 It must be admitted, however, that the information which they contain relating to Southeast Asia is not very detailed, and is often confused, often more confused than many scholars have recognized. The itinerary from Sīrāf to Khānfū in the first book of the Accounts of China and India is as follows: from Sīrāf, ships make for Muscat, where fresh water and provisions are available. Then the route continues to Kollam (Kūlam Malī)Footnote 153 in India, a month from Muscat “if the wind is constant”. Fresh water is again available in Kollam. Ships then sail onwards to islands called “Lanjabālūs”, probably the Andamans or the Nicobars. After this, the next stop is Kalah Bār, a place which has never been certainly located.Footnote 154 It is said to be subject to the kingdom of al-Zābaj, which can reasonably be identified with Srivijaya. To Kalah Bār from Kollam is a voyage of a month.Footnote 155 This makes about two months from Muscat, which agrees well with the statements of Chinese sources. The next place is Tiyūmah, “where there is fresh water for anyone wanting it”. This is very probably Pulau Tioman, off the east coast of the Malay Peninsula, almost opposite the modern port of Mersing. It must be noted that the passage from the Indian Ocean to the South China Sea is not mentioned at all. Was it via the Strait of Melaka, or the Sunda Strait? Then, ten days further on, is a place called “Kanduranj” (or Kadrang, or Ku[n]durung), where fresh water is again available. This place has been variously identified, usually either as Panduranga (modern Phan Rang in Vietnam), or as Pulo Condore (now the Côn Đảo archipelago, Vietnam).Footnote 156 My own view is that it is almost certainly Pulo Condore, which was apparently also known as Kundurung at one time. To make it into “Panduranga” it is necessary to change the initial letter, as well as to insert an ‘n’ (the first ‘n’ is not present in the original Arabic text, and also has to be inserted if “Kundurung” is preferred).Footnote 157 The fact that the vowels are not indicated in the Arabic original of course helps to increase confusion. It can be said with considerable confidence, however, that Pulo Condore is in the right place to be Kundurung or Kanduranj. This group of islands is repeatedly mentioned in Chinese itineraries and was obviously a major landmark when ships were crossing the Gulf of Thailand. This, I feel sure, is because the main island rises to more than 1,500 ft (500 m) above the surface of the sea, and is therefore visible for some distance, while the neighbouring mainland is the flat and low-lying delta of the Mekong River.
Beyond this place (wherever it was), the next port of call, after a voyage of ten days, is Ṣanf, which is Champa. Then, after another ten days, there is an island called Ṣandar Fulat, a month from China across a sea called Ṣankhī, which seems to be a rendering of the Chinese Zhang Hai,Footnote 158 the South China Sea.Footnote 159 All commentators have got into difficulties here. The island called Ṣandar or Sundur (Fulat being a Persian version of Malay Pulau, ‘island’) has defied definite identification.Footnote 160 Tibbetts thought that “some mistake has crept in here”. The mistake was certainly partly his own, as he believed that Sandar/Sundur was “in the vicinity of Hainan”, which of course made a month from there to China much too long. My explanation is a simple one: both Kundurung and Sundur Fulat are Pulo Condore.Footnote 161 The author of this text heard of the island under two different names, and failed to realise that the two were the same place. He therefore inserted both of the names into his route. A month's voyage from Pulo Condore to China is still rather long, but well within the bounds of possibility. Confusion of this kind is by no means uncommon.Footnote 162
The frequent mention of fresh water being available at various places lends an air of authenticity to this account. It is probable that the itinerary was based on information obtained from merchants or sailors who had actually undertaken the journey. They would often have been concerned about their supplies of fresh water for drinking. During the great voyages undertaken by the Chinese in the early 1400s, there were special ships for carrying drinking water.Footnote 163 On the other hand, it is noticeable that only comparatively few places are mentioned. The omission of both Java and Srivijaya (Palembang) as points on the route is surprising (although Srivijaya figures prominently elsewhere in this text).Footnote 164 Far fewer places are mentioned than in the itinerary of Jia Dan,Footnote 165 which dates from about the same period.Footnote 166 Ibn Khurdādhbih's account is confused, quite probably corrupted by copyists.Footnote 167 It is difficult to make good sense of his route in Southeast Asia, especially in the area of the South China Sea. Insofar as anything can be made of it, a great deal depends on the identification and location of toponyms, which are often less than certain. As Tibbetts noted, his sailing times cannot be reconciled with those of the Accounts of China and India.Footnote 168 Tibbetts took a favourable view of Ibn Khurdādhbih's information,Footnote 169 but it seems to me that the Accounts of China and India is more accurate. Buzurg's work contains very little information about places or routes. My overall impression of these Arabic accounts is that they show only limited, and often inaccurate, knowledge of Southeast Asia and the sailing route to China. The evidence suggests that even at the height of maritime contact with China, Arabian knowledge of the sea route across the South China Sea was far from comprehensive. This in turn suggests that Arabian ships can rarely have sailed all the way to China, for if they had done so, Arabian sailors would surely have had much fuller knowledge of the route. Arabian merchants travelling to China during the eighth and ninth centuries no doubt usually travelled in ships of other countries, as is suggested by the Chinese sources of the Song period.
As a footnote to this, issue must be taken with the now commonly-expressed belief that sewn-plank ships were not as inferior as Western observers, including Marco Polo, Friar Jordanus, and Portuguese navigators, often suggested.Footnote 170 If they had any advantages over ships held together with nails, then it must be questioned why the technique of sewing was superseded by nailing more or less everywhere in the world apart from the Indian Ocean region. The sewn-plank construction technique is recorded from almost all parts of the world except Australia,Footnote 171 yet virtually everywhere it was abandoned. It has been noted that sewn-plank vessels had a tendency to leak.Footnote 172 Indeed they did,Footnote 173 and it is only in nonsense rhymes that anyone goes to sea in a sieve. As Jordanus recorded, “they take in water to such an extent that the men always, or almost always, must stand in a pool to bale out the water”.Footnote 174 Marco Polo remarked that: “their ships [in the Persian Gulf] are very bad and . . . dangerous, and many of them are lost”.Footnote 175 That this was true is confirmed by the documents from the Cairo Geniza. In the letter already cited above, from David Maimonides to his brother, there is mention of two ships foundering, during the space of the few weeks that he was in ‘Aydhāb. David himself was lost at sea during his first voyage towards India.Footnote 176 “[S]hipwreck was a recurrent feature of seafaring on the Indian Ocean . . .”.Footnote 177 It may be noted that it was not only Europeans who considered nailed ships superior to sewn-plank vessels. Zhou Qufei recorded that: “In the remote coastal prefectures of Guangdong, it is difficult to obtain iron nails and tung oil, so when ships are built they are all constructed by making holes in the planks and binding them together with rattan”.Footnote 178 Clearly, Zhou considered this to be a backward technique, used only in remote regions where iron nails were hard to find. There can be no question that, for long voyages across open ocean, sewn-plank ships were unsafe. It was never an advantage for the hulls of ships to be flexible. Indeed, the size of wooden ships was limited by the fact that very large wooden hulls could not be constructed with sufficient rigidity. Only the use of iron strapping made it possible to build very large wooden ships. Hulls constructed of wooden planks tended to leak, and even to break up, if the hull was not sufficiently rigid.Footnote 179 It may also be noted that sewn-plank vessels were often strengthened by the use of scarf or tenon-and-mortis joints, dowelling, and treenails.Footnote 180 Once Portuguese ships had arrived in the Indian Ocean, they quickly began to influence local shipbuilding techniques. Factories producing iron nails were set up in western India.Footnote 181 Sewn-plank vessels survived only locally, and were mainly used on coastal and inland waters.
Evidence of Muslims in Champa is sometimes derived from essentially circular arguments. For example, Manguin asserts that: “It is further known that the famous Muslim families Pu of Guangzhou (Canton) and of Quanzhou, both of Arab origin, and who had arrived in China under the Song, had previously lived in Champā”.Footnote 182 However, the principle arguments for these Pu families being Muslims of Arab descent is precisely that their name was allegedly derived from Arabic ‘Abū’, and that they had come from Champa, where it was ‘known’ that there were Muslims.Footnote 183 It must be noted also that the arguments in favour of Pu ShougengFootnote 184 being a Muslim of foreign ancestry have been considerably weakened by the demonstration that the Xin shi Footnote 185 is virtually certainly a late-Ming forgery, not the thirteenth-century work which Kuwabara claimed it to be.Footnote 186 It was the only supposedly contemporary source which said that Pu Shougeng was of foreign descent (though it does not say that he was either an Arab or a Muslim). Whether these Pu families were really Muslims is at least questionable. The only Chinese sources which suggest that they were are late, dating from no earlier than the middle of the fifteenth century. Certainly, the fact that they came from (or may have come from) Champa can no longer be used to support such a claim. While there was apparently good evidence of an Islamic presence in Champa during the eleventh century, it was not unreasonable to argue that people from Champa with names beginning with Pu in Chinese transcription were Muslims. Now that the provenance of Ravaisse's inscriptions has been proved to be incorrect, it is no longer reasonable to make such an argument. The reconstruction of the originals of foreign words in Chinese transcription is often difficult and frequently fraught with ambiguities.
There are lessons that should be learned from the unfortunate affair of these two inscriptions “from Champa”. One is that forgery and faking, not only of objects, but perhaps even more of provenances, is all too easy.Footnote 187 ‘Piltdown Man’ deceived the scientific world for almost fifty years (despite the fact that reservations were expressed about authenticity at an early stage), and caused great harm, “seriously undermining the standing of British palaeontology”.Footnote 188 The Piltdown forgery was finally exposed principally because other discoveries eventually showed that ‘Piltdown Man’ was anomalous.Footnote 189 The Piltdown forgery was a combination of a modern human skull with the jaw of an orang-utan, but humans had not evolved large brains at an early stage, before other primitive features, such as an ape-like jaw, had disappeared. In reality, human ancestors had begun by walking erect and becoming more gracile. Large brains had appeared only at a late stage. It is particularly relevant here that the creator of ‘Piltdown Man’ not only forged the remains but also fabricated a false provenance for them.Footnote 190 Any object that lacks a well-established provenance should always be considered suspect. In the context of inscriptions, it is the case that the place where they are found is not necessarily their place of origin. It may well be imagined that objects such as gravestones did not normally move very far, but this is in fact not true.
Shaped stones are useful, and therefore have some value. They can be used and reused, and frequently were. One use of shaped stones which is particularly relevant in the present context is as ballast in ships.Footnote 191 All sailing ships needed ballast of some kind: there had to be weight in the bottom of the ship to counterbalance the weight of the masts and sails, and the force of the wind pushing against the sails. A sailing ship without ballast was unstable and very likely to capsize. Ballast can take many forms, but it was very common for wooden sailing ships to carry stone as ballast in the lowest part of the hull (the bilge). Squared-off blocks of stone were often preferred as ballast, as they were less likely to shift than rounded pieces of stone. If the ballast shifted to one side while a ship was under way at sea, it would take on a list and might capsize and sink. Ballast was changed quite frequently, partly because it often became foul and stinking after sitting in bilge-water for months, and partly because it was often necessary to lighten the ship to carry out repairs and maintenance. Ships were regularly careened by offloading ballast and allowing them to tilt over to one side so that the bottom of the hull could be cleaned. It was common for old ballast to accumulate in places regularly visited by ships. One town in France, located in an area of marshland, was said to have been built on a foundation of old ballast.Footnote 192 Many of the houses in the town are partly built of ballast stone, easily recognisable because it is clearly different from any local stone. Much of it is granite from northern Europe.Footnote 193 The old town wall at King's Lynn in Norfolk, U.K., probably originally constructed between about 1266 and 1340, contains a significant proportion of rock from the Baltic region. An analysis of the rock types suggests strong trading links with Estonia.Footnote 194
It might be thought that tombstones would have been shown respect, and rarely used for other purposes, but there is no reason at all to believe that the Portuguese, who were inimical to Muslims, or the Spanish or Dutch, would have respected Islamic cemeteries. Indeed, in Portugal itself, King Manuel I instituted a persecution of Jews and Muslims, during which Muslim property was confiscated by the Crown:
In May 1497 Manuel granted the [Muslim] cemetery in Lisbon to the municipal authorities to use as a pasture for animals but gave the tombstones to the newly founded hospital of All Saints for use in its construction works. Two years later, a field and some ruins ‘in the cemetery that used to be that of the Muslims’ was leased to Leonardo Moniz for an annual lease of 50 reais. The field which had been the Muslim cemetery of Colares was leased by the Crown to . . . an inhabitant of the town, in 1499.Footnote 195
It is also recorded that, after the Portuguese had taken Goa for the second time, and had massacred more than six thousand Muslims (“Moors”):
Afonso Dalboquerque turned his attention without delay to the fortifications of the city, and ordered that a great quantity of cement should be prepared, and all the sepulchres of the Moors thrown down, in order to obtain plenty of stone for the works . . . . .Footnote 196
In China, many old tombstones have been recovered from city walls.Footnote 197 It is unlikely that Chinese sailors would have had very much respect for foreign tombstones bearing inscriptions in a language that they could not read. An Islamic tombstone in Java shows clear indication of having been modified for use as an anchor-stone.Footnote 198 In fact, old Islamic tombstones were not immune from being reused by coreligionists. An Islamic gravestone in the collection of the Metropolitan Museum of Art has an inscription in Kufic script on one face, dating from the tenth century. On the other face, and upside down in relation to the Kufic inscription, is another tomb inscription of the mid-thirteenth century in Naskh script.Footnote 199 Apparently, a stone from an old grave was simply taken and reused for a later burial. Overall, there is no good reason to believe that Muslim tombstones would never have been used as ship's ballast.
It was also sometimes possible to replace ballast with saleable goods. One reason why Chinese porcelain became such a major item of trade was that it could be carried as ballast. In bulk, it was heavy, and it was not damaged by bilge-water. Ships trading with China regularly offloaded stone ballast in Chinese ports and replaced it with porcelain.Footnote 200 Even after the secret of making porcelain had been discovered in Europe, European ships still usually carried at least a limited amount of porcelain as ballast when they sailed from China, even though it was no longer very profitable.Footnote 201 Over the centuries, huge amounts of stone must have arrived in Chinese ports as ballast. No doubt much of it was reused as building material. It was very probably as ballast that the earliest Islamic tombstone found in Java arrived in the island.Footnote 202 It must therefore always be suspected that any stone found in or near a port in Southeast Asia or China (or, indeed, elsewhere), might have been carried there from somewhere else, unless there is evidence to the contrary.
Another lesson to be learned from the case of the inscriptions “from Champa” is that, once published, claims are often repeated and circulated more or less incessantly, even if they are refuted. Having been mentioned in Maspéro's book, these inscriptions featured in most accounts of Islam in Southeast Asia, and even in China, long after Finot had cast serious doubt on their provenance. The presence of some kind of Islamic community in Champa during the early eleventh century became an established ‘fact’, on which a number of other hypotheses were founded. Even though the inscriptions have now been clearly shown to have no connection with Champa, at least some of these hypotheses still remain in circulation.
Inscriptions continue to be found. In 1998, a paper by Zhang Shimin was published in Xi'an in China which reported the discovery of an inscription dating from the Tang period. This inscription was cut into the Spirit Way TabletFootnote 203 of the tomb of a certain Yang Liangyao.Footnote 204 Its most startling content is a statement that Yang had been appointed as ambassador to the “black-clothed Arabs”,Footnote 205 that is, the ‘Abbasid Caliphate.Footnote 206 Surprisingly, there was very little reaction to this first publication of such an astonishing revelation. Yang Liangyao remained an obscure personage for several more years. Then, in 2005, there was a further announcement by Zhang Shimin of the discovery of the inscription, this time with a full transcription of its text.Footnote 207 Nonetheless, very little notice was taken of the inscription before 2011, when Rong Xinjiang of Beijing University presented a paper about it at a conference in Guangzhou, which was subsequently published early the next year.Footnote 208
The inscription records that, in 785, Yang Liangyao was appointed to lead an embassy from the Tang court to the ‘Abbasid Caliphate, and that he apparently travelled by the maritime route. This is immediately problematic for a number of reasons. Firstly, no trace of either this embassy or Yang Liangyao can be found in any Chinese source other than this inscription. This does not necessarily indicate that neither can have existed, but it is certainly odd that there is no record at all of the embassy, which was clearly an important enterprise. Secondly, such an embassy would be highly anomalous. There is no record of any similar undertaking until the late thirteenth century, when ambassadors to the Yuan court from the Ilkhanate returned by sea (with Marco Polo), and the early fifteenth century, with the voyages of Zheng He from China into the Indian Ocean. Even then, the first three of Zheng He's voyages did not sail beyond southern India. Only after experience had been gained in navigating the South China Sea and the Indian Ocean did the Chinese ships go further.Footnote 209 It is by no means clear that Chinese ships ever sailed as far as Arabia during the Tang period. The date of the embassy, 785, also seems unlikely. Chinese sources record embassies from Dashi to the Tang court in 772, 773, 774, 791, and 798, but no more during the Tang period.Footnote 210 It would seem, therefore, that, by the last three decades of the eighth century, contacts between Tang China and the Arabs were drawing to a close. The question must therefore arise as to whether the stele and its inscription are genuine. Unfortunately, as is common in Chinese publications, the issue of the authenticity of the inscription is not discussed at all in any of the papers by Zhang Shimin and Rong Xinjiang. It is simply assumed that the inscription is genuine. Since no clear photographs of the stele and inscription have yet been published anywhere, as far as I am aware,Footnote 211 it is impossible to form any very firm opinion about their authenticity, which must remain suspect until further information is available. Grounds to be sceptical are strengthened by the fact that, more than thirty years after its alleged discovery, and eighteen years after the publication of Zhang Shimin's first article about it, the stele remains in a small local museum in the town of Jingyang xianFootnote 212 in Shaanxi province. Despite its apparent importance, it has not been given recognition as a protected cultural relic,Footnote 213 either at national or provincial level. It has been entirely ignored by Chinese scholars of repute, with the exception only of Rong Xinjiang. The conclusion must be that Chinese historians and archaeologists generally doubt its authenticity.
Despite such problems, a number of claims have already begun to spread, based upon the evidence of this inscription. Professor Dr. Angela Schottenhammer has written that:
Yang Liang-yao was sent to the court of the Abbasids . . . in 785 . . . and took the sea-route from Guangzhou, the description of which partly resembles Chia Tan's [Jia Dan's]. This has led to speculation that Chia Tan originally received his information from Yang Liang-yao.Footnote 214
This, however, goes well beyond what the evidence of the inscription supports. In reality, the only concrete information given about the embassy is that, in the fourth month of the first year of the ZhenyuanFootnote 215 reign-period (May/June 785), Yang Liangyao was appointed to lead it, that unnamed assistants to him were also appointed, and that he was provided with various documents.Footnote 216 After that, there is no further clear detail, but merely a succession of conventional Chinese phrases, which are obviously replete with hyperbole, and often fanciful. Thus, we are told that: “At night, spirit lights showed the way, during the day supernatural creatures swam in front”, and that Yang “passed through ten thousand countries”.Footnote 217 There is no certain mention of Guangzhou,Footnote 218 and no detail whatsoever is provided regarding Yang's route. Not a single one of the “ten thousand countries” is actually named. Nor is it specifically stated that Yang reached his destination. Nothing at all is said about his arrival at Baghdad, his reception by the Caliph, or of any results of the embassy. It must be understood that commemorative inscriptions of this kind are eulogistic, and commonly exaggerate the merits and achievements of those commemorated.
The conclusion to be drawn seems obvious. Even if the stele and inscription are genuine, there is no evidence in the inscription that the embassy was ever successfully completed. Perhaps such an embassy was indeed planned, but nothing suggests that it was carried out. This might explain why it is not mentioned in any source other than the inscription. It must be understood that, although he was appointed ambassador in the fourth month, Yang could not have set sail from China before the tenth or eleventh month. He would have needed to wait for the winter monsoon winds. Perhaps he got as far as Guangzhou, but, while making preparations and waiting for the northeast monsoon wind, received notification that the embassy had been cancelled. Perhaps he actually did set sail, but, for some reason, was unable to proceed and returned without completing the voyage. Whatever the case, there is absolutely no evidence that this supposed embassy to the ‘Abbasid Caliphate ever took place. It is out of the question to compare Yang's route with Jia Dan's itinerary, as there is absolutely no information about what Yang's route may have been.Footnote 219 It would certainly be premature to start to rewrite history on the basis of this inscription.
The difficulties involved in disproving false versions of history are indicated by the strange case of Jacob d'Ancona. Like Yang Liangyao, this medieval Jewish merchant was completely unknown, until he made his first public appearance in 1997.Footnote 220 It was claimed that he had travelled from Ancona to Quanzhou as a merchant in 1270–73. This claim was very soon questioned. T. H. Barrett reviewed the account of Jacob's travels more than once,Footnote 221 rejecting it as fake. He was joined by such eminent authorities as Igor de Rachewiltz, Donald D. Leslie,Footnote 222 and Jonathan Spence,Footnote 223 among others. Nevertheless, David Selbourne, who claimed to have seen and translated Jacob‘s manuscript account of his travels, put up a defence against such attacks, and gained support from various quarters, including the curator of the Quanzhou Maritime Museum.Footnote 224 Indeed, although it appears now to be generally accepted that Jacob's account is a forgery, doubts have occasionally continued to be expressed in its favour.Footnote 225 What is important in the present context, however, is that the clearest evidence of forgery has been largely overlooked. It seems that many readers of Jacob's account felt it to be less than authentic, but that they often had difficulty in finding clear evidence to support their suspicions.Footnote 226 Such evidence as was found could, in many cases, be explained as possibly the result of corruption of the original manuscript by a later copyist.Footnote 227
There are, however, several very clear and obvious reasons to believe that Jacob d'Ancona never existed, and that the account of his journey to Quanzhou and back is not merely a forgery, but a modern work. It must have been produced during the twentieth century, almost certainly after 1922. Some of these reasons are relevant in the present context, so it seems worthwhile to examine them here. Firstly, as already pointed out above, ships did not sail eastwards from Arabia during the autumn or winter, but usually during late May or June. Jacob d'Ancona, however, apparently departed from the head of the Persian Gulf on 13 October.Footnote 228 Even more unfeasibly, he reports that he left Sri Lanka on 4 February, sailing “towards the east-north-east”.Footnote 229 This would clearly have been impossible in the face of the prevailing northeast wind of the winter monsoon. Moreover, he set sail from Quanzhou on his return journey on 24 February 1272, “with a favourable north-west wind”,Footnote 230 although this was far too late to set out southwards across the South China Sea. By late February, the winter monsoon wind from the northeast would already have failed. During March and April, the prevailing wind in the South China Sea blows from the east.Footnote 231 Jacob's ship could therefore not have sailed any further than the Gulf of Tonkin, before needing to wait for several months for a wind to carry it further south. Perhaps a copyist changed all the dates, but this would require a substantial rewriting of a large part of the text. In fact, it is unlikely (although perhaps not completely impossible) that Jacob's journey, from Ancona to Quanzhou and back, could have been completed within the space of a little over three years, as claimed.
A fatal flaw is what Jacob says about Champa: “From this place of Sondore [apparently Pulo Condore] we set sail towards the east-north-east, . . . and thus came to a place called Zabai which is in Ciamba in the realm of the Comari. Here they worship the law of Mahomet . . .”.Footnote 232 Leaving aside the facts that “Zabai” was not in Champa,Footnote 233 which was the land of the Chams, not the Khmers (“Comari”), there is absolutely no evidence that the religion of Champa during the 1270s was “the law of Mahomet”. What Jacob says goes much further than Ravaisse's claim that there were Muslims settled in Champa during the eleventh century, who were certainly not the majority of the population. It completely contradicts what Marco Polo reports of the Chams. It must be assumed that the forger based this passage on some exaggerated account of Islam in Champa, which must undoubtedly have been written after the publication of the inscriptions “from Champa” in 1922. Of course, the claim would have been more credible in the 1990s than it is now, after the publication of the articles by Kalus and Abdeljaouad.
There are other indications of the modernity of the fabrication. Many of Jacob's versions of Chinese words show a clear influence of the Wade-Giles system of romanisation. Thus, he writes tentsu for tianzi, “Son of Heaven”Footnote 234 (Wade-Giles t'ien-tsŭ), and ciuntsu for junzi (Wade-Giles chün-tsŭ).Footnote 235 Yet he writes Manci for Manzi,Footnote 236 clearly because the forger failed to realise that the final zi was identical in all three cases. There can be no reason for such a discrepancy, other than that the forger took Manci from Marco Polo (or perhaps from a translation of Rashīd al-Dīn), but drew upon other works for tentsu and ciuntsu. Since the Wade-Giles system of romanisation is of late nineteenth-century origin, the forgery must be modern. It was also very clearly fabricated by someone with no knowledge of Chinese language. Jacob's Ciasuto Footnote 237 is evidently intended for the name of the late Song chief minister, Jia Sidao.Footnote 238 The major problem here is that the initial of the first character was not J- (or Č-) when Jacob was supposedly in Quanzhou. During the thirteenth century, the name would have been pronounced Kya Zɿdaw.Footnote 239 Nor can it be argued that the pronunciation heard by Jacob was local Quanzhou dialect, which would also have had a hard initial K- (as it still has today). Again, the influence of the Wade-Giles system can be discerned in the vowel of the second syllable. My own final judgment of Jacob d'Ancona's The City of Light is that it is an obvious forgery, fabricated by someone with only a very superficial knowledge of Chinese history, culture, and language. Anyone with reasonably good knowledge of thirteenth-century China could not possibly believe that it is genuine.
Despite this judgment, and having read several reviews of The City of Light, I am left with the impression that it would almost certainly have been much more successful, at least with a popular audience, if it had been more readable. Barrett said that he “found insoluble problems every time” he tried to read the entire work,Footnote 240 which I fully understand. Most of it is turgid and tedious, and entirely fails to make captivating reading. Despite this, despite the widespread feeling among scholars that the account did not ring true, and despite the many obvious errors which it contains, there were those who resolutely defended it.Footnote 241 It perhaps came close to being accepted as a historical document relevant to thirteenth-century China. Fortunately, it eventually failed to achieve such acceptance. Jacob d'Ancona and his “City of Light” must be consigned to the dustbin of factitious history, along with the Islamic inscriptions “from Champa”, and very probably also with Yang Liangyao's embassy to the ‘Abbasid Caliphate.
Addendum: An anachronism in the Yang Liangyao inscription
According to the inscription, Yang Liangyao was appointed to the position of Pizhang of the Three Armies of the RightFootnote 242 in the fifth month of the first year of the YongzhenFootnote 243 reign-period (805). He died the following year. At the end of the inscription, it is recorded that it was cut in the tenth month of the first year of the YuanheFootnote 244 reign-period (806).Footnote 245 This is apparently anachronistic, as both the Tang huiyao and the Zizhi tongjian indicate that the title of Pizhang shi,Footnote 246 to give it its full form, was not created until 818.Footnote 247 That the title did not exist until after 806 is confirmed by the Cefu yuangui, which records that: “During the Yuanhe reign-period of the Emperor Xian Zong,Footnote 248 two posts of Palace SecretaryFootnote 249 were created. . . . Later, there were the Pizhang shi of the Three Armies of the Left and of the Right.” The next date mentioned is the fifteenth year of the Yuanhe reign period.Footnote 250 As already noted, the Yuanhe reign-period began in 806. The fifteenth year was its last, equivalent to 820 (ending on 5 February 821). This record is therefore entirely consistent with the creation of the title of Pizhang shi in 818. At any rate, the title did not exist until after 806.
This title appears in a number of Tang-period tomb inscriptions. The abbreviated form Pizhang occurs in at least two. In 820, Xu SuizhongFootnote 251 is reported to have been appointed Pizhang of the Three Armies of the Right (You San Jun Pizhang), exactly like Yang Liangyao.Footnote 252 Liu YinglianFootnote 253 was also appointed Pizhang of the Right (You Pizhang) in 836.Footnote 254 The epitaph of Yang XuanlüeFootnote 255 records that, in 851, he was made Pizhang shi of the Left (Zuo Pizhang shi).Footnote 256 All of these appointments were made after 818. Only Yang Liangyao is supposed to have held the post of Pizhang before this date. An apparent anachronism of this kind of course strengthens the suspicion that the Yang Liangyao inscription is a forgery. It is reasonable to suppose that the text was fabricated using various tomb inscriptions of the Tang period as models. In this case, it seems that the forger(s) failed to realize that the title of Pizhang of the Three Armies of the Right, mentioned in the epitaph of Xu Suizhong, did not exist in 805, when Yang Liangyao was supposedly appointed to the post.
