The Mystery of Fu Lin

1. Introduction: The Rome That Wasn’t?

In Chinese historiography, “Fu Lin” (拂菻) is the name to be given to a great western empire that succeeded “Da Qin” during the Tang and Song dynasties (c. 600-1200 AD). While modern scholars generally identify Fu Lin as the Byzantine Empire (Eastern Roman Empire), the primary sources contain glaring contradictions that make a simple identification impossible.

“Da Qin” (大秦)

“大秦國一號犂靬，在安息、條支西大海之西，從安息界安谷城乘船，直截海西，遇風利二月到，風遲或一歲，無風或三歲。其國在海西，故俗謂之海西。有河出其國，西又有大海。海西有遲散城，從國下直北至烏丹城，西南又渡一河，乘船一日乃過。西南又渡一河，一日乃過。凡有大都三，却從安谷城陸道直北行之海北，復直西行之海西，復直南行經之烏遲散城，渡一河，乘船一日乃過。周回繞海，凡當渡大海六日乃到其國。……自葱領西，此國最大，置諸小王甚多，故錄其屬大者矣。”

——《三國志·魏書》三十 倭人傳

It has been believed that “Da Qin” was a reference to the Roman Empire.

TL;DR, take me to conclusion!

2. The Primary Documents

A. The First Embassy to Tang

拂菻國，一名大秦，在西海之上，東南與波斯接，地方萬餘里，列城四百，邑居連屬。其宮宇柱櫳，多以水精琉璃為之。有貴臣十二人共治國政，常使一人將囊隨王車，百姓有事者，即以書投囊中，王還宮省發，理其枉直。其王無常人，簡賢者而立之…
貞觀十七年，拂菻王波多力遣使獻赤玻璃、綠金精等物，太宗降璽書答慰，賜以綾綺焉。自大食強盛，漸陵諸國，乃遣大將軍摩栧伐其都城，因約為和好，請每歲輸之金帛，遂臣屬大食焉。乾封二年，遣使獻底也伽。大足元年，復遣使來朝。開元七年正月，其主遣吐火羅大首領獻獅子、羚羊各二。不數月，又遣大德僧來朝貢。
——《舊唐書·列傳》第一百四十八 西戎

貞觀十七年，王波多力遣使獻赤玻黎、綠金精，下詔答賚。大食稍強，遣大將軍摩拽伐之，拂菻約和，遂臣屬。乾封至大足，再朝獻。開元七年，因吐火羅大酋獻師子、羚羊。
—— 《新唐書·列傳》第一百四十六下 西域下

Important points:

• In 643 AD (貞觀十七年), the King of Fu Lin sent envoys to Tang.
• Fu Lin submitted to Da Shi (Arabs) after a siege on their capital.
• Another envoy was sent in 667 AD (乾封二年).
• Another envoy was further sent in 701 AD (大足元年).
• In 719 AD (開元七年), the King of Fu Lin sent Great Chief of the Tuhuoluo¹.

B. The Second Embassy to Song

拂菻國東南至滅力沙，北至海，皆四十程。西至海三十程。東自西大食及于闐、回紇、青唐，乃抵中國。歷代未嘗朝貢。

元豐四年十月，其王滅力伊靈改撒始遣大首領你廝都令廝孟判來獻鞍馬、刀劍、真珠，言其國地甚寒，土屋無瓦。產金、銀、珠、西錦、牛、羊、馬、獨峯駝、梨、杏、千年棗、巴欖、粟、麥，以蒲萄釀酒。樂有箜篌、壺琴、小篳篥、偏鼓。王服紅黃衣，以金線織絲布纏頭，歲三月則詣佛寺，坐紅床，使人舁之。貴臣如王之服，或青綠、緋白、粉紅、褐紫，並纏頭跨馬。城市田野，皆有首領主之，每歲惟夏秋兩得奉，給金、錢、錦、穀、帛，以治事大小為差。刑罰罪輕者杖數十，重者至二百，大罪則盛以毛囊投諸海。不尚闘戰，鄰國小有爭，但以文字來往相詰問，事大亦出兵。鑄金銀為錢，無穿孔，面鑿彌勒佛，背為王名，禁民私造。

元祐六年，其使兩至。詔別賜其王帛二百匹、白金瓶、襲衣、金束帶。
——《宋史·列傳》 二百四十九 外國六

Important points:

• Fu Lin never tributed (sent envoys) before.
• In 1081 AD (元豐四年), the King of Fu Lin sent envoys to Song.
• Another envoy was sent in 1091 AD (元祐六年).

Never Tributed?

It is very unnatural. Xin Tangshu was written in 1060 AD, two decades before the envoy arrival. The Song historians must be well aware of the Fu Lin appearance in previous materials. Historian Ma Tuan-Lin (馬端臨) in the 13th was also aware of this anomaly. In Wenxian Tongkao (《文獻通考·四裔考》 十六 拂菻), he wrote:

“按唐史有拂菻國，以為即古大秦也。然大秦自後漢始通中國，歷晉唐貢獻不廢；而宋四朝史拂菻傳則以為其國歷代未嘗朝貢，至元豐時始遣使入獻方物。今以二史兩拂菻傳參之，唐傳言其國西瀕大海，宋傳則言西至海尚三十程，而餘界亦齟齬不合，土產風俗亦不同。恐是其名偶同而非大秦也，今故以唐之拂菻附入大秦，而此拂菻自為一國云。”

He pinpoints that there were inconsistencies in the Tang and Song records about Fu Lin. He concludes that the Fu Lin of Tang was likely related to Da Qin, while the Fu Lin of Song was probably a distinct state that merely shared the same name.

3. The Open Questions

Given the materials, there are two crucial questions waiting to be answered:

1. First of all, what is Fu Lin?
2. Who are the named Kings of Fu Lin mentioned by the two records?

4. Theses and Analysis

A. The identity of “Fu Lin”

A-1. Constantinople

Thesis: The most natural thesis is Fu Lin being Constantinople, thus the state refers to Byzantium. It suggests Fu Lin is a transliteration of “Polin” (eis ten Polin = “to the City”²).

Supports: Tang record explicitly mentioned Fu Lin used the old name Da Qin. With Da Qin generally believed to be Roman Empire, Fu Lin is naturally the other state bearing the name of Rome. Furthermore, there was no better candidate geographically that lay on the western shore of the great sea.

Critique:

1. The Subject Problem: Tang record explicitly mentioned Fu Lin’s submission to Da Shi (Arabs). While Umayyad Caliph Muawiyah I besieged Constantinople (c. 674-678 AD), the campaign was not successful and Byzantium never submitted.
2. Linguistical mismatch: “eis ten Polin” into “Fu Lin” was not linguistically sound for the 6th century.

Hirth’s argument

“The first character 拂, now pronounced fu in the Mandarin, and fat in the Canton dialect, has a final t, according to all the mediaeval authorities quoted by K’ang-hi (Rad. 63, 5). In the Tsi-yün, a work which appeared as late as the Sung Dynasty, its sound is described as 敷勿切, i. e., p(ok-m)ut, or pat.

The second character 菻, now pronounced lin in the Mandarin, and lam in the Canton dialect, was according to the Tsi-yün pronounced 力尋切, i. e., l(ik-k)am, or lam, and K’ang-hi quotes the name Fu-lin (Fat-lam or Pat-lam) as an example of that pronunciation (Rad. 140, 8). As a further example of the old sound ending in m, and not in n, I may quote the name of one of the priests which appears in estrangelo characters as Ephraem (read Abraham by Kircher) in the Syriac part of the Nestorian inscription with the Chinese transcription 拂林, = fu-lin, fat-lam or pat-lam.

In order to express the syllable lin in πόλιν [Polin], a Chinese transcriber of the sixth century would have selected some such character as 鄰, lin, the old final of which is n, rather than a sound ending in m. In the T’ang-shu-shi-yin (chap. 24, p. 3), ad vocem Fu-lin, the sound of the character is described as 力尋切 (lik-nam) = lam. As may be seen from China and the Roman Orient (p. 287, note 2), I do not doubt the correctness of the etymology of the name Istambul = Istan-polin (εἰς τὴν πόλιν); but, even granting the Byzantine Romans of that early period having called their capital ‘Istanpolin,’ this need not force us to identify the name with Chinese ‘Fu-lin.’”

A-2. Syria

Thesis: Proposed by Friedrich Hirth, this thesis argues that “Fu Lin” refers specifically to the Roman Province of Syria, with Antioch as its capital. The name “Fu Lin” is a transliteration of “Bethlehem” (Bat-lam), chosen because the Chinese identified the country by the birthplace of the “Holy One” (Jesus), similar to how they named India (Mo-k’ie-to / Magadha) after Buddha’s birthplace.

Supports:

1. Contradiction in Jiu Tangshu: This is the only theory that resolves the “subject” contradiction in Tang record. Syria was conquered by the Umayyads in the 7th century and became subjects.
2. Linguistic Evidence: Hirth identified that the plant names attributed to Fu Lin in Chinese records are Syriac (Aramaic), not Greek.
   • Olive: Chinese Ts’i-t’un (齊暾) $\rightarrow$ Syriac Zaita.
   • Fig: Chinese Ti-ni (底珍) [Hirth reads as Ti-ni (底泥)] $\rightarrow$ Syriac Tenta.
   • Myrtle: Chinese A-tzi (阿縒) $\rightarrow$ Syriac Asa.

🔥 I personally think this is a very strong evidence. There are still contradictions though.

Critiques: Naming a great empire after the small town of Bethlehem is a leap. Though Hirth defends it with the Buddhist analogy, there is generally a lack of similar evidence.

A-3. Rum over Rome

Thesis: Proposed by Shiratori Kurakichi (白鳥庫吉), this thesis argues that “Fu Lin” had always referred to Rome, and was a transliteration of “Rum” (Turkic pronunciation of “Rome”). Northern Altaic/Turkic tribes pronounced “Rum” with a prosthetic vowel as “Hrum” or “Urum”, which the Chinese transliterated as Fu-lin (Put-lam).

Supports:

1. The Shifting Location: This explains the geographical inconsistencies noticed by Ma Tuan-Lin. “Rum” could have meant the Byzantium in Asia Minor in the Tang era, while it could mean the Rum Sultanate in Anatolia in the Song era.
2. The Tribute Anomaly: This is in line with Ma Tuan-Lin’s assumption, that the Fu Lin of Song was a different state (Seljuk Rum) than the Fu Lin of Tang (Byzantium).
3. Explanation of “Mie-Li-Sha”: Shiratori argues that Fu Lin placed its southeast border at “Mie-Li-Sha” in Song record, which referred to Malatya. This was true for Rum Sultanate in eastern Anatolia, but not Syria nor Constantinople.

Critiques:

1. The Subject Problem$^\textdagger$: Like the Constantinople thesis, it struggles to explain the subject problem. Shiratori resolves this by claiming the King “Bo-Duo-Li” was actually an Armenian Prince (Patricius), who did submit to the Umayyads.
2. Timeline Mismatch: Rum Sultan Suleiman I was killed in 1086 by Seljuk Sultan. From 1086–1092, the Sultanate of Rum was in an Interregnum³ (ruled by governors/generals like Abu’l-Qasim). It is historically awkward for the ‘same ruler’ to send envoys in 1081 (Suleiman) and 1091 (Interregnum/Abu’l-Qasim).

$\textdagger$: The record’s claim that Fu Lin became tributary/vassal to Da Shi after attack on its capital, which clashes with standard Byzantine macro-history.

B. The identity of “Bo-Duo-Li”

B.1 Constans II

The Jiu Tangshu records that in 643 AD (17th year of Zhenguan), the King of Fu Lin, Bo-Duo-Li (波多力), sent envoys. The reigning Byzantine Emperor was Constans II (r. 641–668). However, the name “Bo-Duo-Li” bears no phonetic resemblance to “Konstantinos.”

B.2 The Nestorian Patriarch

Thesis: Proposed by Friedrich Hirth, this thesis argues that “Bo-Duo-Li” is a transliteration of the title “Batrirk” (Syriac/Arabic for Patriarch). Hirth suggests the “King” was actually the Patriarch of Antioch (or the Nestorian Catholicos⁴, but less likely), acting as a “Religious King.”

Supports:

1. The “Tuhuoluo” Connection: As noted in Section 2.A, the “Lord” of Fu Lin sent a mission via the “Great Chief of Tuhuoluo” (Bactria). By 719 AD, the political route between Byzantium and Bactria was severed by the Arab Caliphate. Hirth argues that only a Religious Leader (Patriarch) could command a subordinate Metropolitan in Bactria to send a mission through the church’s internal network, bypassing the political blockade.
2. Etymology: The reconstruction of 波多力 (Pa-ta-lik) matches Batrik perfectly.

Critique: It assumes the Chinese court mistook a religious leader for a secular “King” of a major empire.

B.3 The Armenian Patricius

Thesis: Proposed by Shiratori Kurakichi, this thesis argues that “Bo-Duo-Li” is a transliteration of the Roman title “Patricius” (Prince/Noble). Specifically, he identifies the “King” as an Armenian Prince holding the rank of Patricius.

Supports:

1. The Subject Problem: This solves the subject problem as Armenian principalities were actually conquered and paid tribute to the Umayyads (c. 640s-650s).
2. Historical Context: The timeline matches the Arab conquest of Armenia, which started in c. 639 according to the Arabic sources.

Critiques:

1. Massive Downgrade: It requires downgrading “Fu Lin” - a massive empire with a capital of 400 cities to a vassal principality in Armenia, and no Armenian principalities had the scale. It also implied the Tang court could not distinguish between the Empire and a frontier vassal.
2. Timeline Mismatch: According to Arabian sources, the actual conquest of Armenia did not start until c. 645, succeeding the earlier failures. Tang record, on the other hand, suggests that Fu Lin had already submitted to the Arabs in 643, which does not perfectly match.

Shiratori’s Argument

“唐書の拂林王波多力は取りも直さずアーメニア國の君を云ふものにして即ち Po-toli は Batrik を見はすものなるを知るべし… 大食 (アラビア) に降だりしものはコンスタンチノブル、アンチヲーキャに非ずして正さにアーメニア國を指さすなり”

C. The identity of “Mieh-Li”

C.1 Michael VII Doukas

Thesis: The most intuitive candidate for “Mieh-Li-I-Ling-Kai-Sa” (滅力伊靈改撒) was Michael VII (r. 1071-1078). He abdicated in 1078⁵, but could send envoys before his abdication who arrives in Kaifeng in 1081.

Supports:

1. “Mieh-Li-I-Ling-Kai-Sa” could be a plausible transliteration of “Michael Caesar”.
2. Michael VII was the only Byzantium emperor who could reasonably send the envoys in time for them to arrive in 1081.

Critiques:

1. The 1091 Anomaly: The Song record mentions the same king sent envoys in 1091. Michael VII became a monk and died c. 1090. He never restored power.
2. The “Caesar” Problem: “Caesar” was a subordinate rank, not the title of the Emperor (Basileus). An official imperial envoy would not demote their Emperor to Caesar.

C.2 Melissenos Nikephoros

Thesis: Melissenos Nikephoros was a Byzantine general and later pretender. He remained loyal to Michael VII, rebelled against Nikephoros III, and eventually submitted to Alexios I Komnenos in 1081, accepting the title of Caesar.

Supports:

1. The Title of Caesar: Melissenos actually held the title Caesar. Crucially, he was still Caesar in 1091, making him the only candidate who fits both dates.
2. Linguistic Match: Melissenos Caesar closely resembles the pronunciation of “Mieh-Li-I-Ling-Kai-Sa”. Hirth further argues the character Yi (伊) is a scribal error for Sun (筍), making the reading Mit-lik-sun.
3. The Coin Evidence: The Song record mentions coins with the King’s name on the reverse and “Maitreya Buddha” on the obverse. Hirth matches this to a real silver coin of Melissenos, arguing the obverse is the Virgin Mary mistaken for Maitreya.

Critiques:

1. The Time Travel Paradox: Melissenos became Caesar in April 1081. The envoys arrived in China in October 1081. It is impossible to travel from Anatolia to China in 6 months. The envoys must have departed years earlier, when Melissenos was just a general/rebel, not Caesar.
   • Hirth’s Defense: Hirth argues the Chinese might have applied his current title to the record, or the envoys were merchants who updated their credentials. It’s a weak point worth highlighting.
2. Status: Why would a subordinate Caesar send envoys instead of the reigning Emperor, Alexios I?

❗Hirth’s defense is arguably a very weak point. Assuming overland/combined caravan-maritime transmission, title-updating en route is possible in principle, but a full diplomatic identity shift synchronized with court registration in Kaifeng within months is low-probability.

C.3 Suleiman I of Rum

Thesis: Proposed by Shiratori, this argues the King is the Seljuk Sultan of Rum. The name is a transliteration of the title “Melik-i-Rum Caesar” (Melik = King, Rum = Rome, Caesar = Emperor)⁶.

Supports:

1. Dress Code: The King is described wearing a “cloth wrapped head” (turban) and gold-woven robes. This is distinctly Islamic/Seljuk dress, not the Byzantine crown and purple robes.

Refer to the rest of supports and critiques in Section A.3.

5. Conclusion and Personal Take

The conflicting records of the Tang and Song dynasties are best resolved by accepting that “Fu Lin” was not a static location, but a shifting label for the “Great Western Power”.

Reference

[1] Hirth, F. (1909). The mystery of Fu-lin. Journal of the American Oriental Society, 30(1), 1-31.
[2] SHIRATORI, K. (1956). A New Attempt at the Solution of the Fu-lin Problem. Memoirs of the Research Department of the Toyo Bunko (The Oriental Library), 15, 165-329.

Footnotes

1. Tuhuoluo (吐火羅) referred to Tokharistans inhabited in the Bactrian area (where the Greco-Bactria Kingdom was). It is believed this region was inhabited by a mix of Bactrian and Turkic people ruled by a Turkic lord. They are distinct from the “Tocharian” speakers of the Tarim Basin, a misnomer by modern academia (blame Friedrich W. K. Müller). Importantly, the trade and communication between Bactria and Constantinople had likely been cut or severely limited by the rise of Arabs in the middle. ↩

2. Constantinople was so massive and central at its time, that Greeks (or Romans as they called themselves) simply called it “the City” in most times. ↩

3. Interregnum: noun. a period when normal government is suspended, especially between successive reigns or regimes. ↩

4. The Patriarch of the Church of the East was Eshuyow II (628-644). ↩

5. His successor, Nikephoros III Botaneiates, abdicated as well in 1081. Karma. ↩

6. “malik” (ملك) was an Arabic title meaning “king”. ↩