Heirloom and Exemplar: Family and School Portraits of Confucius in the Song and Yuan Periods

The Song was a crucial period in the formation and circulation of several kinds of visual representation of Confucius (Kongzi 孔子). Sculptural images in temples where he received official sacrifices displayed imperial garb and paraphernalia related to posthumous honors bestowed on him by successive Song and Jin emperors, which remained standard until 1530.1 In addition to temple icons, whose attributes did not reflect Confucius’s social status in life or convey his personality, several pictorial compositions evolved during the Song from portraits possessed by his alleged descendants, the Kong 孔 lineage of Qufu 曲阜, Shandong.2 Although these pictures had no role in official sacrifices, they were venerated in other ways. Reproductions of the images were disseminated and inscribed both by Kong descendants and by the educated elite, who honored Confucius for his teachings. Documents mentioning depictions of Confucius suggest that people held

conflicting views about their reliability as representations of the ancient sage. Some texts reflect contemporary concerns about what constituted true resemblance and whether portraits were appropriate to use in ancestral rituals. In general, Kong descendants placed greater emphasis on the origins of an image in evaluating its accuracy, while men of the scholar-official class initially appear to have been more attentive to its expressiveness, particularly how successfully it captured Confucius's personal qualities. For the Kongs, Confucius was the illustrious ancestor from whom their own prestige and power stemmed. Portraits of him signified this connection, so their authenticity depended largely on their provenance, and idiosyncratic representational features were rarely discussed. For scholars and political figures, steeped in Confucian learning from an early age, Confucius was the wise teacher and moral exemplar. A portrait that conveyed his humane qualities could be inspirational, regardless of when the image was created or by whom, and it might be particularly effec tive to display in a school or office. Commissioning portraits of Confucius conveyed the patrons' endorsement of the values that he stood for, and such images might be used to promote ideological orthodoxy or to encourage group identity. Although the literati sometimes discussed provenance when writing inscriptions for replicas of specific portrayals, they generally did not share the descendants' interest in tracing the image back to the lifetime of Confucius himself. Moreover, some writers expressed skepticism about the feasibility of ever achieving a true physical likeness, and others considered visual portrayals of Confucius superfluous altogether, since he was abundantly represented in texts.
The roles played by members of the Kong lineage were particularly important in creating, preserving, and transmitting pictorial representations of Confucius. (See Chart 1 for the Kong descendants mentioned in this article.). In the Northern Song, various Kongs began reproducing pictures of Confucius on stone tablets, from which rubbings could be made, and these in turn could be-and were-used to carve new stones elsewhere. Portraits also were described and sometimes reproduced in family genealogies. Perhaps the most significant portrayal was a three-quarter-profile depiction of Confucius as an elderly, slightly stooped figure, standing with his hands clasped together at his chest and a sword tucked under his arm. A small composition showing him accompanied by one disciple, a shorter and younger man standing in a similar pose, was replicated in Qufu and at government schools elsewhere in the late eleventh and early twelfth centuries. Confucius later appears p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n Sage (Yansheng gong 衍聖公) began being awarded to the senior male in each generation, starting with Kong Zongyuan 孔宗愿 (fl. 11th c.) in the forty-sixth generation. 5 In 1085, his second cousin and prominent literatus-official, Kong Zonghan 孔宗翰 (fl. mid-late 11th c.), 6 published an important genealogy that formed the core of later such compilations. 7 The following year, Kong Zonghan also memorialized the throne about "deficient ceremonies of the scholars' temple" (ru miao que dian 儒廟闕典) and requested an increase in the number of households assigned to the upkeep of the temple and cemetery in Qufu. 8 His plea to improve the maintenance of the important sites associated with Confucius was favorably received by the Yuanyou regency for the newly enthroned Zhezong (r. 1085-1100), which had brought to power men who advocated sage learning and moral cultivation over utilitarian policy as appropriate means to improve governance.
A diagram of the Qufu temple as it was configured around the time that Kong Zonghan published his genealogy shows an elaborate complex of tile- roofed buildings and enclosed courtyards (Fig. 1). 9 The structures labeled Book Tower (Shu lou 書樓) and Hall of Imperial Encomia (Yuzan dian 御 贊殿) housed books and commemorative inscriptions bestowed by various emperors, who also wrote calligraphy for signboards. 10 Sacrifices to Confucius were performed in the large double-roofed building labeled Main Hall (Zhengdian 正殿), in front of which was a stepped stone platform called the Apricot Platform (Xingtan 杏壇), built in 1022 to mark the supposed site where Confucius lectured to his disciples. 11 Behind the grand ceremonial hall was another two-story building for separate offerings to the wife of Confucius, Madame Qiguan 亓官 (titled Yunguo furen 鄆國服人), and smaller shrines for his son Boyu 伯魚 (Sishui hou 泗水侯) and grandson Zi Si 子思 (Yishui hou 沂水侯) on the east and west sides, respectively. Flanking the main axis of the temple to the west were shrines to the father and mother of Confucius (Qiguo gong 齊國公 and Luguo taifuren 魯國太夫人) and the Five Worthies 9. The Song temple's layout may have been illustrated in Kong Zonghan's 1085 Jiapu and carried over into Kong Chuan's update of the genealogy; if not, then it may have originated with Kong Chuan. Kong Yuancuo probably copied the picture from Kong Chuan's now-lost Zuting zaji 祖庭雜記 of 1124 (discussed in note 19), and it matches Kong Chuan's verbal description of the temple in his 1134 Dongjia zaji, xia 下.16a-17a (133-135); also see note 42. Despite various fires and rebuilding, the configuration was only slightly altered under the Jin; compare Fig. 1 with Kong Yuancuo, Kongshi zuting guangji, preliminary section (11). A slightly rearranged diagram of the Song temple and an enumeration of its buildings are preserved in late Yuan versions of the late Southern Song illustrated compendium Shilin guangji 事林廣記; for example, see the edition in the Naikaku bunko, Japan (whose interior title is given as Xinbian zuantu zenglei qunshu yilan shilin quanbi 新編纂圖增類群書一覽事林全璧; Harvard-Yenching microfilm FC-5829), hou ji 後集 3.2a-b; for further references on Shilin guangji, see note 69. The picture is labeled with the temple's Song name, Zhi sheng wen xuan wang miao 至聖文宣王廟, in other words predating the Yuan emperor Wuzong's (r. 1307-1311) addition of the epithet "Dacheng 大成" in 1307. Yang Huan 楊奐 visited the Qufu temple in 1252 and described its layout in detail; see his "Record of a Journey to the East" ("Dongyou ji" 東遊記), in Huanshan yigao 還山遺稿, rpt. Yingyin Wenyuange Siku quanshu, v. 1198 (Taipei: Shangwu yinshuguan, 1983), shang.20a-23a (1198-233-1198-234).
11. The Apricot Platform was installed by 45th-generation descendant Kong Daofu 孔道 輔 (986-1039), who rebuilt the Qufu temple; see Kong Chuan, Dongjia zaji, xia.2b (106). A comparison of the Song and Jin ground plans of the Qufu temple (see note 9) shows the Jin addition of a stele and pavilion on top. The reference to Confucius teaching his disciples at the Apricot Platform comes from Zhuangzi "Yufu pian" 漁父篇; see Zhuangzi jishi 庄子集釋, Guo Qing fan 郭慶藩, comp. (Beijing: Zhonghua shuju, 1961), 1023. p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n Fig. 1  五賢, 12 and to the east a family temple (Jiamiao 家廟) for worshiping more immediate lineage ancestors. Yet further east are the Kong lineage's administrative offices and facilities for receiving guests. Because the Kong descendants worshiped their own recent ancestors in addition to maintaining the cult of Confucius on behalf of the emperor, the ritual complex and practices in Qufu differed somewhat from those of the typical Confucian temple associated with government schools throughout the empire.
Kong Zonghan's 1085 genealogy included a detailed list of Confucius's physical attributes and summary descriptions of four pictorial representations, 12. According to Yang Huan, the Five Worthies were Meng(zi) 孟子, Xun(zi) 荀子, Yang (Xiong) 揚雄, Wang (Chong) 王充 and Han (Yu) 韓愈; "Dongyou ji," shang. 21b (1198-233). The shrine appears behind that of Confucius's mother on the Song plan (i.e., Fig. 1), and along the west side of the parental compound on the Jin plan (labeled Wuxian tang 五賢堂); see Kong Yuancuo, Kongshi zuting guangji, preliminary picture section (11). starting with a portrait that was currently kept in the family temple (jin jia miao suo cang hua xiang 今家廟所藏畫像). 13 The Jiamiao was an ancestral shrine used only by the members of the senior Kong lineage, unlike the "scholars' temple" (i.e., the grand ceremonial hall labeled Zhengdian), where scholar-officials also participated in sacrificial rituals. 14 According to Kong Zonghan, the picture was called simply the Small Portrait (Xiao ying 小影; literally "small shadow") and was "the truest of the Sage's portraits" (yu sheng xiang wei zui zhen 於聖像為最真). 15 The composition portrayed Confucius dressed in leisure robes and accompanied by Yan Hui 顏回, his favorite disciple (yi yan ju fu, Yan zi congxing 衣燕居服, 顏子從行). The work was probably a painting, although the term hua xiang could also indicate an incised tablet. Kong Zonghan gave no further information about it, nor his reasons for proclaiming that it was the most genuine among all depictions of Confucius.
Besides the Small Portrait, Kong Zonghan described three other pictures that had come down to recent generations. 16 Two showed Confucius seated by a curved low table and holding a jade whisk. In one of the two, he was attended by ten disciples, among them someone holding a canopy and another a jade chime-stone. In the other composition, seventy-two disciples stood in a row, some of them grasping a bow or arrow or opening a scroll. picture showed Confucius riding in a carriage, accompanied by ten disciples. Without saying exactly where these pictures were or what they were called, Kong Zonghan disparaged them as "all re-creations of later people, hardly true portraits of the First Sage" (jie hou ren zhui xie, dai fei xian sheng zhi zhen xiang 皆後人追寫, 殆非先聖之真像). 17 The implication is that later people who had never seen Confucius just depicted their own conceptions of him and thus could not have portrayed him accurately. As discussed below, some of the pictures disparaged by Kong Zonghan were owned by his close relatives, and his comments may be emblematic of intra-lineage competition for prestige and power.
Kong Zonghan's notes on the four portraits were transcribed by his nephew, Kong Chuan 孔傳 (ca. 1059-ca. 1134), 18 into an expanded edition of the genealogy completed around 1124. 19  . Kong Duanchao's 孔端朝 1132-dated postface to Dongjia zaji implies that Kong Chuan compiled two works: a genealogy that he brought to the South, which was scattered during a raid in 1131 but substantially recovered; and an abridged version that excluded collateral branches of the Kong lineage; Kong Chuan, Dongjia zaji, hou xu.1b-2b (166-168). Therefore, it seems possible that, in the South, Kong Chuan revised an earlier, fuller version of the genealogy that he had initially prepared in the Qufu, and Kong Yuancuo subsequently used the earlier version as the basis for his own edition. Careful scrutiny of the National Library of China's early edition of Dongjia zaji, which it calls a "Song edition with ongoing revisions" (Song ke di xiu ben 宋刻遞修本), shows that Song taboos are observed for emperors through Gaozong, supporting a compilation date between 1127-1162; see the facsimile published in Zhongguo zaizao shanben, Tang Song bian, Shi bu 中華再造善本. 唐宋編, 史 部, v. 142 (Beijing, Beijing tushuguan chubanshe, 2006). However, it also includes descendants through the end of the Southern Song, and the pages in various sections are cut in different quoted a panegyric called "Encomium for the Small Portrait (Xiao ying zan 小影贊), which had been composed by Yin Fuzhen 尹復臻 (active late 11th century), probably after his 1089 appointment as a teacher at the school associated with Qufu temple (Queli miaoxue jiaoshou 闕里廟學教授). 20 The first few lines of Yin's ode to "the truest of the Sage's portraits" link the ancient heirloom to Confucius's home and imply that it was made around his lifetime:

夫子之象
The imposing and yet not intimidating, 恭而安 courteous and yet at ease" 21 . . .
Yin Fuzhen's references to the portrait's great antiquity and family provenance probably reflect oral traditions among the Kong descendants and suggest the main reasons for regarding it as a genuine and reliable representation. The image also reminded Yin of a well-known passage in the Analects (Lun yu 論 語) that described Confucius's appearance, a further confirmation that the portrait was accurate. After Yin Fuzhen's encomium, Kong Chuan restated his uncle's verdict on portraits of Confucius: "Of the pictorial portraits transtypefaces, suggesting a compilation from multiple sources. The Qing editors of Siku quanshu zongmu tiyao 四庫全書總目提要 were puzzled by numerous discrepancies between the edition of Dongjia zaji that they saw and the work's characterization in earlier bibliographies; their comments are transcribed at the front of the Kongzi wenhua daquan reprint, 9-12. Finally, Dongjia zaji includes an illustration of Confucius seated with disciples at the Apricot Platform, which is not mentioned anywhere in the text and probably was added later. I discuss all these matters further in "Descendants and Portraits of Confucius in the early Southern Song." 20. Kong Chuan, Dongjia zaji, xia.4a (109). Yin Fuzhen's appointment as teacher in 1089 and his acquisition of 20 qing 頃 (hectares) of land to support the students are noted in a 1299-dated stele recording the history of the temple school; transcribed in Luo Chenglie 駱承烈, Shitou shang de rujia wenxian: Qufu beiwen lu: 石頭上的儒家文獻: 曲阜碑文錄 (Ji'nan: Qi Lu shushe, 2001), 240-242. I have not found an independent text containing Yin's "Encomium for the Small Portrait," nor biographical information on him.
21. This is a line from Lun yu 論語 7.38; translation by Raymond Dawson, Confucius: The Analects, Oxford World's Classics (Oxford: Oxford University Press, 1993), 27. p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n mitted to this generation, except for the Small Portrait, all are false versions" (shi zhi suo chuan, fei xiao ying hua xiang, jie wei yan ben 世之所傳,非小 影畫像,皆為贗本). 22 Next, Kong Chuan quoted several lines from the late Tang literatus Liu Yuxi's 劉禹錫 (772-842) Stele for the New Temple in Xuzhou (Xuzhou xin miao bei 許州新廟碑), dateable to around 836. 23 In his commemorative record for a recently reconstructed temple at the Xuzhou Prefectural School (modern Xuchang 許昌, Henan), Liu Yuxi described its image of Confucius as "A likeness with the head of Yao, the body of Yu, a fine cap and ivory ornaments, acquired from Zou Lu [Qufu]" (Yao tou Yu shen, hua guan xiang pei zhi rong, qu zhi zi Zou Lu 堯頭禹身, 華冠象佩之容, 取之自鄒魯). 24 In the next sentence of the temple stele, not transcribed by Kong Chuan, Liu asserted that the sculptural icons and painted images in the sacrificial hall were in accord with regulations obtained from the Imperial University. 25 The two sentences are structurally parallel, and both invoke authoritative models elsewhere; the ritual layout was correct because it came from the most prestigious school in the capital, and the portrait of Confucius was reliable because it came from the Sage's home region. Liu's text thus endorsed provenance as the guarantor of authenticity. Kong Chuan concluded the entire discussion by declaring that the image described by Liu Yuxi was none other than the Small Portrait that had come down to his own era. 24. My translation is based on the stele text as given in Liu Yuxi ji, 3.28. As transcribed in the edition reprinted in Kongzi wenhua daquan, Kong Chuan's excerpt has the words xiang fu 像 服 (portrait dress) instead of xiang pei 象佩 (ivory ornaments); see Kong Chuan, Dongjia zaji, xia.4b (110). The edition in the National Library of China (see note 12) renders it as xiang pei 像 佩 (portrait ornaments). Earlier in the passage quoted from Kong Zonghan's Jiapu, Confucius's bodily features are described as resembling those of ancient sage-emperors; see Kong Chuan, Dongjia zaji, xia.3b (108).
25. Liu Yuxi ji, 29. The Imperial University was the highest of the "scholars' temples" and an appropriate model for a prefectural school's temple to follow.
26. I suspect that Liu Yuxi was not describing a two-dimensional portrayal of Confucius and Yan Hui, but rather, the Xuzhou temple's main icon, which displayed emblems of posthumous kingly rank and probably was a sculptural image. If so, then Kong Chuan's conclusion would not be justifiable.

Northern Song images
Kong Chuan's Dongjia zaji did not include an illustration of the Small Portrait, but his description of Confucius in leisure clothing with Yan Hui following behind generically fits an image reproduced on a number of late Northern Song stone tablets, such as a 1095-dated stele in Qufu, from which the rubbing in Fig. 2 comes (Fig. 2)  遺像), a reference to the Jiamiao. 30 Moreover, Zhang Chi's 1088 inscription echoes Kong Zonghan and Kong Chuan by concluding, "Of all the portraits of the sage, past and present, only this one is genuine" (gu jin sheng xiang, du ci wei zhen 古今聖像, 獨此為真). 31 However, Zhang says that the image depicted Confucius as Minister of Justice in Lu (Lu sikou 魯司寇), suggesting that the costume was more formal than leisure garb. 32 Zhang also attributes the portrayal to a specific artist, the Tang master-painter Wu Daozi 吳道子 (ca. 689-after 755), who is not mentioned in the genealogical account. 33 Although authorship by Wu Daozi would seem to refute any possibility that the portrait of Confucius with Yan Hui had come down from Confucius's lifetime, the contradiction could be resolved by positing that Wu had copied an ancient work. If that were the case, the Kongs might have understood Wu's role as merely facilitating the transmission of the authentic image. For the literati who sponsored replicas of the portrait, however, it may have been more important that Wu's depiction captured essential qualities associated with Confucius. Discussions of figure painting had long given priority to achieving lifelike vitality (sheng qi 生氣), and Su Shi 蘇軾 (1037-1101) had recently written an essay emphasizing the importance of conveying the subject's inborn nature (tian 天). 34 after 1142), whose inscription below the image consists primarily of a quotation from Yin Fuzhen's "Encomium for the Small Portrait," composed for its prototype (Fig. 3). 38 The two compositions are virtually identical and probably represented the precious heirloom fairly accurately. Its appearance is further confirmed by a line-drawing labeled with two titles, Small Portrait (Xiao ying 小影) and Master Yan Following in Attendance (Yanzi cong xing 顏子從行), which is reproduced in the front section of a later genealogy, Kong Yuancuo's 孔元措 (1181-c. 1251) Kongshi zuting guangji 孔氏祖庭廣記 (Fig. 4). 39 Zhang   successively invited to serve in the states, with carriage, clothing, and figures displaying pomp and ceremony, which was exceedingly fine" (miao bi you Wu Daozi hua xian sheng li pin zhu guo, ju fu ren wu wei yi, ji wei jing miao 廟 壁 有 吳 道 子 畫 先 聖 歷 聘 諸 國 , 車 服 人 物 威 儀 , 極 為 精 妙 ). 43 The seeming contradiction suggests that Kong Chuan appreciated the artistry of this representation, even if it was one of the "re-creations by later people" disparaged by Kong Zonghan. In any case, a composition that corresponds to these verbal descriptions appears as a line-drawing in Kong Yuancuo's Kongshi zuting guangji, under the title Riding in a Carriage (Cheng lu 乘輅) (Fig. 5). 42. Kong Chuan, Dongjia zaji, xia.16a-17a (133-135). Highly consistent with Fig. 1, the entire section reads as a commentary to a pictorial plan that was probably included in an earlier version of the genealogy; see note 9. It is possible that picture and text both came from Kong Zonghan's 1085 Jiapu, as the latest date mentioned is 1048.
43. Kong Chuan, Dongjia zaji, xia.17a (135). p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n Its depiction of Confucius traveling in grand style as a high official with a large retinue presents a sharp contrast to the modest representation in the Small Portrait, where he has only one attendant and appears to be walking (cf. Figs.  2-4). The grandiose image may have appealed to Kong descendants' pride in their ancestor, but it seems not to have inspired replications elsewhere, perhaps because it conflicted with the more prevalent concept of Confucius as a frustrated statesman and humane teacher. Seven months after Kong Duanyou erected his stele juxtaposing the two Song emperors' encomia to Confucius with a replica of the Small Portrait, a forty-sixth-generation descendant named Kong Zongshou 孔宗壽 (fl. 11th c.) set up a stone incised with the same two imperial odes above a different image (Fig. 6). 44 Instead of Confucius standing with Yan Hui, it portrayed the master holding a ruyi scepter, seated on a dais with an armrest and attended by ten disciples, one of whom grasps the pole of the canopy over Confucius's head.  p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n Kong Zongshou's own inscription, incised beneath the scene, claimed that his family had two pictures by Wu Daozi; this one, which he rather confusingly calls the Small Portrait (Xiao ying), and a standing image of Confucius with Yan Hui, which he refers to as Practicing the Teaching (Xing jiao 行 教). 45 Kong Zongshou notes that the latter had already been reproduced in stone, undoubtedly referring to the tablet that Kong Duanyou had erected in the third lunar month of the same year, 1095. Therefore, Kong Zongshou continues, he was having the seated portrait carved on durable polished stone as well, in order to prevent the many imitations from ever becoming confused with the genuine (zhen 真).
Perhaps the reason why Kong Zongshou called the picture he reproduced "the Small Portrait" was to lay claim to the authenticity associated with that title in Kong Zonghan's genealogy and Yin Fuzhen's "Encomium for the Small Portrait." 46 However, the composition of Confucius seated with ten disciples largely fits the description of the first portrayal that Kong Zonghan, his first cousin, had dismissed as one of the "re-creations by later people." 47 The regal canopy and ornate dais again aggrandize Confucius, although somewhat less than the aristocratic carriage of the previous picture. The line-drawing of the composition in Kong Yuancuo's Kongshi zuting guangji, titled Leaning on a Small showing three of the disciples standing in front of Confucius, rather than all ten around the sides and back of his dais. 48 This more interactive arrangement suggests a narrative in which a disciple's question elicits a pronouncement from Confucius, very like (and possibly influenced by) the opening scene of illustrated Song versions of the Classic of Filial Piety (Xiao jing 孝經). 49 Although it is impossible to say whether Kong Zongshou's stele or Kong 48. The one later replication of Kong Zongshou's stele that I have found is a stone erected around 1338 at the Jiangning 江寧 Prefectural School by education officials; see rubbing reproduced in Beijing tushuguan cang huaxiang taben huibian, 1:93. To the right of the image is the notation "respectfully copied by Wang Yuan of Qiantang" (錢唐王淵敬摹). The dedicatory inscription below conflates Kong Zonghan's description of portraits with Kong Zongshou's comments attributing this "Small Portrait" to Wu Daozi, and says that the censor Zhou Shunju 周 舜舉 saw a version at the Yaozhou school and ordered a replica carved.  p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n Yuan cuo's woodblock illustration was more faithful to the family painting they both purported to reproduce, the iconic quality of the stele's figural arrangement is echoed in closely related illustrations of Confucius and disciples at the Apricot Platform, such as the one included at the front of Kong Chuan's Dongjia zaji (Fig. 8). 50 The one portrayal that Kong Yuancuo did not illustrate in his expanded thirteenth-century genealogy was a composition that Kong Zonghan had characterized as showing Confucius seated with his seventy-two disciples arranged in a row, some of them holding a bow, arrow, or handscroll. 51 Although the picture may simply have been too wide to reproduce well in a small book, it also had acquired a problematic new association since Kong Zonghan's time. As described below, the portrayal became closely identified with the Southern Song restoration in Lin'an (Hangzhou), and Kong Yuancuo served the rival Jin dynasty as the fifty-first-generation Duke for Perpetuating the Sage. He not only lost control of Qufu while staying in the Jin capital at Kaifeng, but also another duke was installed when the Southern Song recaptured Qufu in 1225. Kong Yuancuo compiled his ambitious genealogy in part to reassert his claims as senior descendant and rightful duke, an effort that succeeded in 1233. 52 He bolstered his case by augmenting the text with visual representations of portraits and precincts that were significant to the Kongs' history, which were later adopted into a more comprehensive Ming gazetteer, Queli zhi 闕里 誌. 53 Kong Yuancuo would not have gained any advantage from reproducing an image that had become associated with the Southern Song.   p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n

The Move to the South: Southern Song and Jin Images
The Jin invasion and conquest of North China had a far-reaching effect on the Kong lineage and its ancestral cult of Confucius. In the autumn of 1128, Song emperor Gaozong summoned Kong Duanyou, by then the forty-eighthgeneration Duke for Perpetuating the Sage, to perform a sacrifice in the Song temporary capital at Yangzhou. The Jin installed his younger brother Kong Duancao 孔端操 (c. 1062-c. 1133) as duke in Qufu, and Kong Duanyou followed Gaozong south to Lin'an in early 1129, along with other members of the senior lineage. 54 When the advancing Jin armies temporarily drove Gaozong out of Lin'an, the Kongs went on to Quzhou 衢州, a more protected area up a mountain valley to the southwest. 55 Gaozong awarded them offices and titles to provide them with subsistence and enable them to resume their ancestral sacrifices in the South. As performances there were supposed to be just a temporary expedient, Gaozong did not build a grand new temple for the Kong ancestral sacrifices. 56  to serve as the Kong Family Temple, located northeast of the prefecture at Caltrop Lake (Linghu 菱湖). 58 The Southern Song and Jin regimes both patronized the cult of Confucius and invested successive generations of Kongs with the hereditary title of Duke for Perpetuating the Sage, thus creating rival dukes in the South and North. 59 Although the Jin controlled the ancient homeland of Confucius, the Southern Song harbored his senior descendants, and its legitimacy was enhanced by protecting and supporting them. Song Gaozong also used visual media to asso ciate his restorationist regime with the Confucian legacy in order to ensure the allegiance of the educated elite, particularly after making a controversial peace settlement with the Jin in 1142. 60 Among other things, between 1143 and 1146, he transcribed the complete texts of six Confucian classics and had them carved on stone tablets erected in the imperial university, so that rubbings could be distributed to prefectural schools. 61 In 1144, after visiting the imperial university, Gaozong composed a poetic encomium in honor of Confucius and also had it carved for further dissemination. Subsequently, he wrote an encomium for each of the seventy-two disciples, outdoing his imperial predecessors Taizu and Zhenzong, whose officials had composed these subsidiary texts. In 1155, Gaozong re-transcribed his eulogies for Confucius and the seventy-two disciples and had each text paired with a pictorial image. When the imperial university moved into spacious new quarters, the set of texts and pictures was carved on stone tablets and installed there in the twelfth month of 1156 (January 1157 by the Western calendar). 62 Its stated purposes 58. This is the date given in Yang Tingwang comp., (Kangxi) Quzhou fuzhi, 7.4a (153); Xiaobing Wang-Riese 王霄冰 dates the founding to 1253; Nanzong ji Kong 南宗祭孔 (Hangzhou: Zhejiang Renmin chubanshe, 2008), 31. Perhaps a reason why the Kong descendants only gained proper facilities in the South this long after fleeing from Qufu was that the Mongol conquest of the Jin made their return even more unlikely; for additional discussion, see "Nanzong Kong miao" 南宗孔廟, Hudong baike 互動百科, note 56 above).  p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n were to inspire these elite students and facilitate broader distribution through rubbings, which were sent to schools throughout the South (Fig. 9). The composition shows Confucius sitting on a low dais, facing a line of disciples holding various objects, with inscriptions appearing in the intervening spaces. The pictorial array generally corresponds to Kong Zonghan's verbal description of Confucius seated with a jade whisk and accompanied by seventy-two disciples arranged in a row. 63 Although a Southern Song court painter probably designed the figures simply from verbal prompts, it is possible that they were modeled on a painting brought from the North. 64 In any case, the ancillary inscriptions say nothing about the authenticity or antiquity of the portrayals, which must have been deemed satisfactory for the emperor's political goals of displaying his commitment to Confucian ideals of governance and promoting ideological conformity among students. According to family traditions recorded in later sources, the Kong refugees brought with them precious documents and heirlooms from Qufu. These allegedly included a pair of pistache wood (kai mu 楷木) votive statues depicting Confucius (Fig. 10) and his wife, which were believed to have been carved by Zi Gong 子貢, the devoted disciple who had kept a six-year 63. The main differences are that the Hangzhou stones do not show Confucius resting against a low table, and he holds a stiff implement shaped like a back-scratcher, perhaps meant as a ruyi scepter, rather than a jade whisk.
64. In addition to the remote possibility that the Kong refugees brought a painting from Qufu, there is a persistent later tradition that associates the subject with a Northern Song artist, Li Gonglin 李公麟 (c. 1049-1106). Various forgeries purport to be Li's work, such as a colored handscroll in the Beijing Palace Museum, which has a fake signature and date of 1088, reproduced in Jin Tang    Although the original stone seems to have been destroyed at the end of the Southern Song, a replacement was carved in about 1520 (Fig. 11), when a new temple for the Southern Kong descendants was built in the eastern part of the city, its present location. 67 The lifesize Confucius depicted on the tablet looks like an enlarged version of his figure in the Small Portrait compositions (cf. Figs. 2-4). Given the importance of the Small Portrait, which Kong Zonghan and Kong Chuan had called the truest portrayal of Confucius, perhaps Kong Duanyou had brought it from Qufu, or at least a rubbing of his 1095 tablet reproducing it (Fig. 2). In any case, the two-man composition was widely preserved in rubbings and on stone tablets at government schools. 68 Moreover,  Another reason for thinking that the Ming re-cut stele faithfully reflected the lost original is because it resembles Yuan and early Ming paintings and rubbings whose mutual similarities suggest a common archetype (e.g.   74 He immediately recognized the stele as Wu Daozi's portrait of Confucius at leisure (Xian sheng yan ju zhi xiang 先聖燕居 之像) and learned the story of its unusual provenance from the Daoist priests. Jin discussed his discovery with colleagues, and they submitted a petition to move the tablet to the Tower of Revering Culture (Chongwen'ge 崇文閣), a building in the Jiangling Prefectural School. After the stone was transferred, rubbings were made and sometimes used to carve new stelae elsewhere. One of them was the stele at the Huangmei school, for which Chen Hao wrote this commemorative inscription soon after arriving at his post in 1326. 75  75. Chen Hao says that he heard the story from Secretarial Assistant Tao Jingshan 陶景山 soon after starting his job at the Huangmei District School. Tao claimed to have been in Jiangling when Jin Liangshu discovered the portrait and took credit for proposing the petition to move it to the Chongwen'ge. Chen Hao subsequently obtained the portrait (presumably a rubbing) from the Li Hefu Family School (Li Hefu jiashu 李和甫家塾 ; unidentified). When he showed it to the magistrate, a Mr. Li of Jingshan 景山李氏, the latter was overjoyed and ordered a new stone carved to reproduce it, for which Chen wrote the inscription. The connection between the the tale are repeated on a number of stelae that reproduce the large standing image, although they sometimes omit mention of Chen Hao and Huangmei, simply focusing on the miraculously discovered stele in Jiangling. 76 The details of this narrative may have been embellished with stock elements from other tales of images that displayed paranormal signs of their heavenly efficacy. The story suggests that the incised stone Luo Feng found in the bridge did not bear an identifying title, or if it did, the writing was in archaic seal script and not easily read. The tablet probably had been removed from its original location amid some kind of social breakdown, such as the fighting and turmoil associated with the Mongol conquest. Perhaps the portrait stone had belonged to a Southern Song school or academy that was subsequently destroyed. The stone's incorporation into a bridge and the failure of the official who found it to recognize the subject suggest that the cult of Confucius had been disrupted for some time, as was the case in the early Yuan. The supernaturally-tinged account of the tablet's discovery and the prominence it gained from being moved to the Jiangling Prefectural School made the image especially appealing to reproduce elsewhere to spread its efficacious benefit. The resemblance between this portrayal and the Ming replacement stele in Quzhou suggests that they had a common source, which I suggest was the stele originally erected in the early 1130s by Kong Duanyou and Kong Chuan, ultimately based on the family heirloom known as the Small Portrait.
The Quzhou Legacy Portrait of the First Sage differs from the Qufu Small Portrait in three major ways that made its iconography more suitable for the relocated family and state cults in the early Southern Song. In the Quzhou Huangmei stone and the one in Jiangling thus seems indirect. However, the notations "Stele in the Huangmei School" (Huangmei xian bei 黃梅縣碑) and "brush of Wu Daozi" (Wu Daozi bi 吳道子筆) appear on a 1547-dated stone with the same image, under the title Legacy Portrait of the Ultimate Sage (cf. Fig. 13); see rubbing reproduced in Jean Keim, Chinese Art (New York: Tudor Publishing Co., 1961), v. 1 pl. 8. 76. For example, a Yuan military official, Sengjianu 僧家奴, erected a stele titled Legacy Portrait of the Exalted Sage (Xuan sheng yi xiang 宣聖遺像) at the Guangzhou Prefectural School in 1345, based on a rubbing he had acquired in 1341, and his inscription recounts the discovery of the Jiangling stone without mentioning Chen Hao; see illustration and partial transcription in Luo Chenglie, Huaxiang zhong de Kongzi, 27. I obtained a more complete version of Sengjia nu's text by examining the National Library of China's rubbing of a 1440-dated stele that was made to replicate the Guangzhou stele for the Huating 華亭 District School in Songjiang 松 江 prefecture; a poor reproduction appears in Beijing tushuguan cang huaxiang taben huibian 1: 97.  p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n stele, Confucius stands alone, without the attending disciple Yan Hui, who also had a temple and tomb in Qufu receiving Jin patronage. Thus, the portrait was perhaps less likely to evoke the loss of the northern homeland. Also, Confucius appears to be more elderly and wears the simple cloth cap of the scholar-recluse, signifying the last years of his life. By contrast, the Small Portrait shows a flower-shaped cap with a large pin, which was associated with depictions of him traveling around the northern states in search of a worthy ruler to serve. Finally, the enlargement of Confucius's figure rendered him more vividly present, outside the limits of time and place, inviting the viewer to engage and commune with him. This evocative image, attested as a true portrayal handed down by his direct descendants, offered latter-day viewers the possibility of a transformative encounter with the ancient sage.
of Confucius in leisure clothing was eventually replicated on a stele at the Qufu temple, without a title but a notation attributing it to Wu Daozi, it was never illustrated in any edition of the gazetteer, Queli zhi, which always included the portrayal with Yan Hui. 80 And from the mid-Ming onward, the two-man portrait often served as the pictorial frontispiece to serial narrative illustrations of Confucius's life, the Shengji tu 聖蹟圖. 81

Concluding Remarks
Although Kong descendants and unrelated men of the educated class both were involved in transmitting and reproducing portraits of Confucius, they seem to have held different views on the significance of the images. For the Kongs, Confucius was the founding ancestor, and heirloom portraits facilitated familial worship. Moreover, descendants owed the social and economic entitlements they received from successive rulers to his posthumous recognition as a sage, and the possession of authentic portraits symbolized legitimate claims to favored treatment. Perhaps the creation of a lifesize and approachable image of Confucius on a stele in Quzhou was a means of affirming Kong family identity, as well as providing psychological comfort to members displaced from their p o r t r a i t s o f c o n f u c i u s i n t h e s o n g a n d y u a n ancestral homeland. In other situations, Kong descendants endorsed images that depicted Confucius as an honored statesman.
For the scholar-elite, Confucius was the archetypal teacher and the exemplar of both high-minded official service and principled withdrawal. In the brutally factionalized sphere of Song politics, literati viewers may have found images of Confucius in retirement particularly resonant. After years of frustration in his quest to implement his ideals of governance, he had finally given up seeking official appointment, making him a consoling model for later men who suffered career setbacks in adverse political conditions. Failure also had enabled Confucius to become the charismatic figure who inspired legions of disciples and later followers to preserve and spread his teachings. 82 In schools, a large image of Confucius dressed in the informal robes of a scholar invited quoti dian veneration outside the formal sacrificial rites in the temple. The daoxue master Zhu Xi 朱熹 (1130-1200) is said to have led students at the Cangzhou Academy 滄州書院 in offering obeisance to a portrait of Confucius every morning. 83 By contrast, the depictions of Confucius and seventy-two disciples with Song Gaozong's encomia, replicated for government schools from the incised stones in the Hangzhou imperial university, probably did not attract the same kind of reverential engagement. Not only are the figures much smaller, their varied postures and animated facial expressions have the lively and entertaining character of narrative illustrations, rather than the solemn stasis of icons. Moreover, the accompanying texts, composed by the emperor, may have seemed more important.
References to replication repeatedly come up in connection with portraits of Confucius. Favored images were often reproduced, sometimes in other media, thereby becoming more widely known. Kong descendants preserved the compositions of their heirloom paintings by having them copied in stone and woodblock, from which rubbings and prints enabled additional versions to be made elsewhere. I have not found any indication that a replica was considered to represent Confucius less well than its prototype or to have less efficacy. Through reproductions of a true portrait of Confucius, he could be made present anywhere. Nonetheless, it is difficult to match extant images with textual descriptions, and connections must always be considered tentative. Accounts that were originally written for or about a specific portrayal, such as Chen Hao's record of the stele carved for the Huangmei School, were freely appropriated for other images and sometimes excerpted in misleading ways. The rather generic titles given to depictions of Confucius also make it hard to trace individual examples. Filiations suggested by slight variations in iconographical details, such as the direction Confucius faces, the exact position of his hands, the type of headdress he wears, and the way his sword protrudes, are of only limited help in identifying possible prototypes. Although writers sometimes refer to the extraordinary features of his physical appearance, these elements were not incorporated into visual representations. 84 Neither the Kong descendants nor most literati seem to have been particularly concerned about the specific visual details of the portraits; validation as a true image depended on other qualities. 85 Ultimately, the Tang artist Wu Daozi came to be associated with all the variations on the large solo portrait.