Brotherly diplomacy: on the Kitan–Mongol model of pseudo-kinship and the origins of the Kitan emperors’ fictive kinship with Chinese rulers

Abstract

Evidence shows that tenth- and eleventh-century Kitan (Liao) emperors used pseudo-kinship to cement diplomatic relations with foreign powers as well as for internal affairs. Similarities between this practice and twelfth- and thirteenth-century Mongol anda (sworn friendship) were previously highlighted by Wang Guowei but have yet to be the focus of further study. Kitan emperors used pseudo-kinship as a preferred political tool to establish alliances and reinforce their position in both external and domestic policies. A comparison of Kitan and Mongol traditions also shows a high degree of similarity. However, although they share concepts of sworn friendship and common oath rituals, the establishment of pseudo-kinship occurred in different contexts and often for different purposes. This article attempts to show that Kitan rulers successfully continued the pseudo-kinship diplomacy that existed since the Tang between the hegemons of the steppe and the Central Plain. They achieved this by making regular use of pseudo-kinship diplomacy, in addition to seeking ways to make the practice more acceptable to the Chinese court. These adaptations included a progressive estrangement of diplomatic pseudo-kinship from its original form, casting away oath rituals and adopting a new system of kinship terms.

Introduction

During the Song 宋 Dynasty (960–1279), literati ‘revived’ and expanded concepts of environmental, natural, and moral separation between Chinese (Hua 華) and Barbarians (Yi 夷). One of these scholars, Ouyang Xiu 歐陽脩 (1007–1072), in his Xin Tangshu 新唐書 (New Book of the Tang), criticised, among other things, the brotherhood that tied the first Han Emperor Liu Bang 劉邦 with the Xiongnu ruler Modun 冒頓 after the siege of Baideng 白登 in 200 B.C.:

What are we to make of the Son of Heaven's dignity, if he allies with the Xiongnu as brothers? of the title of the emperor's daughter, if she rides in the same carriage as Barbarian hags? There, incestuous mothers marry their sons; how can we follow these filthy customs? The difference between China and the Barbarians is our distinction between father and son, man and woman. For the pleasant and seductive beauty [of these Chinese women] to be destroyed among the alien brood—this is foul disgrace in the extreme! But none of the Han rulers or ministers were ashamed of it.¹

奈何以天子之尊，與匈奴約為兄弟？帝女之號，與胡媼並御；蒸母報子，從其汚俗？中國異於蠻夷者，有父子男女之別也。婉冶之姿毀節異類，垢辱甚矣。漢之君臣，莫之恥也。

By using this critique to introduce the ‘Chronicles of the Türks’ (Tujue zhuan 突厥傳) and setting the Han–Xiongnu brotherhood as a bad precedent, Xiu² criticised the ongoing diplomatic ‘brotherhood’ between Song and Kitan (Qidan 契丹) emperors. Implicitly drawing a parallel between past and present, he believed an infamous Chinese historical moment was repeating itself.

From the Han to the Tang, the Dynasties of Chang'an and Luoyang negotiated peace with steppe rulers through marriage. The Chinese called the proactive diplomacy of marriages between emperors and qaghans heqin 和親—a concept that Nicola Di Cosmo translated as ‘peace through kinship relations’.³ While modern historians agree that heqin diplomacy solely revolved around marriage (real kinship), Song thinkers also included fictive kinships in this definition. In this citation, Ouyang Xiu considered both types of ‘diplomatic families’ as similarly shameful for the Son of Heaven, supposedly equal to no other. Song intellectuals such as Sima Guang 司馬光 (1019–1086) and Zhu Xi 朱熹 (1130–1200) took this idea a step further and referred to brotherhoods between Chinese and steppe rulers as heqin, despite not being marriage-based.⁴ This may have been a way to mirror the Song diplomatic situation with those of the previous dynasties; in other words, the Tang marriage-based heqin with the Türks gave way to the Song pseudo-kinship heqin with their ‘barbarian’ neighbours—the Kitans (Qidan 契丹).

Ouyang Xiu's criticism arose at a time when the official stance of the Song towards the Kitan state was regulated by the Chanyuan Covenant (Chanyuan zhi meng 澶淵之盟),⁵ in which, in 1005, both courts agreed to normalise their relations. Both emperors officially recognised each other and agreed to maintain a semblance of equality in official discourses. They implemented a form of diplomatic correspondence in which Shengzong 聖宗 of the Kitans (Yelü Longxu 耶律隆緒, 972–1031, r. from 982) and Zhenzong 真宗 of the Song (Zhao Heng 趙恆, 968–1022, r. from 997) referred to each other as ‘elder brother’ (xiong 兄) and ‘younger brother’ (di 弟) according to their respective ages.⁶ Also known as the ‘Liao Dynasty’ (Liaochao 遼朝) in historiography, the Kitan empire, which was founded in 916 by Abaoji 阿保機 (872–926, qaghan in 907, emperor in 916), adopted a ‘Kitan–Chinese’ dual administrative system and bilingual political rhetoric. In Chinese sources, Kitan leaders are often seen to be establishing alliances with the rulers of the Central Plain and obtaining recognition as holding equal status with the Son of Heaven. Modern research has often emphasised the political, military, and economic reasons for these diplomatic activities and their consequences. However, the pseudo-kinship relations upon which the sovereigns systematically agreed have not yet attracted much academic attention.⁷ Wang Gungwu has suggested that rulers used it as a tool to assert dominance or express equality as situations required.⁸ The Kitans had continuously sought pseudo-kinships with emperors of the Five Dynasties from 905 to 979 or later, but met with limited success. They negotiated the Chanyuan Covenant after less than three decades of border tensions, during which they engaged in skirmishes and short wars with the Song.⁹ Among ancient steppe hegemons, they succeeded in maintaining the longest relations with Central Plain rulers, effectively creating a century of fragile but uninterrupted pseudo-kinship between the North and the South from 1005 onwards.

Ritual kinship is also present in Chinese culture. Even today, almost everyone in China knows the ‘Oath of the Peach Orchard’ (Taoyuan jieyi 桃園結義) between Liu Bei 劉備 (161–223), Guan Yu 關羽 (d. 220), and Zhang Fei 張飛 (d. 221), as it is discussed in school books and featured in popular novels, series, and films. Depicted in Luo Guanzhong's 羅貫中 (circa 1330–circa 1400) Romance of the Three Kingdoms (Sanguo yanyi 三國演義), this oath remains the most famous Chinese-style sworn brotherhood, which is the meaning of jieyi 結義.¹⁰ However, pseudo-kinship rarely occurred in China before the Tang Dynasty and was never systematically used in the political discourse after the Warring States period.¹¹ Sworn brotherhoods and friendships during the Tang period often involved Türks, Uyghurs, Shatuo, and Tibetans. Jonathan Skaff emphasised that fictive kinship rhetoric was common diplomatic practice across medieval Eurasia, citing the Byzantine emperors and their neighbours as examples.¹² It should be added that pseudo-kinship played a structural role in Turco–Mongol societies, as different clans that belonged to the same ‘tribe’ were prone to seeing themselves as one kinship unit.¹³ Despite their importance, fictive kinships in Sui and Tang diplomacy are often overshadowed by heqin and are rarely studied as an individual phenomenon. Furthermore, little attention has been paid to the continuation of this practice after the fall of Tang.¹⁴ To my knowledge, only Wang Gungwu has interpreted the kinship terms that were used by Kitan and Song emperors as an extension of Tang rhetoric.¹⁵ This article demonstrates that Song–Kitan diplomatic language did not merely copy the Tang precedent, but derived from Kitan practices of sworn friendship and pseudo-kinship.

Wang Guowei highlighted similarities between Kitan pseudo-kinship and the Mongol tradition of anda, usually described as a sworn brotherhood.¹⁶ By using the Secret History of the Mongols (Mongγol-un niγuča tobčiyan) and the Shengwu qinzheng lu 聖武親征錄 to build a definition of anda, Wang listed four examples from the chronicles of Shengzong and Daozong 道宗 (Yelü Hongji 耶律弘基, 1032–1101, r. from 1055) in the Liaoshi 遼史 (History of the Liao)¹⁷ and concluded that Kitan brotherhoods ‘were completely similar to the Mongol custom of concluding anda (pacts), thus the word anda in Mongolian may have been taken from the Kitans’.¹⁸ Although Paul Pelliot considered this particular note to be highly valuable,¹⁹ no further discussion resulted from it. Isono Fujiko wrote an article in which she questioned the meaning of anda, showing that its connection to pseudo-kinship was only indirect.²⁰ In sum, contributions to the study of pseudo-kinship and anda among Mongols are limited. Furthermore, Wang Guowei's discussion on the Kitan custom remains without posterity.

The rediscovery of multiple inscriptions in Kitan scripts and the progress made in their decipherment through the last and current centuries have increased our understanding of Kitan history and society, especially concerning the eleventh century and the first half of the twelfth century.²¹ These Kitan script sources are supplemented by an even greater number of Chinese epitaphs and other inscriptions. Parallel readings of the Liaoshi and rediscovered inscriptions provide rich information on the pseudo-kinships that were associated with the sixth and seventh Kitan monarchs Shengzong and Xingzong 興宗 (Yelü Zongzhen 耶律宗真, 1016–1055, r. 1031). Surviving Song historical works such as the two Wudaishi 五代史, the Zizhi tongjian 資治通鑑, the Cefu yuangui 冊府元龜, and the Zizhi tongjian kaoyi 資治通鑑考異 document various alliances between emperors that were made during the Five Dynasties and Northern Song periods. Their testimonies are complemented by the two main Yuan works on the Kitan empire: the Qidan guo zhi 契丹國志²² and the Liaoshi. Similarities between the relationships that are displayed in these sources and thirteenth-century Tatar–Mongol pseudo-kinships invite comparison.²³

This article not only seeks to underline common traits between Kitan and Mongol traditions, but also aims to explain the many differences to be found between them. These differences were products of political and cultural circumstances, and, in some cases, appeared as the result of cross-cultural contacts.

Following Wang Guowei's discussion, this article first examines the thirteenth-century Mongol anda, as this practice is better documented and thus constitutes a model against which Kitan practices may be compared. The discussion then moves on to examples of pseudo-kinship in Kitan society. Having established the common patterns of this practice among different steppe cultures, pseudo-kinship in Kitan diplomacy can be explained from its first appearance through its later evolution.

Thirteenth-century Mongol anda and sworn kinships

What does anda mean?

Although there is evidence of the existence of pseudo-kinship and sworn friendship as early as the Five Dynasties, such practices start to be well documented in Turco–Mongol society from the time of Činggis Qan (Temüǰin) onwards. Tatar family chiefs occasionally swore to be anda and to treat each other as brothers. Historiography commonly holds that oath makers had to be of equal status to become sworn brothers.²⁴ The Secret History describes the personal alliances of Yesügei Ba'atur (Temüǰin's father) and Ong Qan (named Toγril), and then Temüǰin and J̌amuqa (circa 1160–1205) in detail. This strongly indicates that becoming someone's anda was at the time an established social practice among most Tatars. Although its purpose was to seek strategic partners outside the restricted family unit, in which brothers were considered natural allies, the oath of the anda was not to be undertaken lightly. Like real brothers, anda would remain connected for the rest of their lives. Since there was no written contract to officialise it, anda friends needed to swear the oath in front of witnesses. This limitation meant that J̌amuqa and Temüǰin had to renew their pledge twice. Similarly, Temüǰin had to remind Ong Qan of his former brotherhood with Yesügei when asking for support.

Tatar–Mongol societies designed pseudo-kinship to be theoretically independent of matrimonial or actual familial ties, as marriage was neither a requirement to become sworn brothers nor an obligatory goal between anda. Therefore, the oath takers’ families could rarely prevent the formation of the pact.²⁵ After Ong Qan agreed to let Temüǰin become an ‘elder brother’ with his son Ilqa Sengün, the latter refused a marriage proposal to strengthen their alliance.²⁶ Temüǰin also never let his relatives marry into J̌amuqa's family, nor did he ever become one of his matrimonial allies—a concept called quda. Although Tatar–Mongols distinguished anda (ritual kinship) from quda (real kinship), they sometimes opted to cement the anda oath with marriage and to become anda–quda. According to a passage in the Jāmi‘ at-tavārīkh of Rashīd ad-Dīn, just after Toru Qajar Bahadur and Sartaq Bahadur (تورو قجر بهادر و سرتاق بهادر) became anda–quda with the Mangγud, Qada'an Daldurqan of the Tarγud decided to do the same, declaring:

We must be family and brothers for one another. As Mongols marry their daughters between them, we shall marry ours the same way and when one of us takes a daughter from another group, we all shall treat [them] as son-in-law and daughter-in-law.²⁷

در آن وقت سوگند خورده اند و در آمده و عهد کرده که مانند اوروغ و برادر یکدیگر باشیم . و چنانکه مغولان دختران یکدیگر بخواهند ما نیز بخواهیم و هریک از ما که دختری از قومی دیگر بخواهد همدیگر را به راه عروسی و دامادی ادب نگاه داریم

Since the Tarγud was a subject clan of the Tayiči'ud Mongols, this speech hints that not all Tatars combined anda with quda, even when they were close to the Mongols who allegedly practised it. The absence of marriage between Temüǰin's family members and ritual kin also indicates that Mongol communities themselves did not always agree on becoming anda–quda. As Chih-Shu Eva Cheng explained, Tatar–Mongols viewed quda as a weaker relationship than anda and natural kinship.²⁸

Both the fact that anda never called each other by any conventional kinship terms other than ‘anda’ and the fact that oath makers could consider a marriage alliance show one essential aspect of this practice: anda did not consider each other as real brothers. They were not full fictive kin, for their respective families would not accept the other as one of them. Constantin d'Ohsson's translation of anda as ‘sworn friend’ (ami juré) appears to be the most accurate. As Isono Fujiko formulated it, anda were ‘brother-like friends’ (kyōdai noyōni nakayoku suru 兄弟のように仲よくする).²⁹ This specific category of alliance could be paired not only with quda, but also with another form of pseudo-kinship: the fictive father–son relationship.

Fictive adoptions

An anda relationship could result in the formation of other pseudo-kinship alliances; as Ágnes Birtalan said: ‘[Temüǰin] inherited the anda-relationship of his father […] with Togril’.³⁰ Based on this assumption, we know that a pact can potentially involve two families indirectly, and lead their members to treat each other with specific obligations. Therefore, beyond the pacts that involve two individuals, an anda relationship can lead to the establishment of long-lasting associations between two families.³¹ Some were successful, such as Qaidu and Baraq, who, according to Rashīd ad-Dīn, after being reconciled, ‘made peace between them, and they swore an oath and became anda to each other—and to this day their descendants are also anda to one another’.³² However, none of these ‘adoptions’ resulted in legal equality between natural sons and ‘adopted’ sons—a particularity that the idea of a matrimonial alliance between Temüǰin and Ilqa Sengün (son of Ong Qan) illustrates. Therefore, these ‘adoptions’ obeyed the same principle behind anda ‘brotherhoods’: they were fictive.

Other fictive adoptions can be observed in the ancient Mongol world. One of the most famous is the ‘adoption’ of the Qočo king (ïduq qut)³³ Barčuq Art Tekin as Činggis Qan's ‘fifth son’ in 1211.³⁴ Thomas T. Allsen pointed out that, albeit he was a ‘son’ of the Mongol ruler, Barčuq also married the latter's daughter Al-Altan,³⁵ hence showing that this filial status was simply ‘honorary in nature’.³⁶ According to the epitaph of Qitai Šari (Qitai Sali 乞台薩里) that was written by Zhao Mengfu 趙孟頫 (1254–1322), ‘the Uyghurs were the strongest and the earliest to join, then an imperial decree accorded to their ruler (Barčuq) the title of ïduq qut and Fifth son. Him [Barčuq] and the imperial sons swore brotherhood, he [his kingdom] was then extraordinarily favoured as the first among all countries’.³⁷ No text speaks of the nature of the relationship between Barčuq and his ‘brothers’; however, we can hypothesise that it was similar to the anda that once tied Činggis Qan and J̌amuqa. Therefore, Mongols closely associated fictive adoptions and sworn brotherhoods (anda). We may point out that this was the only known time at which Činggis Qan applied pseudo-kinship in his diplomacy with other polities. Unlike the Kitans, Mongol rulers quickly abandoned it. The establishment of a quda relationship between the imperial family and the kings of Qočo was quickly followed by the creation of matrimonial kinship between the two clans, making the renewal of pseudo-kinship unnecessary.³⁸

The honorary nature of the ‘Fifth son’ title can also be verified when compared with the early adoption of Šigi-Qutuqu (circa 1178–1260, also named Šigiken-Qutuqu) by Börte (1161–before 1227), the wife of Temüǰin and then still childless. Temüǰin (Činggis Qan) himself came to consider the adopted son as his fifth son, despite Šigi-Qutuqu's being older than all his natural-born sons.³⁹ Boris Vladimirtsov pointed out that, although adopted sons of Mongols did not enjoy equal status with natural sons, they could receive a part of the inheritance.⁴⁰ Šigi-Qutuqu, to whom this case applied, could not become one of the full sons of the qan despite his being admitted into the household. His case differed significantly from that of Barčuq Art Tekin, who took part in Činggis Qan's household only as an imperial in-law.

Manifesting the oath: ritual performances and discourse

In the Shengwu qinzheng lu 聖武親征錄, a note on the first reference to anda 按答 gives its definition: ‘friendship of the exchanged objects’ (jiaowu zhi you 交物之友).⁴¹ The Secret History of the Mongols made it clear that anda must be agreed through a particular ritual: sharing personal items and permitting the other to use them.⁴² Sharing something to which sentiment was attached or possessions that were viewed as very personal was considered to contribute to making the bond stronger. When Temüǰin was 11 years old, he and J̌amuqa exchanged knucklebones and played together, ‘then they declared themselves anda’ (tende anda ke'eldüle'ei 田迭 安荅 客額_勒都列埃).⁴³ Years later, they decided to renew their oath, exchange arrows, and attack the Merkit together.⁴⁴ Once they were victorious, they exchanged the most prestigious pieces of their respective loot. Temüǰin made J̌amuqa ride the horse and wear the sash that he had taken from Toqto'a, and agreed to ride and wear those that J̌amuqa had taken from Dayir Usun.⁴⁵ When Temüǰin went before Ong Qan to ask him to fulfil his duty and become his adoptive father, he gave him the black sable jacket (qara bula'an daqun ^中哈^舌剌 不䶉罕 荅^中忽因) that Börte Üǰin's parents had previously offered as dowry (śitkül 失_惕坤_勒 or emüsgel 額木思格_勒).⁴⁶ According to Elizabeth E. Bacon, Ong Qan, by accepting the dowry, had taken the role of the lost father of Temüǰin, and thus publicly displayed the nature of their negotiated kinship.⁴⁷ In other words, a personal father–son relationship reinforced the pre-existing suzerain–vassal relationship.

The Mongol narratives also depict sworn friends as ‘blood brothers’ who cut their thumbs and, even in today's Mongolia, women shamans cut their fingers to swear sisterhood.⁴⁸ Drinking each other's blood was one of the distinctive features of the anda, which had to be a ‘blood brotherhood’ to be considered authentic.⁴⁹ According to the Secret History, the oath ritual sometimes required animal blood. When the chiefs elected J̌amuqa as the new Tatar emperor, they sacrificed a stallion (aǰirγa 阿只^舌児^中合) and a mare (gegün 格昷)⁵⁰ to become anda allies (andaqol 安荅^中合_勒).⁵¹ The use of human blood in Tatar–Mongol oaths seems to have been inconsistent or subject to unknown rules and evolutions. Pelliot and Hambis said in their commentary of the Shengwu qinzheng lu that, although anda rituals among Uyghurs and Mongols theoretically needed blood, actual pacts rarely featured it, and it eventually became a poetic reference.⁵² Dang Baohai discussed a parallel practice of drinking gold powder mixed with alcohol, showing an adaptation that was practised by the elite from the thirteenth to the seventeenth centuries.⁵³ Whenever the ritual featured blood or the drinking of gold, these consumables enhanced the ceremony and sincerity of the oath.

Mongols admitted consuming gold and drinking each other's blood as two coexisting practices. Yet, as far as we know, they avoided the association of the two. This research has found only one exception to this rule. After the jealous Baha’ al-Din (Baoheding 寶合丁) poisoned Hügeči (Hugechi 忽哥赤), son of Qubilai, in Yunnan, Zhang Lidao 張立道 and 13 others sought vengeance and made an oath together. ‘They pricked their arm to blood and drunk it with gold dust’ to swear brotherhood.⁵⁴ That Zhang had served the Mongols for a long time might not have been unrelated to this, as he imitated the practice of his masters to achieve personal vengeance and combined two distinct rituals.

Tatar–Mongols also relied on the display of mutual feelings to publicly prove the reality of their fraternal friendships. Often lost during the transmission of sources, these details survived only in some documents. According to the Secret History, Temüǰin and J̌amuqa ‘loved each other; they enjoyed themselves revelling and feasting, and at night they slept together, the two of them alone under their blanket’.⁵⁵ Expression of feelings between oath makers seems to have been used as a concrete manifestation of pseudo-kinship between individuals and served to formalise the pact.⁵⁶ In a military context, it aimed to convince soldiers under the command of both generals to cooperate and fight for a common goal.

As the examples above illustrate, and as Birtalan insisted upon, Mongol anda and related pacts of pseudo-kinship always conflated political and military purposes.⁵⁷ Birtalan and Isono also pointed out that the practice became sporadic after the collapse of the Mongol empire.⁵⁸ Nonetheless, anda appears to us as a rare occurrence even during the pre-imperial and Činggis Qan period Tatars. It is unclear whether this impression was caused by a lack of documentation or by the actual rarity of such pacts during this period. Nonetheless, as sources began to grow numerous for later periods, the presence of anda faded away, which hints at a significant decrease in its use along with the pacification and unification of Mongol communities under large domains (ulus).

On pseudo-kinship in Kitan society

A Kitan anda: ‘pricked-to-blood friendships’

While discussing Wang Guowei's note on ‘Kitan anda’, Isono Fujiko pointed out that the friendships that were described in the Liaoshi were contracted between people of unequal status, and that their equivalence with anda still lacks decisive proof.⁵⁹ Because of the scarcity of sources at our disposal, it is indeed difficult to determine the degree of similitude between the eleventh-century Kitan and late-twelfth-century Mongol brotherly friendships. However, as explained below, they shared common roots, which can be observed through the similarity of their implicit rules and rituals.

Linguists and historians point out that the Mongol anda came from the Turkic ant that meant ‘oath’, ‘contract’, or ‘to pledge’.⁶⁰ It is therefore not surprising to observe that this word was absent in Kitan vocabulary, in which the only equivalent that has been identified so far is *nugur (pl. *nugji), which shares its etymology with the Mongol nökör (friend).⁶¹ As Jirüke understood it in his 2014 article, *nugur was the equivalent of ‘companion’ or ‘partner’. In a broad sense, it can be used to describe an attendant, a wife, or a close friend, and it can sometimes mean ‘friendship’.⁶² Although it is the closest known Kitan term to anda, it does not have the same implications and is too vague to express the concept of sworn friendship or pseudo-brotherhood in the way in which the Mongol terms did.

The Kitan Small Script and Chinese epitaphs of Yelü Renxian 耶律仁先 (1013–1072, Kitan name *Tioriń Cara⁶³) were discovered together on Lianhuashan 蓮花山, or Lotus Mountain (Qinghemen District, Fuxin, Liaoning).⁶⁴ The Kitan text mentions the special relationship that his father Yelü Sizhong 耶律思忠 (*Caran Kuin⁶⁵) had with the seventh Kitan ruler Xingzong.⁶⁶ The following is an interpretation of the text that was made by Ji Shi and translated by me with slight modifications: ‘The posthumous promotion to the title of Prince of Yan (燕王) of his second son, who pricked (?)⁶⁷ to bleed with *Au Ordula'ar Hung Ti (i.e. Shengzong).’⁶⁸

Ji Shi sees cis *pir-er, where cis means ‘blood’, as the equivalent of the Chinese cixue 刺血 ‘prick to blood’ found in the Liaoshi. Han Baoxing made a connection between this epitaph and the ‘Biography of Yelü Xinxian’ in the Liaoshi. Yelü Xinxian 耶律信先 (*Niarġun⁶⁹) was the brother of Renxian, and his biography also recorded the event described in the epitaph:

Because his [Yelü Xinxian's] father Guiyin (i.e. Yelü Sizhong), was a ‘pricked-to-blood friend’ with Xingzong, he was raised in the imperial palace. […] The emperor asked him what he wished for, Xinxian said: ‘Although my deceased father Guiyin was like a brother to Your Majesty, he was never granted any princely title. If imperial benevolence could reach the underworld, I would desire this done.’ The emperor said: ‘This is something we have neglected.’ He [Guiyin] then was granted post-mortem the title of Prince of Yan.⁷⁰

興宗以其父瑰引為刺血友，幼養於宮。[…] 上問所欲，信先曰：「先臣瑰引與陛下分如同氣，然不及王封。儻使蒙恩地下，臣願畢矣。」上曰：「此朕遺忘之過。」追封燕王。

Not only do these two texts prove that the Kitans admitted some kind of sworn brotherhood tradition, but the term cixue you 刺血友,⁷¹ ‘pricked-to-blood friend’, also explicitly implies that oath makers went through a blood pact. The use of you 友 hints at one aspect of the Chinese point of view on this practice, which categorised it as a kind of ‘friendship’.

As we can infer from the term cixue you, blood that was obtained from self-inflicted cuts played a key role in Kitan rituals. The Liaoshi informs us that ‘Shengzong once pricked his arm to blood with Honggu 弘古 (Kitan name *Qudugin)⁷² to ally himself with him as a “friend”, he was treated with unusual privilege, becoming Grand Counsellor of the Southern Administration and viceroy of Shangjing’.⁷³ Song Qi 宋琪, a man from You 幽 (the region around Beijing) who served Taizong 太宗 of the Song (Zhao Jiong 趙炅, 939–997, r. from 976), once said: ‘The Qai (Xi 奚) and Xi 霫 tribes (neighbours of the Kitans with whom they shared a similar culture), at the time of Liu Rengong 劉仁恭 and his son Liu Shouguang 劉守光 (both died in 914), used to cut their faces to become foster sons.’⁷⁴ Song Qi confirms here that ritual kinship required a blood pact that involved all oath makers. The relative absence of blood pacts in available sources about Tatar–Mongol anda contrasts with its almost systematic presence in accounts about Kitans. Therefore, rituals that involved blood seem to have been a signature of Kitan oaths. As they used self-inflicted bloodletting to convey solemn and sincere resolution, the ‘blood tears’ rite was also widespread among the Kitans and Qai aristocracy.⁷⁵ This expression of mourning was similar to the Turkic ritual that is described in the Zhoushu 周書 (Book of the Zhou, p. 636): ‘One must cut his face with a knife while crying, so blood and tears would run down together, repeat this seven times, then stop.’⁷⁶

Aside from blood rituals, the Kitan oath also involved an exchange of personal objects in a similar manner to the Mongol anda. One instance is the case of Shengzong, whose mother made him become friends with General Yelü Xiezhen 耶律斜軫, one of her most trusted aids, after he ascended to the throne in 983. The young emperor and the veteran general ‘exchanged their bow and arrows, and their saddle and horse in front of the Empress Dowager (Chengtian 承天, 953–1009)’.⁷⁷ In 1015, ‘he [Shengzong] honoured Madugu's achievements and exchanged with him his clothes and horses to be on good terms (i.e. friends)’.⁷⁸

Chinese chroniclers under the Jin 金 (1117–1234)⁷⁹ and the Yuan might have considered the mention of rituals to be superfluous and chose not to include them in the Liaoshi or in other previous accounts, which may explain why the official history also has instances of friendship oaths with no mention of a ritual.⁸⁰ In any case, descriptions of Kitan sworn friendship or brotherhood rituals in Chinese sources all involve either the consumption of blood or the exchange of objects, which links the practice with that of Tatar–Mongols. Based on extant records, the consumption of gold remains absent from Kitan oaths. More differences between Kitan and Tatar–Mongol pseudo-kinship can be found through a closer examination. However, these differences manifest as variations that were imposed in the context of the respective socio-political backdrop: while the times of Činggis Qan were of war and military covenants, the rule of the Kitans saw flourish a time of peace under an unchallenged leadership.

A well-documented case: the pseudo-kinship between the Yelü and Han families

The most well-known example of pseudo-kinship between two families in the Kitan empire is the one that connected the Han family 韓 lineage from Yutian 玉田 (descendants of Han Zhigu 韓知古) with the Yelü 耶律 imperial family.⁸¹ Scholars have paid close attention to the Han family, identifying it as the most favoured Chinese clan under the Kitan regime. Pamela Crossley showed that, among the elite, ethnic differences tended to be ignored; consequently, the Han presented themselves as an integrated part of the Liao nobility.⁸² Han Kuangsi 韓匡嗣 and the second emperor (Taizong, Yelü Deguang 耶律德光, 902–947, r. from 927)'s younger brother and official heir Lihu 李胡 swore brotherhood, and later his son Han Derang 韓德讓 came to be considered a brother of Shengzong.⁸³ Derang became known as Yelü Longyun 耶律隆運 after the death of the empress dowager.⁸⁴ The Liaoshi does not mention the brotherhood between Shengzong and his powerful minister. However, the presence of such an oath can be deduced from the contents of the recently discovered funerary inscription of Yelü Longyun. The epitaph states that, during the two years that separated the deaths of his mother and the minister (between 1009 and 1011), Shengzong often said about Derang: ‘Despite being the Son of Heaven, I must have an elder, so I call him my elder brother.’ The emperor ‘then linked him to the imperial personal name, bestowing upon him the name of Longyun’.⁸⁵ The introduction of the character long 隆 in Derang's name in 1009, according to the Liaoshi, confirms this year as the date of the oath.⁸⁶ The three sons who were born from Emperor Jingzong 景宗 (Mingyi 明扆, 948–982, r. from 969) and Empress Chengtian—Shengzong and his two younger brothers—bore the Chinese names of Longxu 隆緖, Longqing 隆慶, and Longyu 隆裕. Therefore, the new name that was given to Han Derang aligned him with the imperial brothers. Yet, as Kitans never bore surnames (Yelü and Xiao are only found in Chinese-language material),⁸⁷ the bestowal of the imperial surname (cixing 賜姓) had little meaning in the eyes of the elite.⁸⁸ The Kitan ruler gave the imperial patronym to Han Derang/Yelü Longyun to make ritual kinship, which was then alien to Chinese tradition, known and understood by Chinese subjects. Thus, the surname bestowment served as an equivalent to the Kitan sworn brotherhood. However, as Han Kuangsi's descendants obtained the Yelü surname together with Derang, later oaths only resulted in the change of the personal name.

The best-documented case of pseudo-kinship involves Xingzong and Yelü Longyun's nephew, Han Dilu 韓滌魯 (*Sengin Tirug).⁸⁹ Although it had a clear political purpose, their sworn brotherhood allegedly originated from genuine friendship, as Dilu was a childhood friend of Xingzong. Three independent Chinese sources record this story: Han Dilu's biography in the Liaoshi, his epitaph, and the funerary inscription of Lady *Urbin.⁹⁰ These three texts not only affirm that Dilu considered Xingzong to be his brother, but also that he was adopted by Shengzong before this:

1. 1) Liaoshi:

Dilu bore Zunning as his surname. He was raised in the imperial palace during his childhood and became a Lesser General. […] Dilu was quick-witted and sharp, he was considered a son of Shengzong, and Xingzong treated him as an elder brother. He remained humble despite his noble status.⁹¹

滌魯，字遵寧。幼養宮中，授小將軍。[…] 滌魯神情秀徹，聖宗子視之，興宗待以兄禮，雖貴愈謙。

1. 2) Epitaph of Han Dilu/*Sengin Tirug:

The prince's name was Zongfu and his first name came from the imperial family. […] In the middle of the Tonghe era, Emperor Shengzong exceptionally raised him and gave him the same dignity as his own sons. Today, we consider that he and Xingzong treated each other as brothers. Raised in the Forbidden Quarters and honoured to be linked to the emperor by name, he received extraordinary favours.⁹²

王諱宗福，氏出國姓。[…] 時統和中，特蒙聖宗皇帝升于子息之曹，令与興宗叅于昆弟之列。貴處宸禁，榮連御名，寵也。

1. 3) Epitaph of Lady Wuluben/*Urbin au'ui:

The Chamberlain's second son Xunning Diligu (*Sengin Tirug) was Southern Chancellor and Prince of Han […]. He used to be an adopted son in Emperor Shengzong's palace. His Highness bestowed on him a name linked with Xingzong's name: Zongfu.⁹³

侍中次子遜寧迪里姑南宰相、韓王 […]。曾在聖宗皇帝宮中為養子。御賜與興宗連諱宗福。

According to these texts, the ‘adoption’ of Han Dilu by Shengzong resulted in his friendship with Xingzong, or vice versa. The modification of Dilu's Chinese name from Han Yuanzuo 韓元佐⁹⁴ to Yelü Zongfu 耶律宗福 (the character zong 宗 linked together the sons of Shengzong) follows the same logic as Han Derang's name change. As for the name ‘Han Dilu’, it is the product of later historiography in which Chinese patronyms were systematically added before the transcription of Kitan names.

In Chinese thought—especially in the neo-Confucian discourse—morality required a clear-cut separation between brotherhood and friendship. This went against the ambiguous Kitan and Tatar–Mongol model of sworn friendship that was designed to blur the frontier between these two kinds of relationships.⁹⁵ As the following quote from the Qidan guo zhi illustrates, the Chinese viewed Kitan brotherhoods between men of different statuses as immoral:

After he [Xingzong] began to reign by himself, he started to indulge in debauchery and let himself to boundless extravagancies, he was dissolute and unrestrained. He once swore to be a brother to Wang Shuiqing from the Musician's Quartier and ten other people, who were in and out of his house and even courteously visited his parents. He changed his clothing to go incognito and often went to taverns where he indulged in obscene speech and indecent language. He returned only when fully content.⁹⁶

既親政，後始自恣，拓落高曠，放蕩不覊。嘗與教坊使王稅輕等數十人約為兄弟，出入其家，至拜其父母。變服微行，數入酒肆，褻言狎語，盡懽而返。

By using this ‘barbarian’ form of brotherhood as an argument against the monarch, the author of the Qidan guo zhi provides a valuable description of how sworn friends were treated by each other's families. Recognition of a sworn brother by the family was far from being a real adoption. As far as available sources tell, they were treated as close friends, maybe with additional privileges. However, the establishment of a sworn friendship between both individuals did not significantly alter their legal status towards each other. These kinds of sworn friendships may have been quite rare, and the way in which the families would have interacted with each other must also have varied. Ultimately, the extent of these variations in both Kitan and Tatar–Mongol customs of pseudo-kinship is impossible to determine, as the available sources are scant, often incomplete, and tend to focus on the oath and its immediate political consequences.

The Kitan pseudo-kinship diplomacy

From father–son to uncle–nephew relationships: an evolution of kinship terms

Different traditions of ritual kinship coexisted during the Tang period. One was imported from the Central Asian martial tradition of čākar, mainly through Sogdian generals. Čākar were soldiers who devoted their lives to their master and were treated like sons by him. They were usually numerous and allowed the generals to federate large armies.⁹⁷ Another practice came from the ‘Turco–Mongols’, who concluded personal alliances between men, often as a sworn brotherhood, but sometimes as father–son or uncle–nephew relationships.⁹⁸ Far different from the čākar tradition, this kind of political alliance involved fewer oath makers—typically only two. Kitan emperors’ pseudo-kinships with Five Dynasties and Song leaders, and contemporary social practices that were observed among Kitans and Mongols all belong to this second category.

The establishment of father–son relationships stratified authority in the ancient steppe. Türk qaghans, for example, called their vassals ‘sons’ (oγlan) in the Orkhon inscriptions.⁹⁹ After many years of war with the Tang, Bilgä Qağan asked for peace and to become the ‘son’ of Emperor Xuanzong 玄宗 (Li Longji 李隆基, r. 745–756).¹⁰⁰ The father–son relationship continued when Tengri Qağan succeeded his father in 734. The official letter that was written by Zhang Jiuling 張九齡 (678–740) for Xuanzong to the new qaghan expressed the will to maintain the father–son relationship—a gesture to avoid a grandfather–grandson relationship that could prove to be humiliating to the Türks:

I and the parent of you, the qaghan, shared the deep affection of blood kinship, so that he was with me like a son, therefore, you, the qaghan, will from now on be a grandson. […] If you were to be my grandson, we could grow distant from each other, therefore, I wish that you, the qaghan, will now become my son for our filial bond to be stronger.¹⁰¹

朕與可汗先人， 情重骨肉，亦既與朕為子， 可汗即合為孫。[…] 若以為孫漸成疏遠，故欲可汗今者還且為兒，義結既深。

Incidentally, Xuanzong's explanation shows that, under normal circumstances, his relationship with Tengri Qağan would have been that of grandfather–grandson. This obeys the rules that were observed by Ong Qan when he accepted Temüǰin as his son almost five centuries later. Yin Lei interpreted the previous examples for the Türks as political projections of a paradigm in which the Heaven fathers all beings: the ‘Celestial qaghans’ (tian kehan 天可汗) acted as paternal figures to the other lords.¹⁰² Yin also situates the Kitan efforts to establish pseudo-kinships with Five Dynasties’ rulers as an attempt to bring the Central Plain under the paternity of the Celestial qaghan. The similarities between the aforementioned strategies also show that Türks and Kitans shared a common concept of imperialism.

After the Chanyuan Covenant in 1005, emperors of the Kitan ‘Northern court’ (beichao 北朝) and Song ‘Southern court’ (nanchao 南朝) greeted each other with kinship terms. The best-known study in a Western language is the work of David C. Wright, who described in detail the pattern of the Song–Kitan diplomatic routine and its general rules.¹⁰³ In Chinese historiography, in 1940, Nie Chongqi first suggested that pseudo-kinship alliances between Kitans and Five Dynasties emperors and the alliance that resulted from the Chanyuan Covenant shared similarities.¹⁰⁴ Tao Jing-shen, Zhang Guoqing, and Mōri Eisuke further explained how age and generation dictated the kinship term that one emperor would call the other.¹⁰⁵ In other words, although pseudo-kinship terms that were set at the time of the alliance reflected the power balance between the parties involved, subsequent changes happened that followed the implicit rule of pseudo-family without altering the political balance. This can be interpreted as the logical shift from an agreement between two individuals to a predetermined relationship between two families.

Empress Chengtian and later her son Shengzong used these friendships as a political strategy to ensure the aristocracy's support by displaying proximity and feelings towards them. This shows how some nomadic societies welcomed and valued these brotherhoods, and how the Kitans adapted Chinese political rhetoric to their views and benefit. On the other hand, the Song government would only reluctantly resort to pseudo-kinship to normalise relations with entities whom they viewed as inferior ‘Barbarians’—initiators of such diplomacy could only have been polities that embraced steppe culture. The Chinese epitaph that is dedicated to Shengzong praises his treaty with the Song, saying: ‘he honoured the sage and good men, and wished to call them son and nephew; he cherished friendship and wanted to be a brother [with him]’.¹⁰⁶ Song subjects did not share the same enthusiasm, as they only accepted this form of association out of pragmatism, treating it as a lesser evil.¹⁰⁷

Kinship terms that emperors of the Five Dynasties and Song agreed to use in their relations with the North changed over time. Although all of the terms that were used imply a familial relationship, none of these pacts required marriage and no matrimonial union took place between imperial families. As shown in Table 1, Wright's observations on the Kitan–Song diplomacy follow my observations on the Five Dynasties period.

Table 1. Kitan rulers and their diplomatic kinship with Chinese and Shatuo rulers

^a According to the Zizhi tongjian (Shanghai: 1956, 269: 8810, 275: 8989), Abaoji called Li Cunxu his son in front of the Tang envoy Yao Kun 姚坤, and Li Cunxu used to call Abaoji his ‘uncle’ (shu 叔). However, no source testifies that the two used these honorifics in their diplomatic correspondence.

^b Mōri Eisuke 毛利英介, ‘Sen'en no mei no rekishiteki haikei: Unchū no kaimei kara Sen'en no mei e’ 澶淵の盟の歴史的背景mei eeei ee m, Shirin 史林 LXXXIX (2006), p. 84.

^c Both the Jiu Wudaishi 舊五代史 (Beijing: 2016, 43: 591) and the Cefu yuangui (Beijing, 1994, 170: 2058–2) put the following words into Li Siyuan's mouth: ‘I and his father (Tuyu's father Abaoji) swore brotherhood’ (吾與其先人約為兄弟). Such an alliance between Abaoji and Li Siyuan seems inconceivable, as the Kitan ruler died only a few months after Li Cunxu. Being the envoy that informed Abaoji of the Tang emperor's passing, Yao Kun was still in the Northern court when the ruler died. It is therefore highly possible that ‘xianren’ 先人 in the Chinese text was added during the compilation of the Jiu Wudaishi and that the authors of the Cefu yuangui copied from it, or that the mistake was found in a text from the Five Dynasties. The original meaning was perhaps that Li Siyuan and Taizong tied themselves as brothers and that Tuyu benefited from it.

^d According to the Nantang shu 南唐書 that was written by Lu You 陸游 (1125–1210), a Kitan message that treated Li Bian ‘according to the fraternal etiquette’ (以兄禮事帝) was received in Jinling 金陵 (Nanjing, Jiangsu) in 938. Nantang shu (liang zhong) 南唐書 (兩種) (Nanjing, 2010), 1: 219; see also Cao Liu 曹流, Qidan yu Wudai guo zhengzhi guanxi zhu wenti 契丹與五代國政治關係諸問題 (unpublished PhD dissertation, Peking University, 2010), p. 54. However, it does not seem that the Southern Tang court agreed to officialise this relationship.

^e As the negotiations between Southern Tang and the Kitan in 947 were about military and political alliance, it is not impossible that Taizong wished to be ‘father’ or ‘brother’ with Li Jing. Cao, Qidan yu Wudai guo zhengzhi guanxi zhu wenti, pp. 64–66.

^f The draft of an official letter that was discovered in Dunhuang testifies that Shi Zhonggui considered himself and Taizong as two monarchs with equal status. Yang Lien-Sheng, ‘A ‘posthumous letter’ from the Chin emperor to the Kitan emperor in 942’, Harvard Journal of Asiatic Studies, 10 (1947), p. 418. Moreover, no sources confirm that the two emperors had officially called themselves ‘grandfather’ and ‘grandson’ before 947. Shi Zhonggui acknowledged his status of grandson only when he submitted to Taizong following the Southern campaign of 946–947. Jiu Wudaishi, 85: 1306–1307; Xin Wudaishi 新五代史, Beijing: 2016, 17: 204–205.

^g Since the official kinship between Shizong and Liu Min was one generation apart, the first Northern Han ruler is considered here as being from the ‘second generation’.

^h Despite being a short-lived relationship, this alliance is recorded in the Jiu Wudaishi (135: 2110), the Xin Wudaishi (70: 978), the Zizhi tongjian (290: 9460), the Liaoshi (5: 72, 85: 1450), and the Songshi (Beijing: 1977, 482: 13934). Only the epitaph of Liu Jiwen 劉繼文 (?–981), who was a cousin of the last Northern Han emperor, and the Songshi say that it was a father–son relationship. The epitaph was written a few months after Liu Jiwen's death and was rediscovered in 1926 in the county of Jianchang (Chaoyang, Liaoning); see the edited text in Xiang Nan 向南, Liaodai shikewen bian 遼代石刻文編 (Shijiazhuang, 1995), pp. 71–73.

ⁱ The two monarchs may have had a pseudo-kinship relationship, but available sources are silent on the matter.

^j Xin Wudaishi, 70: 981.

^k If any, it was a short-lived relationship.

^l Only known indirectly through one of the biographies of the Liaoshi (72: 1338).

^m This table was made according to D. C. Wright, ‘Parity, pedigree, and peace: routine Sung diplomatic missives to the Liao’, Journal of Song-Yuan Studies XXVI (1996), p. 70.

As Mōri Eisuke explained, Kitans and Later Tang based their pseudo-kinship relations more on generation and less on pragmatic considerations, though the latter played a decisive role. Mōri further suggested that the father–son relationship that Taizong established with the first emperor of the Later Jin was decided under the influence of previous alliances with Later Tang rulers. Indeed, Jingtang was Li Siyuan's son-in-law and therefore one generation younger than Taizong.¹⁰⁸ With this being the case, pseudo-kinship alliances with Later Jin may have been perceived by Kitans as a logical continuation of the alliances with Later Tang. Such ritual kinships seem to have abided by apparent rules: the two ancestors who swore brotherhood were counted as the first generation, and the pseudo-kinship relation between their descendants depended on the generation difference, according to which they were either brothers, father and son, or grandfather and grandson.

As the table shows, a change in the early diplomatic tradition appeared after Taizong failed to conquer the Central Plain in 947. The new Kitan monarch, Shizong (Wuyu 兀欲, 919–951, r. from 947), agreed upon an uncle–nephew alliance with the Northern Han ruler Liu Min, which added a new stage to the hierarchy of pseudo-kinship. However, the father–son relationship between Muzong (Shulü 述律, 931–969, r. from 951) and Liu Chengjun went against the generational order that had previously been dictated. Muzong, as the cousin of Shizong, and Liu Chengjun, as the son of Liu Min, should have been considered as being two generations apart. However, this is the only known breach of the rule.

During the Chanyuan Covenant, the empress dowager sought to revive the pseudo-kinship relationship between the Kitan and Central Plain emperors. When Zhenzong died in 1022, both courts agreed that, instead of calling Shengzong ‘father’, the newly enthroned Renzong emperor would call him his ‘uncle’ (bo 伯) and that he should be treated accordingly as a ‘nephew’ (zhi 姪).¹⁰⁹ When Zhezong inherited the throne from his father in 1085, another generation difference was added. Zhezong and Huizong called Daozong their ‘granduncle’ (shuzu 叔祖) and called themselves ‘grandnephews’ (zhisun 姪孫).¹¹⁰ This replacement of the father–son relationship by an uncle–nephew model differentiates the post-Chanyuan period from the previous one. By this point, the Kitan–Song pseudo-kinship had already been heavily influenced by a century of evolution and was distinctly unlike the original brotherhood between Abaoji and Li Keyong that had taken place in 905. Not only did the use of kinship terms change, but the ritual behind the oath and the oath itself were also significantly transformed.

It should be noted that, from 1031, Kitan empress dowagers also took an active part in this pseudo-kinship diplomacy. The Kaifeng court had to dispatch separate embassies, gifts, and letters for both the Northern ruler and his mother. Song rulers called the Liao dowager ‘aunt’ (shen 嬸) or ‘great aunt’ (shuzumu 叔祖母) according to the official kin terms that were used with the emperor.¹¹¹ Between 1058 and 1063, when Daozong of the Kitans called himself the ‘nephew’ of Renzong of the Song, the Southern emperor was, for the first time, of the same fictive generation as the Northern dowager. Thus, he referred to Empress Zongtian 宗天 (also called by her posthumous title Renyi 仁懿) as his ‘sister-in-law’ (difu 弟婦).¹¹² The Song court objected that it was inappropriate for a man to directly communicate with his sister-in-law. Daozong agreed with this view and halted official communication between his mother and his fictive uncle.¹¹³ This new adaptation to the pseudo-kinship diplomacy illustrates the Kitans’ readiness to adjust their diplomatic practices to fit the social requirements of the Song. This behaviour undoubtedly permitted the continuity of the pseudo-kinship diplomacy and, as the following part will show, its shift from the original rituals of sworn brotherhood or friendship.

From the oath to the treaty: a changing ritual

Most early pseudo-kinship alliances that were made by emperors involved the rituals already described above for the Kitan and Tatar–Mongol oaths. The Kitans based their early diplomacy according to their own rules and imposed them on the courts of Luoyang and Kaifeng. However, the pact and the ritual evolved alongside the change in kinship terms.

From the beginning, Kitan alliances with Shatuo leaders followed the practices that were found among themselves and, later, among Tatar–Mongols. According to the account of the event that was found in the Liaoshi, when Li Keyong and Abaoji allied themselves near present-day Datong in 905, ‘they exchanged their robes and horses, and became brothers’.¹¹⁴ This exchange is different from the lavish gifts that they gave each other after the pact was concluded. These robes and horses might have been personal goods that were similar to those that the Tatar–Mongols would later exchange during Temüǰin's time. When Taizong and Shi Jingtang swore to be father and son, the Kitan ‘took off his white sable coat and made the emperor [Shi Jingtang] wear it’.¹¹⁵ The court scholar Wang Pu 王溥 (922–982) noted, in his Wudai huiyao 五代會要, an important detail about the alliance between Taizong and Shi Jingtang in 936:¹¹⁶

In the first year of the Tianfu era of the [Later] Jin, […] during the eleventh month Gaozu (Shi Jingtang) seized the throne, and, Deguang (Taizong) having the strength to support him, smeared blood on their mouth (shaxue) to ally themselves as father and son.

晉天福元年……十一月，高祖踐位，以德光有援助之力，歃血盟結為父子。

The word shaxue 歃血 refers to alliances in which the oath takers smeared blood on their mouths, and possibly drank it, during the Warring States period. Wang Pu used this term because of its literary value rather than its accuracy.¹¹⁷ Therefore, we should not interpret this word anachronistically according to its classical definition. In the present case, the Kitans were clearly the initiative and there is little doubt that the ritual behind shaxue was identical to the aforementioned cixue pact.¹¹⁸ According to the Jiu Wudaishi, Taizong and Shi Jingtang ‘held hands while crying, for a long time they could not part each other’.¹¹⁹ Such public manifestation of mutual feelings of friendship can be linked to the expression of friendly bonds that can be found in Mongol narratives: Abaoji with Li Keyong, Taizong with Shi Jingtang, Xingzong with Dilu, Temüǰin with J̌amuqa—in all these cases, we see the same ritual gestures, behaviours, and implications.

Despite the Northern origins of the alliances between Kitan and Five Dynasties’ rulers, the methods for contracting these diplomatic pacts evolved alongside the socio-political background. Agreements that were made between Abaoji and Li Keyong, and thereafter between Taizong and Shi Jingtang, initiated a hereditary transmission of the pseudo-kin relationship. This meant that Li Siyuan did not need to perform a ritual in order to treat the Kitan prince Tuyu as his pseudo-kin (maybe his brother)¹²⁰ and Shi Zhonggui was able to recognise his status as a ‘grandson’ of Taizong through a letter.¹²¹ After the agreement with Liu Min in 947, new pseudo-kinship relations stopped being initiated through personal meetings. Rituals such as the exchange of objects or blood pacts became superfluous or impossible to perform. The geographical distance between the oath takers caused a shift from the personal encounter to the exchange of embassies. The adoption of an intense written diplomatic activity by the Kitan court also enabled the introduction of pseudo-kinship in official correspondence, which made the shift possible. As a result, although both Song and Kitan rulers were geographically close to each other during the Chanyuan Covenant, they set the terms of the alliance solely through the exchange of diplomats, translators, and texts.¹²²

The pseudo-kinship diplomacy under the Jin: a transformed heritage

Pseudo-kinship diplomacy did not disappear with the Jurchen conquest of the Kitan empire and Northern China between 1117 and 1128. The Liaoshi and Jinshi 金史—the official history of the Jin—both state that the Jurchen conqueror Aguda 阿骨打 (Taizu of the Jin, 1068–1123, emperor in 1117) attempted to negotiate peace with Emperor Tianzuo 天祚 (Yelü Yanxi 耶律延禧, 1075–1128, r. 1101–1125) of the Kitans through a mutual recognition as ‘brothers’.¹²³ However, the Kitan refused and later lost the war. The Jin then utilised pseudo-kinship with surrounding states. They conquered Northern China and set the puppet state of Qi 齊 (1130–1137) there, with Liu Yu 劉豫 (1073–1146) as its emperor. The Jin ruler Taizong (Wuqimai 吳乞買, 1075–1135, r. from 1123) required that he should ‘be hereditarily treated according to sons etiquette’ (shi xiu zili 世修子禮).¹²⁴ However, immediately after the passing of Taizong in 1139, the Jin court issued an edict that ordered the Qi ruler to call itself the ‘servant’ (chen 臣) of the third Jurchen ruler Xizong 熙宗 (Wanyan Dan 完顏亶, 1119–1150, r. from 1135).¹²⁵ After several decades during which Song rulers acknowledged themselves as the ‘servants’ of the Jin, both sides reached a new agreement through the Longxing 隆興 Treaty of 1165 and ensured an uncle–nephew relationship between both emperors.¹²⁶ Finally, as the pressure of the Mongols grew, the Jin emperor requested an alliance with the Tangut emperor of Xia 夏 (1038–1227) in 1225. They negotiated that both emperors should be ‘brothers’.¹²⁷ The alternative use of pseudo-kinship terms and a ‘suzerain–subject’ relationship sets the Jurchens apart from the Kitans.

A closer examination of these treaties shows that the Jurchen diplomacy did not follow the same rules as the Kitan diplomacy. First, the treaty that established Liu Yu as the ‘son’ of Taizong specified that his status would be ‘hereditary’ (shi 世), which could imply that it would not be affected by the gaps in fictional generations. According to Xu Mengshen's 徐夢莘 (1124–1207) historical compendium, the Sanchao beimeng huibian 三朝北盟會編, the Jin ‘did not know whether to call each other as “brothers”, “uncle and nephew” or “good friends”’ when they started to negotiate the new treaty in 1123.¹²⁸ The Song advised that the relationship should imitate that of Song–Liao and then settled for an uncle–nephew kinship.¹²⁹ However, the Jin emperor called his Chinese counterpart ‘nephew’ (zhi), while the Song emperors used either bo or shu to refer to the Jin emperor, depending on the age difference.¹³⁰ In the third and last case of diplomatic brotherhood, although the last Jin ruler Aizong 哀宗 (Wanyan Shouxu 完顏守緒, 1198–1234, r. from 1224) was more than 30 years younger than the Xia ruler Shenzong 神宗 (Weiming Zunxu 嵬名遵頊, 1163–1226, r. 1211–1223), the first became the elder brother of the latter.¹³¹ The disregard of Jurchen emperors for the conventions that were set up by Kitans shows that their ideas of fictive kinship significantly differed.¹³² For the Jin and the Song, the pseudo-family of the Chanyuan Covenant became a diplomatic norm that they simply adapted to their needs, disregarding the kinship dynamic that had originally presided over it. Thus, it was the collapse of the Kitan empire that ended the dominance of the Kitan–Mongol model of pseudo-kinship in the East Asian diplomatic world.

Conclusion

The two centuries of pseudo-kinship diplomacy between Kitans and the Chinese dynasts was a unique phenomenon during the long history of relationships between the steppe and the Central Plains polities. Its apparition and later development took shape alongside the formation of a post-Tang order. The establishment that was reached after the Chanyuan Covenant qualified the new special relationship that tied together both rulers and their domains. Modern analyses of this alliance tend to follow the perception that was held in the Song court, which accepted the use of kinship terms as a mere decoy that was hiding the unequal nature of the treaty. This point of view emptied the pseudo-kinship of its original intent and marginalised it, hence its relative absence in Song discourse on diplomacy with the North. The importance that this status had for the Kitans is, by contrast, relatively unknown. As this article has shown, the emperor and the imperial clan used pseudo-kinship to foster long-lasting and stable alliances with other important noble families. From a Kitan perspective, the Yelü imperial clan chiefs were pseudo-kin of both the Han family and the Zhao emperors of Song.

Less than two years before his death, the ageing Xingzong emperor asked to receive the portrait of Renzong, saying: ‘Us and the Song master swore brotherhood. For many years we enjoyed happiness and peace, thus I wish to see his painted portrait. We shall instruct this to his envoy.’¹³³ This text is the only extant evidence to show the direct link that the Kitans drew between sworn brotherhood, peace, and international order. Through such small hints, we can still recognise that the relationship that Northern and Southern emperors endorsed in 1005 found resonance within social practices and representations of alliance among the Kitans and other steppe communities.

Not only did Kitans and Mongols borrow from real or fictional precedents and rituals to form these alliances, but they also naturally adapted the assumed relationship to their own concepts. These pseudo-kinship alliances were designed with two major and related purposes. The first was to create a personal alliance that convinced their surrounding communities and the second was to constrain oath makers into a commitment that was expressly determined or tacitly understood. Both Kitan and Mongol oath takers had the alliance take the shape of a ‘better friendship’ that elevated strictly strategic alliances into a quasi-familial moral commitment. Equality and inequality of status could be shown by adopting a fictive brotherly or father–son relationship, respectively. Kitans and Mongols also commonly understood that the pseudo-kinship relationship could be extended by family members of both sides in such a way that ‘independent’ or ‘original’ pseudo-kinship had to be distinguished from ‘secondary’ ones. Therefore, a secondary father–son pseudo-kinship that continued an original sworn brotherhood often had different implications from an independently established father–son pseudo-kinship. Kinship terms that were used between oath takers in secondary pseudo-kinship rarely reflected actual equality or a difference in status, but were of ritual and memorial significance that was ultimately tied to the original oath.

The Kitan–Mongol model of pseudo-kinship that only theoretically appears through the observable similarities that are presented in this article might be close to the concepts behind the alliances that had been initiated by Türks and Uyghurs several centuries prior. Better understanding of the poorly documented pseudo-kinships that involved Tang China's Northern neighbours would be beneficial to determine not only what the period that lasted from the tenth to the fourteenth centuries meant in the longue durée of diplomacy between the steppe and the Central Plain, but also the forms that friendship could take in ancient steppe societies.