As I have said a million times by now, there is a placename in Chinese historical sources, Sanfoqi 三佛齊, that in the early twentieth century, French scholar George Cœdès claimed indicated a place called “Srivijaya,” a polity supposedly based at Palembang on the island of Sumatra.
As I have also said a million times, Sanfoqi was not on the island of Sumatra, and in the previous post, I documented exactly where I believe it was – in Cambodia, somewhere around Angkor Borei, with an international port at what is now Hà Tiên.
I think there are a lot of people who do not believe me because there is over a century of scholarship on Srivijaya, and no one has pointed this out before, so how could I possibly be right?
Ok, with that in mind, let’s take a look at some of the information that gets discussed in the scholarship on Srivijaya.
Since Cœdès claimed that the “Sanfoqi” in Chinese sources was “Srivijaya,” then obviously it would be important to look at Chinese sources to see what they say about Sanfoqi.
It was during the Song dynasty period (960-1279) that this name came into use. So, let us take a look at some of the information about Sanfoqi in the History of the Song. Here is what that text records about the first few tribute missions from Sanfoqi.
1) 建隆元年九月，其王悉利胡大霞里檀遣使李遮帝來朝貢。
2) 二年夏，又遣使蒲蔑貢方物。
3) 是冬，其王室利烏耶遣使茶野伽、副使嘉末吒朝貢。
4) 其國號生留，王李犀林男迷日來亦遣使同至貢方物。
5) 三年春，室利烏耶又遣使李麗林、副使李鵶末、判官吒吒璧等來貢，迴，賜以白犛牛尾、白甆器、銀器、錦線鞍轡二副。
6) 開寶四年，遣使李何末以水晶、火油來貢。
7) 五年，又來貢。
8) 七年，又貢象牙、乳香、薔薇水、萬歲棗、褊桃、白沙糖、水晶指環、瑠璃瓶、珊瑚樹。
9) 八年，又遣使蒲陁漢等貢方物，賜以冠帶、器幣。
10) 太平興國五年，其王夏池遣使茶龍眉來。
1) In the ninth lunar month of the first year of the Jianlong era [960], its king, Xi-li Hu-da-xia-li-tan 悉利胡大霞里檀, dispatched the envoy, Li Zhedi 李遮帝, to come and present tribute.
2) In the summer of the second year [961], again an envoy, Pu Mie 蒲蔑, was dispatched to present local goods as tribute.
3) That winter, its king Shi-li Wu-ye 室利烏耶, sent envoy Cha Yejia 茶野伽 and deputy envoy Jia Mozha 嘉末吒 to present tribute.
4) King Li-xi-lin-nan-mi-ri-lai 李犀林男迷日來, of its kingdom called Shengliu 生留, also sent envoys that arrived at the same time to present local goods as tribute.
5) In the third year [962], Shi-li Wu-ye again sent envoy Li Lilin 李麗林, deputy envoy Li Yamo 李鵶末, and official Zha Zhabi 吒吒璧 to come and present tribute. When they returned, they were granted what yak tail, white porcelain, silver utensils, and two sets of golden-sewn saddles and bridles.
6) In the fourth year of the Kaibao era [971], envoy Li Hemo 李何末 was sent and presented quartz and flammable oil [paraffin?] as tribute.
7) In the fifth year [972], they again came with tribute.
8) In the seventh year [974], they also presented ivory, frankincense, rose water, longevity jujubes, flat peaches, white sugar, quartz rings, glass bottles, and coral trees.
9) In the eighth year [975], they again sent envoy Pu Tuohan 蒲陁漢 and others to present tribute. They were granted caps, belts, ritual vessels, and coins.
10) In the fifth year of the Taiping Xinguo era [980], its king, Xia-chi 夏池, sent the envoy, Cha Longmei 茶龍眉.
Ok, so in the first three years that Sanfoqi sent tribute to the Song court, there were no less than five missions from three different “kings.” That is actually not surprising. In fact, the same thing happened later when the Ming first established tributary relations with Sanfoqi, as well as when the Ming first established tributary relations with Ayutthaya.
As I see it, I think what happened in all of these cases is that various powerful people, from kings to merchants, all saw the tributary relationship as a chance to enrich themselves through trade with China, and they all tried to establish a connection.
Let’s see who from Sanfoqi did this at the beginning of the Song dynasty period.
First, we have Xi-li Hu-da-xia-li-tan 悉利胡大霞里檀. The first two characters are “Sri.” That’s easy to see.
Beyond that, another character that I think we can identify is the “xia 霞.” This should represent the sound “ha” as we find this same character in the History of the Song representing the Old Malay title “haji” (xiachi 霞遲) and we see a character with the same sound in #10 above also indicating this same Old Malay title (xiachi 夏池), and no, this was not the same as the Muslim title, “haji,” and yes, there is a reason why there is a Malay title here even though Sanfoqi was in Cambodia (I’ll write about that later).
The “li” 里 that follows it could be “ri.” That would create “hari,” and that’s a sound that we can find in Indic names. Looking through Philip N. Jenner’s A Dictionary of Angkorian Khmer, I found “vyavahāri,” meaning “merchant” or “trader,” and that would fit the “hu-da-xia-li” part well. Chinese doesn’t have a “v” so “hu-da” could be “vyava” and then the next two characters can, as we just saw, easily be “hāri.”
That would leave a final character, “tan” as a given name, to create “Sri Vyavahāri Tan,” meaning something like the “Venerable/Royal Merchant Tan.” I’m obviously not 100% confident in this, but I think that is one possibility.
Whatever this person’s title/name was, he sent a tribute mission in 960 and another in 961.
Then we have “Shi-li Wu-ye” 室利烏耶. This name again begins with “Sri,” but it’s not clear what “Wuye” might be. The “Wuye” is oddly short for the name of a king, but let’s assume that this person was in fact a king. Did Venerable/Royal Merchant Tan work for this man? It’s not clear.
Then things get really interesting. . .
In 961, there was another king who sent an envoy to the Song court, King Li-xi-lin-nan-mi-ri-lai of “its kingdom called Shengliu.”
The wording here is a bit confusing. However, one way to try to understand this is to think of Sanfoqi as a mandala polity. Shengliu was perhaps one of its vassals, and in the same year that a person who was possibly the king of Sanfoqi, Sri Wuye, presented tribute, so did the king of one of Sanfoqi’s vassals, Shengliu.
As for the name of the king of Shengliu, Li-xi-lin-nan-mi-ri-lai 李犀林男迷日來, I have no idea what this is.
In any case, at the most basic level, what we can see here is that the Song received missions from multiple people from Sanfoqi in these first years of contact.
If we put the above information into the setting of the area around Hà Tiên, we can imagine a Song envoy making contact with people in the port, such as the Venerable/Royal Merchant Tan. Did this guy represent a king? It’s unclear, but he did make two missions to China.
Then someone reported to be a king by the name of Sri Wuye sent a mission, as did another king, perhaps one of Sri Wuye’s vassals (or if we consider that such vassal polities were at times ruled over by members of the same royal family, this could have been someone who was related to Sri Wuye). And then Sri Wuye sent another mission after that.
This all makes sense, and it made me wonder how people who have written about Srivijaya discussed this same information, because since George Cœdès claimed in 1918 that the place referred to in Chinese sources as Sanfoqi was Srivijaya, then therefore, historians must have looked at this information carefully.
Or so one would think. . .
In fact, that’s not the case at all. Instead, we can see in the scholarship on Srivijaya that Cœdès repeated incomplete information about this passage in his scholarship and subsequent historians repeated what Cœdès wrote.
Cœdès could not read Chinese. Why anyone would believe a claim that the place in Chinese sources called “Sanfoqi” was “Srivijaya” that was made by a man who could not read those sources is a mystery to me, but that’s another issue.
Not knowing Chinese, Cœdès had to rely on the work of others, and the person he relied on for his understanding of what was written in the History of the Song was Dutch Sinologist W. P. Groeneveldt who translated parts of that text in his 1876 Notes on the Malay Archipelago and Malacca Compiled from Chinese Sources.
This is how Groeneveldt translated the above passage:
In the 9th month of the year 960 Si-ri-hu-ta-hia-li-t’an sent an envoy to bring tribute, which he repeated in the summer of the next year. In the winter of the same year tribute was offered by a king of the name Si-ri-wu-ya.
In the spring of the year 962 the king Si-ri-wu-ya sent an embassy of three envoys to bring tribute. They got back tails of the Yak (Bos grunniens or poëphagus), white porcelain, silver utensils, silk thread and two sets of saddle and bridle.
In the year 971 one of the former envoys was sent to present crystals and lamp-oil; in the next year he came again and in 974 they brought as tribute ivory, olibanum, rosewater, dates and flat peaches, white sugar, crystal finger-rings, glass bottles and coral trees. The next year new envoys came who were presented with caps and girdles.
In the year 980 they king Ha-ch’i (Hadji or Adji) sent an envoy. . . [64]
This is not a “translation.” This is a “summary translation” as there is information that Groeneveldt omitted.
In summarizing this passage, Groeneveldt decided to leave out things like the names of the envoys, as well as the names of that other kingdom, Shengliu, and its king.
As I said, the wording in that passage is a little confusing, and perhaps Groeneveldt couldn’t figure it out and decided to just leave it out. Ideally, he should have indicated that he was skipping over some information (. . .), but he didn’t.
That is why it is always essential that historians CHECK THE SOURCES. Historians should never just believe what someone else has written.
Cœdès, however, had no ability to check the Chinese sources, so he just repeated what Groeneveldt had written, and this is what Cœdès wrote in a work that was first published in French and then later came out in English translation as The Indianized States of Southeast Asia:
Among the Śrivijayan kings, the History of the Sung tells us of Si-li Hu-ta-hsia-li-tan in 960 and of Shih-li Wu-yeh in 962; both these names are probably transcriptions of the same name, Sri Udayäditya(varman). The accounts of embassies to China in 971, 972, 974, and 975 do not give any king’s name; embassies of 980 and 983 are said to come from a King Hsia-ch’ih, in Malay Haji, which is simply a royal title. It was during the reign of this king in 983 that “the priest Fa-yü, returning from India where he had been seeking sacred books, arrived at San-fo-ch’i and there met the Indian priest Mi-mo-lo-shih-li [i.e., Vimalaśri], who after a short conversation entrusted him with a petition in which he expressed the desire to go to the Middle Kingdom and translate sacred books there.” [131-132]
Cœdès got the idea from a scholar by the name of J. L. Moens that the names of the first two kings, Xi-li Hu-da-xia-li-tan 悉利胡大霞里檀 and Shi-li Wu-ye 室利烏耶, were both the same name: Udayäditya(varman).
There’s just no way that those two terms can be made to render the name “Udayäditya(varman).” However, it’s important to point out here that I think Cœdès wanted to see one king. He wanted those two very different names to be referring to the same person. Why is that?
This is something you can see throughout his work. Cœdès viewed the world in terms of single kingdoms with single genealogical lines of kings, and he worked very hard in all of his research to try to get the historical information from the Southeast Asian past to fit that model.
However, not all historical information does fit that model, and in fact, this very passage provides a perfect example of when that model does not work, because it talks about simultaneous tribute coming from Sanfoqi and “its kingdom called Shengliu.”
But even if Cœdès could have moved beyond his intellectual biases, he would not have been able to know about this, because he relied on Groeneveldt’s translation, and Groeneveldt didn’t translate that part, and Cœdès could not read the original Chinese text.
So Groeneveldt’s incomplete translation went on to serve Cœdès’s flawed ideas, and then. . .
And then I was wondering if anyone has looked at this information in the History of the Song more recently. There was a book that was published in 2009 called Nagapattinam to Suvarnadwipa: Reflections on the Chola Naval Expeditions to Southeast Asia. A place called Srivijaya/Srivisaya was attacked by the Chola kingdom in 1025, and this volume theoretically should constitute our most updated knowledge about that topic.
That Chola attack took place during the time of the Song dynasty, so I thought that surely someone in that volume must have checked to see what the History of the Song had to say about Sanfoqi, the place that people who believe in Cœdès’s Srivijaya narrative think was the Chinese name for Srivijaya.
As it turns out, I only found one reference, and this is what it said:
Records from the Song dynasty period say that in AD 960 the Sriwijaya (San-fo-qi) king was Se-li Hu-ta-hsia, and in 962, She-li Wu-yeh. Both names may be equated to Sri Udayadityawarman. In the years 971, 972, 974, and 975, several envoys were sent to China, but they never mentioned the king’s names. However, envoys from 980 and 983 said that their king’s name was Hsia-She. In 983 Chinese priest Fa-Yu visited San-fo-qi on his return from India where he had studied holy books. In this country he met the Indian priest Mi-mo-lo-shi-li (Vimalasri) who wished to travel to China to translate holy books (Coedès 1968, pp. 131–32; Soemadio, ed. 1984, p. 66). [229]
Ok, I know people are debating a lot these days about what exactly constitutes plagiarism. . . so let’s not go in that direction. Instead, we can simply note that this author completely repeated what Cœdès had written roughly a half century earlier. . .
Wow! Think about that. 133 years after Groeneveldt made his incomplete translation of the information about Sanfoqi in the History of the Song, that same information is still getting repeated, and the parts he left out, continue to be invisible to readers.
Let that sink in.
People don’t check the sources.
In fact, most of the people who have worked on the topic of Srivijaya can’t check the sources because they can’t read either Old Malay (for the inscriptions) or classical Chinese (for information in Chinese texts).
I have gotten plenty of pushback or dismissal regarding my ideas over the past few years. However, look at the extant scholarship on Srivijaya. Who has actually ever gone back and looked closely at the Chinese sources that Cœdès claimed were about Srivijaya?
O. W. Wolters wrote two books that deal with Srivijaya, but he did not talk about the period that coincides with the Song dynasty in those books. Why not? That’s when the most information about Sanfoqi was recorded.
Wouldn’t it make sense that if you knew certain historical sources were about a topic that you were interested in, that you would go through those sources carefully and document what they record?
That makes sense to me. So, why do we still Groeneveldt in scholarship in the twenty first century?
That is another mystery to me.
However, hopefully this post helps people see why there are major problems with the scholarship on Srivijaya, and why it is possible to completely transform that scholarship by doing the one thing that historians are supposed to do, but in the case of the scholarship on Srivijaya, have not done – CHECK THE SOURCES.
This Post Has 5 Comments
On Shengliu see J. Kurz, “(Mis)Reading Chinese Texts in Southeast Asian History: ‘Moliu’, Duoluomo and the Lure pf ‘Srivijayan’ Conquests”, Journal of Asian History 55.1 (2021): 67-88.
I am not convinced that Groeneveldt deliberately left out certain passages from his translation. I have noticed him missing text in other cases too, so he may have had a defective copy of the Song history (perhaps from a commercial book seller/printer). The two modern editions of the Songshi most popular are the Bonaben edition of the 1930s (Commercial Press, Shanghai) and the one published by Zhonghua shuju in 1977 (which, as I understand, Piet van der Loon did not hold in high regard).
Thanks for letting me know, Johannes, but I only have access to this journal up to 2018.
As for Groeneveldt, that’s possible, and I’ve considered that. However, it’s a little odd that in the case of Sanfoqi he also doesn’t include the statement at the beginning that it was “neighboring Champa” (與占城為隣). He believes Sanfoqi was on Sumatra, and for some reason a statement that contradicts that is left out. So was his version of the text defective? Or did he leave out parts that did not fit his understanding of what the text should be saying (or parts he couldn’t figure out)? I don’t know. I suppose one could identify all of the instances and try to see if there is some kind of pattern.
Also, I just saw your comments on my working paper last night. Thanks for that. I will go through it closely when I return to that topic, but I definitely agree that there’s more to the Java issue, and that I simplify it too much in that working paper. I don’t know if you saw Andrew Chittick’s recent translations (https://storymaps.arcgis.com/stories/987f5e0085c04ec998dc76d2dcced5de), but that also got me thinking about that issue too.
Thank you for the link to the Chittick site, I have not been aware of it previously. As for Groeneveldt, it comes down to the texts he had. During his time there was no talk of ‘Srivijaya’ yet, so he would not have had a reason to skip passages in the text. The problem is not so much Groeneveldt, but scholars relying on his limited textual material, and having no idea where relevant information came from. The phrase you noticed missing from his translation of the Songshi (a 14th century compilation) – 與占城為隣 – for instance, is found in at least four Song dynasty texts dating to the 1080s, 1101, early 12th cent., 1210, respectively, predating the Songshi by more than a hundred years. In contrast, the Taiping huanyu ji of the late 10th century has nothing on Sanfoqi.
Scholars of Southeast Asian history still refer to Groeneveldt because that is the one translation available in English. The French translation of the description of Sanfoqi in the Wenxian tongkao (early 14th cent.) by Saint-Denys (1883) includes the phrase missing from Groeneveldt’s translation, but scholars rarely if ever refer/red to it. Incomplete knowledge of the Chinese source material and knowledge of modern languages – in addition to English – appear to be the key issues with the study of early Southeast Asian history.
Hi Liam, thank you for your intriguing theory regarding the location of Srivijaya. Have you taken into account other sources of evidence (inscriptional and archaeological) that point to Sumatra and Palembang as the location of Srivijaya? Here I’m thinking in particular of John Miksic’s paper, “Archaeological Evidence for Esoteric Buddhism in Sumatra, 7th to 13th C.” (in ‘Esoteric Buddhism in Mediaeval Maritime Asia’, ed. Andrea Acri, ISEAS 2016.).
Hello Dominic, thanks for the comment!!
Actually, that article perfectly demonstrates the point I have been making. Take a look at what Miksic says on the first page:
“Since George Cœdès first identified this kingdom in 1918, Śrīvijaya has been envisioned by many scholars as a major thalassocracy which dominated the Strait of Melaka from the late 7th to the late 13th century. Archaeologists and art historians have been less confident about the nature and geographical extent of the kingdom.”
So, what Miskic says here, is that there is a disconnect between what historians say and what the material evidence demonstrates.
That the material evidence does not support the ideas that historians have developed from written sources is because, as I have been showing for a few years now, the historical sources are not about a place on Sumatra at Palembang.
When you take away the historical sources, you are left with very little evidence for anything. There are some inscriptions that mention “Srivijaya” from a limited time period (~one century) and a limited geographic range (Nakhon Si Thammarat – Palembang/Bangka).
We also have archaeological evidence of things like Esoteric Buddhism on Sumatra (although as Miksic points out, no major Buddhist structures at Palembang).
That extremely limited evidence does not come anywhere close to demonstrating the existence of “a major thalassocracy which dominated the Strait of Melaka from the late 7th to the late 13th century.”
The idea for this “major thalassocracy” has been developed from historical sources, particularly Chinese historical sources, and as I have been showing over and over and over, those sources are not about Sumatra. And that should not surprise us, because Coedes could not read those sources, and therefore, did not have a solid understanding of what he was talking about.
One problem that I think a lot of people have is that they have been talking about Srivijaya for so long that they no longer are aware of what information comes from (Chinese) historical sources and what information comes from material evidence.
So, for instance, you can find Miksic in this article matter-of-factly saying things like the following:
“In Kedah, one of the bastions of the Buddhist kingdom of Śrīvijaya”
“Chinese and Indians viewed Śrīvijaya as an active participant in the evolution of Buddhist doctrines and rituals.”
“The first person known to have written the name Śrīvijaya (as Śrībhoja) was a Chinese Buddhist monk named Yijing.”
However, these are all ideas that come from historical sources that do not talk about a place called “Srivijaya” (Even if you want to render Shilifoshi as “Sribhoja,” that’s still not “Srivijaya,” and there is no way to render Sanfoqi to anything even close to “Srivijaya”) or the island of Sumatra (it’s easy to demonstrate that both Shilifoshi and Sanfoqi were not on Sumatra).
That is what I have been demonstrating. However, for people who cannot read the historical sources, I think it is difficult for them to realize what parts of their knowledge come from the historical sources that Coedes claimed were about Palembang but which actually are not.
For such people, I think the easiest thing is do is to look at the topic from the other direction. Identify the material evidence: a few inscriptions from one century that mention “Srivijaya” and are in two locations, and various artifacts across Sumatra and the Malay Peninsula.
What story can we develop from that evidence? . . .
In the end, I think that there is historical evidence which corresponds with the place mentioned in those inscriptions, and I think that place was centered around Lake Songkhla. And I talk about that in this working paper:
https://ias.ubd.edu.bn/wp-content/uploads/2023/06/working_paper_series_73.pdf