Saturday, 31 March 2018

Jurchen Inscription on the River Arkhara

This post discusses a recently-identified inscription in the Jurchen script on the bank of the River Arkhara in Russia. The main source of information is a book on the subject by Aisin Gioro Ulhicun 愛新覚羅・烏拉熙春 and her husband Yoshimoto Michimasa 吉本道雅 which was published last year. Viacheslav Zaytsev has kindly provided me with additional information from Russian sources.


The River Arkhara (река Архара) flows through the Arkharinsky District of the Amur Oblast in Russia, and enters the River Amur which forms the border between Russia and China. About 70 km upstream of the town of Arkhara, in the middle of the Siberian forest, there is a bend in the river (49°38'22.7" N 130°36'35.9" E). On the north bank of the bend there are rocky outcrops up to 10 metres in height, on which numerous petrographs showing people, animals and abstract art are drawn using ochre. The rock art at this point of the river has been known since at least 1902 ("Expedition of 1902" with an unknown surname is written in Russian in one place), and the rock was known as the "Rock of writing" (писаный камень) by local people since the early 20th century. The rock art has been described in publications many times since 1930, and the site has been a designated protected area of regional significance since 1978.

Location of site on the River Arkhara

Click to go to my Google Maps map of Tangut, Khitan and Jurchen inscriptions

Although the rock art on the River Arkhara had been studied by several Soviet and Russian teams during the 20th century, the first person to notice the presence of a faint inscription in an unknown language written in black ink was Prof. Andrey Pavlovich Zabiyako (Андрей Павлович Забияко) during his expedition to the site in 2003 (Zabiyako's initial hand-drawn reproduction of the inscription from 2003 is reproduced in Aisin Gioro 2017 p. 30). He provisionally identified the inscription as being in the Khitan or Jurchen language, and after consultation with several Russian linguists he concluded that the inscription was most likely written in Jurchen. In November 2014 Zabiyako showed photographs of the inscription to Jin Shi 金適 (elder sister of Aisin Gioro Ulhicun, and an amateur Jurchenologist) at a conference in China, and she confirmed his suspicions that the inscription were indeed Jurchen, although she was unable to translate it on the spot. Later that month Aisin Gioro Ulhicun contacted Zabiyako, and he provided her with photographs, drawings and other material from the 2003 expedition. The discovery of an early Jurchen inscription at Arkhara was announced in the Russian press in February 2015, and then in August 2015 Aisin Gioro participated in a joint expedition with the Amur State University to study the inscription in situ.

According to information received from Professor Zabiyako, there had been plans to produce and publish a joint study of the inscription by Russian and Japanese scholars. However, it seems that Aisin Gioro Ulhicun decided to publish her research on the Jurchen inscription on her own. Her research appeared in a chapter of the book co-authored with her husband that was published in August 2017. Although the book is titled ロシア・アルハラ河畔の女真大字墨書 [Jurchen large character ink writing on the bank of the Arkhara River in Russia], in fact only a single chapter is devoted to the Arkhara Jurchen inscription.

Rock face on the bank of the River Arkhara with red painted petrographs and Jurchen inscription

The Jurchen inscription is just to the left of the Russian writing in white

Source: @cosmicore

Aisin Gioro Ulhicun's Reading of the Jurchen Inscription

The Jurchen inscription is situated on an area of the rock face about 15 cm × 25 cm, and comprises three vertical lines (running from right to left) written in black Indian ink.

Detail of rock face showing the Jurchen inscription

Source: @cosmicore

Having been exposed to the elements for hundreds of years, the writing is very unclear, and it is extremely difficult to identify the individual characters, some of which are now little more than a smudge. Zabiyako's 2003 drawing of the text illustrates how difficult a task it is to interpret and read the inscription in its current state.

Drawing of the Jurchen inscription made by Prof. Zabiyako in 2003

From Aisin Gioro 2017 p. 30

Aisin Gioro is able to make out a total of twenty-four characters: seven characters on line 1 (on the right), ten characters on line 2 (in the middle), and seven characters on line 3 (on the left). With the benefit of many years of studying Jurchen, and having examined the actual inscription in person, she is able to identify every single Jurchen character, and read the entire inscription. Her transcription of the three lines of Jurchen text using a computer font is shown below (but here using a font modified by me from one designed by Prof. Jing Yongshi).

Transcription of the Arkhara inscription

Based on the transcription in Aisin Gioro 2017 p. 30

I am unable to verify the accuracy of this transcription because the Jurchen writing is not sufficiently clear in the best photograph that I have access to (shown above). For me the inscription as shown in the photograph is virtually illegible, and I can only identify one or two characters unaided (the 6th character on line 2 is clear, and it is the only character in Zabiyako's drawing that exactly matches Aisin Gioro's transcription). Aisin Gioro's transcription of the first five characters of line 1 looks plausible, and the 1st, 4th and 6th characters of line 2 seem reasonable, but the whole of line 3 is a complete blur to me. To be honest, I am quite surprised that Aisin Gioro is able to read this inscription with such apparent certainty. Her reading and translation into Japanese of the Jurchen inscription is shown below, with my translation into English.

Aisin Gioro's Reading and Translation of the Arkhara inscription

Line 1

pulan imula ʃunʤa ania

丁 未 五 年

5th year [which is] the ding wei [red goat] cyclical year [1127]

Line 2

tarɣando i oson muə pərgilə gai-man

タルアンド の 小 水 下流に 得たる

Attained the lower reaches of a small river in Targhando

Line 3

ʤua bia oniohon inəŋgi ʃin-tərin

十 月 十九 日 申忒鄰

10th month 19th day [written by] Shin Terin

Readings from Aisin Gioro 2017 pp. 31–32

Japanese translation from Aisin Gioro 2017 pp. 42, 47, 51

English translation by Andrew West

If Aisin Gioro's reading of the inscription is correct then this is a remarkable record of an expedition along the Arkhara River in 1127, made by a Jurchen owing allegiance to the Jin dynasty (1115–1234). 1127 was just 12 years after the founding of the Jin dynasty, and less than ten years after the creation of the Jurchen script in 1119 or 1120. This date also means that it is the earliest known dated inscription in the Jurchen script (see List of Jurchen inscriptions for dates of other Jin dynasty Jurchen inscriptions).

However, I think it is necessary to strike a note of caution. The current condition of the inscription, written in ink on rock and exposed to rain and shine for hundreds of years, means that it is far from easy to identify the Jurchen characters, and identification of unclear characters can be a subjective process.

Discussion on the Date of the Inscription

"5th year" refers to the 5th year of an unstated reign era (it was not uncommon to omit the name of the era when it was obvious to the author), and ding wei (represented as "red goat" in the Jurchen calendar) refers to the 44th of the series of sixty cyclical years used to number the years of the Chinese calendar. Aisin Gioro interprets this date as the 5th year of the Tianhui era (天會五年) of the Jin dynasty, which corresponds to the year 1127. As this is the only 5th year of a Jin dynasty era which is also a ding wei cyclical year this would appear to be a sound interpretation. The only other date in later Chinese history that is both the 5th year of an era and a ding wei cyclical year is the 5th year of the Yongzheng era (雍正五年) of the Qing dynasty, exactly 600 years later in 1723. However, this is 200 years after the last documented use of the Jurchen script, and it is very unlikely that the Jurchen script was still in use during the Qing dynasty.

Unusually, the year is given on the first line, and the month and day are given on the third line, separated from each other. The 19th day of the 10th month on the third line presumably belongs to the 5th year mentioned on the first line. Together, they give the exact date of the 19th day of the 10th month of the 5th year of the Tianhui era, which corresponds to 24 November 1127 in the Julian calendar (my Daily Calendar for the Liao and Jin Dynasties can be downloaded as an Excel archive or text archive).

Unfortunately, I find it difficult to be convinced that the date of 1127 deduced by Aisin Gioro is correct. There are two main problems. The first problem is the "5th year". Aisin Gioro's transcription of the first five characters of line 1 seem plausible based on the photograph of the inscription, but I find it very difficult (although not quite impossible) to recognise the 6th character on line in the photograph as the Jurchen character for "five". And the 7th character appears to me as no more than a remnant smudge, in which case Aisin Gioro may be reading it as "year" contextually on the basis that it is preceded by "five". I think that we have to admit that the reading of line 1 characters 6–7 as "5th year" is at best doubtful.

The second problem is the cyclical year. Without an era name being given, the cyclical year is essential to pin down which "5th year" is referred to in the inscription. The sexagenary cycle used in China and elsewhere in Asia comprises a combination of ten "heavenly stems" (天干) and twelve "earthly branches" (地支) which repeats every sixty years. In Khitan, Jurchen, Mongolian and Manchu, the ten heavenly stems are represented as five colours (blue/green, red, yellow, white and black), and the earthly branches are represented by the twelve animals of the Chinese zodiac. Aisin Gioro reads line 1 characters 1–2 as meaning "red" and characters 3–5 as meaning "goat", which would indeed correspond to the cyclic year ding wei 丁未. The reading of characters 1–2 as "red" looks correct to me, but I am very unsure about the reading of characters 3–5 as "goat".

Aisin Gioro regards character 3 as a variant of *i (a phonetic component in many words). However the glyph form identified by Aisin Gioro does not occur in any other Jurchen sources, and it could equally be any of a half dozen similar-looking characters, including (variant of *ai) which is what Zabiyako shows in his drawing of the inscription.

Character 4 has a reading of *əm(u) and is a special character for *əmu "one" (cf. Manchu emu "one") that is used in some compound words such as *əm(u)-dʒï "same" and *əm(u)-xuŋ "only". As far as I know, this character is not attested in any word where it does not have a meaning related to "one".

Character 5 has a reading *ula and means "river", specifically a large river (cf. Manchu ula "large river"). As far as I know, the Jurchen character is not attested in any word other than "river".

Aisin Gioro takes characters 3–5 (*i-əmu-ula) together as a phonetic transcription imula corresponding to Ming dynasty Jurchen imala (*imara) meaning "goat". The phonetic match is not great, and I know of no other Jurchen word where an initial i precedes and displaces a different vowel in the following character (əmu + ula = əmula is normal but i + əmu = imu is not expected). Not only is the phonetic match doubtful, but it is probable that the Jurchen word used to represent the earthly branch wei 未 was *xoni "sheep" (cf. Manchu honin "sheep" used for wei; as well as *xoni used for wei in Khitan). Perhaps my biggest objection to Aisin Gioro's interpretation is that characters 4 and 5 are only attested as special characters with particular meanings of "one" and "river" respectively, and there is no other evidence I know of that either character was ever used for general phonetic transcription. For these reasons I am unable to agree with Aisin Gioro's reading of characters 3–5 as "goat", and if this reading is not accepted then the cyclical year ding wei vanishes, and the identification of "fifth year" as the 5th year of the Tianhui era (1127) can no longer be supported.

If characters 3–5 do not mean "goat", then what could they mean? The obvious answer is that characters 4 and 5 mean exactly what they mean in Ming dynasty Jurchen, namely "one, single" and "(large) river" respectively, and refer to the river by which the inscription was found. An intriguing possibility is that *əmu ula = əmula "single river" was the Jurchen name for the River Amur. The origins of the river's name are unknown, but it may derive from a Tungusic language related to Jurchen, where amur is cognate to Jurchen əmula "single river". Characters 1–2 meaning "red" may be a modifier to the river name, perhaps referring to the River Arkhara as the "red Amur". The identification of the 3rd character is uncertain, and I will not hazard a guess.

Notes on the Jurchen Characters in the Arkhara Inscription

These are my notes on the individual Jurchen characters in the Arkhara inscription, as identified by Aisin Gioro Ulhicun. The glyph shown in the column labelled "AGU" is the form shown in Aisin Gioro's transcription, although it is quite possible that in some cases the glyph shown by Aisin Gioro reflects her Jurchen font rather than the exact glyph form used in the rock inscription. The column labelled "HYYY" gives the corresponding glyph form of the character as used in the Jurchen Vocabulary of the Chinese–Barbarian Dictionary (Huá-Yí Yìyǔ 華夷譯語) [HYYY] compiled by the Bureau of Translators (四夷館) during the Ming dynasty (1368–1644). Reconstructed Jurchen readings are taken from the 1984 Dictionary of Jurchen written by Jin Qizong (father of Aisin Gioro Ulhicun), and are marked with an asterisk. Aisin Gioro Ulhicun's reconstructed readings are given in italics with no asterisk.

Position AGU HYYY Notes
1.1 HYYY: *fula in the word *fula-giɛn {弗剌江} "red colour" (cf. Manchu fulgiyan, fulgan "red").
1.2 HYYY: phonetic suffix *-(g)an {岸}. Here suffixed to fula to form the unattested word *ful(g)an "red", cognate with Manchu fulgan "red".
1.3 HYYY: phonetic component *i {一} in many words, including *imara {一馬剌} "goat".
1.4 HYYY: *əm(u) {厄木} in the words *əm-dʒï {厄木只} "same" and *əm-xuŋ {厄木洪} "only" (cf. Manchu emhun "alone"). This character is a special way of writing *əmu {厄木} "one" (cf. Manchu emu "one") in a few compound words. The Jurchen character is not attested in any word that is not related to "one".
1.5 HYYY: *ula {兀剌} "river" (cf. Manchu ula "large river"). The Jurchen character is not attested in any other word.
1.6 HYYY: *ʃundʒa {順扎} "five" (cf. Manchu sunja "five").
1.7 HYYY: *ania {阿揑} "year" (cf. Manchu aniya "year").
2.1 HYYY: *ta in the words *faita {肥塔} "eyebrow" (cf. Manchu faitan "eyebrow") and *ta-(x)un {塔温} "fat" (cf. Manchu tarhūn "fat").
2.2 HYYY: phonetic component *-(g)an {安} in many words.
2.3 HYYY: phonetic component *do {朶} in many words. Aisin Gioro takes *ta-(g)an-do (tarɣando) to be the name of the Jin dynasty administrative region (mouke 謀克) that included the place where the Arkhara inscriptions are located. She suggests that it may be a Jurchen name or possibly derived from another language spoken there. There does not seem to be any other evidence for a place with this name.
2.4 HYYY: genitive particle *i {以}. Sometimes used as a phonetic component *i.
2.5 HYYY: *oso in *oson {斡速灣} "small" (cf. Manchu osohon "small"). Occurs as a single character *oso "small" contrasting with *amba "big" in the Yongning Temple Stele of 1413 (erected near the mouth of the Amur River in present-day Russia).
2.6 HYYY: *muwə {没} "water" (cf. Manchu muke "water, small river"). The Jurchen character is not attested in any other word. Aisin Gioro takes *oso muwə (oson muə) to mean "small river", referring to the River Arkhara.
2.7 HYYY: *fə in the word *fədʒïlə {弗只勒} "down, below, under" (cf. Manchu fejile "under"). The word *fədʒïlə also occurs twice in the Yongning Temple Stele of 1413. The Jurchen character is not attested in any other word. Aisin Gioro equates the word *fədʒïlə with a word with the same meaning that is written *fədʒï-gi (cf. Manchu fejergi "under, below"), which leads her to conclude that the original Jin dynasty word for "down, below, under" was pərgi, cognate with hərgi "under" in other Tungusic languages. So she reads here as pərgi "down, below, under".
2.8 This character is not used in HYYY, but does occur as in the manuscript Jurchen Character Book (Nǚzhēn zìshū 女真字書), but with unknown reading and meaning. Aisin Gioro says that it is here a locative suffix , added to to make pərgilə "at the lower part", i.e. downstream section of the "small river".
2.9 HYYY: phonetic component *gai {該} in many words. Aisin Gioro says that it is here used as a verb meaning "to get, to take, to lead" (cf. HYYY *gai-bie "to take"; Manchu gaimbi "to take (away)").
2.10 HYYY: *ma in the word *ma-na-ra "to be broken" (cf. Manchu manambi "to be worn-out, to be tattered"). The glyph form given by Aisin Gioro occurs once as part of a personal name in a monument recording the names of successful candidates for the degree of jinshi in 1224, which Jin Qizong gives the very uncertain reading of *busu. Aisin Gioro reads the character as a verbal suffix man, although I do not know of any other examples where it is used as a verbal suffix.
3.1 HYYY: *dʒua {撾} "ten" (cf. Manchu juwan "ten").
3.2 HYYY: *bia {必阿} "moon, month" (cf. Manchu biya "moon, month").
3.3 HYYY: *onioxon {斡女歡} "nineteen" (no cognate word for "nineteen" survives in Manchu).
3.4 HYYY: *inəŋgi {一能吉} "day" (cf. Manchu inenggi "day").
3.5 HYYY: *ʃïn in the word *ʃïn-gə {申革} "rat" (cf. Manchu singgeri "rat"). Also *ʃïn in the word *ʃïn-ai "to wait" in the 1185 monument commemorating the victory of Emperor Taizu of Jin over the Khitans in 1114. Aisin Gioro takes this to be an otherwise unrecorded Jurchen family name, Shin.
3.6 HYYY: *mədə in the word *mədə-əri = *mədəri {脉忒厄林} "sea" (cf. Manchu mederi "sea").
3.7 HYYY: *əri in the word *mədə-əri = *mədəri {脉忒厄林} "sea". Also occurs as a phonetic component in many other words. Aisin Gioro reads as tərin "sea", based on the gloss of "sea" (海) for telin 忒鄰 in the Glossary of Jin Words (Jīnguó yǔjiě 金國語解) appended to the History of the Jin Dynasty. She suggests that telin was the original Jin reading of the two characters meaning "sea", and mədərin was the Ming dynasty reading of the same two characters. The History of the Jin Dynasty records two people with the name Telin 忒鄰, most notably Prince Ge (葛王), a son of Emperor Zhangzong. Aisin Gioro therefore takes Shin Tərin in the Arkhara inscription to be the name of the author of the inscription.


  • 愛新覚羅烏拉熙春 Aishingyoro Uruhichun [Aisin Gioro Ulhicun] and 吉本道雅 Yoshimoto Michimasa, Roshia, Aruhara kahan no Joshin daiji bokusho: Joshin, Kittan moji iseki o tadotte ロシア・アルハラ河畔の女真大字墨書—女真・契丹文字遺跡をたどって— [Jurchen large character ink writing on the bank of the Arkhara River in Russia: following the remnants of Jurchen and Khitan writing]. Kyoto: Hōyū shoten 朋友書店, 2017. ISBN 978-4-89281-162-3

No comments: