Tang Taizong in Korea: The Siege of Ansi

Tineke D’Haeseleer


In the lengthy history of Sino-Korean relations, only two Chinese emperors ever set foot on Korean territory: Sui Yangdi 隋煬帝 (r. 604–17) in the campaign against Koguryŏ (Chin. Gaogouli 高句麗) in 611, and Tang Taizong 唐太宗 (r. 626–49) in a similar effort to subdue the kingdom in 645. No other emperor would, in person, travel to Koguryŏ or one of the successor states Koryŏ 高麗 (918–1392) or Chosŏn 朝鮮 (1392–1910).[1] Yet East Asian historiographical tradition presents these two emperors in quite different ways. Sui Yangdi was cast as the stereotypical bad last ruler of a dynasty; his relentless efforts to conquer Koguryŏ were considered part of the cause of the dynasty’s downfall.[2] Tang Taizong had a markedly different reception. He served across East Asia as a model for a great ruler in a multiethnic empire. His Zhenguan 貞觀 reign was synonymous with a long period of peace and stability across Tang China. Rulers and their chosen successors in many East Asian countries studied the Zhenguan zhengyao 貞觀政要, compiled by Wu Jing 吳競 (670–749) in the early years of the eighth century. The book was translated into Khitan, Jurchen, Mongolian and Manchu, and was read in Korea and Japan.[3] Emperor Chongzong 崇宗 (r. 1086–1139) of the Tangut Xia 夏 kingdom in northwestern China tried to recapture some of the glory of Tang Taizong’s reign when he proclaimed his own Zhenguan era (1101–03).[4] Difan 帝範, Taizong’s advice to his son and heir apparent, was read on the Korean peninsula from the late Koryŏ period onwards, and the Japanese emperors-to-be were given lectures on the text.[5] The study and translations of these texts confirmed Taizong’s place in the East Asian world as an exceptional ruler.

Yet Taizong’s personal involvement in the campaign against Koguryŏ invites the question if and how that action influenced the image of this emperor in Koryŏ and Chosŏn, and how this may differ from the general consensus across East Asia. Mixed with the conventional praise for this ruler’s reign, between the habitual cautions that Taizong’s behaviour in his final years did not live up to the high standard of the early years, Koryŏ and Chosŏn literati emphasised the 645 campaign against Koguryŏ in their own way. In their writings, ranging from editorial comments in histories to poetry, they focus strongly on the siege of Ansi 安市 (Chin. Anshi), where the Koguryŏ commander of the fortress successfully withstood a four-month long siege by the imperial army. In this way, Tang Taizong was directly connected to Korea’s own history, through Taizong’s personal acknowledgement, and the historians’ repeated reminders, that the commander of a single fortress in a small kingdom had bested him.

The Image

The fascination with Tang Taizong is not limited to the pre-modern East Asian world. Biographies (zhuan 傳), historical appraisals and studies, as well as historical novels and fictionalised TV dramas about the second emperor of the Tang dynasty, are numerous in East Asia.[6] The Zhenguan period likewise continues to attract the attention of scholars in the West.[7] Scholarship has shown that the representation of Taizong in the historical record is highly idealised, in a process which started as early as the eighth century. The overall image of Tang Taizong in the Chinese sources is positive, though with an important caveat: the essence of the Zhenguan period’s good government (Zhenguan zhi zhi 貞觀之治) is found in the first half of his reign, much less so in the second half. The early years of Taizong’s reign were characterised by an atmosphere of openness. The emperor was open to admonitions from officials, and he was cautious. He exercised great care in the selection of officials, and strived to rectify the many problems that plagued the early Tang. This changed when the Tang economy became more stable in the 630s, when Taizong began to show another side of his personality. Casting caution and frugality aside, he ordered the building of new palaces, set out on large-scale hunting expeditions, and followed an active and interventionist policy in foreign relations.[8] Added to this was the problem of finding a suitable successor, unpleasantly reminiscent of the way in which Taizong himself had come to the throne: by killing his brothers in a struggle for the position of heir apparent.[9]

Howard Wechsler pointed out that the idealised portrayal of the Zhenguan period was influenced by two factors. First, Tang Taizong was a highly self-conscious ruler, preoccupied with his image, and his public persona was created by his desire to be appreciated by later generations as a model of a Confucian-inspired monarch. Second, because many of Taizong’s actions in the early Zhenguan period conformed to such a model, the scholar-officials who wrote the historical verdict of this reign were easily persuaded to follow Taizong’s lead, and idealise it.[10] Wu Jing’s Zhenguan zhengyao is perhaps the earliest example of that tendency.

Taizong’s deviation from this ideal during the second half of his reign, however, means that the historiography of the Zhenguan period is fraught with contrasts. Historians have struggled with this friction between the idealised picture of the reign, and the emperor’s shortcomings equally clearly documented in the official historiographical record. The campaign against Koguryŏ was no exception, as an analysis of the Chinese and Korean sources will show: opposition to the campaign and in particular the emperor’s personal participation, was strong, yet the portrayal of Taizong’s actions is generally positive. But before turning to a close textual analysis of the historiography of Taizong’s actions in Koguryŏ, it may be helpful to give an overview of the events surrounding the 645 expedition.

Invasion and Defeat in Koguryŏ

In the first lunar month of 645, Tang Taizong led his armies from Luo-yang 洛陽 on a punitive expedition against Koguryŏ, the kingdom in southern Manchuria and northern Korea.[11] The campaign was triggered by events dating back to 642. In that year, Yŏn Kaesomun 原蓋蘇文 (d. 666), an official of Koguryŏ’s eastern circuit,[12] had murdered the king and a host of high-ranking officials, put a nephew of the king on the throne and appointed himself as mangnichi 莫離支. This office gave him control over the civil and military affairs of the state.

For centuries, Koguryŏ’s kings had received Chinese recognition of their status through a system of enfeoffment from the emperors, although this was largely a nominal relationship. Ko Kŏnmu 高建武 (r. 618–42) received his investiture from Tang Gaozu 唐高祖 (r. 618–26) and was thus a vassal of the Tang emperors. When the news of the coup reached Taizong in December 642, he considered it his duty to avenge his vassal’s death, punish the perpetrator and restore order in Koguryŏ. Although in theory all investitures entailed this responsibility, few emperors acted upon it. Taizong, too, first attempted a diplomatic solution. Only when it became clear in the early months of 644 that Yŏn Kaesomun would not co-operate, did the emperor state his intention to lead a punitive expedition in person. All but one of Taizong’s advisors vehemently opposed his personally leading the army, but he disregarded their counsel. Preparations for the expedition started in the autumn of 644. These included building ships for transporting provisions, and gathering intelligence on the enemy’s situation in Liaodong 遼東.

In the spring of the 645, the vanguard reached Koguryŏ’s territory. Initially the Tang army was successful: Kaemok-sŏng 蓋木城 was taken easily and the naval force was victorious at Sabi-sŏng 沙卑城.[13] Yodong-sŏng 遼東 proved to be a tougher nut to crack. Here, Tang Taizong, leading the main host, joined Li Shiji 李世勣 and his troops, and together they conquered the fortress. The next target, Ansi-sŏng 安市城 controlled the main supply route for any advance on the capital P’yŏngyang and the emperor and his most senior military advisor deemed Ansi’s conquest a sine qua non for success. Yŏn Kaesomun was also aware of the strategic importance of this location, and sent a substantial force of Koguryŏ and Malgal 靺鞨 (Chin. Mohe) troops to intercept the Tang army as it approached the fortress. The Tang won a resounding victory in what became known as the battle of Chup’il Hill (Chup’ilsan 駐蹕山). During this battle, Xue Rengui 薛仁貴 first caught the attention of Tang Taizong. The emperor would in hindsight remark that the discovery of the military talent of Xue was one of the few positive points to the campaign.[14]

After this success, the campaign stalled. Only once did the besieged Koguryŏ forces attempt a sortie under the cover of darkness, but this was foiled. The Tang forces built over the course of two months an earthen rampart to try and scale the walls. During the attack phase, a Tang commander lost his nerve when the structure collapsed together with the wall, and Koguryŏ forces captured it, and successfully defended against Tang attempts to retake the rampart. Finally, after four months, Taizong lifted the siege on 15 November 645, because supplies were running low, and the bitter winter was approaching. The inhabitants of Ansi, perhaps fearing a ruse, did not dare to leave their fortress when the Tang troops formed ranks to retreat. From the safety on top of the walls, the Koguryŏ commander bowed respectfully to the emperor, and in return Tang Taizong congratulated him on his staunch defence: he rewarded the commander with one hundred bolts of silk in recognition of his loyal service to the king of Koguryŏ. On the return journey, the Tang army was caught in a severe blizzard, which killed numerous soldiers before they reached Yingzhou 營州 (mod. Chaoyang 朝陽, Liaoning), in the Tang’s northeastern frontier area.

The tangible results of the long military campaign were the conquest of ten Koguryŏ fortresses north of the Yalü River; three new prefectures were created in this area. Seventy thousand people were relocated to Tang territory (neidi 內地).[15] Taizong also ransomed and released up to fourteen thousand prisoners who were destined to be given as slaves to the soldiers for reward. With the main objective of the campaign, that is the removal of Yŏn Kaesomun, not achieved, and the severe loss of life in the Tang army, the emperor considered himself defeated by the ‘petty barbarians’ (xiao yi 小夷).[16] Taizong did not give up the ambition to bring Yŏn Kaesomun to justice, and until the emperor’s death in 649, officers would lead forays into Koguryŏ territory. In 648, he even ordered the construction of a large fleet for transporting supplies for a war that would ‘in one campaign annihilate [Koguryŏ]’ 一舉滅之.[17] But one year later, Taizong died and in his political testament he ordered the crown prince to halt all preparations for the campaign.[18]

Chinese Historical Sources

These events were known to Korean literati from a number of sources. Foremost among these were the official Chinese histories, the Jiu Tangshu 舊唐書, Xin Tangshu 新唐書, and the Zizhi tongjian 資治通鑑. These were also the main sources consulted by Kim Pusik 金富軾 (1075–1151) for his narration of the 645 campaign in the Samguk sagi 三國史記, the oldest extant Korean history. Kim and his team also used a small number of apocryphal stories from other Chinese sources.[19] The Samguk sagi further contains snippets of information from Korean sources that covered this period, but which are no longer extant, for instance when providing extra meteorological information and omens.[20] The full extent of materials available to Kim and Korean literati of later periods is not clear. The Samguk sagi does not always indicate which sources it uses, and where they are signalled, we can often only hazard a guess at the contents, as these original peninsular sources have been lost.

Epigraphic material, for instance, in the form of inscriptions on commemorative monuments marking sites of historical importance, or on walls of temples, and epitaphs may also have been available.[21] The practice of making rubbings or copies of epigraphic texts existed, but there is no indication that such materials pertaining to the Koguryŏ campaigns circulated in Chinese literary collections in the same way as did, for instance, epitaphs (muzhiming 墓誌銘) composed by famous writers. Some anecdotes, perhaps originating in oral accounts about Taizong’s campaign against Koguryŏ, were preserved in collections such as the Sui Tang jiahua 隋唐嘉話 and Tang Yulin 唐語林. Many of these in turn found their way into the Xin Tangshu and Zizhi tongjian. Much later, in the late Ming 明 (1368–1644), these and other stories formed the foundation of works of historical fiction, known as Romances (yanyi 演義), but it is very difficult to ascertain the origin of the story cycles, which may have circulated orally or in other written versions, before we see the earliest extant traces in these Romances. Nevertheless, these works were very popular in Ming and Qing 清 (1644–1911), and in Chosŏn.[22]

Because the Chinese histories were the foremost source of information for Koguryŏ history, it is useful to start with an analysis of the Chinese historiographical process. This will clarify the bias found in the Chinese materials, and offer a useful foundation for a comparison with Koryŏ and Chosŏn writings.

Historians have generally agreed that the 645 campaign ended in defeat, and Taizong himself admitted as much.[23] Yet the narrative in the Jiu Tangshu, Xin Tangshu and the Zizhi tongjian give at first sight a rather positive evaluation of Taizong during these events. The texts are certainly not free from criticism, but the emperor is depicted as a concerned ruler, and a capable military commander during this expedition. This runs counter to expectations, because in the first half of the dynasty many officials vociferously opposed territorial expansion through military means.[24] The writing of history was one of the ways open to them to voice their concerns. Although there had always been a close connection between politics and historiography, the two became institutionally more intertwined from the early Tang period. The writing of the standard histories (zhengshi 正史) became more formalised during this period, with the establishment of the History Office (shiguan 史館) in 625, when officials were appointed to keep records of daily activities at the court. These materials would then be used to produce Shilu 實錄 and Guoshi 國史.[25] In their capacity as court officials, many of the historians at the History Office assisted the emperor in formulating policies and making decisions about issues that closely resembled those they covered as historians. Fang Xuanling 房玄齡 (579–648) and Chu Suiliang 褚遂良 (597–658), two of Taizong’s close advisors, took a strong stance against military adventurism and territorial expansionism in policy debates.[26] They were also appointed by Taizong as editors of the Jin History 晉書. Michael Rogers argues that the prospect of a campaign against Koguryŏ had a significant effect on their editing of the Jin History, which they presented to the throne in 644. The biography of Fu Jian 苻堅 (r. 357–85) of the Former Qin 前秦 (351–94) was, argues Rogers, manipulated with the specific aim of providing a cautionary example for Tang Taizong.[27] This suggests that the raw materials that served as the basis for the official history accounts of the Koguryŏ campaign, the Shilu, were similarly open to editorial intervetion to formulate an anti-expansionist point of view point of view, but this is by no means certain. Zhangsun Wuji 長孫無忌 (d. 659) was the director for the Zhenguan shilu 貞觀實錄.[28] He assisted Taizong closely during the campaign, and was generally supportive. Only once in the extant written record did he suggest that the personal presence of the emperor in Koguryŏ had a negative influence, because the emperor’s personal safety was paramount and required a more careful approach in offensive warfare.[29] These records were part of the material that was tampered with by Xu Jingzong 許敬宗 (592–672), an official at the History Office during the early years of Gaozong’s reign.[30] If, or how, Zhangsun Wuji put his editorial stamp on the records of these events remains unknown.

For Koryŏ and Chosŏn literati, Sima Guang’s 司馬光 Zizhi tongjian was also a significant source of information. Its narrative of the Koguryŏ invasion follows the Xin Tangshu closely, and both are more detailed than the Jiu Tangshu. Sima Guang’s historical outlook, in particular as seen in the Zizhi tongjian, is linked to his political views. As a conservative statesman, Sima Guang held the view that achievements were hard to gain and easy to lose, for an individual as well as for the state.[31] The behaviour of the ruler in particular was key to understanding the rise and downfall of states.[32] In the light of these general ideas that come to the fore in the Zizhi tongjian it is somewhat surprising that Sima Guang does not have a separate editorial comment on the 645 campaign: the venture waisted precious resources, achieved little of lasting endurance, and put the imperial enterprise at risk thanks to Taizong’s personal presence as leader of the expeditionary force. However, Sima Guang did not need to resort to such explicit editorial comments. Hoyt Cleveland Tillman and Mark Strange have revealed how Sima Guang was able to present his opinions and verdicts about these events through the selective presentation of facts and careful choice of wording, rather than through twisting descriptions.[33]

The portrayal of Tang Taizong and his role in the 645 campaign is rather positive in the official historical record, in view of these observations on the Jiu Tangshu, Xin Tangshu and the Zizhi tongjian. The historians refrained from strong condemnations in editorial comments, and despite the officials’ opposition to the emperor’s plans they did not use the opportunities they had to present the narrative of the campaign in a negative manner. To the observant reader, the extant texts are peppered with little hints of respect for the emperor. These can be divided into four types: 1) the emperor’s desire to spare his subjects the hardships caused by warfare; 2) his desire to be treated as an equal of the soldiers, not as the emperor; 3) demonstrations of his keen insight into the strategic situation; and 4) expressions of concern for the wellbeing of the population of Koguryŏ.

In the first category there are specific references to the burden a military campaign would bring to his subjects. In 641 Taizong suggested that a combined attack over land and sea would allow an easy victory over Koguryŏ, the only polity neighbouring Tang that still flouted the emperor’s supremacy. However, for the time being Taizong preferred to allow Shandong 山東 to recover.[34] When the news of the coup d’état by Yŏn Kaesomun reached the court in late 642, Taizong had ample justification for a military campaign to avenge his vassal, the late king of Koguryŏ Ko Kŏnmu. Yet he hesitated: it would be improper to undertake military action during the period of mourning for the king, and the emperor was still not prepared to burden the people of Shandong.[35] The following year he referred once more to the difficulties a military expedition would impose on the common people. Thus he suggested the use of non-Han troops such as Khitan and Malgal.[36] The underlying logic was that these were units of specialist cavalry, whose lifestyle would be less interrupted by warfare. No further plans for an attack were made at that time, but in the 645 invasion these troops formed part of the vanguard under the command of Li Shiji 李世勣.[37]

The decree announcing the punitive campaign against Koguryŏ issued in the tenth month of 644 spans two of the aforementioned categories: in the name of frugality (jianjie 儉節) officials were urged not to observe the usual protocol for an imperial progress. While this lessened the burden of the local people where the imperial cortège passed, it also allowed Taizong to present himself in the persona of soldier, rather than emperor. In this way, he may well have hoped to recapture something of the camaraderie among the soldiers which he had experienced in the early days of the Tang’s bid for power. The decree stated that the encampments should not be embellished and food should merely fill against hunger; building bridges was not necessary if rivers could be forded, and roads should not be repaired if they were still passable. Furthermore, the encampments should be far away from county or prefectural seats, so the students and elderly people were not required to come out to meet the emperor.[38] Supplies should be taken along, rather than relying on local provisions. These measures also emphasised the different approach compared to Sui Yangdi’s disastrous attempts to conquer Koguryŏ.

During the campaign, Taizong managed to maintain this mien of ‘comrade in arms’, according to the sources. After the siege of Ansi, the emperor refused to change the brown robe he had been wearing for the duration of the entire campaign. ‘The clothes of the soldiers are very ragged, how would it be permissible that I alone wear new clothes?’ 軍士衣多弊,吾獨御新衣,可乎?[39] Taizong also personally provided medical care for three of his officers wounded during the campaign, an act which combines a gesture of compassion with the companionship he sought among his fellow officers.[40]

A third group of anecdotes concerns Taizong’s ability as a military commander. Taizong’s fellow commanders showed great respect for his superior strategic insight. For instance, when the Koguryŏ and Malgal relief force showed up at Ansi, Taizong predicted correctly the course of action his adversaries would take. Zhangsun Wuji encouraged him to devise the strategy for the battle, referring to his tried and tested capacities as a tactician and commander:

When Your Majesty was not yet twenty, you personally led the ranks, and always came up with unconvential [strategies], and created [the conditions for] victory. In all cases you issued sage stratagems, all the commanders simply received the preconceived plans. In today’s matters, we beg Your Majesty for instruction! 陛下未冠,身親行陣,凡出奇制勝,皆上稟聖謀,諸將奉成算而已. 今日之事,乞陛下指蹤![41]

The emperor needed little persuasion, and the Koguryŏ force was defeated comprehensively in the battle of Chup’il Hill, which played out completely according to Taizong’s plan.[42]

The final and fourth category contains anecdotes illustrating his concern for the population of Koguryŏ. Taizong argued in early 644 that the punitive campaign against Koguryŏ was expected of him: ‘Kaesomun bullied his seniors and abuses his subordinates, the people crane their necks waiting for rescue’. 蓋蘇文陵上虐下,民延頸待救.[43] During the campaign of 645, he refused to accept the services of the defeated auxiliary troops from Kasi-sŏng 加尸城 during the siege of Kaemok-sŏng, fearing that Yŏn Kaesomun would retaliate by killing their families. ‘I could not bear to obtain the strength of one person [at the price of] having their family exterminated’ 得一人之力而滅一家,吾不忍也.[44] Instead, he sent them on their way with provisions.[45] After the siege of the Paegam-sŏng 白巖城, all the male inhabitants and refugees of the fortress were to be given as rewards to the Tang soldiers who had distinguished themselves, but the emperor ransomed them by rewarding his soldiers with items from his personal treasury.[46] Among those captured was — at least in the Zizhi tongjian’s narrative — also the vice-commander of the Yodong (Chin. Liaodong 遼東) fortress, who had taken refuge at Paegam fortress with the widow and children of the Yodong commander after the latter was killed. Taizong rewarded the vice-commander for his upright conduct and sent him back to P’yŏngyang, with a reward of five bolts of silk.[47] When, shortly thereafter, Koguryŏ’s reinforcements were defeated in the battle of Ch’upil Hill near Ansi, the majority of the troops was sent back to P’yŏngyang, while the 3,500 commanding officers were to be taken back to Tang China and given military ranks.[48] A captured Koguryŏ spy who said he had not eaten for several days was fed, before being sent back to inform Yŏn Kaesomun that ‘if he wishes to know information about the [Tang] army, he can just send somebody directly to Our location’ 欲知軍中消息,可遣人徑詣吾所.[49] Another example of Taizong’s love for grand gestures of generosity towards the Koguryŏ people came upon his return to Youzhou 幽州 (present-day Beijing). He freed the inhabitants of the newly created prefectures of Liao 遼, Gai 蓋 and Yan 巖 (from the conquered fortresses of Yodong, Kaemok and Paegam respectively), who were destined to become slaves as rewards for the Tang officers, and as before Taizong provided the soldiers with rewards from the imperial treasury.[50]

In short, despite the failed campaign, the Chinese sources do not present Taizong in a negative way. They ascribe to him the qualities of a good leader, a competent strategist and a compassionate emperor. Yet they could not deny there were problems with this image. Juxtaposed with these aforementioned anecdotes, the sources record instances of Taizong’s less glorious behaviour. Although he pardoned many enemy officers and soldiers after the battle of Chup’il Hill, he punished more than three thousand Malgal enemy warriors by having them buried alive.[51] A Tang officer was executed after making a critical error in the siege of Ansi, and the Prince of Jiangxia was only spared the same fate because of substantial merits earlier in the campaign.[52] It is difficult to form a straightforward, uniform judgement of Tang Taizong’s character.

The final paragraph of Taizong’s ‘Benji’ 本紀 in the Xin Tangshu contains the Appraisal (Zan 贊) of the historian. It is worth quoting extensively, because it suggests how Ouyang Xiu 歐陽修 as the chief editor of the Xin Tangshu tried to solve the contradictions in Taizong’s behaviour.

When the Tang possessed the empire, it was handed down for twenty generations, and three of the rulers are worth praising. But Xuanzong (r. 712–56) and Xianzong (r. 805–20) were unable to carry this through to the end of their reign. How magnificent were the achievements of Taizong! In the manner he cleared out the disorder of the Sui, he followed in the footsteps of Tang and [Zhou king] Wu; in bringing about the ideals of government, he approximated [the Zhou kings] Cheng and Kang. Since Antiquity merit and virtue have combined into greatness, but since the Han this had not occured. As for being led by what one cares about greatly, reinvigorating Buddhism, admiring what was grandiose, taking pleasure in achievement, and having the soldiers toil in distant lands, that is something which ordinary rulers with mediocre talents commonly do. But then the rules of [praise and blame in] the Spring and Autumn Annalscommonly place high demands on the worthy. Because of this, everybody sighed in admiration at the ideal of a gentleman of later generations aspiring to becoming a perfected man.
唐有天下,傳世二十,其可稱者三君,玄宗、憲宗皆不克其終,盛哉,太宗之烈也!其除隋之亂,比迹湯、武;致治之美,庶幾成、 康. 自古功德兼隆,由漢以來未之有也. 至其牽於多愛,復立浮圖,好大喜功,勤兵於遠,此中材庸主之所常為. 然春秋之法,常責備於賢者,是以後世君子之欲成人之美者,莫不歎息於斯焉.[53]

With the explanation that exceptional rulers are measured against more exacting standards, the commentary attempts to reconcile Taizong’s faults with the ideal of the ‘good government of the Zhenguan period’. The campaign against Koguryŏ was normal, perhaps even acceptable, behaviour for a mediocre ruler, and in this case a ground for criticism only because Taizong was in many other aspects considered belonging to a superior category. This implied that Taizong provided an acceptable model from the recent past, rather than remote Antiquity, for rulers who aspired to greatness.

Historical Record of the Three Kingdoms

The Chinese sources fell on the positive side of the judgement, and this did not escape the attention of the Koryŏ and Chosŏn scholars, but they did not slavishly follow that tone. The oldest extant source for Taizong in peninsular historiography is the Samguk sagi. As indicated above, Kim Pusik and his collaborators used Chinese sources extensively to describe the military campaign of 645. This did not prevent Kim from presenting a different image of Taizong, a sharper critical judgement. He included two editorial commentaries (nonchan 論贊) about Tang Taizong.[54] Both of these borrow the language of the Appraisal (Zan 贊) found in the Xin Tangshu, but latch onto the laudatory element as a basis for criticism. The first one is inserted into the narrative at the end of the fourth year of King Pojang 寶藏 (645), that is, immediately after Taizong’s return from the Liaodong campaign.

Tang Taizong was a monarch of surpassing and uncommon wisdom. In his suppression of disorder he can be likened only to Tang and Wu, and in his governance to Cheng and Kang. In his use of military tactics his singular use of strategy knew no bounds and in every engagement he was unequalled. Yet in the battles of the Eastern Expedition and his defeat at Ansi fortress, the fortress chief can be called heroic and exceptional. However, his name has been lost from the histories. It is just as Yangzi[55] noted, the names of the great officials of Qi and Lu are not recorded in the histories. This is truly lamentable.
唐太宗聖明,不世出之君,除亂比於湯武,致理幾於成康,至於用兵之際,出奇無窮, 所向無敵,而東征之功,敗於安市,則其城主可謂豪傑非常者矣,而史失其姓名,與楊子所云齊魯大臣史失其名無異甚可惜也.[56]

With one simple comparison between the two men, Kim elevates the anonymous commander of Ansi to a level as equally deserving of historical fame as Tang Taizong himself.

The second comment is inserted immediately after the death of Taizong.

When [emperor] Taizong first waged war in Liaodong, those who remonstrated were many. Also, after the return of the troops from Ansi, [Emperor Taizong] himself, on account of not having been succesful, regretted it deeply and lamented: ‘Had Wei Zheng been alive, he would not have allowed me to behave like this’. When he was going to launch an expedition [against Koguryŏ] once more, the Minister of Works, Fang Xuanling, sent in a petition and remonstrated [...] from his sick bed.
When he was about to die, the words of the Duke of Liang [=Fang Xuanling] were emphatic like this, but the emperor did not accept them. Instead he delighted in thinking about the destruction of the eastern regions and only his death ended it. The historian[57] has commented, ‘He [Emperor Taizong] admired what was grandiose, took pleasure in achievement, and let the soldiers toil in distant lands’. Is this not what is meant? Liu Gongquan’s story says: ‘In the battle at Chu-p’il Hill, Koguryŏ and the Malgal joined forces, [their ranks] extended over forty li. When Taizong saw this, he looked frightened’. It also said: ‘When the six [Tang] armies were put under pressure by Koguryŏ to the point of being unable to hold out, a scout reported to the Duke of Ying’s staff that the black standard was surrounded, and the emperor grew very troubled.’ This was the nature of his fears, although in the end he escaped, yet the Jiu Tangshu, and Sima Guang’s Zizhi tongjian make no mention of this. Were they not concealing this for their country?
初,太宗有事於遼東也,諫者非一. 又自安市旋軍之後,自以不能成功,深悔之,歎曰:「若使魏徴在,不使我有此行也. 」及其將復伐也,司空房玄齡病中上表諫. 〔...〕梁公將死之言諄諄若此,而帝不從. 思欲丘墟東域而自快,死而後已. 史論曰:「好大喜功,勤兵於遠」者,非此之謂乎?柳公權小説曰:「駐蹕之役,高句麗與靺鞨合軍,方四十里. 太宗望之,有懼色. 」又曰:「六軍爲高句麗所乘,殆將不振,候者告英公之麾. 黑旗被圍,帝大恐. 」雖終於自脱,而危懼如彼,而新舊書及司馬公通鑑不言者,豈非爲國諱之者乎?[58]

This comment contains incisive criticism of Taizong and the historians who praised him. Fang Xuanling’s memorial, the lengthy quotation omitted in my translation, was a very sharply worded condemnation of Tang Taizong’s plans for a second campaign against Koguryŏ. Fang accused the emperor of injustice, because he sent his soldiers to a certain death in a distant country, whereas criminals condemned to death could count on three reviews before the execution would take place. Such an accusation bordered on lèse-majesté, and Fang could only write such harsh words because he was at death’s door anyway. Kim Pusik used this famous memorial[59] to point out Taizong’s faults, and to bring the emperor down from the generous praise given in the Xin Tangshu. The combination of Kim’s critical tone about Taizong’s achievements, and his praise for the anonymous hero of Ansi, resonated with Koryŏ and Chosŏn literati, as an investigation of the peninsula’s literary output on the 645 campaign demonstrates.

Koryŏ and Chosŏn Literary Works

Regular embassies from Koryŏ and Chosŏn to the Yuan 元 (1260–1368), Ming and Qing courts brought many literati as diplomatic envoys close to the sites of the historical events of 645. The itinerary for the delegates was not very different from the one taken by Taizong, and they drew inspiration from their surroundings, using their historical imagination. It was customary for these men to produce a travel diary, and many of these texts have been preserved, generally known as Yŏnhaengnŏk 燕行錄.[60] They consist of descriptions of the day’s events, musings on a variety of topics, and poems. They were written mainly in literary Chinese but occasionally in Korean,[61] and were available in an edited format to later generations through their collected writings (jip 集). Poems were important in these travel writings. The compact style of si (Chin. shi 詩), for instance, lent itself very well to an impressionist composition, an emotional snapshot of the author’s response to a historical site he encountered. Such poems are often also packed with a multitude of historical references within the limited space offered in this very traditional genre, in a demonstration of erudition.

An early example that touches upon two of the recurring themes is Chŏnggwanŭm 貞觀吟 by Yi Saek (李穡, 1328–96). Yi travelled multiple times to the Yuan court in the mid-fourteenth century, and his route took him past Yulin Pass (Yulin guan 榆林關, later known as Shanhai Pass (Shanhai guan 山海關), strategic for the defence of the Yanjing 燕京 region from the northeast. Yi Saek described it in another poem as a bleak and desolate (sojo 蕭條) location.[62] At the time, its main claim to fame was that Taizong had set up camp here, on the way back from Koguryŏ, and the heir apparent had come out to Yulin Pass to welcome him back.[63] For Yi Saek the location triggered a reflection on the Zhenguan reign. The poem starts with a celebration of Taizong’s achievements, but Yi keeps a special place for Koguryŏ. The kingdom was the only polity contiguous with Taizong’s empire which did not readily accept the status of vassal, and ultimately became the place of Taizong’s only defeat as a military commander. In sharp contrast with the first half of the poem, the closing lines describe the despondency experienced at Yulin, by the poet and his imagined emperor Taizong:

He thought that [the conquest of the Koguryŏ] was a done deal; who could have know that the white feather would land in the dark flower?
The Duke of Zheng [Wei Zheng] was already dead, and the way to remonstrate was blocked; ironic that the [Duke’s] headstone, first torn down, was restored.
I turn my head and cry out three times for the Zhenguan years; to the end of the sky a sad wind sighs.
謂是囊中一物耳. 那知玄花落白羽.
鄭公已死言路澁. 可笑豐碑蹶復立.
回頭三叫貞觀年. 天末悲風吹颯颯.

This short passage touches on two important themes: the arrow that allegedly landed in the emperor’s eye, and the rehabilitation of Wei Zheng’s status, on Taizong’s sobering introspection in the immediate aftermath of the expedition against Koguryŏ in 645. The former story is spurious: there is no record of Taizong being wounded during the expedition, but this is the earliest mention of an anecdote that endured through the ages on the Korean peninsula.[64]

In contrast, Wei Zheng’s role in Taizong’s reflections after the campaign is attested in historical records. Wei Zheng had died a few years before the campaign against Koguryŏ. His relationship with Taizong was far from easy, because Wei never hesitated to remonstrate against Taizong. This relationship was in later times idealised as the epitome of the loyal minister and the enlightened ruler who was open to admonition. In fact, Wei was not the closest of Taizong’s advisors — that honour went to Fang Xuanling — but this idealised image gave Wei Zheng a very prominent place in Tang history.[65]

When Wei died, Taizong personally composed the text for the epitaph, a very high honour. But before long, allegations of Wei’s role in factionalism came to the surface. As a result, various imperial favours were withdrawn: the gravestone with the emperor’s composition was torn down, and plans for Wei Zheng’s son’s marriage to an imperial princess were cancelled. Only when Taizong returned from Koguryŏ, did he come to regret these decisions, as he realised Wei Zheng would correctly have advised the emperor against the expedition. Taizong made amends and ordered the headstone be restored, and provided Wei’s family with a stipend.[66]

The image of the emperor coming to that realisation was a strong one for literati-officials, those from the Korean peninsula included. Wei Zheng’s idealised persona provided an attractive model of an upright minister who placed the dynasty’s and the people’s needs before his own or his ruler’s personal desires, and who had a close relationship with the emperor. Wei Zheng’s posthumous restoration of his status provided an example and perhaps some comfort to many officials who felt that while their well-intended advice fell on deaf ears, they might one day be vindicated.

The ‘Chŏnggwanŭm’ was added to the Tongmunsŏn 東文選, and may have inspired Sŏ Seyang 蘇世讓 (1486–1562) to write a similar poem, with short introduction, when he travelled through Funing 撫寧 county just north of the Yulin Pass in 1533, during his journey to the Ming capital.

About seven or eight li west of Funing county there is a small hill overlooking a mountain stream. Its top is round and flat. There are very mysterious and ancient pine trees there. People say that it is the place where Tang Taizong stayed overnight when he campaigned in the east. 撫寧縣西七八里許. 有小峯俯臨溪水. 其頂圓平. 有松樹甚奇古. 人言唐太宗征東時駐蹕之處.
On the short hilltop the pines are old, with curled and bent branches; it is recalled that the Civil Emperor [=Taizong] stationed his coloured banner here.
In the end, he turned back, and praised the general of Ansi; how about not tearing down the headstone for the Duke of Zheng [=Wei Zheng]?
The flow of the stream has not yet washed away the regrets of those times; the mountains still harbour a myriad of ancient sorrows.
Alone I stand facing the wind, with many regrets; in the dust at dusk on the flat plain my horse moves slowly.[67]
短峯松老偃虯枝. 記得文皇駐綵旗.
畢竟反褒安市將. 何如不踣鄭公碑.
川流未洗當時恨. 山意猶含萬古悲.
獨立風前多憾慨. 暮煙平楚馬行遲.

The final line can also be read as if Taizong were the subject, and the poet imagined the emperor in a contrite mood, reflecting on his mistakes. Here again, the terse reference to Wei Zheng’s headstone suffices as a reminder that his harsh criticism always had the empire’s best interests at heart.

The defeated emperor riding off into the sunset appeared again a few years later, in Ch’oe Yŏn’s 崔演 (1503–49) poem ‘Chup’ilsan yugam’ 駐蹕山有感), written during his mission to China in 1548.[68] In the preface he sets the historical background:

This mountain is about fifteen li southwest of Liaocheng (K. Yosŏng). When Tang Taizong attacked Ko[gu]ryŏ, he once stayed the night here and inscribed a stone to commemorate the merit, subsequently [the name of the hill] was changed to its current name. 本首山在遼城西南十五里. 唐太宗伐高麗. 嘗駐蹕勒石紀功. 因改今名.

The last two lines closely echo Sŏ’s poem, and contain once more the reference to Wei Zheng:

Exasperated that Duke Wei [Zheng][69] has been dead for a long time; in the slanting [rays of the setting] sun, he reins in his horse, and is lonely, overcome with emotion.
最恨魏公今死久. 斜陽立馬獨含情.

In the seventeenth century the tone of the literary works dealing with Taizong in Koguryŏ changed. The hitherto anonymous commander of Ansi suddenly acquired the name Yang Manch’un (Chin. Liang Wanchun 梁萬春 or Yang Wanchun 楊萬春), and the focus shifted from Taizong’s remorse and Wei Zheng’s reminder from beyond the grave to the heroic effort of the commander of Ansi. The name Yang Manch’un appeared first in the Ming-dynasty historical novel Tangshu zhizhuan tongsu yanyi 唐書志傳通俗演義 by Xiong Damu 熊大木, also known as Xiong Zhonggu 熊锺谷 (1506–79).[70] It is not entirely clear if Xiong invented the name of the commander, or if this was part of a story cycle about Tang Taizong which served as the base for the novel.[71] In any case, this development marked a significant turning point in the way Chosŏn literary works looked back on the campaign of 645. The attention moved away from emperor Taizong and how his defeat at Ansi led to his personal re-evaluation of his reign, and was instead directed towards the commander of the fortress, Yang Manch’un.

The earliest extant reference in Chosŏn writings comes from Yun Kŭnsu 尹根壽 (1537–1616). He explained how he was introduced to this new piece of information through a Ming officer who came to Chosŏn in the wake of the first Japanese invasion of the Imjin war (1592–98). The officer referred Yun to the chapter ‘Taizong dongzhengji’ 太宗東征記, in the Tangshu zhizhuan tongsu yanyi by Xiong Damu.[72] Yun Kŭnsu refers to this event a second time in a preface (, Chin. xu 序) for a collection of poems presented as a parting gift to a friend who was about to embark on the journey to the Ming capital.[73] Yun went on that journey himself in 1589, and seems eager to share his knowledge of the sights on the way, and he provides background information in the lengthy preface. As he recounts the site of Ansi, and the siege, he exclaims:

Only the commander of Ansi succeeded in safeguarding the isolated fortress, his fame resounded in the world, but the [Ko]ryŏ historians lost his name; there is no trace in the written documents, until now. A story from China now says he was Yang Manch’un, and so we have obtained his name, isn’t this exceptionally fortunate?
且也安市城主獨全孤城. 名聞天下. 而麗史顧逸其名. 文獻無徵. 一至於此. 中朝小說乃謂梁萬春其人. 因此得其姓名. 豈非千古之一快乎?[74]

The wording here brings to mind Kim Pusik’s commentary about the commander of Ansi, whose ‘name has been lost from the histories. This is truly lamentable’.[75]

Later authors were less inclined to accept this claim. Kim Siyang 金時讓 (1581–1643), who travelled to the Ming court in 1610, observed that the source for the name is not given in this Tangshu zhizhuan tongsu yanyi, and he thought it unlikely that such important information would not have been recorded anywhere else, if it had been known.[76] Other writers also raised doubts about the apocryphal anecdote concerning Taizong being hit in the eye, and even the location of the Ansi fortress itself was subject to speculation. But the repeated doubts over all these points suggest that such apocryphal stories stubbornly refused to disappear over the centuries.

In a short prefatory note to his akpu (Chin. yuefu 樂府) poem ‘Sŏngsang pae’ 城上拜, Sim Kwangse 沈光世 (1577–1624) likewise states that it is not clear if the name of the commander is really Yang Manch’un.[77] Sim also refers to Yi Saek’s poem ‘Chŏnggwanŭm’, mentioned above, and casts doubt on the reference it contains to the anecdote of Taizong being hit in the eye by a stray arrow.[78] Yet despite these reservations about the historical accuracy of the legend emerging around Ansi, neither Kim Siyang nor Sim Kwangse wished to detract from the heroic stance they attributed to the commander of the fortress.

Sim Kwangse’s initial offering was the trigger for a number of poems of the same akpu genre with the title ‘Sŏngsang pae’ in the late seventeenth and eighteenth centuries.[79] In contrast to the poems mentioned so far, most of those were written by men who had not travelled to China. Indeed, it seems that Sim himself had not travelled abroad. In other words, the siege of Ansi had moved from a topic prompted by physical proximity to historic sites, to an independent topic. The commander of Ansi’s observation of ritual propriety as the imperial host broke off the siege, spoke to the imagination of many Chosŏn literati.[80]

The popularity of the theme of the defence of Ansi is also seen in Kang Paengnyŏn’s 姜栢年 (1603–81) writings. As he travelled past the alleged site in 1660 on the way to the Qing court, he wrote a ‘Song’ (ŭm 吟), echoing the title of Yi Saek’s ‘Chŏnggwanŭm’,[81] but he also imagined what the imperial edict, issued by Taizong as he lifted the siege, must have looked like.[82] No such edict is preserved, but it is indeed likely that a similar document was proclaimed when Taizong rewarded the commander of Ansi with one hundred rolls of silk for his steadfast defence of the fortress in difficult circumstances.[83]

Kim Ch’anghŭp 金昌翕 (1653–1722) composed a set of fifty poems dedicated to his younger brother who was setting out on a journey to the Qing court in 1712.[84] In the seventh poem he picks up the theme of Yang Manch’un:

One thousand years, the bravery of Yang Manch’un! He hit the dragon’s whis-ker and [his arrow] landed in the eye. 千秋大膽楊萬春. 箭射虬髥落眸子.[85]

Ch’anghŭp himself did not travel to the Qing, but like the authors of the akpu on the title ‘Sŏngsang pae’, he was conversant with the literature produced in connection with these journeys and the main sights along the road. For instance, the set also contains poems about the ‘Chup’il t’aeja ha’ 駐蹕太子河, where Taizong defeated the relief force for Ansi, and Shanhai Pass 山海關, earlier known as the Yulin Pass. The many Yŏnhaengnŏk had created a textual double of that world, which was now familiar even to those who had not travelled to the Qing court.

In the second half of the eighteenth century, Yi Tŏngmu 李德懋 (1741–93) wrote a critique of Kim Ch’anghŭp’s poem. Such continued conversation with inter-textual references suggests robust circulation of these literary collections across generations. Here Yi traces the various strands of the anecdote back via Chosŏn literati Yun Kŭnsu and Sŏ Kojŏng, and Koryŏ authors Yi Saek and Kim Pusik, all the way back to the earliest extant Chinese sources, such as the Xin Tangshu and Zizhi tongjian:

When Samyŏn 三淵 [Kim Ch’ang-hŭp] saw No Kajae 老稼齋 [Kim Ch’ang-ŏp 金昌業, 1658–1721, younger brother of Ch’anghŭp] off on his journey to Yanjing, he said in the poem:
One thousand years, the bravery of Yang Manch’un! He hit the dragon’s whisker and [his arrow] landed in the eye.[86]
When you investigate it, the name of the commander of Ansi is Yang Manch’un, and that came from the Tangshu zhizhuan tongsu yanyi. Dilettantes have turned this into his [actual] name, but it is not sufficiently credible. More details are given in [Yun Kŭnsu’s] Wŏljŏng mannok[87]and Sŏ [Kŏjŏng]’s Saga jip. Mogŭn’s [=Yi Saek] ‘Chŏnggwanŭm’ reads
He thought that [the conquest of Koguryŏ] was a done deal; who could have known that the white feather would land in the dark flower?
The dark flower refers to the eye, the white feather to the arrow. For generations it has been transmitted that when Tang Taizong attacked Ko[gu]ryŏ, and he came to Ansi, he was hit by an arrow in the eye and turned back. When one investigates, neither the Xin Tangshu or Zizhi tongjian have recorded it. The historians of the time undoubtedly considered it unmentionable for China so it is not strange it is not noted.
But Kim Pusik’s Samguk sagi likewise does not record it, so we don’t know where Mog[ŭn] and No [Kachae] obtained this information.
三淵送老稼齋入燕詩曰. 千秋大膽楊萬春. 箭射虬髥落眸子. 案安市城主爲楊萬春. 出於唐書演義. 好事者爲之作姓名. 不足取信. 詳見月汀雜錄. 徐四佳云. 牧隱貞觀吟曰. 謂是囊中一物耳. 那知玄花落白羽. 玄花言其目. 白羽言其箭. 世傳唐太宗伐高麗. 至安市城. 箭中其目而還. 考唐書通鑑. 皆不載. 當時史官. 必爲中國諱. 無恠其不書也. 但金富軾三國史. 亦不載. 未知牧老何從得此.

Pak Chiwŏn (朴趾源, 1737–1805) also entered the conversation with Kim Chang’hŭp and Yi Saek’s work. In his well-known Yŏrha ilgi 熱河日記 he writes on the occasion of his visit to the alleged site of the Ansi fortress in 1780:

For generations it has been transmitted that the commander of the Ansi fortress was Yang Manch’un, and he shot the emperor in the eye. The emperor paraded the troops at the foot of the fortress, and bestowed one hundred bolts of silk on [the commander of Ansi] as a reward for the commander’s steadfast defence.
又世傳安市城主楊萬春. 射帝中目. 帝耀兵城下. 賜絹百匹. 以賞其爲主堅守. [88]

Pak suggests that Kim Ch’ang-hŭp and Yi Saek may have obtained the information from old peninsular legends or oral stories, and criticises Kim Pusik’s attitude to the sources:

Kim Pusik simply regretted that the histories had lost [the commander’s] name. Possibly when Pusik made the Samguk sagi, he only consulted Chinese histories and the documents copied and transcribed one side only, in order to make [a narrative of] the events. He went so far as to cite Liu Gongquan’s story, in order to give evidence for the encirclement at Chup’il, but neither the Jiu Tangshu, Xin Tangshu nor Sima [Guang’s] Zizhi tongjian show a record thereof, and he suspected it was because the Chinese concealed it. But if that is so, then when it comes to such things as old anecdotes from our own land, he did not dare to record one sentence briefly. Between [deciding on] transmitting what is credible and suspicious, it is likely he omitted it. I say that Taizong lost his eye at Ansi, although I cannot prove it, but that this town [Feng huang cheng 鳳凰城] is likely Ansi, is, in my opinion wrong.
金富軾只惜其史失姓名. 葢富軾爲三國史. 只就中國史. 書抄謄一番. 以作事實. 至引柳公權小說. 以證駐驆之被圍. 而唐書及司馬通鑑. 皆不見錄. 則疑其爲中國諱之. 然至若本土舊聞. 不敢略載一句. 傳信傳疑之間. 葢闕如也. 余曰. 唐太宗失目於安市. 雖不可攷. 葢以此城爲安市. 愚以爲非也. [89]

While Yi Tŏngmu, Pak Chiwŏn and others cast doubts on the veracity of this anecdote, even as late as the nineteenth century it had its supporters. Yi Kyukyŏng 李圭景 (1788–1856) wrote about the same line of Kim Ch’anghŭp’s poem:

This did not come from a standard history, but it really happened. The [Xin] Tangshu avoided mentioning it, as there was no trace of evidence. But then Mr Samwŏn [=Kim Ch’anghŭp] brought it into a poem, and it should indeed be considered as real; it is not a random story.
蓋此非出於正史. 實有其事也. 唐史諱之而無信蹟. 然三淵先生旣入於詩. 則亦
可以爲實. 非浪傳者也.[90]


Over the course of more than a millennium, the siege of Ansi moved from history to legend. Koryŏ and Chosŏn historians and writers adopted and adapted the information that came to them from Chinese sources, and they remoulded the narrative of Taizong’s campaign in 645, and in particular the siege of Ansi, to give it significance in a new cultural context.

Kim Pusik’s Samguk sagi made ample use of Chinese sources. He worked within the same framework of classicist (yu, Chin. ru 儒) historiography as the compilers of the Chinese materials, but he did not leave the master narrative of these Tang and Song historiographers unchallenged. While his description of events included many of the elements that helped to portray the emperor in a positive light in the Tang and Song sources, his judgement of Tang Taizong was far more critical in the editorial commentaries (nonchan 論贊). Emphasising the sense of defeat for Taizong at Ansi then allowed later writers to present the 645 campaign in an introspective mode with the rehabilitation of Wei Zheng as a major theme. This move ultimately made the commander of Ansi the real hero.

The legend of Taizong being blinded in one eye by an arrow seems to be Korean in origin, but here too Chinese materials were added in. When the name Yang Manch’un as commander of Ansi was first introduced in Chosŏn in the late sixteenth century from a Chinese historical novel, the focus shifted sharply from the somber reflections of Taizong to a celebration of the commander of Ansi. Before long he was also identified as the archer who wounded Taizong. The story of Ansi now focused on a Koguryŏ hero who combined martial skills (mu 武) as military leader with civil virtues (mun文); in particular the commander’s observation of the ritual protocol when the imperial army retreated was highlighted. If Taizong could be seen as a model of such a leader in East Asia in general, there was also an example on ‘our’ (a 我) side, from a Koryŏ and Chosŏn perspective. Although many Chosŏn writers pointed out the enormous time lag between the events of 645 and the first mention of the name Yang Manch’un in the late Ming, and hence the lack of reliability of this information, their repeated denunciations only serve to underline the durability and popularity of the legend of Yang Manch’un. In present-day historical fiction, and indeed on some internet forums with questionable historical research, the name Yang Manch’un is still bandied about, and the legend lives on.[91]

The steady stream of poems about Ansi furthermore indicates that not only Koryŏ, but also Chosŏn considered Koguryŏ as one of its charter polities.[92] According to John Duncan, memories of Koguryŏ formed an important part of the Korean collective historical memory ‘for at least the past 1,000 years’.[93] That this memory was shaped among other things by mistaken notions of the heroic exploits of the defender of the Ansi fortress does not detract from the strength of the feeling.

Returning to the initial question of this article — if the presence of Tang Taizong on Koguryŏ territory has influenced the image of the emperor in Koryŏ and Chosŏn — it is clear the answer is affirmative. In Chinese sources, the emperor was presented as a great ruler and strategist, and this was well known on the peninsula. This allowed for a direct comparison with a homegrown hero, and emphasised even more the enormity of the achievements of ‘Yang Manch’un’. Koryŏ and Chosŏn were heavily indebted to Chinese culture, but the reshaping and reinterpretation of the siege of Ansi is one example of how in these circumstances a distinct identity could still develop.

Tineke D’Haeseleer
Princeton University

1 Successor states in the sense that Koguryŏ was together with Paekche 百濟 (trad. 18 BC – AD 660) and Silla (trad. 57 BC – AD 935) part of the historical conscience expressed by the Koryŏ and Chosŏn elite. See R.E. Breuker, Establishing a Pluralist Society in Medieval Korea, 918–1170: History, Ideology and Identity in the Koryŏ Dynasty (Leiden: Brill, 2010), and John B. Duncan, ‘Historical Memories of Koguryŏ in Koryŏ and Chosŏn Korea,’ Journal of Inner and East Asian Studies 1 (2004): 118–36.

2 See Arthur F. Wright, ‘Sui Yang-Ti: Personality and Stereotype,’ in The Confucian Persuasion, ed. Arthur F. Wright (Stanford, California: Stanford University Press, 1960), pp.60ff., for an analysis of the stereotype of Sui Yangdi as a ‘last ruler’.

3 Herbert Franke and Denis Twitchett, ‘Introduction,’ in Cambridge History of China, Vol.6: Alien Regimes and Border States, ed. Herbert Franke and Denis Twitchett (Cambridge: Cambridge University Press, 1994), p.33; Denis Twitchett, ‘How to be an Emperor: T’ang T’ai-Tsung’s Vision of His Role,’ Asia Major 9 (1996): 1–102, at p.6.

4 Ruth Dunnel, ‘The Hsi Hsia,’ in Cambridge History of China, Vol. 6, p.198.

5 The earliest mention of the Difan on the Kor-ean peninsula is in Yi Chŏm 李詹, ‘Chebŏm pal’ 帝範跋 in Ssangmaetang hyŏpjang munjip 雙梅堂篋藏文集 25: 382b. (Unless specified otherwise, all Korean texts are taken from ‘DB of Korean Classics’: <http://db.itkc.or.kr>; the number before the colon is the kwŏn/juan, followed by the page number.) Twitchett, ‘How to Be an Emperor,’ pp.44–48.

6 For an overview, see Jack W. Chen, The Poetics of Sovereignty: On Emperor Taizong of the Tang Dynasty (Cambridge: Harvard University Asia Center, 2010), p.11, who talks about ‘a small industry of biographical studies in Chinese, Japanese, and English’. Many Chinese and some Japanese biographies have been translated into Korean. Korean authors usually focus on Taizong in connection with the Koguryŏ campaign, or as the adversary of Yŏn Kaesomun, in academic articles and fiction alike. Some recent examples are Kim Rakki 김락기, ‘Sipch’il-sipku segi Koguryŏ Ansisŏng insik-kwa ‘Sŏngsang pae’: Yŏnhangnok-kwa munjip-ŭl jungsim-ŭro› 17–19 세기 고구려 안시성 인식과 ‘城上拜’: 「연행록」과 「문집」을 중심으로, Yŏksa minsokhak 역사민속학 42 (2013): 135–67; Sŏ Yŏngkyo 서영교, ‘Tang T’aejong-ŭi Koguryŏ ch’inchŏng-kwa chakchŏn waegok: Yodongsŏng chŏnt’u chŏnhu-ŭi sanghwang-ŭl chungsim-ŭro’ 당태종의 고구려 親征과 작전 歪曲 - 요동성 전투 전후의 상황을 중심으로, Tongbuga munhwa yŏn’gu 동북아 문화연구 36 (2013): 211–25; and in fiction Chŏng Su-in 정수인, Koguryŏ: Chŏng Su-in daeha sosŏl 5: Tang T’aejong Yi Simin: Kkum-kwa tojŏn 고구려 : 정수인 大河小說. 5 : 당태종 이세민, 꿈과 도전 (Seoul: Saeum, 2005).

7 In addition to Jack Chen’s own monograph and the works he mentions in note 1, p.13, see also Arthur Wright, ‘T’ang T’ai-tsung: The Man and the Persona,’ in eds J.C. Perry and B.L. Smith Essays on T’ang Society: The Interplay of Social, Political and Economical Forces (Leiden: Brill, 1976).

8 Howard J. Wechsler, ‘T’ai-Tsung (r. 626–49) the Consolidator,’ in Cambridge History of China, Vol.3: Sui and T’ang China (Cambridge: Cambridge University Press, 1979), pp.189–93; 220ff.

9 For the Xuanwu 玄武 incident, see Wechsler, ‘T’ai-Tsung (r. 626–49) the Consolidator,’ p.186; Sima Guang (comp.), Zizhi tongjian 資治通鑑 (hereafter ZZTJ) (Beijing: Zhonghu shuju, 1976) 192: 6008–11. For Taizong’s problems in finding a suitable successor, see Wechsler, ‘T’ai-Tsung (r. 626–49) the Consolidator,’ pp.236–39.

10 Wechsler, ‘T’ai-Tsung (r. 626–49) the Consolidator,’ pp.189 and 191.

11 Multiple Tang and Song sources cover the campaign in detail, see for instance Liu Xu 劉煦 (comp.), Jiu Tangshu 舊唐書 (hereafter JTS) (Beijing: Zhonghua shuju, 1975) 199a: 5322–26; Ouyang Xiu 歐陽修, Song Qi 宋祁 et al. (comp.), Xin Tangshu 新唐書 (hereafter XTS) (Beijing: Zhonghua shuju, 1975), 220: 6189–194; ZZTJ 197–98: 6207–231; Cefu yuangui 冊府元龜 has many entries about the campaign, spread out over numerous categories: Wang Qinruo 王欽若, Cefu yuangui (hereafter CFYG), 12 vols. (Beijing: Zhonghua shuju, 1960). Compare with Kim Pusik (comp.), Samguksagi 三國史記 (hereafter SGSG), ed. Yi Pyŏngdo 李丙燾 (Seoul, Ŭlyu munhwasa, 1980) 20: kwŏn 20, p. 189 and kwŏn 21, pp. 190–97.

12 Sima Guang explains in the notes on divergent sources (kaoyi 考異) to ZZTJ that he follows the Shilu 實錄 instead of JTS 199a: 5322, which states it was the western circuit. (ZZTJ 196: 6181).

13 JTS 199a: 5323; XTS 220: 6190. ZZTJ 197: 6220 gives Bishacheng 卑沙城. SGSG 21: 193 follows ZZTJ. For locations of the Koguryŏ fortresses, see Tan Qixiang 譚其驤, Zhongguo lishi dituji: Shiwen huibian: Dongbei juan 中國歷史地圖集:釋文匯編:東北卷 (Beijing: Zhongyang minzu xueyuan, 1988), pp.67–71.

14 Xue Rengui’s exploits in Liaodong became fictionalised, see for instance the anonymous Qing-dynasty novel Rengui zheng dong 薛仁貴征東. He was also a well-known figure among Chosŏn readers of fiction and history and even surfaces in legends in locations south of P’yŏngyang. See Yi Ki-hyŏng 이기형, ‘Sŏl In-kwi chŏnsŏl-ŭi pigyo koch’al’ 설인귀 전설의 비교 고찰, Hanguk minsokhak 韓國民俗學 44 (2006): 369–401.

15 SGSG 21: 197; ZZTJ 198: 6230. XTS and Tongdian state that people from Liaozhou (formerly Yodong-sŏng) and Gaizhou (formerly Kaemok-sŏng) were relocated, but give no figures. See XTS 220: 6193; Du You 杜佑, Tongdian 通典 (Beijing: Zhonghua shuju, 1988), 186: 5017.

16 ZZTJ 198: 6235; SGSG 21: 197.

17 ZZTJ 199: 6258.

18 Song Minqiu 宋敏求, Tang da zhaoling ji 唐大詔令集 (Beijing: Shangwu yinshu guan 1959) 11: 67.

19 Kim refers to Taizong being encircled at Chup’il Hill by Koguryŏ and Malgal reinforcements — a story found in Liu Su 劉餗, Sui Tang jia hua 隋唐嘉话 (Beijing: Zhonghua Shuju 1979), p.10, as well as in Wang Dang 王讜, Tang Yulin 唐語林 (Beijing: Zhonghua shuju, 1987), 5: 429. See below, p.15.

20 A number of the peninsular sources possibly available to Kim and his team are discussed in Kenneth H.J. Gardiner, ‘The Samguksagi and Its Sources,’ Papers on Far Eastern History 2 (1970): 1–42; and Kim Pusik, The Koguryŏ Annals of the Samguk Sagi (Second Edition), ed. Edward J. Shultz and Hugh H.W. Kang (Seongnam-si: Academy of Korean Studies, 2012), pp.24–25. Some examples: red snow was seen in P’yŏnyang in the winter of 644/45 (SGSG 21: 219); a shamanness predicted (incorrectly) that the fortress of Masu 馬首 (between present-day Anshan 鞍山 and Liaoyang 遼陽, east of the Liao River) would be safe (SGSG 21: 193).

21 Taizong ordered Xu Jingzong 許敬宗 to compose a text commemorating the victory at Chup’il Hill, to be carved on a stele (JTS 199a: 5325). Neither text nor monument are still extant.

22 See p.13ff.

23 Wechsler, ‘T’ai-tsung (r. 626-649) the Consolidator,’ pp.233 and 235; Mark Edward Lewis, China’s Cosmopolitan Empire: The Tang Dynasty (Cambridge: Belknap Press of Harvard University Press, 2009), p.154. For the emperor’s own reaction, see XTS 97: 3881; ZZTJ 198: 6230; Liu Su 劉肅, Da Tang xinyu 大唐新語 in Tang Song shiliao biji congkan 唐宋史料筆記叢刊 (Beijing: Zhonghua shuju, 1984), 1:4 (regret, and rehabilitation of Wei Zheng); ZZTJ 198: 6234–35, parallel in Sui Tang jia hua, p.11 (Taizong enquiring after the reason why the Tang army was defeated by Koguryŏ).

24 See for instance Li Daliang’s 李大亮 (586–644) petition on his appointment as Grand Commissioner for Appeasement in the Northwestern Circuit (xibeidao anfu dashi 西北道安撫大使) in 630. He memorialised to the emperor that China (Zhongguo 中國) is like the trunk, and the four ‘barbarians’ (siyi 四夷) are the branches and leaves. Investment in these outlying areas would be costly and without any benefit for China itself (ZZTJ 193: 6081–82). The economic argument was also emphasised by Wei Zheng 魏徵 (580–643) in 640 when Taizong wished to put the newly conquered state of Gaochang 高昌 under the Tang’s regular territorial administration (ZZTJ 195: 6155-56). In 642, Chu Suiliang 褚遂良 (597–658) argued in the same manner for the abolition of Gaochang as a regular prefecture (JTS 80: 2736). Exactly the same arguments were employed by Di Renjie 狄仁傑 in 699, when he suggested the abolition of the Andong Protectorate (Andong duhufu 安東都護府) in the former territory of Koguryŏ (Wang Pu 王簿, Tang huiyao 唐會要) (Shanghai: Shanghai guji chubanshe, 1991) 73: 1562–563).

25 For more information on this process, see Denis Twitchett, The Writing of Official Historiography under the T’ang (Cambridge: Cambridge University Press, 1992), pp.5–30.

26 See note 24 above for Chu Suiliang. Fang Xuanling’s last memorial to the emperor was a plea to halt the preparations for Taizong’s second campaign against Koguryŏ (JTS 66: 2464–66); see also pp.23–24.

27 Michael C. Rogers, The Chronicle of Fu Chien: A Case of Exemplar History, (Berkeley and Los Angeles: University of California Press, 1968), pp.3–4, pp.40–51.

28 Denis Twitchett, The Writing of Official History under the T’ang, p.126 suggests 654 instead of 650 as the date for the presentation of these records.

29 ZZTJ 198: 6228; SGSG 21: 196.

30 Twitchett, The Writing of Official Historiography under the T’ang, pp.128–29.

31 Xiao-bin Ji, Politics and Conservatism in Northern Song China: The Career and Thought of Sima Guang (A.D.1019–1086) (Hong Kong: Chinese University Press, 2005), pp.38–39.

32 Peter K. Bol, ‘Government, Society, and State: On the Political Visions of Ssu-Ma Kuang (1019–86) and Wang an-Shih (1021–86),’ in eds. Robert P. Hymes and Conrad Schirokauer, Ordering the World: Approaches to State and Society in Sung Dynasty China (Berkeley: University of California Press, 1993), pp.149 and 158.

33 Hoyt Cleveland Tillman, ‘Textual Liberties and Restraints in Rewriting China’s Histories: The Case of Ssu-Ma Kuang’s Reconstruction of Chu-Ko Liang’s Story,’ in ed. Thomas H.C. Lee, The New and the Multiple: Sung Senses of the Past (Hong Kong: The Chinese University of Hong Kong, 2004); Mark Strange, ‘A Reading of Hou Jing’s Rebellion in Zizhi Tongjian (Comprehensive Mirror to Aid Government): The Construction of Sima Guang’s Imperial Vision,’ in ed. Daria Berg, Reading China: Fiction, History and the Dynamics of Discourse (Leiden: Brill, 2007).

34 XTS 220: 6187; ZZTJ 196: 6169–70; SGSG 20: 189. Some of the fiercest fighting during the Sui-Tang transition took place in this region, referring to modern Hebei 河北 and Shandong provinces.

35 XTS 220: 6188; ZZTJ 196: 6181–82.

36 XTS 220: 6188; ZZTJ 197: 6202; SGSG 21: 190.

37 XTS 220: 6189, 6192; ZZTJ 197: 6215; SGSG 21: 191.

38 Song Minqiu, Tang da zhaoling ji, 13: 703. Parallel text in CFYG, 117: 1398 (2).

39 XTS 220: 6194; ZZTJ 198: 6231.

40 Li Simo 李思摩 JTS 199a: 5323; ZZTJ 197: 6221; Qibi Heli 契必何力 JTS 109: 3293; XTS 110: 4119; ZZTJ 198: 6223; Daozong, Prince of Jiangxia 江夏王道宗 JTS 60: 2356; XTS 78: 3516; ZZTJ 198: 6229; SGSG 21: 196.

41 ZZTJ 198: 6225.

42 JTS 199a: 5325; XTS 220: 6193; ZZTJ 198: 6225; SGSG 21: 195.

43 ZZTJ 197: 6207; parallel version XTS 220: 6189.

44 ZZTJ 198: 6223; parallel versions JTS 119a: 5324; XTS 220: 6191; SGSG 21: 194.

45 Ibid.

46 JTS 199a: 5324; XTS 220: 6191; ZZTJ 198: 6222-23; SGSG 21: 194.

47 ZZTJ 198: 6223; SGSG 21: 194.

48 JTS 199a: 5325; XTS 220: 6192; ZZTJ 198: 6226; SGSG 21: 195-96.

49 ZZTJ 198: 6227; parallel version in XTS 220: 6193 and CFYG 181: 490–2.

50 JTS 199a: 5326; XTS 220: 6194; ZZTJ 198: 6231.

51 JTS 199a: 5325; ZZTJ 198: 6226; SGSG 21: 195. XTS 220: 6193 says they were ‘executed’ (zhu 誅).

52 JTS 199a: 5325–26; XTS 220: 6193; ZZTJ 198: 6229.

53 XTS 2: 48–49.

54 SGSG 21: 197 (after Taizong’s return to Tang territory); 22: 200 (after the death of Taizong).

55 Yang Xiong 揚雄, Yangzi Fayan 揚子法言 8.5 (Beijing: Zhonghua shuju, 1992), p.175.

56 SGSG 21: 197. Translation based on Kim Pusik, The Koguryŏ Annals, pp.248–49.

57 Referring here directly to the Appraisal of the New Tang History, XTS 2: 48–49.

58 SGSG 22: 200-01. Based on translation from Kim Pusik, The Koguryŏ Annals, pp.258–60.

59 Listed by Du You as one of the three insightful texts about Tang foreign policy. Du You, Tongdian, 200: 5502. Text of memorial given in Tongdian 186: 5017–18.

60 They are conveniently collected in Im Ki-chong 林基中, Yŏnhaengnok chŏnjip 燕行錄全集 (Seoul: Tongguk Taehakkyo Ch’ulp’anbu, 2001), 100 vols.

61 Hong Taeyong 洪大容 (1731–83) wrote two separate accounts. The one in literary Chinese, known under the title Tamhŏn Yŏngi 湛軒燕記, is organised by topic. The Korean counterpart, written in han’gŭl, and known as (Ŭlbyŏng) Yŏnhaengnok 乙丙燕行錄 is chronologically organised.

62 Mogŭn sigo 牧隱詩藁, 2: 9b.

63 JTS 2:58; XTS 2:44; ZZTJ 198: 6231.

64 This will be dealt with in more detail later in this paper.

65 Howard J. Wechsler, Mirror to the Son of Heaven (New Haven: Yale University Press, 1974), p.87.

66 XTS 97: 3881; ZZTJ 198: 6230.

67 Sŏ Seyang, untitled poem in Yanggok sŏnsaeng jip 陽谷先生集, 3: 9b–10a.

68 Ch’oe Yŏn, ‘Chup’ilsan yugam’ 駐蹕山有感, in Kanjae sŏnsaeng jip 艮齋先生集 9: 14a.

69 Wei Zheng’s posthumous title was Duke of Zheng, not Wei. It is possible that Duke of Wei refers to Fang Xuanling, who had the title Duke of the state of Wei 魏國公, but in view of the latter’s last memorial with the strong objections to the Koguryŏ campaigns, the sentiment expressed in the poem remains the same.

70 Xiong Zhonggu, Tangshu zhizhuan tongsu yanyi, juan 8, chapter 83 (in Guben xiaoshuo congkan Vol.4.2 (Beijing: Zhonghua shuju), p.745). The story covers events from 617 to 645 AD, including the campaign against Koguryŏ and the retreat from Liaodong.

71 There were popular stories about Tang Taizong in late Ming China, as illustrated by the important role he plays in the early part of the Xiyouji 西遊記. The story of his descent into hell in chapters 10 and 11 is based on a much older story. The first documented version of this is found in the Dunhuang manuscripts and possibly dated to the ninth century (see translation by Arthur Waley in Classical Chinese Literature: An Anthology of Translations, Vol.1, From Antiquity to the Tang Dynasty, ed. John Minford and Joseph S.M. Lau (Hong Kong/New York: The Chinese University Press/Columbia University Press, 2000), pp.1081–88). From the Xiyouji Taizong’s descent into hell eventually became adopted by the shaman tradition in the Korean peninsula (Boudewijn Walraven, Songs of the Shaman: The Ritual Chants of the Korean Mudang, (London: Kegan Paul International, 1994), pp.103–104).

72 Yun Kŭnsu, Wŏljŏng manp’il 月汀漫筆 in Taedong yasŭng 大東野乘 (Seoul: Seoul University Press, 1968), Vol.11, p.123; parallel passage in ‘Wŏljŏng mannok’ 漫錄, in Wŏljŏng sŏnsaeng pyŏljip 月汀先生別集, 4:17a.

73 Yun Kŭnsu 尹根壽, ‘Pongsong Cho ch’ŏmch’u yŏkyŏng sŏ’ 奉送趙僉樞如京序, Wŏljŏng sŏnsaeng jip 5: 11a–b.

74 Ibid.

75 See above, p.23.

76 Kim Siyang, Pugye kimun 涪溪記聞, in Taedong yasŭng Vol.13, p.534.

77 Sim Kwangse, ‘Sŏngsang pae’, in Hyu-ong jip 休翁集 3: 6b–7.

78 Ibid.

79 For instance, Yi Ik 李瀷, Sŏngho sonsaeng jip 星湖先生集 7:22a–b; Im Ch’angt’aek, 林昌澤 Songak jip 崧岳集 1: 24a–25b; O Kwang’un, 吳光運 Aksan mango 藥山漫稿 5: 9a–9b; Yi Kwangsa, 李匡師 Wŏngyo jip 圓嶠集 1: 16a.

80 The increased popularity of this topic occurred in the aftermath of the devastating Japanese invasions in the late sixteenth century, and the two Manchu invasions in 1627 and 1636. Partly this was due to the fact that the commander’s name only came to the attention of the peninsular writers during the Imjin war; and once the hero was given a name and a more fully developed story in the Tangshu zhizhuan tongsu yanyi, the story spread even more easily. On the other hand, one may wonder if the calamitous events of those years also supported the desire and psychological need for a native exemplary military leader. The appearance of Yang Manch’un in late Ming historical fiction coincided fortuitously with that need. But no Chosŏn author has suggested such an explicit link. Literature did play a role in the Chosŏn recovery from that trauma, as Jahyun Kim Haboush shows, but the narratives focused on a direct confrontation and reworking of the recent events, not a reimagining of the more distant past. (Jahyun Kim Haboush, ‘Dead Bodies in the Postwar Discourse of Identity in Seventeenth-Century Korea: Subversion and Literary Production in the Private Sector,’ The Journal of Asian Studies 62:2 (2003): 415–42.

81 Kang Paegnyŏn 姜栢年, ‘Ansisŏng tojung u ŭm’ 安市城途中偶吟 in Sŏlbong yugo 雪峯遺稿 14: 3.

82 Kang Paegnyŏn, ‘Ŭi Tang T’aejong chang yu Ansi chujang cho’ 擬唐太宗安市班師時諭政府詔, Sŏlbong yugo 22: 13a–b.

83 A second example of such literary historical re-enactment is from the late nineteenth century, by Kim Yunsik 金允植 (1835–1922), see ‘Ŭi Tang T’aejong Ansi-si pansa si yu chŏngbu cho’ 擬唐太宗安市班師時諭政府詔, in Unyang jip 雲養集 15: 31a–32b.

84 Kim Ch’anghŭp 金昌翕, ‘Song taeyusu Paksi pu Yŏn’ 送大有隨伯氏赴燕, in Samyŏn jip 三淵集11: 11a–21a.

85 Kim Ch’anghŭp, ‘Song taeyusu Paksi pu Yŏn, ki ch’il’ 三淵集, in Samyŏn jip 11: 12b.

86 Yi Tŏngmu 李德懋, ‘Tang T’aejong myomok’ 唐太宗眇目 in Ch’ŏngjangkwan chŏnsŏ 青莊館全書, kwŏn 32, no page number.

87 Now preserved in Wŏljŏng manp’il or Wŏljŏng mannok, see note 72.

88 Pak Chiwŏn 朴趾源, ‘Tokangok’ 渡江錄 Yŏrha ilgi 熱河日記 in Yŏn’am jip 燕巖集 11: 16b.

89 Ibid., 11: 17a. Fenghuangcheng, or Phoenix Fortress, 鳳凰城 (also Feng fortress 鳳城, modern Fengcheng 鳳城, Jilin) was long associated with the site of Ansi. Pak explains that this was based on the (erroneous) folk etymology that in the Koguryŏ language, a ‘large bird’ 大鳥 was pronounced ansi. A similar etymology was given for the Paegam (Chin. Baiyan 白巖) fortress, allegedly derived from the sound of the Koguryŏ word for snake.

90 Yi Kyukyŏng, ‘Saye’ 射藝, in Oju yŏnmun changjŏn sango 五洲衍文長箋散稿, in the ‘Insa p’yŏn: Kisullyu’ 人事篇: 技藝類 (no kwŏn or page number).

91 See, for instance, the thread ‘Yang Manch’un in History’ on the ‘Chicago Korean Drama Fan Club’ website, where history and fiction are freely mixed together. <http://deiner.proboards.com/thread/6407>.

92 For the concept of a charter polity in the Korean context, see Breuker, Establishing a Pluralist Society, p.17.

93 Duncan, ‘Historical Memories of Koguryŏ in Koryŏ and Chosŏn Korea,’ p.122.