[August 2022 update: I don’t believe what I wrote here anymore. I’m now convinced that Shilifoshi referred to a place in what is now southern Thailand around Surat Thani province and that this polity has “tributaries” on the western side of the Malay Peninsula.
I develop that idea here and here.]
Srivijaya has always bothered me.
The story goes that in 1918 French scholar George Coedès “discovered” that there had been an “empire” by this name in island Southeast Asia (from roughly the 7th to 13th centuries) based in the area of Palembang on the island of Sumatra when he determined that the name “Srivjaya” in some inscriptions from Sumatra and the Malay Peninsula (in an area that is now in Thailand) was referring to a place that is recorded in Chinese sources as “Shilifoshi” and “Sanfoqi.”
And with that, Voilà! The “empire” of Srivijaya, with its “capital” at Palembang on the island of Sumatra, became a topic of scholarly research.
Over the past century, much has been written about Srivijaya. Scholars like Hermann Kulke have produced excellent studies based on information in inscriptions that temper some of the early claims of Coedès and see Srivijaya as a loosely-structured “kadatuan” with a ruler in a “kraton” or palace in a city claiming nominal sovereignty over surrounding polities rather than an empire.
That makes sense to me, but what has never made sense to me is the statement that gets repeated over and over, which is that “the Chinese referred to Srivijaya as Shilifoshi or Sanfoqi.”
Shilifoshi 室利佛逝 and Sanfoqi 三佛齊 are NOT Srivijaya. There is absolutely no way on earth that either of those names can possibly be interpreted to signify the term “Srivijaya.”
As obvious as that is, I have not seen any scholar ever address that issue. Instead, scholars just repeat the claim made by Coedès in 1918 that “the Chinese referred to Srivijaya as Shilifoshi or Sanfoqi” and then move on.
That the terms “Shilifoshi” and “Sanfoqi” have not been investigated is due in part to the fact that many of the scholars who have worked on Srivijaya have not possessed adequate knowledge of classical Chinese, and have therefore just had to assume that Coedès was correct.
There are readers who will counter this claim by stating: “But wait! O. W. Wolters wrote about Srivijaya. He knew Chinese!”
To which I would respond: “Not really!” I pointed this out long ago in an article (pgs. 322-330), and I will demonstrate this again below.
Terms like Shilifoshi and Sanfoqi beg to be examined from a Sinological perspective, and that is what I propose to do in this post. I have been inspired to do this after attending a presentation rich in information by my colleague Johannes Kurz where the early writings on Srivijaya were subject to his critical Sinological analysis, and needless to say, Johannes found much to question.
One of the points that Johannes brought up was that scholars have tended to speculate about the meanings of Chinese terms for Southeast Asian places based on modern Chinese pronunciations. This can lead to misunderstandings. The Sinological approach would be to employ the reconstructed pronunciations of the era being examined by referencing a work like Edwin G. Pulleyblank’s Lexicon of Reconstructed Pronunciation: In Early Middle Chinese, Late Middle Chinese and Early Mandarin.
I completely agree, but in the case of some of the names in Chinese for places in Southeast Asia, I would argue that there is another approach that is fruitful to investigate.
In this post, I will examine the first of the two Chinese terms that scholars argue indicate Srivijaya – Shilifoshi 室利佛逝. This term appears in the writings of the Chinese Buddhist monk Yijing 義淨, a man who traveled from Tang Dynasty China to India in the late seventh century and who spent 11 years in Nalanda studying and translating Buddhist texts into Chinese.
In journeying to India, Yijing spent a few months in Shilifoshi, and then he spent at least a couple of years there (~689) on his return journey.
We can see from Yijing’s writings that Shilifoshi was a center of Buddhist learning and that there were numerous Chinese monks who lived there, as well as monks from Korea (Silla) and Vietnam (Giao Châu).
What is more, we can also see that at least some of these monks were multilingual. They knew Sanskrit (梵語) and some of these men were also conversant in the “Kunlun language” (崑崙語 or 崑崙音), a reference to the local language in Shilifoshi, which I think it is safe to assume must have been an Austronesian language.
Given this context, and given that these men were devoting years of their lives to studying and translating Sanskrit texts, it makes sense to look at the place names that Yijing documented through the prism of his, and his colleagues’, linguistic abilities.
To do this, I have examined the names that Yijing recorded, with the words that use the same characters in A Dictionary of Chinese Buddhist Terms by William Edward Soothill and Lewis Hodous (1937), to gain a sense of how Sanskrit terms were translated into Chinese, as I will argue below that Shilifoshi is a Chinese rendering of a Sanskrit term.
As Charles Muller explains in his introduction to the online version of this text, this dictionary has its limitations. Nonetheless, I think the exercise I engage in below by employing this text can point us in a new direction in our thinking about “Srivijaya.”
Yijing mentioned Shilifoshi in two books that he wrote, A Record of the Inner Law Sent Home from the South Seas (Nanhai jigui neifa zhuan 南海寄歸內法傳) and The Great Tang Biographies of Eminent Monks who Sought the Dharma in the Western Regions (Da Tang Xiyu qiufa gaoseng zhuan 大唐西域求法高僧傳).
He also mentioned Shilifoshi in an annotation to a Sanskrit text that he translated, the Mulasarvastivada-ekottarakarmasataka 根本說一切有部百一羯磨.
In The Great Tang Biographies of Eminent Monks who Sought the Dharma in the Western Regions, Yijing mentions Shilifoshi together with some other place names at one point when he talks about the itinerary from China to India in a section about a monk named Zen Master Wuxing (無行禪師).
Examining these names is a good way to begin our investigation of Shilifoshi. I’ll transcribe the place names that Yijing mentions here first in modern Mandarin and then we’ll examine them in other ways below.
Yijing states that it took Zen Master Wuxing about a month to reach Shilifoshi from China. He then traveled another 15 days to Moluoyu 末羅瑜. After that, he journeyed an additional 15 days to Jiecha 羯茶, and then roughly another 30 to India.
Let’s start by looking at the second place mentioned here: “Moluoyu.” That word sounds a lot like “Melayu,” a term that now refers to people and things “Malay.”
However, if we look at how these characters were used in transcribing Sanskrit terms, we find that the first two, 末羅, were used to transcribe the sounds “mala” and “mara.” The final character, meanwhile, was used to indicate the “yo” in “yoga” 瑜伽.
So, from a Chinese Buddhist perspective, it is easy to see 末羅瑜 as indicating “Malayo” or “Marayo,” which is strikingly similar to the “Muaro” in the name “Muaro Jambi,” a famous historical Buddhist site in southern Sumatra.
Further, if we look at Sanskrit words that the characters 羯 and 茶 are used to transcribe, it is easy to see that 羯茶 refers to “Kada,” that is, “Kedah,” another early Buddhist center in Southeast Asia, located on the western side of the Malay Peninsula.
So let’s now turn to Shilifoshi. Unlike these other two terms, Shilifoshi clearly is not a transcription of the sounds of a local name.
First of all, “Shili” is unequivocally the Sanskrit honorific term “sri,” meaning “great” or “prosperous” or “auspicious,” etc. Sometimes Yijing uses that term, while at others he simply refers to this place by the final two characters, “foshi.”
As for “foshi” 佛逝, the first of these two characters is the character used to refer to the Buddha 佛.
When Chinese Buddhist translated Sanskrit texts, they generally did not use this character to represent sounds. It was almost exclusively used for terms relating to the Buddha.
There are a couple of examples in Soothill’s dictionary where it is used to represent the sound “pur” at the beginning of a word (where it is interchangeable with the similar character of 弗). However, the vast majority of terms that contain this character do so with the meaning of “the Buddha” and are used to represent what is written in Sanskrit as “Budh” or “Bodh” or “Buddha.”
Finally, there is shi 逝. This is used at times to represent the sounds “je” and “ja” in Sanskrit.
These ways in which Chinese monks historically transliterated Sanskrit terms were undoubtedly what led Japanese scholar Junjirō Takakusu to render “Shilifoshi” as Sribhoga, when he translated Yijing’s A Record of the Inner Law Sent Home from the South Seas into English in 1896.
That was a good guess, but I think that there is another way to read Shilifoshi 室利佛逝.
When Chinese monks translated Sanskrit texts, sometimes they used Chinese characters to transliterate Sanskrit consonant-vowel sounds. At other times, the meaning of Sanskrit terms was translated through the meaning of Chinese characters.
So, for instance, the two characters, shanshi 善逝 (note that the second character here is the same as the final character in Shilifoshi), were used to represent the Sanskrit term “sugata.” These two characters literally mean “good” and “to pass away,” respectively, and they thus fit the meaning of the Sanskrit term “sugata,” where “su” is a prefix that means “good” and “gata” is the past participle of the verb “to go.”
Hence, sugata/善逝 means something like “the well-departed,” and it is a term that was used to refer to the Buddha.
So the character 逝 could be used to transliterate sounds like “je” and “ja,” but it could also be used to translate the term “gata.”
At other times, transliterations and translations were combined. We can see this, for instance, in one of the ways that the term “Jetavana” was rendered into Chinese. Jetavana was a forested area that was donated to the Buddha by a man named “Jeta” and it became the site of an important monastery during the Buddha’s lifetime. “Jetavana” literally means something like “Jeta’s forest” and it is also referred to in English as a “grove” or a “garden.”
In rendering this term into Chinese, sometimes the entire word was transliterated, while at other times part of it was transliterated and part of it was translated, thereby creating 逝多林. Here the first two characters (逝多) represent the sounds “je” and “ta” (with the first character being the same one that we’ve been talking about here, that is, the final character in Shilifoshi), while the third character (林) is the character for “forest.”
With all of this in mind, I would argue that Shilifoshi 室利佛逝 could be read from a seventh-century Chinese Buddhist perspective in a couple of different ways.
First, it could be read as Sri Budhga[ta], meaning Sri Buddha-gata, or the “Great Buddha’s Departure.” I don’t really think this is likely, however, as there are established terms for referring to the Buddha’s departure, and this is not one of them.
Second, it could be read as Sri Budhje[ta], meaning Sri Buddha-Jetavana, or the “Great Buddha’s Jetavana [monastery/garden].”
This, in any case, is something that I was wondering about, but then I came across a reference to Shilifoshi in an annotation to Yijing’s translation of the Mulasarvastivada-ekottarakarmasataka where Yijing makes the meaning of this term clear.
Yijing mentions Shilifoshi three times in this annotation. In each case, he omits the first two characters, those that transliterate “sri.” This is something he does in other places in his writings as well.
In other words, Sri Budhje could just be referred to as “Budhje,” and Yijing uses this term two times in this annotation. The third time, however, he adds an additional character to the name, one that is pronounced “ta” (多), to create 佛逝多 or “Budh-je-ta.”
Not surprisingly, this “ta” is the exact same “ta” that is combined with this exact same “je” in the term “Je-ta-vana” (逝多林).
Therefore, Yijing made it 100% clear here that this term is meant to indicate “Budhjeta” with the meaning of “the Buddha’s Jetavana [monastery].”
“Shilifoshi” was therefore definitely not “Srivijaya.” Instead, it was “Sri Budhjeta.” But what on earth was Sri Budhjeta?
Yijing makes a couple of other interesting points in this annotation. First, he says that Kedah “belonged to” Budhjeta (屬佛逝), and he states that Muaro “is now the polity [‘guo’ – more on this below] of Budhjeta already” (今為佛逝多國矣).
Then in the same annotation in his The Great Tang Biographies of Eminent Monks who Sought the Dharma in the Western Regions that we have just discussed, Yijing states of Muaro that “It has now changed to Sri Budhjeta” (今改為室利佛逝也).
These comments appear to have been added during Yijing’s return journey, as they note changes that have taken place, and these comments have led various scholars to argue that this is a reference to the expansion of Srivijaya, since to them (and thanks to Coedès) Shilifoshi = Srivijaya.
However, I have hopefully made it clear that we cannot equate Shilifoshi with a place called Srivijaya (and I’m not sure that such a place ever existed). Nonetheless, what we can see though is that something changed during the years that Yijing was in India, such that he needed to annotate the information that he had recorded.
What exactly had happened?
The statement that Kedah belonged to Budhjeta is written in the political language of a suzerain-vassal relationship. In particular, this character meaning “to belong” (屬) means to belong in the capacity of a vassal state.
However, there is nothing political about the statements regarded Muaro’s change of status. In stating that Muaro “is now Budhjeta” or that it “is now changed to Budhjeta,” Yijing gave no sense of military conquest or political expansion or anything like that.
In the information that Yijing provided about the itinerary of Zen Master Wuxing’s journey to India, it would appear that originally Sri Budhjeta was somewhere up the eastern coast of the Malay Peninsula, as it was an equal distance from Muaro as Kedah was (each was a journey of 15 days away).
If Muaro’s change in status was because of military conquest, then it would mean that someone went from the area of around Talat Chaiya or Nakhon Si Thammarat in what is now southern Thailand down to Muaro on the island of Sumatra, conquered it, transferred a capital there, and then made Kedah a vassal.
I suppose that’s possible, however, Yijing’s wording does not suggest any of this at all.
Meanwhile, the wording of the term “Sri Budhjeta” suggests another explanation. I would argue that Sri Budhjeta was perhaps the name of a monastery, just as Jetavana was one of the most famous Buddhist monasteries in India. When Yijing first traveled to India, Sri Budhjeta was located somewhere on the eastern coast of the Malay Peninsula. When he returned more than a decade later, it had relocated to the island of Sumatra.
Alternately, and this is the explanation that strikes me as the most realistic, it could have been more of an honorific title that referred to a place of Buddhist learning – “The Buddha’s Jetavana [garden].”
Whatever it was, how did it move? It could have been due to the patronage of the ruler of Muaro. The ruler may have been seeking to enhance his power, and in the process, he may have somehow brought Kedah under his nominal suzerainty. Enticing Buddhist monks to take up residence in his capital may have been another aspect of this attempt to enhance the power and prestige of his realm.
Sri Budhjeta was thus probably the name of a monastery or a center of Buddhist learning rather than a “kingdom” or “empire.” However, Yijing did at times use the term “guo” 國 (“kingdom”) in referring to Sri Budhjeta. So how do we explain that?
This is where we get to O. W. Wolters. . . In 1986, Wolters published an article in the journal Indonesia entitled “Restudying Some Chinese Writings on Sriwijaya” in which he looked at the same sources that I am discussing here.
These sources employ different political/spacial terms for Kedah, Muaro, and Sri Budhjeta – namely, “guo 國” (“polity/kingdom”) and “zhou 洲” (“islet”) – and Wolters analyzed these terms in an effort to determine what the political relationships were between these places.
Here the annotation in the Mulasarvastivada-ekottarakarmasataka, contains relevant information. It says that “Kedah guo” belongs to Budhjeta. It also says that “Muaro zhou” is now “Budhjeta guo.”
So what the heck does that mean?
Wolters spilled a lot of ink trying to differentiate the territorial and political meanings of these terms, guo and zhou, but I would argue that none of that is really necessary.
While “guo” can mean “kingdom” it can also just refer to the “important part” of a kingdom – its walled center. Indeed, at one point in his annotation, Yijing refers to Sri Budhjeta as a 廓, a term that is also pronounced “guo” and which, as an annotator has indicated, is a variant of 郭, a character that refers to a wall surrounding a citadel, meaning a secondary containing wall.
Wolters translates this term as “suburbs” and has monks living in the suburbs. . . However, the expression in the text just indicates that they were living within the city walls (郭下).
As for zhou, it literally means “islet,” but I think that in this context we can understand it as indicating a settlement along the water without walls. Therefore, as I see it, what Yijing is recording are simply the differences between kampongs (zhou 洲) and fortified settlements (guo 國/廓/郭).
As such, what Yijing is saying here is that in the years since his previous visit, Muaro had built itself up from a kampong into a fortified settlement, and its ruler had also enticed the monks from the Budhjeta monastery or center of Buddhist learning to relocate there. In its broadest sense, “Budhjeta guo” was, therefore, “the fortified settlement of the Buddha’s Jetavana monastery/garden” (佛逝多國).
For this to make sense, however, one has to understand that 佛逝多 (Budhjeta) is a three-character rendering of 佛逝 (Budhje), and that both terms refer to the concept of “Buddha-Jetavana.”
Wolters did not understand this. He, therefore, translated the passage that contains the three-character rendering of Budhjeta as “Malayu [what I am referring to as Muaro] Zhou has now become one of Sriwijaya’s [what I am calling Bhudhjeta] many guo” (末羅遊洲今為佛逝多國矣).
As mentioned above, the character that Yijing added to “Budhje” (ta 多), so that it more clearly reads as “Budhjeta,” was used in translating Sanskrit texts to indicate the sound “ta.” However, its literal meaning in Chinese is “many” or “a lot” (duo 多).
As such, there are two characters next to each other in this passage (the final character of Budhje-ta [多] and the following character which means guo/kingdom[國]) which, when taken out of that context and read literally on their own together, mean “many guo/kingdoms” (多國).
This is how Wolters read this passage (see the image above to visualize this). He didn’t understand that “ta 多” was transliterating the final character of the name Budhjeta. Instead, he put it together with the following character, “guo 國,” and read it for its literal meaning (“many”), and thereby came up with the idea that the text was saying something about “many guo/kingdoms.”
That said, to get this passage to say something about “many guo/kingdoms” still requires inserting a lot of information that is not in the text. That should have been a sign to Wolters that he was heading in the wrong direction.
For instance, there is nothing indicating “one of” in this passage. To read this passage the way Wolters did, the characters indicating “many” and “kingdom” would come out sounding like a compound, so the sentence would read something like “Malayu Zhou is Sriwijaya’s manykingdom” and “manykingdom” is not a word.
So Wolters translation doesn’t make any sense. But just to verify this, I searched through a database of classical Chinese texts for “多國” (“many guo/kingdoms”) to see if such a term did ever exist, and I found basically no evidence of that expression ever having been used (see the image above).
One can find it in Buddhist texts where it appears in contexts like the one here, that is, where “ta 多” is the final character of a kingdom’s name. And the database also brought up numerous returns in the Korean chronicles, but it did so simply because there were many cases were 多 was the final character in one sentence and 國 the first character in the next sentence (ex: 亦且不多，國家備禦之策).
By contrast, I also did a search for “zhuguo” 諸國, a term that means “multiple kingdoms” or “various kingdoms,” and a term that I’ve come across frequently in classical Chinese writings, and sure enough, there were many instances of this term in the database.
It is obvious that Wolters relied heavily for his understanding of this passage on an existing translation (always a bad idea!) in the above-mentioned work by Junjirō Takakusu. In his introduction to Yijing’s A Record of the Inner Law Sent Home from the South Seas, Takakusu translates this passage.
Takakusu was obviously also confused by this passage, and translated it as, “. . . we come after a month to the country of Malayu, which has now become Bhoga; there are many states (under it).”
As mentioned above, Takakusu translated Shilifoshi as “Sribhoga” (or “Bhoga” when it appears without “Sri”), which I commend him for, as that is a more accurate guess than Coedès and Wolters’ Sriwijaya.
However, because he did not understand the real meaning of this term, he did not realize that the character “多” was part of the name, and tried to read it literally as “many,” as Wolters later did, which resulted in the disconnected phrase “there are many states (under it).” Further, as was the case with Wolters’ translation, Takakusu had to inject information into the text that simply is not there (“there are” and “under it”) in order for his translation to work.
As far as I know, this is the first time that anyone has ever seriously investigated the term “Shilifoshi.”
For over 100 years, so many scholars have taken Coedès at his word that Shilifoshi = Srivijaya. And more recently scholars have taken Wolters at his word that “Malayu Zhou” became “one of Sriwijaya’s many guo.”
However, none of that is true.
What is true is that there was a Sri Budhjeta in Southeast Asia. And that is something that I don’t think anyone has ever realized before. . . except of course for Yijing.
After all, he wrote about it. 🙂
This Post Has 5 Comments
Using modern Mandarin pronunciations to figure out the Middle Chinese transliteration of Sanskrit is probably not a good idea, I’m afraid.
Thanks!! I agree. That’s why I didn’t do that.
eightroomofelixir probably meant instead of rendering 國 & 郭 as simply “guo” (which is apparently a Standard Mandarin pronunciation), you should add a Middle Chinese pronunciation alongside.
Interesting topic. Did you already read this?https://storymaps.arcgis.com/stories/39bce63e4e0642d3abce6c24db470760
what do you think about that translation?
Thanks for the question. Yes, I’m familiar with that website. What do I think of it? Well, the translator did a good job of explaining that text based on the extant scholarly knowledge, but the problem is that there are countless problems with the extant scholarly knowledge, so. . . in the end it’s unfortunately not an accurate or reliable translation.