Datasets:

todosi
/

asdf

Dataset card Viewer Files Files and versions Community

Split (1)

train · 324 rows

text stringlengths 42 3.85k	source stringclasses 1 value
88 enshrinement of the Buddha, of his body and his wisdom. 94 The central pillar is further conflated with the imaginary geographical center of the world--the Sumeru in the Buddhist cosmography and its Chinese equivalent, the Kunlun Mountain. 95 The Abhidharmakosa (Ch. Abidamo jushelun 阿毗达磨俱舍論 ) depicts a world system centered around the Sumeru, the highest mountain emerging from the eight seas and encircled by seven concentric rings of lesser mountains, with an additional ring of iron wall, the cakravala (Ch. tieweish an 鐵圍山 ), at the periphery. 96 This is represented by an intriguing diagram from the Dunhuang Library Cave (fig. 42): a four-story pavilion-type pagoda extends from the foot to the waist of the mountain, on top of which is a monaster y where an extremely tall, twenty-four-story pagoda occupies the center, standing for the twenty-four levels of heavens superimposed above the Sumeru. The Sumeru motif also appeared in the revolving sutra case where it was represented on the dais, giving i t the name of the “Sumeru dais ( xumizuo 須彌坐). ”97 For a sutra case that is turnable, the archetype of the central pillar is even more accentuated by the pivot. It becomes an amalgam of multiple interrelated images and their cultural underpinnings, alluding to the medieval monastic layout, central-pillar cave temples, the Sumeru, the Kunlun Mountains, and the axis mundi in a more general sense. 94 Seckel 1980, 249. As Seckel observes, the stupa in East Asia still serves as a n embodiment of the Buddhist ideal of nirvana, as a relic container, a nd as a commemorative monument. 95 Stein 1990, 223-46, 248. The world pillar motif in Stein's discussion is a shared Pan-North Asia cultural phenomenon. For China, the Sumeru is sometimes conflated with the mythical mountains of Kunlun and Penglai. It can even relate to the spinal column of the human body. 96 Kloetzli 1983, 24-25. 97 Stein 1990, 254-56. In the Yingzao fashi, the dais is simply referred to as “zuo. ”	di-luo-a-grain-of-sand-full-draft-3.pdf
89 The wheel The Sumeru motif leads us to the discussion of the third archetype for the Longxingsi sutra case, and that is the wheel--the most distinctive cha racteristic of this rotating device. Though the mechanism of the wheel is straightforward enough, the purpose of integrating it with a receptacle of sacred images and books has often puzzled scholars. Carrington-Goodrich approaches this question from both practical and ritual angles, suggesting that the revolving bookcase must have helped to save space and accelerate the work of the translators and copyists of Buddhist sutras, who would have the entire set of the Tripitaka close at hand so that they did not have to run around the room to locate a certain scripture. 98 In terms of ritual, he suspects that the turning sutra case functioned in a similar manner as the prayer wheel ( zhuanjingtong 轉經筒 ) widely used in Tibetan Buddhism. 99 As I have elaborated earlier in this chapter, the rotating wheel granted the receptacle miraculous powers to enlighten the mass. Turning the wheel not only became an “expedient means” to salvation, but it was also st rongly reminiscent of the practice of circumambulation--the clockwise, circulatory movement around a Buddhist monument. 100 What can be added is the symbolic aspect of the wheel associated with the metaphor of the Sumeru as the world pillar. Randy Kloetzli ha s explicated that in the single-world Buddhist cosmos, the Sumeru is literally the immutable cosmic “axle” around which the sun, the moon, and the stars 98 Carrington-Goodrich 1942, 157. If it were for the convenience of the translators and copyists, the historical records presented earlier in this chapter should have mentioned it, but none of them has. This could mean that the revolving mechanism is main ly a powerful religious and ritual symbol not necessarily for practical purposes. 99 Ibid., 152-55. According to Carrington-Goodrich, neither of the two Chinese pilgrims to India, Faxian and Xuanzang, mentioned any kind of prayer wheel or revolving device i n their travel logs. The Tibetan wheels also present some problems: they have an unclear origin, and all dated ones are much later than the Chinese repositories. 100 Ibid., 158. I thank Professors Richard von Glahn and Katsuya Hirano for their helpful comme nts on the possible connections between Chinese repositories, Indian rituals of circumambulation, and Japanese Buddhist practices of nenbutsu念仏, which they made based on a conference presentation of the earliest draft of this chapter.	di-luo-a-grain-of-sand-full-draft-3.pdf
90 rotate constantly. 101 The “ cakra ” in the “ cakravala ” means “wheel;” this image and the momentum of rotati on are further strengthened by the seven concentric rings of mountains encircling the Sumeru. 102 What corresponds to the wheels of celestial bodies and terrestrial features is an even more intangible kind of wheel--the wheel of Buddhist law, or dharmacakra (Ch. falun 法輪 ), which, as eternal and constant as the movement of the stars, represents the timeless, inexhaustible, and invincible Buddhist teaching. Turning the wheel hence becomes a rhetoric of expounding, preaching, and continuing Buddhism, which has been consid ered the greatest accomplishment of the cosmic Buddha. 103 To turn the sutra case, therefore, was to mimic and reenact the wheeling of dharma and of the heavenly orders. The fascination with the wheel runs deep in human history. Wheelwrights of the Warring States China were already equipped with a highly developed set of theory about wheel-making. 104 Perhaps first applied to chariots, the wheel soon became indispensable for the production of pottery, watermills, and time-keeping devices. Most curiously, wheels m ight have been incorporated into architecture as early as the Western Han dynasty. According to the Liexianzhuan 列仙傳 (Hagiography of celestial beings, attributed to Liu Xiang 劉向, 77-6 BCE), a certain carpenter called Lupigong 鹿皮公 once built a “suspended revo lving pavilion ( zhuanlun xuange 轉輪懸閣 )” in the 101 Kloetzli 1983, 43-45. This is one of the four Buddhist cosmologies in Kloetzli's exposition. It is the simplest one and also the “module” for multiple-world sys tems. The notion of the universe revolving around a cosmic axle is found in the Abhidharmakosa as well as in the writings of Plato and Aristotle. The Abhidharmakosa further compares the trajectory of the planets to a great wheel across the night sky, comme nting that “the stars turn about [Sumeru] as though caught in a whirlpool. ” 102 Ibid., 46. To Kloetzli, the seven rings of mountains in fact represent “the 'planets' of Antiquity, i. e., Sun, Moon, Mercury, Venus, Mars, Jupiter and Saturn. ” 103 Ibid., 49. 104 Wen Renjun聞人軍, Kaogongji yizhu 考工記譯註 (The Kaogongji interpreted and annotated) (Shanghai: Shanghai guji chubanshe, 2008), 17-28.	di-luo-a-grain-of-sand-full-draft-3.pdf
91 mountain alongside a divine spring, the water of which extended his lifespan for several hundred years. 105 A more famous example is Nero's rotating banquet hall (d. 64-68 CE) which used hydraulic power as the mai n drive; it was originally thought as a legend but has recently been excavated by archaeologists in the ruins of Domus Aurea. A comparable structure is Sui Yangdi's 隋煬帝 (r. 604-618) “palace on wheels”--the Guanfeng xingdian 觀風行殿--which was said to be a spec tacular, moving piece of architecture able to accommodate several hundred imperial guards. 106 Su Song's 蘇頌 (1020-1101 ) astronomical clock presents an extraordinary case where the wheel and the miniature wooden pagoda were combined. Formally similar to the rev olving sutra case yet functionally distinct, Su Song's pagoda was embedded in a large clock-tower (fig. 43). It contained eight interconnected wheels inside, which were driven by the dripping water, and on each level of the five-storied pagoda was one or three open doors through which the jacks (human figurines) would appear to report the hours (fig. 44). 107 It is not clear if the pagoda was octagonal, but unlike the sutra case which could be turned, it was a fixed architectural facade with a much more complex rotating core. 105 http://ctext. org/lie-xian-zhuan/lu-pi-gong/zh. It is also possible that the zhuanlun xuange be interpreted as “a suspended pavilion on wheels” like Gongshu Ban's cloud-ladders. 106 Tanaka Tan 田中淡, Chugoku kenchikushi no kenkyu 中國建築史の研究 (Research on Chinese architectural history) (Tokyo: Kobundo, 1989 ), 214-16. 107 Xin yixia ng fayao 新儀象法要, by Su Song 蘇頌 (1020-1101), 3. 2a, http://ctext. org/library. pl?if=gb&file=86840&page=89. The structure and mechanism of this clock has received meticulous examination in Joseph Needham, Ling Wa ng, and Derek J. De Solla Price, Heavenly Clockwork: The Great Astronomical Clocks of Medieval China (Cambridge: Cambridge University Press, 2008 ), 28-45. Also see Forte 1988; Wu Hung, “Monumentality of Time,” in Monuments and Memory: M ade and Unmade, ed. Robert S. Nelson and Margaret Olin (Chicago: University of Chicago Press, 2003 ), 107-32.	di-luo-a-grain-of-sand-full-draft-3.pdf
92 Conclusion: the Revolving World in a Nutshell The Longxingsi sutra case is a typical example of miniature architecture. An analysis of its scale and formal features reveals that strict d ownscaling procedures have been followed. When it comes to dating, relying on textual evidence and our established knowledge about full-scale structures could be misleading, whereas identifying the woodwork as a miniature helps to clarify the issue. In fac t, a comparison with the jingzang prescribed in the Yingzao fashi suggests that the sutra case is a Northern Song remain. The high standardization of miniature-making in the Northern Song not only produced these excellent woodwor ks, but also commenced a trend of “progressive miniaturization” in Chinese architecture, foreshadowing the gradual diminution of scale and degeneration of structural members in the wooden buildings of later dynasties. According to legend, the revolving sut ra case was invented in the sixth century so that the illiterate could, by constantly rotating the bookcase, accumulate the same merit as those reading and studying the scriptures. In Tang monasteries, the sutra case was involved in the all-important ritua l of ordination; it was looked upon as an expedient means to enlightenment, through which the preservation of dharma, social harmony, and personal salvation could be achieved. The surviving sutra cases from the Song, Ming, and Qing dynasties were each a sp ectacular assemblage of profuse Buddhist icons and motifs encapsulated in a miniature pagoda. Such extravagance attracted many followers to come to turn the wheel seeking personal and familial blessings, while their donations also turned out a sizable part of the monastic income. The charm of the sutra case has a long-lasting effect in history; today, it has been remembered and recreated as a cultural icon of Chinese art and civilization. A deconstructive interpretation of the Longxingsi sutra case has perm itted us to excavate more about its formal and functional qualities as a miniature. Even though a resemblance to real architecture was essential, the miniature was free to disregard structural integrity and durability	di-luo-a-grain-of-sand-full-draft-3.pdf
93 because it was no longer supposed to f unction as a real building. This detachment of form from the assumed function necessarily entailed a dissolution of “meanings” and modes of signification; even though the sutra case is often described by scholars as modeled on an octagonal wooden pagoda, t he miniature form has in fact greatly problematized any idea of such a direct imitation. One could, alternatively, deconstruct the sutra case into three distinctive archetypes--the octagon, the central pillar, and the wheel--each of which carrying an array of religious and cultural significances and technological implications that were not limited to Buddhist ideologies. The coming together of the three archetypes, therefore, created a largely self-referential entity which refused any definitive, unequivoca l reading.	di-luo-a-grain-of-sand-full-draft-3.pdf
94 3. Miniatures as Sacred Repositories, Part II: The Huayansi Sutra Cabinets The set of sutra cabinets at the Huayansi, while also serving as a receptacle for the Buddhist Tripitaka like the Longxingsi sutra case, displays several distinctive forma l attributes. It does not occupy the center of the library hall but is attached to the walls, forming a U-shape periphery. The cabinets cannot be turned, which means that the rituals associated with the turning of the wheel (and dharma) discussed in the pr evious chapter are not applicable in this case. The elongated facade features a more varied combination of miniature wooden architecture and architectural elements. All these suggest that in comparison with its Longxingsi counterpart, the Huayansi library has implemented a fairly different spatial program, wherein the sutra cabinets generate a different set of meanings and symbolisms. To address these issues, this chapter starts with an analysis of scale. This not only helps to reveal the significant connec tions between the Longxingsi and Huayansi miniatures in terms of design, woodworking technology, and cultural interactions, but also transforms our narratives of Chinese architectural history. Additionally, it problematizes the notion of “Liao architecture ” as a derivative of or antithesis to what has been considered “classical” and “canonical” Chinese architecture. Contextualizing the sutra cabinets within the discourse of architectural history reveal only partially their distinct nature. One has to probe into the history of miniature cabinets and take into account other types of architectural-shaped shrines and repositories to comprehend their functions and range of applications in the society, especially in a religious setting. To this end, a reference to traditional furniture, especially cabinetry, as well as to similar examples found in Japan helps to deepen and broaden our understanding.	di-luo-a-grain-of-sand-full-draft-3.pdf
95 The intellectual interest of this chapter lies in the kind of visual experience brought forth by miniature architectu re. I argue that in the case of the Huayansi sutra cabinets, mini halls and pavilions have come from a system of highly developed visual trope which stresses multiplicity and multiplication. The vision of the miniature and the myriad pertains to the Flower Repository Universe described in Huayan Buddhism, and it is the acquirement of such a vision that is essential to enlightenment. The creation of miniature architecture, it seems, was meant to evoke a vivid imagination of this heavenly world of the myriad. The Bizang (Wall Repository) at the Huayansi One's first impressions of the Huayansi sutra cabinets include awe, admiration, wonder, but also bewilderment and perhaps a certain sense of displacement. An ironic feature of this exquisite woodwork is that i t can never be viewed as a whole--the elevations drawn by Liang Sicheng 梁思成 and Liu Dunzhen 劉敦楨 in their 1930s survey of the monastery, therefore, can be perceived on paper only, but the “big picture” forever eludes the eye since it is obstructed by the la rge group of statues in the middle of the library hall (fig s. 45, 46). 1 A visitor can view the cabinets section after section by strolling along the nine-feet-wide ambulatory between the central altar an d the walls (fig. 47). 2 The extremely dim light in the 1 Liang Sicheng 梁思成 and Liu Dunzhen 劉敦楨, “Datong gujianzhu diaocha baogao 大同古建築調查報告 (Survey report on the ancient architecture in Datong),” Zhongguo yingzao xueshe huikan 4 (1933. 3/4): 1-168, reprint in Zhongguo gujianzhu diaocha baogao 中國古建築調查報告 (Survey reports on ancient Chinese architecture), by Liang Sicheng, vol. 1 (Beijing: Sanlian shudian, 2012), 204-400. See below for an evaluation of this source. 2 The ambulatory is bordered by additional wooden frames surrounding the central statues, which seem to be much simpler shrines for smaller statues probably added in a later period ; these are shown in the photographs in Sekino Tadashi関野貞 and Takeshima Takuichi 竹島卓一, Ryo-Kin jidai no kenchiku to sono Butsuzo 遼金時代の建築と其仏像 (Liao-Jin architecture and Buddhist sculpture), 2 vols. (Tokyo: Toho bunka gakuin Tokyo Kenkyujo, 1934 ). According to Mr. Zhang 張, Head of the Office of Conservation of Cultural Relics at the Huayansi, these additions were from the Qing and used for hanging drapes and banners. Since at least 2003, during my first trip to the monastery, “strolling”	di-luo-a-grain-of-sand-full-draft-3.pdf
96 ambulatory dramatizes this “paralyzed” viewing experience, yet it works perfectly well for creating an atmosphere of an innermost shrine which is intimate, dar k, obscure, and sacred (fig. 48). 3 Entering the tunnel-like ambulatory, one passes a series of wooden cabinets shaped in miniature architecture where Buddhist scriptures and images are kept; on the other side is a waist-high altar where thirty-odd statues are organized around three Tathagata Buddhas. This “chancel-and-ambulatory” formula of the interior encourages circumambulation, a ritual somewhat reminiscent of the wheeling mechanism of the revolving sutra case. The two-storied miniature archit ecture has been built using the most majestic form and style possible, representing the highest building standards and the most complex techniques which normally would be permitted for imperial projects only. The imperial solemnity and augustness, nonethel ess, has been successfully conveyed through miniature-making. By twisting the size of real structures, the miniaturists worked on an otherworldly scale and was emancipated from restrictions in real practice. This allowed them to materialize their vision of the transcendental realm and share it with the audience, but their creativity and imagination were never entirely unbridled. Only by mastering the art of mimesis, the principles of scaling, and by carefully blending the familiar with the unfamiliar could such an extraordinary miniature world be created. along the ambulatory has not been an option available to tourists. Now only the two corner sections of the repository next to the entrance of the hall remain in sight. 3 The Buddhist images originally placed on the second level of the repository, too, would have had only minimal exposure to the sight of the viewers.	di-luo-a-grain-of-sand-full-draft-3.pdf
97 Scale and form The sutra cabinets have adopted a cai measuring 4 by 3 centimeters, a value fairly close to the 4. 5 by 3 centimeters at the Longxingsi. 4 Regarding the overall scaling scheme, the observation s based on the Longxingsi sutra case are largely applicable here, especially in terms of the brackets. The consistency between the two miniature woodworks can be found not only in how individual bracket sets are downscaled, but also in the controlled way t hey are positioned in relation to each other. According to the Yingzao fashi, the distance between two adjacent bracket-sets ( puzuo zhongju 鋪作中距, measured from center to center) of the jingzang is regulated to be 100 fen, or ten times the thickness of cai. 5 This standard number is used to control the breadth of the bay, hen ce the overall size of the entire woodwork. 6 In our case, the spacing of the bracket-sets of the Huayansi cabinets is approximately 36 centimeter, equivalent to 120 fen (fig. 49). 7 This number, though deviating from the theoretica l value, is nonetheless comparable to the 32. 57 centimeter (108. 6 fen) observed in the 4 My own measu rement in the summer of 2014, however, suggests an average of over 4. 2 by 3 centimeters as cai. But since my measurement was not systematic, here I keep to Liang and Liu's results. The cai adopted by both the Longxingsi and Huayansi examples are slightly l arger than the standard dimension given in the Yingzao fashi. 5 The 100-fen module has been dictated (though not explicitly) in the Yingzao fashi for another type of miniature--the fodaozhang. This consistency reconfirms that zang and zhang need be examined as belonging to the same group and are interchangeable forms, a point attempted by this and the preceding chapter. See Chen Tao, “ Yingzao fashi xiaomuzuo zhangzang zhidu fanying de moshu sheji fangfa chutan 營造法式小木作帳藏制度反映的模數設計方法初探 (A preliminary research on the modular design of the making of miniature shrines and repositories in the Yingzao fashi ),” Jianzhu shilun huikan 4 (2011): 244-51. The 100-fen module regulates not only the spacing of the bracket-sets, but also that of th e lotus-petal ornaments ( furongban 芙蓉瓣 ), the “tortoise legs ( guijiao 龜腳 ),” and certain elements of the tiangong louge. As Chen notices, the 100-fen module is not always kept; sometimes it has to be adjusted within a certain range to better suit different d esign s. 6 Scholars have been debating whether or not the same principle was applied by contemporary buildings, as no set formulas are offered in the Yingzao fashi. Chen Mingda 陳明達 argues that the breadth of the bay was a set valu e of 125 fen; while some scholars remain suspicious of the application of strict proportioning rules. In the Qing, however, the breadth of the bay was regulated to be 110 fen, though this fen was defined differently. Further explorations of miniature archi tecture should shed some ligh t on this issue. See Chen 2010, 239-40, 252. 7 This and the next numbers were measured not from the real repositories but from the digital models I developed according to my survey and Liang and Liu's drawings. For the Huayansi cabinets, the spacing of the pingzuo brackets along the south-wall is increased to 160 fen (48 centimeters). The result of my survey has been digitized and accessible online at https://sites. google. com/site/sourcesofchinesearchitecture/shanxi/huayansi.	di-luo-a-grain-of-sand-full-draft-3.pdf
98 Longxingsi sutra case. 8 Indeed, the similarity in the value of cai and certain other dimensions suggests an intriguing affinity of the two examples in both chronological and technological senses. Such a numerical consistency, however, is hard to perceive by the eye, since the Huayansi cabinets appear so fundamentally different in form. Here the miniature architecture constitutes a border, not a center: it is a long, conti nuous series of buildings instead of a self-enclosed structure, providing a linear viewing experience as in the unrolling of a Chinese landscape painting. What we see is not one, but several miniature pavilions, some as corner towers and some accentuated b y an elevated center flanked by two wings, all connected by long galleries and an arching bridge over the opening on the back wall (fig. 50). This configuration might have been derived from the contemporary monastic design; an alt ernative source of inspiration could have been the streetscape of the Lia o Western Capital (modern-day Datong 大同 ) where the Huayansi was located. Interestingly, in recent years, the city of Datong opened up a plaza named the “Huayan Guangchang 華嚴廣場 ” in front of the monastery as a new tourist attraction and service center, the s urrounding ancient-style buildings of which are claimed to have been modeled on none other t han the sutra cabinets (fig. 51). 9 Evidently, the sutra cabinets evoke the experience of walking into a meaningful “place” rather than of standing and admiring a single building. 8 The spacing of the pingzuo brackets in this case is slightly smaller: 31. 76 centimeter (105. 9 fen). For large-scale buildings in the Song-Liao-Jin period, the spaci ng of the bracket sets ranged between 110-150 fen, with an average of 125 fen. Similarly, this number was used to control the breadth of the building. See Fu Xinian 傅熹年, “Zhongguo gudai mugou jianzhu sheji 中国古代木构建筑设计 (Design of ancient Chinese wooden architecture), ” Sheji yu yanjiu (2016. 39), http://mp. weixin. qq. com/s?__biz=Mz A5OTgw NDAz MA==&mid=402361954&idx=1&sn=71736814947c62f2430fd 0239b6c8a02&scene=2&srcid=0229Wf1q AKQQ1da Jo QRYod N7#rd. 9 The buildings on the plaza are not exact copies of the cabinets ; considerable redesigns had to be done to make them fully functional.	di-luo-a-grain-of-sand-full-draft-3.pdf
99 Discovery, dating, and identification The first comprehensive survey of the architectural characteristics of the Huayansi sutra cabinets was carried out by Liang Sicheng and Liu Dunzhen and published in their co-authored field report in 1933. 10 The report recounts a brief history of the library hall by piecing together evidence from local gazetteers, steles, and inscriptions on the wooden structure. Though the official history, Liao shi 遼史, records that the monastery was founded in 1062,11 the hall had been built twenty-four years earlier, in 1038 (the seventh year of the Chongxi 重熙 period) under Emperor Xingzong 興宗 (r. 1016-1055), according to the gazetteers. 12 The year 1038 is further verified by an inscription undern eath a beam of the hall, proving that the wooden frame still stands as an early eleventh-century original. 13 The earliest stele on site, “Record of Repairing the Bojia jiaozang of the Grand Huayansi in the Western Capital of the Great Jin Empire ” (大金國西京大華嚴寺重修薄伽教藏記, d. 1162), 10 See n. 1. The main points of the report regarding the cabinets have been summariz ed in Bai Yong 白勇, “Datong Huayansi Bojia jiaozang dian jianzhu fengge luelun 大同华严寺薄伽教藏殿建筑风格略论 (A brief discussion on the architectural style of the Bojia jiaozang Hall at the Huayansi in Datong),” Wenwu shijie 3 (2011): 15-18. 11 Liao shi, 41. 2b, http://ctext. org/wiki. pl?if=gb&chapter=112993#p3. “In the eighth year of the Qingning Period, the Huayansi was built, where stone and bronze statues of the past emperors were venerably placed 清寧八年建華嚴寺，奉安諸帝石像，銅像. ” 12 See Liang and Liu 1933, 207-08. One of their major source s is the Datong xianzhi 大同縣志, juan 5, which states that “the Bojia jiaozang on the southeast corner of the monastery... was built in the seventh year of the Liao Chongxi Period 寺之東南薄伽教藏... 遼重熙七年建. ” According to the Shanxi tongzhi 山西通志, 169. 44b-45a, http://ctext. org/wiki. pl?if=gb&chapter=976962&remap=gb#p221, “In the seventh year of the Jin Chongxi Period, the Bojia jiaozang was built to the southeast of the main hall 金重熈七年, 建薄伽教藏於殿東南. ” But there were no known “Chongxi Periods” during the Jin. The library hall might had originally been the main structure of a smaller local monastery and was later appended to a larger monastery under the imperial patronage. 13 Liang and Liu 1933, 207-08, 232-33. Two inscriptions are found written in ink on the underside of two beams to the left and right of the central bay. The one on the left reads “ 推诚竭节功臣，大同军节度，云、弘、德等州观察处置等使，荣禄大夫，检讨太尉，同政事门下平章事，使持节云州诸军事，行云州刺史，上柱国，弘农郡开国公，食邑肆仟户，食实封肆佰户，杨又立 ” (杨又玄 in Bai 2011 ). The one on the right concerns the date of construction: “ 維重熙七年歲次戊寅，玖月甲午朔十五日戊申午時建 ” (Bai 2011 missing the second “ 午”), proving “Chongxi year seven, ninth month, fifteenth day ” to be the date of construction.	di-luo-a-grain-of-sand-full-draft-3.pdf
100 claims that the monastery received heavy damage during the war at the end of the Liao and that the hall, formerly containing 579 volumes of the Khitan Tripitaka, received several restorations together with other buildings during 1140-1162. 14 However, to Liang and Liu as well as most other scholars, the 1038 inscription remains the most reliable evidence that the hall and its sutra cabinets can be confidently identified and discussed as Liao structures, even though the monastery underwent significant changes in the Yuan, Ming, and modern times. 15 The hall is named “Bojia jiaozang 薄伽教藏,” bojia being a transliteration of bhagavat in Sanskrit, meaning “the blessed one” or “world-honored one” and often interpreted as shizun 世尊 in Chinese. 16 The meaning of the second half of the name, “jiaozang,” however, is less straigtforward. While zang denotes a repository of the Buddhist Tripitaka, jiao indicates “teaching,” “religion,” and “instruction;” hence “Bojia jiaozang” could be tentatively rendered as “the repository preserving the teachings of the World-honored One. ”17 The sinograph zang, as ambivalent as its English equivalent 14 Liang and Liu 1933, 209 ; Bai 2011, 15. According to Bai, the stele states that the Khitan Tripitaka “was collated and scrutinized during the Chongxi Period of the Liao, and according to convention compiled into five hund red and seventy-nine volumes... The Grand Huayansi of this day has also possessed [a copy of] the said Tripitaka since the past 及遼重熙間，復加校證，通制為五百七十九帙... 今此大華嚴寺，從昔以來亦有是教典矣. ” Today the cabinets hold some 180,000 fascicles of Ming and Qing Buddhist scriptures. See Xie Yubao 解玉保, “Datong Huayansi Bojia jiaozang dian de Liaosu ji jingchu 大同华严寺薄伽教藏殿的辽塑及经橱 (Liao sculpture and sutra cabinets in the Bojia jiaozang at the Huayansi in Datong),” Journal of Shanxi Datong University (Social Science) 23 (2009. 4): 36. 15 Liang a nd Liu 1933, 210-11, 232-33. In terms of style, the hall displays great similarities with the Sandashi dian 三大士殿 at the Guangjisi 廣濟寺, an additional piece of evidence Liang and Liu draw to refute Ito Chuta's judgment that the Bojia jiaozang is a Jin struct ure. Information on major repairs and reconfigurations of the monastery in later periods is given on a Yuan stele, “ Xijing Da Huayansi fori yuanzhao minggong heshang bei 西京大華嚴寺佛日圓照明公和尚碑銘 (d. 1350), ” and the quoted gazetteers. In the Ming, the monastery was split into two, forming an upper and a lower precinct, and the main hall was once converted to a warehouse during 1370-1391, in the Ming Hongwu Period ( Shanxi tongzhi, 169. 45a). The Lower and Upper Huayansi have only recently been reunited as one monaster y by the municipal government of Datong in 2009. 16 A. Charles Muller, Digital Dictionary of Buddhism, http://buddhism-dict. net/ddb/index. html. 17 The use of the name “ jiaozang ” appears in, for instanc e, Ye Shi 葉適 (1150-1223), “Preface to the Buddhist Repository in the Famingsi 法明寺教藏序,” in Shuixinji 水心集, 12. 28a-b. Here it seems to denote the collected works on Buddhism, not necessarily a repository or building.	di-luo-a-grain-of-sand-full-draft-3.pdf
101 “repository,” could mean either a building or a receptacle for the purpose of storage, or the two combined. This is to say that the architectural-shaped wooden cabinets inside the hall might have been indiscriminately referred to as the “Bojia jiaozang” in history, before they became termed by modern scholars as a type of bizang based on the resemblance and contemporaneousness to the template found in the Yingzao fashi, which allows researchers to distinguish small-scale repositories from larger ones. Liang and Liu have identified the miniature five-bay pavilion on top of the arching bridge (fig. 52) as a tiangong louge, reasoning that “ the tiangong louge in the Yingzao fashi... are invariably set upon skirting eaves and mezzanines; this structure [of the pavilion over the bridge] is suspended in the air, which properly conforms to what tiangong louge means. ”18 While this judgment is based on the specific location of the structure--that it has to be from “above”--and their claim is seemingly supported by the illustrations of the “ tiangong louge fodaozhang 天宮樓閣佛道帳 (Buddhist/Daoist shrine with the Heavenly Palace motif)” (see fig. 1) and the “ tiangong bizang 天宮壁藏 (Wall repository with the Heavenly Palace motif) ” (see fig. 2) in the Yingzao fashi, it is in fact a misinterpretation. The correc t way of identifying the tiangong louge is by its scale. While the sutra case is itself miniaturized already, a tiangong louge to decorate it has to adopt an even smaller scale. 19 This makes the tiangong louge a miniature in the miniature world. The pavilio n-and-bridge complex in 18 Liang and Liu 1933, 221. “法式天宮樓閣... 皆設於腰檐平坐之上, 此則臨空結構, 適符天宮樓閣之意義. ” This interpretation is reiterated in Bai 2011. 19 The Yingzao fashi dictates three types of woodwork where this motif should be used: fodaozhang, zhuanlun jingzang, and bizang, all falling into the categ ory of miniature woodwork s. The cai of the tiangong louge needs to be reduced to 0. 6 by 0. 4 cun for shrines and further to 0. 5 by 0. 33 cun for repositories, which is only one-third and a half, respectively, of the regular cai applied to the main body of th e miniatures. See Yingzao fashi, vol. 1, 199, vol. 2, 6, 26. A list of the different values of cai in the Yingzao fashi is given in Chapter 2 of this dissertation. The discussion on tiangong louge and its role in religion and material culture will conti nue in Chapter 4.	di-luo-a-grain-of-sand-full-draft-3.pdf
102 question, however, shares the same cai with the rest of the sutra cabinets, a fact that disqualifies it as a tiangong louge defined in the Yingzao fashi. 20 The cai of the sutra cabinets measures 4 by 3 centimeters and is applied throughout the entire structure. 21 It is perhaps this extraordinary consistency in scale that prompted Liang and Liu to regard the cabinets as “the most appropriate model” wh ich “fully displays the truth characters of Liao architecture. ”22 Clearly it was within the authors' conscious efforts to incorporate this particular miniature into the narratives of Chinese architecture; the structural and stylistic correspondences between the miniature and the full-scale, wherever possible, are detailed and emphasized. To this end, they meticulously measured and recorded the dimensions of all major components of the cabinets, from beams and brackets to the span of the roof and the incremen t of the eaves-line in height, deducing that the miniature must have been based on the same design scheme, especially in terms of proportioning, of the pavilions and halls of other Liao monasteries such as the Dulesi 獨樂寺 (d. 984). 23 The authors also conscio usly juxtapose Liao, Song, and Jin architecture in their exposition of how architecture “evolved” in temporal and geographical terms. The fact that the height-width ratio 20 Similarly, it might be problematic to call the bridge a “yuanqiaozi 圜橋子 ” as Liang and Liu do on p. 231, since the term in the Yingzao fashi denotes a flight of stairs, not a fully suspended walkway. 21 Liang an d Liu 1933, 223. My own measurement indicates slightly different values; see n. 4. 22 Ibid., 221. “ 壁藏與天宮樓閣之結構, 系模仿木造建築, 故可視為遼式建築最適當之模型... 較薄伽教藏殿本身, 及同時諸建築屢經修葺者, 尤足表示遼式建築之真狀. ” The authors believe that the importance of the sutra cabinets is comp arable to the Tamamushi Shrine in Japan. 23 Ibid., 223-24. The report lists three features that demonstrate the sutra cabinets to be a Liao woodwork: 1) the equivalence of the guazigong 瓜子栱 and linggong 令栱 in length; 2) the elongation of th e nidaogong 泥道栱 and mangong 慢栱; and 3) the span of the second and above steps ( tiao 跳) being shorter than that of the first step. These features also appear on the Bojia jiaozang. T he measurement and comparison with other Liao structures are on pp. 224-32. My observation and modeling, however, suggest that proportionally, it is the guazigong (62 fen long in the Yingzao fashi ) that has been shortened, while the lengths of the linggong and nidaogong mostly stay the same (72 and 62 fen, respectively).	di-luo-a-grain-of-sand-full-draft-3.pdf
103 of the cai for both Liao and Song architecture was generally kept at 3:2--in agreemen t with the Yingzao fashi--suggest that the two regimes derived their building methods from a common source, that is, the Tang; whereas the variances shown in the dimension of the zhi 栔 (the secondary standard unit of Chinese architecture, proportional to cai) bespeak the t rend of downsizing timber units following the decline of the Tang. 24 This “biological” model from nativity to maturity and decline was typical of the first generation of Chinese architectural historians pioneered by Liang and Liu in their understanding of t raditional buildings. Parallel to this biologism was a sense of ethnocentrism stimulated by the nationwide patriotism in the chaotic years after the downfall of the last dynastic rule in China: though wholeheartedly praising the virtuosity of Liao archite cture, the authors assert that the Khitans had a low level of civilization and thus could not have exercised any substantial influence on the scaling techniques inherited from the Tang. 25 The Liao, therefore, was seen as a passive receiver, not an innovator ; and Liao architecture was interpreted as mainly inherited from the great tradition without its own regional or ethnographic characteristics. 26 In comparison, the Song, because of its cultural sophistication, represented a more subtle line of inheritance, whereas the Jin, blindedly absorbing elements from both the Liao and the Song, displayed complex and sometimes self-contradictory features of architecture. 27 24 Ibid., 282-83. 25 Ibid., 287. 26 Ibid. Liang and Liu do point out that the use of the xiegong 斜栱 (brackets with diagonal members) was regionally distinct. But they assert that it was more likely a feature of the traditional Yan-Yun 燕雲 area, not a n ethnographic feature associated with the Khitans. 27 Ibid., 288, 292. One comparison given in the report focuses on the shuatou 耍頭 (also juetou 爵頭, lit. sparrow-head; being the topmost transversal member of a bracket-set, intersecting with the linggong ) of Liao architecture, which often features the simple scheme of vertical cut known as the pizhu'ang 批竹昂, lacking the more subtle curves and moulds featured in the Yingzao fashi. Chinese scholars sometimes categorize and d iscuss Liao and Jin architecture as one entity, that is, Liao-Jin architecture, as the latter half of Liang and Liu's report shows. Other than the factor that the two	di-luo-a-grain-of-sand-full-draft-3.pdf
104 The modern “rediscovery” of the Bojia jiaozang and its cabinets was made by a group of Japanese sch olars. 28 One of the key figures, Takeshima Takuichi 竹島卓一, did not conceal his utter frustration toward the fact that the Chinese were always faster in publishing reports, even though they were informed of the structures and their probable dating from the Japanese first. 29 In the early 1930s, surveying traditional architecture in the northern provinces of China was a matter of national contest, and the surveyors' judgment would, perhaps unconsciously, be swayed by ideals of nationality. It turns out that the all-impo rtant inscriptions under the beams of the Bojia jiaozang were first noted and transcribed by Takeshima on the exact date of the twenty-seventh of June, 1931, two years before Liang and Liu's report, which contains no credit to Takeshima for his finding at all. 30 When interpreting Liao structures, Takeshima has also arrived at several different conclusions. He argues that the Liao displayed a rather distinctive architectural style, which received regimes were closely related in terms of dynastic succession and territorial control, the y have been bonded together perhaps also for the reason that both are considered “alien” and “conquest” regimes representing the barbarian and later sinicized cultures, which once stood as the rivals of the Song. 28 Japanese expeditions in North China in th e early twentieth century have introduced many long-forgotten structures to scholars and the public. Besides Dulesi 獨樂寺, Huayansi, and Fengguosi 奉國寺, also rediscovered were the W ooden Pagoda at the Fogongsi 佛宮寺 and the Chongfusi 崇福寺. The se expeditions grea tly instigated the scholars of the Society for the Research in Chinese Architecture (Zhongguo yingzao xueshe 中國營造學社 ) to carry out investigations of their own, who have rediscovered the Guangjisi 廣濟寺 and Kaiyuansi 開元寺. See Takeshima Takuichi, Ryo-Kin jidai no kenchiku to sono Butsuzo 遼金時代の建築と其仏像 (Liao-Jin architecture and Buddhist sculpture) (Tokyo: Ryubun shokyoku, 1944), 12. 29 Takeshima 1944, i-vi. The preface is an especially detailed account of this matter. Though plans to publish their surveys had been made soon after Sekino and Takeshima visited the Dulesi and Huayansi in 1931 and the Fengguosi in 1932, and the plates of photographs were published first in two volumes in 1934 and 1935, the texts, which would constitute the third volume, was delayed by S ekino's untimely death, by other ongoing projects, and again by the breaking of the Sino-Japanese war in 1937. 30 Ibid., 76. Here Takeshima gives the exact date to claim his credit. Before his finding, the dating of the Bojia jiaozang was uncertain because official histories and local gazetteers conflict each other. Though no solid evidence confirms that the cabinets were built in the same year as the hall, they have been generally dated to 1038 according to the inscriptions found by Takeshima and the Jin st ele mentioned above. This dating proves essential, since the Fengguosi Main Hall, another representative Liao structure, was consequently dated to 1020 based on a comparison of the complexity of its brackets with that of the cabinets. See I bid., 83-84; 103, n. 1.	di-luo-a-grain-of-sand-full-draft-3.pdf
105 more influences from the Kingdom of Bohai, the Five Dynasties, and perhaps even some of its sixty tributary states in the west. 31 Redefining Liao architecture The evaluations by Chinese and Japanese scholars has had a clear impact on Western discourses. Alexander Soper, in his general history of Chinese architecture ( first published 1956), traces the Huayansi sutra cabinets to “the Sung ideal of crisp, well-organized richness. ”32 The woodwork “is so much like the mature Northern Sung style that it is tempting to imagine that the Liao architects made a special effort to reproduce some Chinese prototype, either by bringing in a consultant or by referring to some manual that preceded the Ying-tsao Fa Shih. ”33 Here the preconception of the superiority of “Chinese-ness” resumes, together with a mistrust of the ingenuity and in tellect of the “barbarians. ” What also stands out is the lack of attempt to distinguish miniatures from large-scale structures; instead, any boundaries between the two have been completely ignored. Clearly, it becomes problematic when structures are strict ly labeled by race or ethnicity and taken as pure “models” and “imitations” of certain prototypes originated from an allegedly more sophisticated culture. 34 It is hence not surprising that the cabinets and other masterpieces of Liao architecture are grouped under the chapter title “The Barbarian Empires: Liao, Chin, and Yuan,” which Soper uses to characterize an unexciting era when architecture “lapsed into stagnation” and builders “worked 31 Ibid., 4-6. 32 Sickman and Soper 1984, 455. Soper's notes tell us that he mainly relies on Takeshima's and Liang and Liu's field reports. 33 Ibid., 457. 34 Liao architecture, in miniature or full-scale, is generally treated as imitations of Song pro totypes in Soper's discussion.	di-luo-a-grain-of-sand-full-draft-3.pdf
106 under the handicap of depressed morale” whose alien patrons “can have contributed nothing except the ambition and the means to build with a naive boastfulness. ”35 Such prejudiced examination has received a major revision and critique in Nancy Steinhardt's Liao Architecture. Her goal is evident in the name of the book--to hig hlight architecture under the Liao as a corpus of distinctive material culture to be studied. Liao architecture certainly had some sources but could still possess a character of its own, not subsidiary to earlier or other contemporary cultures. 36 The author 's consciousness of ethnicity is still distinct, but it bears a less political undertone and engages mainly with aesthetic concerns. Based on an exhaustive examination of the fifteen major Liao timber structures, including the Huayansi cabinets, Steinhardt defines Liao architecture first and foremost by its unique “somber power” and visual impact. 37 The fact that many Liao wooden halls look intrinsically interesting, she points out, is due to the ingenuity of Liao carpenters and the details of their work: no t only do brackets display a much greater variety, but exquisite small-scale woodworks--including zaojing ceilings and miniature cabinets--exhibit truly remarkable designs. 38 It is the richness of the interior that makes a Liao hall exciting and unsurpassab le by any contemporary Song buildings; the latter usually feature a simpler and rigid structure that tends to dampen any dramatic atmosphere of a religious space. 39 35 Ibid., 439. 36 Steinhardt 1997, 22. “... many of those who conducted literary or archaeological research on Northeast Asia and the Qidan harbored an innate dislike for the Liao, for North Asian peoples, or for the country in which they were engaged in research. ” She explains that other factors of such negative sentiments include d political reasons, nationalism, and wars. Many Chinese scholars today still shun Japanese materials because of “the political scars of Japanese occu pation;” see 25-27. 37 Ibid., 184-85. For a discussion on the Huayansi cabinets, see ibid., 130-133. 38 Ibid., 185-86. Steinhardt contrasts the simplicity of the exterior and the fantastic interior of Liao wooden halls, which seems to suggest that Liao archi tecture should be recognized mainly for its small-scale woodwork, such as cabinetry. See Chapter 4 of this dissertation for more discussion on zaojing. 39 Ibid., 222-23. The comparison between the Dulesi main hall and the Cishige of the Longxingsi, however, is not entirely convincing, since the latter is a side structure of the monastery and supposedly should have been made simpler	di-luo-a-grain-of-sand-full-draft-3.pdf
107 Steinhardt further proposes to contextualize Liao architecture in North Asian traditions, demonstrating that while some forms could be traced back to the Tang capitals of Chang'an and Luoyang (such as the stone sarcophagi in the shape of miniature wooden buildings), certain other practices, most notably the “lantern ceilings” of the tomb chamber s, bear a striking resemblance to the ones found in the Korean Peninsula and the Mogao caves in China's west. 40 The art historical perspective While the Huayansi sutra cabinets have been continuously written into the history of Chinese architecture, Jeehee Hong's recent study shows just how it can be investigated from an art historical perspective. 41 Being perhaps the first to have devoted a paper-length writing to this exemplary woodwork, Hong effectively brings together wooden miniature-making and funerary art into discussion. As she points out, miniature representations can be traced back to the mingqi of the Warring States (ca. 475-221 BCE), a tradition carried on in the medieval period by small-scale furniture and utensils often found in tomb chambers an d Buddhist relic deposits. 42 Hong argues that all miniatures must achieve a “sanctioned absence of the link between their form and mechanism”-- than buildings on the central axis. Understandably, such a comparison might be the best one can find since the original Northern Song structure Foxiangge 佛香閣, the main building at the Longxingsi, had been lost and was reconstructed in modern times. 40 Ibid., 363-64, 374-75. 41 Jeehee Hong, “Crafting Boundaries of the Unseeable World: The Ontology of the Bhagavat Sutra Repository,” Art History (forthcoming 2016). 42 Ibid. Hong further observes that there existed a “reciprocal emulation” between tomb burials and Buddhist relic deposits; the use of multiple coffins and architectural-shaped containers, for instance, are largely shared prac tices.	di-luo-a-grain-of-sand-full-draft-3.pdf
108 the sutra cabinets, for instance, have assumed an architectural form but do not function as a shelter for people. 43 According to Hong, the sutra cabinets have two distinctive spatial qualities: its U-shape plan transforms the architectural facade into the surface of an interior, creating an “inverted space” similar to those found in contemporary tomb chambers. 44 Second, the cabinets look as if they are emerging out of the walls while the other half is left hidden behind, creating a liminal moment when the cabinets cross the boundary between this and the other world. The idea of concealing is closely associated with t he word “ repository ( 藏 zang, or cang when used as a verb),” and it might not be coincident that it is a homophone of zang 葬, burial. 45 As Hong observes, the tradition of building repositories in China started in Confucius's time. The Han official history recou nts how ancient manuscripts were found in the walls of Confucius's residence where they might have been hidden during the First Emperor's burning of the books. 46 Though historical records regarding wall repositories are scarce, Yelu Chucai's 耶律楚材 (1190-1244 ) “Record of the Establishment of the Sutra Repository in the Dajuechansi in Yanjing ” (燕京大覺禪寺創建經藏記 ), written in 1233, gives us a glimpse of how exceedingly extravagant such 43 Ibid. Hong argues for the “virtual functionality” of the miniature architecture (“the Tower Pavilion”) on the upper level of the repository, where Buddhist statues were enshrined, whereas the cabinets on the lower level are for “real” use. She und erstands the upper part to be a representation of the Buddhist “heavenly realm. ” 44 Ibid. Alternatively, the interior is comparable to a typical, self-enclosed courtyard layout in traditional Chinese architecture, which might be described as an introverted space. Such a layout has been widely adopted for palaces, Buddhist monasteries, and residences alike. 45 Ibid. 46 The wall repository is known as the “Kongshi bizang 孔氏壁藏,” which included the ancient version of the Shangshu 尚書 (Classic of documents).	di-luo-a-grain-of-sand-full-draft-3.pdf
109 repositories might have been at the time. 47 The text describes “a circuit of wall re pository, small shrines with truncated pyramidal canopies, and dragon niches, in a total number of twenty jia 架 (sections of cabinets?), all ornamented with gold and painted with various colors, which have exhausted the workers' skills and ingenuity to bri ng forth a splendid new look; the cost of the project amounted to a hundred ingots of white gold ” (壁藏斗帳龍龕一周, 凡二十架, 飾之以金, 繢之以彩, 窮工極巧, 煥然一新, 計所費之直白金百笏 ). 48 Though no further detail is given, the repository in the text appears somewhat similar to the Huayansi s utra cabinets, as both feature miniature shrines, niches, and a circulatory layout along the walls. Repositories, Shrines, Cabinets What has not been attempted so far is to examine the miniature repositories at the intersection between architecture and fu rniture especially cabinetry. The modular design of the Huayansi sutra cabinets has been noted and taken as the grounds for understanding it as an integral part of Chinese architecture, but the formulas and schemes it observes never perfectly tally with th ose of real buildings. 49 The gap between architecture and its miniature, and the internal coherence among the miniatures themselves in general (such as between the Longxingsi and the Huayansi repositories), alert us of extra influences from other domains of carpentry. The following addresses to this 47 Yelu Chucai, Zhanran jushi wenji 湛然居士集, 8. 28b-30b, http://ctext. org/library. pl?if=gb&file=5 2170&page=58&remap=gb. 48 Ibid., 8. 29b. 49 The most direct evidence for this comes from the Yingzao fashi. While certain parts of miniature woodworks use cai; others are based on a set of different proportioning rules. For instan ce, the width and thickness of the miniature columns are proportional to the height of the cabinet, unlike those of full-scale columns, which are controlled by cai and zhi. This new proportioning rule has been summarized thus: “The width and thickness of e ach component are proportioned to the height of the corresponding level 其名件廣厚，皆取逐層每尺之高，積而為法 (Yingzao fashi, vol. 2, 16 ). ”	di-luo-a-grain-of-sand-full-draft-3.pdf
110 problem by exploring closely-related small-scale woodworks found in different settings: in worship halls, monastic living quarters, and houses. Bringing together these examples will illuminate how wooden reposito ries, shrines, and cabinets were structurally and aesthetically interrelated and, under some circumstances, largely interchangeable. In worship halls Early images of the miniature repositories have been preserved in many Buddhist cave temples. The earlies t wooden evidence comes from the Binglingsi 炳靈寺 Cave 172 in Gansu, where a box-like shrine (ca. sixth century) contains three sets of Buddhist triad statues (fig. 53). The “box” measures 2. 57 by 2. 60 meters in plan and 2. 18 meters in height; even though the structure appear s rather simple, it has comprised clearly identifiable architectural features including four octagonal-based columns and a coffered ceiling. 50 Similar forms of miniature architecture must have been used in Buddhist monasteries prior to the sixth century. Th e central pillar of Yungang Cave 6 (ca. 471-499), for instance, showcases an impressive, two-storied structure with niches of Buddhist icon on all four sides (fig. 54). The lower level is supported by four dharani pillars and the upper level by four nine-storied, miniature pagodas; most interestingly, there is also a layer of densely arrayed “rafters” and “roof tiles” between the two levels, as if to suggest that the structure as a whole is modeled after a wooden building. 51 In the eleventh to thirteenth centuries, miniature repositories and shrines expressed much more prominent architectural features. In addition to the Longxingsi and Huayansi examples, there 50 Wang Long 王瀧, “Xin faxian de Beichao mugou jianzhu: Binglingsi shiku 172 ku fozhang 新發現的北朝木構建築 : 炳靈寺石窟 172窟佛帳 (A newly discovered Northern Dynasty wooden structure: the Buddhist canopy shrine in the Binglingsi Cave 172),” Meishu yanjiu (1979. 3): 72-77. 51 Tokiwa Daijo and Sekino Tadashi, Shina bunka shiseki 支那文化史蹟 (Historical remains of Chinese culture), vol. 1 (Tokyo: Hozokan, 1939 ), 24-28.	di-luo-a-grain-of-sand-full-draft-3.pdf
111 are many contemporaneous remains to illustrate this point. The Daoist shri ne at the Erxianmiao 二仙廟 (Temple of the two goddesses, d. 1097) in Jincheng 金城, Shanxi, for instance, displays some notable resemblances to the Huayansi cabinets (fig. 55). Its main body is a three-bay miniature hall; at the front are two double-story gate-towers standing on the left and right, the upper levels of which are connected by an arching bridge with a small pavilion on top. 52 The entire miniature is adorned with well-articulated bracket sets, the mezzanines, and the flying eaves. 53 The accentuation of structural features is also found in slightly later miniature shrines such as the one at the Yuhuangmiao 玉皇廟 in Jincheng 晉城, Shanxi. Erected on the central altar of the Daoist worship hall is a miniature double-story pavilion exhibiting rigorous applications of Northern Song architectural style and scaling schemes (fig. 56). 54 In mon astic living quarters Some visual evidence fairly accurately illustrates the type of miniature shrines and repositories used in a semi-public setting--the monks' living quarters. The Wushan shichatu 五山十剎圖 (Drawings of 52 Nancy S. Steinhardt, “A Jin Hall at Jingtusi: Architecture in Search of Identity,” Ars Orientalis 33 (2003 ): 87. See also Zhongguo wenwu dituji: Shanxi fence 中國文物地圖集 : 山西分冊 (Atlas of Chinese cultural relics: Shanxi), vol. 1 (Beijing : Zhongguo ditu chubanshe, 2006), 449. 53 Other known examples during this period are found at the Chongqingsi 崇庆寺 (d. 1016) in Zhangzi 長子, Jindongsi 金洞寺 (d. Northern Song) in Xinzhou 忻州, Zetianmiao 則天廟 (d. 1145) in Wenshui 文水, and Taiyinsi太陰寺 (d. 1170) in Jiangxian 绛县, all in Shanxi. 54 The worship hall is dated to the Yuan ( ca. 1335 ), but the form and scale of the miniature shrine (the number of bracket sets, their spacing, etc. ) suggest a Northern Song or Jin remain. The cai, 2. 25 by 1. 5 centimers, is only half of that of the Huayansi cabinets, a fact which implies a date later than the eleventh century. See Yin Zhenxing 尹振興, “Jincheng Yuhuangmiao Chengtangdian muzhi shenkan xingzhi jianjie 晉城玉皇廟成湯殿木質神龕形制簡介 (A brief introduction to the wooden shrine in the Chengtangdian at the Yuhuangmiao in Jincheng),” Wenwu shijie (2014. 2): 31-34; Zhongguo wenwu dituji: Shanxi fence, vol. 1, 466-67. Other Yuan and later miniature cabinets and shrines are found at the Upper Guangshengsi 廣勝寺 (d. Yuan) in Hongtong 洪洞, Shanxi; Qianfo'an 千佛庵 (d. 1629?) in Xixian 隰县, Shanxi; Confucius Temple in Qufu, Shandong, and the Niujie 牛街 Mosque in Beijing.	di-luo-a-grain-of-sand-full-draft-3.pdf
112 the Five Buddhist Mountains and Ten Mo nasteries), dated 1248 and sketched by Japanese pilgrim monks to Southern Song China, provides reliable information on how these miniatures looked at the time. 55 One example, labeled as the “ shengseng gongdian 聖僧宮殿 (palatial hall of the Holy Monk),” appears to be a portab le piece of furniture (fig. 57). The base is a squarish altar table, on which stands a miniature hall composed of balustrade, lattice windows, three layers of eaves supported by three-tiered brackets, and the shanhu a jiaoye 山花蕉葉 (mountain flowers and banana leaves) roof ornament. 56 The icon enshrined is the Bodhisattva Guanyin in the guise of a monk, and the shrine is shown placed in the center of a dormitory ( zhongliao 眾寮 ), a large hall where each monk is assigned a seat (a section of a long couch) to study Buddhist scriptures. 57 As the typical p lan of the hall shows (fig. 58), the couches are aligned along the four walls and around the four “sky-wells ( tianjing 天井 )” so as to receive maximum natural light for reading. 58 The central shrine (here represented as a square box containing an image inside), on the other hand, is accompanied by a separate altar table, an incense burner, two candlesticks, and ten square mats at the front. It is not the only min iature architecture in the dormitory: to the left and right of the entrance, similar settings of incense burners 55 Zhang 2000, 3-5, 8-10. The dating of 1248 is mainly based on an inscription, “Eighth year of the Chunyou Period of the Southern Song 南宋理宗淳祐八年,” written on the scrolls. The original drawings no longer remain, but there are many copies, most of which from the Muromachi 室町 Period (1336-1573). As Zhang notes, since most of the monasteries illustrated in the scrolls are now gone, the d rawings serve as the most reliable evidence of Southern Song Buddhist architecture and furniture, thanks to the detailed specifications they include. 56 Ibid., 83. The altar table is stylized as any other detached altar tables seen in the same scrolls. Zha ng assumes the shrine to be originated from the famous Jingshansi 徑山寺. 57 Ibid., 49. The various activities and rituals in the zhongliao, including reading, drinking tea, and taking medicines, are detailed in the Chixiu Baizhang qinggui 敕修百丈清規, T48. 2025, an d in the Yongping qinggui 永平清規, T82. 2584. The se sources inform us of the layout and furnishings of these buildings. 58 The plan however is the zhongliao of Jinshansi 金山寺, not necessarily where the illustrated shrine comes from; but it is the only example av ailable.	di-luo-a-grain-of-sand-full-draft-3.pdf
113 and candlesticks are placed for two additional furniture pieces, each with a sloped roof and a high dais, and the one on the viewer's right is labeled as the “Huayanjing 華嚴經 (Flower Garland Sutra ),” perhaps indicating a sutra repository. Interestingly, this layout of having the main “image hall” at the center and the library at its side perfectly corresponds to the typical plan of Southern Song m onasteries. 59 Another example of the shengseng gongdian comes from the Jingshansi 徑山寺, the most eminent of the Five Buddhist Mountains (monasteries) of the Southern Song (fig. 59). The shrine proper is seated on top of a high Sumer u dais and is connected to the ground with an arching stairway. The overall design bears a remarkable resemblance to the Yingzao fashi template, including certain dimensions. 60 Like the previous example, it enshrines the image of the “Holy Monk” Manjusuri and is placed in the center of a dormitory (fig. 60). 61 These shrines seem to have been designed strictly for the monastic group to perform daily and monthly rituals, and were largely out of the sight of t he general public except on certain days of the year. 62 This semi-public nature was a 59 Zhang 2000, 37-42, 114-15. Though usually the library halls were on the west side of the central avenue, just like in the case of the Longxingsi. Both the Jingshansi and the Lingyinsi 靈隱寺 have adopted this scheme. 60 Ibid., 83, 142. Several dimen sions marked on the drawing include : zhangzuo 帳坐 (dais), 46 cun tall; zhangshen 帳身 (body), 69 cun wide and 73 cun deep; zhangzhu 帳柱 (columns), 100 cun tall and 3 cun across. Dimension-wise it is fairly close to the fodaozhang in the Yingzao fashi, which features a 45-cun tall dais and 125-cun tall columns. 61 Ibid., 47-48, 118. As the drawing shows, the Diamond Sutra (Ch. Jingang jing 金剛經 ) is stored behind the main icon Manjusuri, perh aps within certain receptacles. T o the east of the altar are assigned the seats for the Dongzang 東藏 and Xizang 西藏, monks who were in charge of the east and west wings of the repository of the Tripitaka. According to another drawing of the jiela pai 戒臘牌 (seniority placards) in Zhang 2000, 145, the hall was able to hold “a total of eight hundred and fifty-four monks 清眾共八百五十四員 ” during an assembly. 62 Ibid., 101. On every third, eighth, thirteenth, eighteenth, twenty-third, and twenty-eighth day of each month, a chanting ritual called the “ sanba niansong 三八念誦” was held in the hall, where monks circumambulated along the couches. According to the Baizhang qinggui, T48. 2025. 1152a, http://www. cbeta. org/result/normal/T48/2025_007. htm, both the sengtang 僧堂 (monks' hall) and zhongliao would be open to the laity on the afternoon of the fourteenth day of every July, when all seniority placards were displayed and wor shippers came to offer incense.	di-luo-a-grain-of-sand-full-draft-3.pdf
114 determinative factor that they needed not be built as large and grandiose as those displayed in worship halls. In houses Not all architectural miniatures served religious purposes. The stone reliefs in several late Eastern Han tombs, most notably Tomb 1 at Dahuting 打虎亭 in Mixian 密縣, depict a certain type of wooden cupboards with a distinctive sloped roof (fig s. 61, 62). 63 Judging from the pictorial context, these cupboards were used mainly for food storage and perhaps also clothes, and they might have been modeled o n granaries and barns. 64 It is likely that by the early third century, architecture had been adopted as a belov ed form for furniture pieces especially cabinetry, which turned out to be a major prototypical source for the religious shrines and repositories in later periods. 65 Extant examples of household miniatures, unfortunately, are rare and scattered. Nonetheless, the Lu Ban jing 魯班經 (Carpenter's classic ), a fifteenth-century carpenter's manual compiled by a Ming official, shed some light on this issue from the perspective of furniture-making. 66 It introduces 63 A detailed excavation report and preliminary study is in Mixian Dahuting Hanmu 密縣打虎亭漢墓 (Han-dynasty tombs at Dahuting in Mixian) (Beijing: Wenwu chubanshe, 1993), which contains descriptions and photographs of the reliefs. There are two instances where architectural-shaped cupboards are represented. One is t he roofed kitchen cupboard on the north wall of the east side-chamber ( p. 139); the other is on the north wall of the north side-chamber, where a very similar sloped-roof cupboard with two door-leaves appears ( p. 172). The tomb has been dated to late Easte rn Han; see pp. 340-344. 64 Ibid., 26-27. Miniature granaries made of bronze have been found in Warring States burials; see C hapter 5 of this dissertation. 65 Li Zongshan 李宗山, Zhongguo jiajushi tushuo 中国家具史图说 (A pictorial history of Chinese furniture) (Wuhan: Hubei meishu chubanshe, 2001), 164, 177, 182-83. In his discussion of the chuwu 櫥屋 (architectural-shaped cabinets), Li refers to Eastern Han tombs at Dahuting and Ban gtaizi 棒台子 in Liaoyang 遼陽, proposing that the particular architectural shapes might h ave come from granaries. 66 Lu Ban jing, its full title being Xinjuan jingban gongshi diaozhuo zhengshi Lu Ban jing jiangjia jing 新鐫工師雕斫正式魯班木經匠家鏡 (The newly carved, authent ic classic of woodworking and guidance of carpentry of Lu Ban), is a fifteenth-century carpenter's manual compiled by the Ming official, Wu Rong 午榮. One work pioneering the study on the Lu Ban jing is Klaas Ruitenbeek, Carpentry and Building in Late Imperi al China: A Study of the Fifteenth-century Carpenter's Manual Lu Ban	di-luo-a-grain-of-sand-full-draft-3.pdf
115 a certain portable shrine called shenchu 神廚 (spiritual ca binet) used in a private setting such as the residence (fig. 63). 67 Situated on a four-legged table, the cabinet-like shrine is fashioned into a simple miniature hall with hanging posts, balustrades, lotus-based columns, and flame-patterned screens. 68 As no brackets are used, the shrine seems to be a less sophisticated (and much smaller) version where miniaturization is nonetheless achieved by incorporating various architectural elements. Such simplification allows the miniaturist to cohere with cabinetry conventions without necessarily applying any specific rules of scaling. The same building material and the same post-and-lintel, mortise-and-tenon structural framework shared between architecture and furniture must have significantly facilitated his work. 69 Household shrines like this were also used in imperial palaces (fig. 64). Jing (E. J. Brill, 1996), which consists of a facsimile of the manuscript, a full annotated translation of the entire text, and an analysis of the historical and social background. As Rui tenbeek details, The manual was compiled by court officials of the Board of Industry ( gongbu 工部 ) during the Yongle Period (1403-1424) soon after the move of the capital from Nanjing to Beijing. It has incorporated the full text of the Lu Ban yingzao zhengshi 魯班營造正式 (Authenticate building standards of Lu Ban), a manuscript circulated in the Yuan, and excerpts from several different encyclopedic works, household handbooks, and almanacs of the Yuan and the Ming, which focus on rituals associated with building activities. The compiled work includes guidelines and illustrations for furniture-making, an d the language is highly colloquial, suggesting a likely origination from orally transmitted principles and techniques, such as those of local workshops. Ruitenbeek's study also reveals Lu Ban jing 's probable references to the Yingzao fashi. See pp. 25-33, 129-39. 67 Ruitenbeek 1996, 202. The original entry title is shenchu chashi 神廚搽式 ; the character cha 搽 seems out of place and might be corrupt. The meaning of tuchu 土廚 in the entry is not clear; Ruitenbeek interprets it as shangch u 上廚 (upper shrine), denoting the miniature shrine on the upper level of the cabinet. Some of his interpretations are tentative but the overall dimens ions should be fairly accurate. 68 Ibid., 199-202. The term huanmei 歡眉 might be a corrupt of huanmen 歡門. The bipartite (furniture below and architecture above) pattern is also found in Shao Xiaofeng 邵晓峰, Zhongguo Songdai jiaju: yanjiu yu tuxiang jicheng 中國宋代家具 : 研究與圖像集成 (Furniture of Song China: A Collection of Research and Images) (Nanjing: Southeast University Press, 2010), 176, figure 5-2-3, showing illustrations of portable shrines in the Zhuzi jiali 朱子家禮 (Family rituals of Master Zhu Xi), though I have not found the original source of the illustrations. 69 Shao 2010, 169-76. Shao's book includes a chapter on the connections between Song architecture and furniture. In terms of material and structural logic, both traditions are based on the timber frame work using mortises and tenons. Functionally, the legs of a chair or table is comparable to the columns, the st retchers to the lintels; the enclosing members such as cabinet doors and panels are like doors and lattice windows of a building, whereas the shuyao 束腰 (middle ornamental section of the base of some furniture) might have inspired the Sumeru dais. Shao's summary is very brief and general, but he points to several directions that invite further exploration s. While the standardization and modularization of Northern Song architecture brought significant changes to contemporary furniture design, distinctive ornamental motifs used in furniture also became appropriated by buildings. There were several ways in which Song	di-luo-a-grain-of-sand-full-draft-3.pdf
116 Among the fifty-odd types of furniture described in the Lu Ban jing, two types--yichu 衣廚 (clothes cupboards) and yaochu 藥廚 (medicine cupboards)--display certain numerical conventions that are similarly discernible in both the Yingzao fashi and the Huayansi sutra cabinets. The terminology of the latter, on the othe r hand, appears to have been largely borrowed from that of cabinetry, incorporating items such as cheng 棖 (rails) and yazi 牙子 (stretchers), which are not applicable to large-scale woodworking. Most intriguingly, a repository should consist of seven shelves exactly a s a medicine cupboard (fig. 65). 70 Each cabinet at the Huayansi measures approximately 151. 4 by 62. 7 centimeters in plan and 135 centimeters in height; this dimension is smaller than what the Yingzao fashi decrees but closer to the size of the cupboards recorded in the Lu Ban jing. 71 Despite these comparable dimensional data, one disparity between the two domains--miniature woodworking and cabinetry--is that the former is strictly premised on a scali ng scheme whereas the latter clearly lacks one, so far as the text suggests. architecture brought changes to furnitu re design: 1) changes in the size of interior space demanded corresponding changes furniture size; 2) entasis and tilts of vertical supports became adopted and exaggerated by chair legs; 3) shape and ornamentation of the roof ridge were mimicked by the top rail of chairs; 4) cap blocks ( ludou 櫨枓 ) used in bracketing were borrowed and used atop chair legs; and 5) concepts of standardization and modularization were embraced by manuals of furniture-making such as the Yanji tu 燕幾圖 (Diagrams of combinative tables ; d. 1194). 70 Seven appears to be considered a favorable number in the Lu Ban jing. Another case where seven shelves (panels?) are used is in a granary ( hecang 禾倉 ); see Ruitenbeek 1996, 207. The seven shelves as a rule is also adopted by the zhuanlun jingzang. The dimension of the shelves differ, however. For medicine cupboards, the space between each shelf is 5 cun; for the repositories, the dimension is not given directly but should be at least 6 cun because of the height of jingxia 經匣, the sutra coffers. See Yingzao fashi, vol. 2, 15, 22; Takeshima 1971, 732-34. The relation between medicine cupboards and the wheel-turning mechanism is exposed by a Ming medicine box (78. 8 by 57 by 94. 5 centimeters, ca. 1573-1620 ), which contains a revolving octagonal center. See Lu Jimin 呂濟民, Zhongguo chuanshi wenwu shoucang jianshang quanshu : muqi 中國傳世文物收藏鑑賞全書 : 木器 (Connoisseurship of Chinese cultural relics: wooden artifacts ), vol. 1 (B eijing: Xianzhuang shuju, 2006), 125. 71 This measurement is extracted from my digital model and does not include the dais.	di-luo-a-grain-of-sand-full-draft-3.pdf
117 The Tamamushi Shrine: a distant echo from Japan The connection between shrines and cabinets is further detectable philologically in Japan, where miniature Buddhist shrines are ge nerally termed zushi; the Tamamushi Shrine, for instance, is such a zushi. The kanji zu is derived from the sinograph chu, alternatively written as 橱 in Chinese, meaning “cabinets. ”72 The zushi in Japan often take the form of architecture, applying a scaling technique not unlike the Yingzao fashi formulas, and they have been treasured as epitomes of contemporary wooden structures. The Tamamushi, dated to the seventh century, is not only the earliest zushi that have come down to us but also probably an antecedent to all surviving wooden buildings in East Asia. 73 Measuring more than two meters in height (including the pedestal), the shrine is made into a miniature wooden hall with distinctive structural features ( see fig. 8). 74 The small scale of the shrine encourages intimacy by potentially shortening the physical distance betwe en the beholder and itself, while its portability bespeaks the need and convenience of transportation, just as any regular cabinet. 75 72 In Japan, there are also zushidana 厨子棚 (cabinets for books) and zushigame 厨子甕--stylized containers made of clay or stone for storing the bones and ashes of the dead, which are often miniature buildings intriguingly similar to some spiritual urns and sarcophagi in China. 73 Walley 2012, 267-68. The dating of the shrine is based on its architectural style and an inventory (d. 747) of the Horyuji 法隆寺, which refers to a shrine made in the shape of a palace hall and decorated with the Thousand Buddha motif, believed to be referring to the Tamamushi. Walley further narrows down the dating to 630s-650 based on her reading of the style of the paintings on the pedestal, which she argues to have shown an affinity to Northern Qi and Sui Buddhist art; see pp. 270-72. 74 The dimension of the shrine is 2,266 (height) by 1,367 (width) by 1,191 (depth) millimeters, according to Walley 2012, 267. The shrine has been discussed again and again in the discourse of Ch inese architectural history as a witness to Tang and pre-Tang architectural styles; see, for instance, Sickman and Soper 1984, 396. 75 Walley 2012, 319-20. Walley believes the shrine to be originally “an object of worship in a private residence,” as its siz e suggests intimacy and private devotion, and would be fit for the use of a small group of devotees. She further argues that the mountains painted on the back side of the pedestal is “inviting us to consider the entire shrine, in a sense, as one large moun tain”--this symbolism of the miniature architecture as the world pillar echoes my discussion in Chapter 2.	di-luo-a-grain-of-sand-full-draft-3.pdf
118 Ono Satoshi's 大野敏 study of ancient and medieval Japanese zushi has demonstrated that miniature shrines in East Asia were ma de in more diverse forms and styles than those covered in the Yingzao fashi. He categorizes the zushi of the Asuka and Nara periods (592-794) into four major types: 1. The “palace” type, or kyuden 宮殿, which emulates the Buddha ha ll; 2. The “canopy” type, or chobo 帳房, composed of a squarish dais, columns, and a flat “ceiling” or canopy above, its prototypes being the canopied couches and beds of the aristocrats; 3. The “round baldachin” type, reminiscent of the use of a round tengai 天蓋 (lit. heavenly cap) over Buddhist icons, which is in this case emulated by the pyramidal roof of either a hexagonal or octagonal hall, its corner rafters stylized in the curvature of bracken shoots (warabite 蕨手 ); 4. The “cabinet” type, either box-like or cylindrical. 76 These four types, especially the “palace” type, had been applied to miniature shrines persistently till the late sixteenth century. 77 Ono's diagrams (fig. 66) tellingly expose the interchange between miniature ar chitecture and cabinetry: the “palace” type presents the closest emulation of architecture and must have applied certain scaling techniques to achieve this formal resemblance, whereas the “cabinet” type appears not so different from regular cabinets, bookc ases, and cupboards. The 76 Ono Satoshi, Muromachi chuki ~ koki ni okeru kyudenkei zushi no kenchiku yoshiki ni kansiru kenkyu 室町中期 ~後期における宮殿系厨子の建築様式に関する研究 (Research on the arc hitectural style of palace-type miniature shrines during the middle to late Muromachi Period) (Shikaban, 2002), 3. The boundaries between the four types are blurry, as there have been examples showing features of more than one category, which gave birth to certain eclectic ( setchu 折衷 ) types, such as the tengai chobo setchu 天蓋帳房折衷 (combination of the “round baldachin” and “canopy”), as Ono calls it, in the Heian period. In some early Japanese texts, the term kyuden has been borrowed to indicate miniature shrines, before zushi came in to use. It is noticeable that the term cho 帳 appears here, just like how miniature shrines are referred to as zhang in the Yingzao fashi. Also noticeable is that tengai can be classified as a type of zushi; for its association wi th zaojing ceilings and the concept and materialization of the “dome of heaven” in Chinese architecture, see Chapter 4 of this dissertation. For zushi, see also Ono Satoshi, “Chusei zushi no keishiki bunrui ni tsuite 中世厨子の形式分類について (On the typology of the m iniature shrines in Medieval Japan),” Nihon kenchiku gakkai keigakukei ronbunshu 日本建築学会計画系論文集 505 (1998): 191-98. 77 Ono 2002, 2, 5-6. Ono argues that later types, except but one “ sanka shoyo 山花蕉葉,” have all been generally based on the four original types. The sanka shoyo type, being derived from the shanhua jiaoye shrines introduced in the Yingzao fashi, might have been imported with Zen Buddhist architecture from Song China to Japan.	di-luo-a-grain-of-sand-full-draft-3.pdf
119 “canopy” and “baldachin” types, on the other hand, are somewhat in between: the “canopy” emulates canopied furniture for sitting and sleeping, while the “baldachin” basically adds a roof-or parasol-like structure directly on top o f an ordinary cabinet. None of these types, however, features the tiangong louge motif in Yingzao fashi, though the “palace” and “canopy” types display a recognizable structural affinity to Northern Song and Liao miniatures. The Miniature and the Myriad All religious space is inherently theatrical, as theatricality allows a swift transition of space, time, identity, and purpose, so that like in any drama, a psychological process of empathy and catharsis is effortless generated. 78 The performers in this “drama” include various Buddhist deities, the practitioners, and sometimes even the worshippers (who were simultaneously the audience). In the Northern Song and Liao, the stage for such a drama was set up by the interior instead of the exterior of architecture, and none could have been more impressive than a backdrop showing a panoramic, all-embracing vi ew of the universe created in miniature form. Miniature-making to this end addressed to aesthetic as well as theological concerns: it had to be more than a spectacle but one that fulfilled certain exegetical and soteriological functions. This was achieved through the image of a world of the myriad evoked by miniaturization. 78 As discussed in Chapter 1 of this dissertati on, in Song China, miniature structures were erected sometimes as makeshift stages for drama. Jeehee Hong's study of Song and Jin tomb art has exposed a subtle connection between miniature architecture and theatricality; see Hong 2011. Buddhist “drama, ” as all types of drama, similarly assumed pedagogical and didactic purposes.	di-luo-a-grain-of-sand-full-draft-3.pdf
120 The Flower Repository Universe While many Buddhist scriptures expound on cosmography, the one that is the most relevant to our study is the Flower Garland Sutra (Huayanjing 華嚴經, Sk. Avatamsaka sutra ). 79 It speaks of a “Flower Repository Universe ( huazang shijie 華藏世界, also translated as “Lotus Repository World”)” consisting of a multitude of seas, lands, and cities: In the land masses of this ocean of worlds are se as of fragrant waters, as numerous as specks of minuscule dust in unspeakably many buddha-fields. All beautiful jewels adorn the floors of those seas; gems of exquisite fragrances adorn their shores. They are meshed by the Vairocana king of the jeweled tre asure into a net... Stairways of ten kinds of precious substances are set out in rows, with balustrades of ten kinds of jewels surrounding them. White lotuses ornamented with jewels, as many as specks of minuscule dust in four continents, are spread over the waters, in full bloom. There are unspeakable hundreds of thousands of billions of trillions of sila banners of ten precious elements, banners of belled gauze of raiments of all jewels, as many as sand grains in the Ganges river, jewel flower palaces of bo undless forms, as many as sand grains in the Ganges river, a hundred thousand billion trillion lotus cities of ten precious substances, forests of jewel trees as many as specks of minuscule dust in four continents, networks of flaming jewels, as many sanda lwood perfumes as grains of sand in the Ganges, and jewels of blazing radiance emitting the sounds of Buddha's speech; unspeakable hundreds of thousands of billions of trillions of walls made of all jewels surround all of them, adorning everywhere. 此世界海大地中, 有不可說佛剎微塵數香水海, 一切妙寶莊嚴其底, 妙香摩尼莊嚴其岸, 毘盧遮那摩尼寶王以為其網... 十寶階陛, 行列分布 ; 十寶欄楯, 周匝圍遶 ; 四天下微塵數一切寶莊嚴芬陀利華, 敷榮水中 ; 不可說百千億那由他數十寶尸羅幢, 恒河沙數一切寶衣鈴網幢, 恒河沙數無邊色相寶華樓閣, 百千億那由他數十寶蓮華城, 四天下微塵數眾寶樹林--寶焰摩尼以為其網, 恒河沙數栴檀香, 諸佛言音光焰摩尼, 不可說百千億那由他數眾寶垣牆, 悉共圍遶, 周遍嚴飾. 80 79 The name of the monastery, Huayansi, seems to single out the utter importance of this sutra and of the Huayan School. According to Bai 2011, 15, Emperor Daozong 道宗 (r. 1055-1101), under whose reign the Huayansi was established, has himself authored ten rolls of Huayan jing suipin zan 華嚴經隨品贊. In the Sui and Tang, the Huayan School earned much imperial favor, especially during the reigns of Sui Wendi (r. 589-605) and Empress Wu Zetian (r. 684-704). The translation of the Avatamsaka was an imperially funded project under Wu Zetian: it was first carried out by Siksananda 實叉難陀 (652-710) in 695 at the Dabiankongsi 大遍空寺 in the imperial palace in Luoya ng, and was completed by Fazang in 699 at the Foshoujisi 佛授記寺 (aka. Jing'aisi 敬愛寺 ). An account of this is in Eugene Wang, Shaping the Lotus Sutra: Buddhist Visual Culture in Medieval China (Seattle, Washington: Univer sity of Washington Press, 2005), 133; a lso see Song gaoseng zhuan 宋高僧傳, T50. 2061. 732a, http://www. cbeta. org/result2/normal/T50/2061_005. htm. 80 Dafangguang fo huayan jing 大方廣佛華嚴經, T10. 279. 40b, http://www. cbeta. org/result/normal/T10/0279_008. htm ; Thomas F. Cleary, trans. The Flower Ornament Scripture: Translation of the Avatamsaka Sutra (Boulder: Shambhala Publications, 1984), 207, with minor changes. The term huazang 華藏 is interpreted as the “Flower Bank” by Cleary,	di-luo-a-grain-of-sand-full-draft-3.pdf
121 The description evokes a seri es of vivid architectural images--the stairways, balustrades, banners, bells, jewel flower palaces, lotus cities, and the encircling walls--which have all rendered the Flower Repository Universe palpable and “measurable. ” It is notable that this universe i s a self-enclosed system with every part of it interconnected through a certain network, which resembles a mandala with a forever expanding interior. 81 More awe-inspiring are the numerals: while the billion and trillion are too abstract and inadequate, spec ks of dust in the continents and sand grains in the Ganges have been brought in to calculate the myriad. To add to the intricacy of this system, there are numerous “world seeds ( shijie zhong 世界地種; Sk. lokabija )” dispersed in the universe, and each “world s eed” is a capsule of a world system as complex as its parent system. 82 The “world seeds” are variably “shaped like high mountains, rivers, whorls, whirlpools, wheel rims, altars, forests, palaces, mountain banners, geometric figures, wombs, lotus blossoms, baskets, bodies of sentient beings, clouds, the distinguishing features of Buddhas, spheres of light, webs of various pearls, doors, and various ornaments. Their shapes, if fully told of, number as many as specks of minuscule dust in an ocean of worlds. ”83 It is perhaps due to this which is alternatively translated as the “Lotus Treasury” or “Lotus Repository” by scholars. The numeral nayuta (Ch. nayouta 那由他 ) is usually rendered as “billions, trillions, incalculable. ” 81 The enclosure is stressed again in the verse after the narration, especially, “Walls surround everything/ With facing towers and pavilions arrayed on them 垣牆繚繞皆周匝, 樓閣相望布其上 ” (T10. 279. 40c; translation after Cleary 1984, 208). This line can actually serve as a description of the Huayansi sutra cabinets. 82 T10. 279. 41c; translation after Cleary 1984, 213. “In these seas of fragrant waters, numerous as specks of minuscule dust in unspeakably many buddha-fields, rest an equal number of world seeds. Each world seed also contain s an equal number of worlds 此不可說佛剎微塵數香水海中, 有不可說佛剎微塵數世界種安住 ; 一一世界種, 復有不可說佛剎微塵數世界. ” The world system is similar to a fractal curve which can be infinitely zoomed in and maintain a certain similarity in structure on all scales, such as the Koch snowflake. The use of the term “seed” implies that the worlds self-generate and self-multiply, a basic point of view in the Huayan philosophy. According to Francis H. Cook, Hua-yen Buddhism : the Jewel Net of Indra (University Park: Pennsylvania State University Press, 19 77), 3, “The cosmos is, in short, a self-creating, self-maintaining, and self-defining organism. Hua-yen calls such a universe the dharma-dhatu (法界 ). ” The “seed” is also related to the concept of tathagatagarbha, the womb of the “Thus Come One,” or 如來藏 ; the same connotation might also exist for the term huazang. See Cook 1977, 45-46. 83 T10. 279. 41c-42a; translation after Cleary 1984, 213. Further, “all of these worlds in each of these world seeds rest on various adornments, connecting with each other, forming a network of worlds, set up, with various differences,	di-luo-a-grain-of-sand-full-draft-3.pdf
122 variety in shape that one can equate any material form--as huge as the Sumeru and as small as a seed, a lotus flower or a miniature building--to a world seed and the multiple worlds it encapsulates. Indra's Net The image of the m inute and the myriad, other than to impress the audience, is brought up to explicate the central tenets of Huayan Buddhism. Cosmology in this sense serves preaching, and the meticulous delineation of the multiple world system seeks to locate the path to ul timate truth and hence ultimate salvation. 84 The metaphor of “Indra's Net”--analogous to the web of “world seeds” quoted above--has been the favorite trope of Huayan literature to convey the interdependence and intercausality of all beings. 85 This concept is elaborated in a work attributed to Dushun 杜順 (557-640), the first patriarch of the Huayan School: The jewels [of Indras' Net] are shiny and reflect each other successively, their images permeating each other over and over. In a single jewel they all appea r at the same time, and this can be seen in each and every jewel. There is really no coming or going. Now if we turn to the southwest direction and pick up one of the jewels to examine it, we will see that this one jewel can immediately reflect the images of all of the other jewels. Each of the other jewels will do the same. Each jewel will simultaneously reflect the images of all the jewels in this manner, as will all of the other jewels. The images are repeated and multiplied in each other in a manner tha t is unbounded. Within the boundaries of a single jewel are contained the unbounded repetition and profusion of the images of all the jewels. The reflections are exceedingly clear and are completely unhindered. 以寶明徹遞相影現涉入重重, 於一珠中同時頓現, 隨一即爾, 竟無去來也. 今且向西南邊, 取一顆珠驗之, 即此一珠能頓現一切珠影, 此珠既爾, 餘一一亦然. 既一一珠一時頓現一切珠既 throughout the Flower Garland ocean of worlds 此一一世界種中, 一切世界依種種莊嚴住, 遞相接連, 成世界網 ; 於華藏莊嚴世界海, 種種差別, 周遍建立 ” (T10. 279. 51b; translation after Cleary 1984, 242 ). 84 This point is stressed in Kloetzli 1983, 50. “Clearly the cosmos represents the map of the path to enlightenment. ” See below. 85 See Cook 1977, 8-16, for an excellent explanation of the con cepts of interdependency and intercasuality.	di-luo-a-grain-of-sand-full-draft-3.pdf
123 爾, 餘一一亦然. 如是重重無有邊際, 有邊即此重重無邊際珠影皆在一珠中, 炳然高現. 餘皆不妨此. 86 One may read this passage as a revelation of the totalistic view that all things are related to and reflected by each other, and that the “self” of the indivi dual is essentially empty except that it is the container of everything else. 87 Other than the ontological and epistemological messages, Dushun is perhaps more concerned about praxis, as he continues, If you sit in one jewel, you will at that instant be sit ting repeatedly in all of the other jewels in all directions. Why is this? It is because one jewel contains all the other jewels. Since all the jewels are contained in this one jewel, you are sitting at that moment in all the jewels. The converse that all are in one follows the same line of reasoning. Through one jewel you enter all jewels without having to leave that one jewel, and in all jewels you enter one jewel without having to rise from your seat in the one jewel. 若於一珠中坐時, 即坐著十方重重一切珠也. 何以故. 一珠中有一切珠故. 一切珠中有一珠時, 亦即著一切珠也, 一切反此. 準以思之. 既於一珠中入一切珠, 而竟不出此一珠 ; 於一切珠入一珠, 而竟不起此一珠. 88 But how does one “enter” and “sit in” a jewel from the first place? If we consider all to be essentially empty and everything inherently interrelated, it is conceivable that such an “en trance” and “sitting” might be achieved by the unhindered mind, which is able to penetrate all things. 89 In this 86 Huayan wujiao zhiguan 華嚴五教止觀 (Calming and contemplation in the Five Teaching s of Huayan), T45. 1867. 513a-b. Translation by George Tanabe in William Theodor e de Bary and Irene Bloom, Sources of Chinese Tradition, vol. 1, From Earliest Times to 1600 (New York: C olumbia University Press, 1999), 473. 87 Cook 1977, 13, “... a universe which is nothing but the complete mutual cooperation of the entities which make it up. ” And pp. 15-16, “Hua-yen insists on a totalistic view of things. Totalism has two meanings. First, it means that all things are contained in each individual... It is for this reason that Hua-yen can make the seemingly outrageous claim that the whole universe is contained in a grain of sand. However, not o nly does the one contain the all, but at the same time, the all contains the one, for the individual is completely integrated into its environment. ” 88 T45. 1867. 513a-b; translation in de Bary and Bloom 1999, 473. 89 Cook 1977, 36, 68. Huayan Buddhism preache s the “interpenetration of all things ( shishi wu'ai 事事無礙 ),” a concept intertwined with interdependence and intercausality. This is evident in Fazang's Xiu Huayan aozhi wangjin huanyuan guan 修華嚴奧旨妄盡還源觀, T45. 1876. 640a: “The mind discussed here is the unhindered mind, which the Buddhas actualized to attain th e dharma-body; the realm is the unhindered realm, which the Buddhas actualized to create the Pure Land 言心者謂無礙心，諸佛證之以成法身；境者謂無礙境，諸佛證之以成淨土,” paralleling the mind to the “world/dharma-field. ” Hence, “In one pore there are numerous Buddha fields/ Each of which contains the four continents and four seas/ The Sumeru and the Cakravala mountains/ Both appear inside with no hindrance 一毛孔中無量剎，各有四洲四大海，須彌鐵圍亦復然，悉現其中無迫隘 ;” and, “Of all the specks of dust in the Flower	di-luo-a-grain-of-sand-full-draft-3.pdf
124 sense, the image of Indra's Net is evoked to assist meditation, a state in which the meditator visualizes himself being one with all things in t he entire universe. 90 Sudhana's epiphany in the Tower of Vairocana The theme of meditation is continued in an exposition by Faz ang 法藏 (643-712), a master of syncretism and the third patriarch of the Huayan School. When meditating on Indra's Net, Fazang explains, one might think of Sudhana's (Ch. Shancai tongzi 善財童子 ) visit to the Tower of Vairocana as recounted in the Flower Garlan d Sutra. 91 Upon his entrance to the tower, Sudhana “saw hundreds of thousands of other towers. And in each one of these hundreds of thousands of towers there were further hundreds of thousands of towers. In front of each one of these towers was Maitreya Bod hisattva, and in front of each Maitreya Bodhisattva was Sudhana. ”92 This, in Fazang's exposition, “manifests the multiple interrelationships in the dharma universe and is like the unending connections in Indra's Net. It also makes clear that Sudhana had a s udden, ultimate insight into the dharma universe as a result of his practice according to the principles of the Flower Repository Universe. Thinking of one tower as the master and all the other towers within it as the retainers is Repository world/ The Buddha enters into each and every one of them 華藏世界所有塵，一一塵中佛皆入. ” More on Fazang below. See also Kloetzli 1983 for the interpretations of the relationship between the universe and the mind. 90 Du Shun himself was a great master of meditation. Cook 1977, 26, quotes D. T. Suzuki that “Hua-yen is the philosophy of Zen and Zen is the practice of Hua-yen. ” During its formation, the Huayan School also absorbed much Daoist elements, especially the “totalistic view of existence” in the Zhuangzi 莊子 (pp. 26-27). 91 Recounts of the travels of Sudhana con sist of the Gandavyuha 入法界品 (Entering the dharma-realm), originally a separate Mahayana sutra and later incorporated into the Avatamsaka as its final chapter, which has been a legendary piece of literature sometimes compared to Dante's Divine Comedy. See C ook 1977, 22. 92 T45. 1876. 640b; translation after Tanabe in de Bary and Bloom 1999, 474.	di-luo-a-grain-of-sand-full-draft-3.pdf
125 the meditation on Indra's Net in which masters and retainers manifest each other. This is also the meditation on the unobstructed interrelatedness of things with all things. ”93 Sudhana's epiphany is strongly reminiscent of Daoxuan's description of the lotus repository discussed in Chapter 2, a miniature woodwork which would open automatically to those receiving perfect ordination, who would be able to see inside “84,000 towers and pavilions. ” It is hard not to believe that the miniature architecture installed inside worship halls, such as the Huayansi sutra cabinets, was not created out of the same purpose of conjuring up a similar image of the myriad. And it is hard not to imagine that a medieval Chinese, upon entering such a stage-like, sacred space surrounded by arrays of small “ towers” and “pavilions,” would not, in the slightest, marvel at the possibility that he or she, a Sudhana in a different space-time, might have penetrated the magnificent multiplicity of worlds and of the dharma universe. 94 To be sure, rarely has a group of miniature architecture been created to literally correspond to the “hundreds of thousands of towers” in the Flower Garland Sutra, but the magic of miniaturization in evoking the myriad is indubitable. 95 The tiangong louge motif in the Yingzao fashi is a tacit evidence to this point; though adopted at neither the Huayansi nor the Longxingsi, it existed probably as the most ideal (and extremely costly) approach to representing the Huayan universe, where small worlds always contain e ven smaller worlds. 93 Ibid. “In the meditation on Indra's Net, the principal master [i. e., the one jewel] and the subordinate retainers [i. e., the other jewels] are manifestations of each other... As soon as one thing is designated master, both the master and retainers are equally brought together in relationships that multiply without end. This indicates that the nature of things lies in multiple relationships reflecting each other unendingl y in all things 主伴互現帝網觀... 隨舉一法即主伴齊收，重重無盡，此表法性重重影現，一切事中皆悉無盡. ” This reminds one of Zhiyi's 智顗 (583-597) “ yinian sanqian 一念三千 (three thousand world-systems within an instant of thought). ” 94 One textual evidence for this is the “Tiantongshan qianfoge ji 天童山千佛閣記 (Record of the Thousand-Buddha Pavilion in the Tiantong Mountain) by Lou Yao 樓鑰 (1137-1213), which alludes to the pavilions witnessed by Sudhana while eulogizing the lofty architecture of the monastery; see Tiantongsi zhi 天童寺志, 2. 8a-b. A quotation is pr ovided in Zhang 2000, 108, n. 2-17. 95 Many literary works on miniature architecture bespeak this point. See, for instance, the stele recording Liang Shouqian's wheel-turning repository discussed in Chapter 2, which extolls the “countless flowery banner-pillars” and the “thousands of tower-pavilions” it contained.	di-luo-a-grain-of-sand-full-draft-3.pdf
126 The relationship between the miniature and the myriad is actually far more substantial than one might have assumed. As Randy Kloetzli has pointed out, the attempt to accurately measure time and space by minute particles found in nature has started at the dawn of human civilizations. 96 In Northern Song China, drops of water, for instance, was used for chronometry in Su Song's famous invention, shuiyun yixiangtai 水運儀象臺, a device combining an astrolabe and a clock (see fig. 43). 97 A more ancient example is Archimedes' (d. 212 BCE) experimental computation of the volume of the universe recorded in “The Sand Reckoner”: the way he determines th e vastness of the universe is by estimating how many grains of sand will be needed to fill its spherical space, and the result indicates a total of 1063 grains of sand to be needed. 98 The revelation of Archimedes' calculation is that even the greatest infin ite can be measured and understood, even though such a measurement necessarily involves the use of infinitesimals expressed in such humongous numbers that they become nearly as inconceivable. 99 Did the Buddha ever apply a similar infinitesimal thinking in h is reckoning of time and space? In the Flower Garland Sutra, as noted above, “specks of minuscule dust” and “grains of sand” serve as units of reckoning, but the most curious case where an Archimedean method is adopted is in the “Parable of the Mirage City ” in the Lotus Sutra (Ch. Miaofa lianhua jing 妙法蓮花經 ): 96 Kloetzli 1983, 113-19. 97 For Su Song's invention and its probable precursor in Wu Zetian's mingtang complex, see Forte 1988, and Needham, Wang, and Price 2008. As remark ed in Kloetzli 1983, 114, n. 3, “Keeping in mind the chronometric significance of these images, we may wonder if the 'sands of the Ganga' do not in some sense cons titute a 'river of time. '” 98 Kloetzli 1983, 115-17. This number is reportedly the exact value of one asankhyeya (Ch. asengqi 阿僧祇 ) in Buddhist cosmology. Further, “... we are driven unavoidably to the conclusion that the value of an asankhyeya is precisely that of the number of sands in the Ganga river understood in its cosmic sense. Since the world is essentially a speck of s and in the perspective of the fixed stars, each of the grains of sand which make up the cosmic river must also be a world, a universe unto itself (p. 121). ” 99 Meanwhile, a provisional limit has to be set up to make calculations and estimation operable. Klo etzli believes that Archimedes's attempt has in essence demonstrated “a fundamental principle of the infinitesimal calculus ( p. 120),” a concern widespread in the Hellenistic world.	di-luo-a-grain-of-sand-full-draft-3.pdf
127 Suppose, for example, that someone takes all the earth seeds ( dizhong 地種 ) in the thousand-millionfold world and grinds them up to make ink powder, and as he passes through the thousand lands of the east, he drops one grai n of the ink powder no bigger in size than a speck of dust... Suppose he goes on in this way until he has finished dropping all the grains of ink made from the earth seeds... And suppose that one speck of dust should represent one kalpa. The kalpas that have elapsed since that Buddha entered extinction would still exceed the number of the grains of the ink powder by immeasurable, boundless, hundreds, thousands, ten thousands, millions of asankhya kalpas. 譬如三千大千世界所有地種, 假使有人磨以為墨, 過於東方千國土乃下一點, 大如微塵... 如是展轉盡地種墨... 一塵一劫, 彼佛滅度已來, 復過是數無量無邊百千萬億阿僧祇劫. 100 Here, the infinitesimal functions to kindle the imagination of a timeless and boundless universe. Such a mental bridge between the infinitesimal and the infinite can be built precisely because in practice, humans have attempted to measure the universe by particles. The Buddha's reckoning, therefore, is not some personal whim but has a solid scientific basis. 101 Parallel to the infinitesimal thinking is an atomic view hel d by people of the ancient world: the Buddhist cosmology tells of a final apocalypse when the world is destroyed and reduced to dust, back to its atomic state. 102 In this light, the universe can be measured by the minute and the myriad only because it is by nature a collection of these minute and myriad particles. 103 100 T9. 262: 22a; Burton Watson, trans., The Lotus Sutra (New York: Columbia University Press, 1993), 117-18. See also Kloetzli 1983, 118-19. It might not be a problem to bring up this sutra here because of its high popularity and wide acceptance in Medieval China. And it was not uncommon for Buddhist scriptures to borrow ideas fro m each other, especially considering the syncretic nature of the Huayan School. 101 Kloetzli 1983, 16, 21, “The power of mathematics which allows the astronomers to measure the motions of the heavens also enables the faithful to comprehend the theological a nd mystical implication of th ese measurements. ” This rationalizes the tireless inclusion of various numbers and numerals in the Buddhist exposition of the universe. 102 More precisely, the world is in a constant cycle of destruction and regeneration known as the “four-eons ( sijie 四劫 ),” which includes phases of formation ( cheng 成), existing ( zhu 住), decay ( huai 壞), and disappearance ( kong 空) (Huayan yuanren lun 華嚴原人論, T1886. 45: 709b). The atomic view was held by many classical philosophers including Democritus, and is still central to today's particle physics. See Kloetzli 1983, 120. 103 The atomic nature of the universe also leads to the understanding of the illusionism and evanescence of all objects and phenomena--as a mirage on the horizon and in constant tran sformation s, they are echoed by the term “ huacheng 化城,” which I render as “Mirage City” here. Huacheng is evoked in the inscription on Liang Shouqian 's repository introduced in Chapter 2 of this dissertation, probably as a me tonym for Buddhist monasteries.	di-luo-a-grain-of-sand-full-draft-3.pdf
128 Fazang's mirror hall: the art of Huayan Buddhism The extraordinary vision of the universe advocated by Huayan Buddhism must have created some difficulties for visual representation. The Flower Gar land Sutra speaks of “the mind as a skillful painter capable of picturing the myriad worlds ” (心如工畫師能畫諸世間 ), but to embody the myriad worlds in forms of architecture and art is quite a daunting task. 104 To efficiently convey the grand, nearly unrepresentable worldview, carpenters and painters had to devise a special system of visual language, which wou ld incorporate traditional motifs but also develop something new and distinct. 105 One exemplary Huayan artwork is a Five Dynasty silk painting of the “Seven Locations and Nine Assemblies ” (qichu jiuhui 七處九會 ) from the Dunhuang Library Cave (fig. 67). The composition of the painting consists of a simple grid of nine squares, each occupied by a Buddha presiding over a single-story wooden hall preaching to a group of audience, and the bottom of the painting shows a giant lotus flower con taining multiple cities (also in a grid plan) as a representation of the Flower Repository Universe. The same composition has also been found in Mogao Caves 61 and 85 (fig. 68). 106 A different example possibly alluding to the Huayan worldview is the seventh-century transformation tableau ( bianxiang 變相 ) on the north w all of Mogao Cave 321 (fig. 69). Against the extended blue sky, to the left and right of the central pagoda, two pavilions seem to be floating i n 104 T10. 279: 102a. Translation in Wang 2005, xix. 105 Wang 2005, xiii-xiv. Wang regards the creation of the transformation tableaux ( bianxiang ) in Medieval China as a “world making” process, in which a “mental topography or imaginary world” is engendered and projected onto the picture. As the structure of the “pictorial universe” was at best hinted at by the scriptures, the painters had to rely on their own judgment and creatively use what pictorial vocabulary they had (p p. 68, 75). 106 See Dorothy C. Wong, “The Art of Avatamsaka Buddhism at the Courts of Empress Wu and Emperor Shomu/Empress Komyo,” in Avatamsaka Buddhism in East Asia: Origins and Adaptation of a Visual Culture, eds. Robert Gimello, Frederic Girard, and Imre Hamar (Wiesbaden: Harrass owitz Verlag, 2012), 254-57. The qichu jiuhui are places where the Avatamsaka has been preached, according to the eighty-fascicle Avatamsaka ; alternatively, the sixty-fascicle Avatamsaka speaks of “Seven Lo cations and Eight Assemblies. ” Historical r ecords indicates tha t transformation tableaux of Avatamsaka also existed in the Tang capitals of Chang'an and Luoyang.	di-luo-a-grain-of-sand-full-draft-3.pdf
129 the air, each of their five stories occupied by two smaller wooden halls with the typical green-glazed hipped roof, a scene strongly suggestive of Sudhana's vision of the myriad pavilions. 107 One commonality shared by these examples is the repetition and m ultiplication of a single visual motif. The world thus represented is not looked through a bird's eye as in most mural paintings at Mogao (hence the panoramic view), but as if through the c ompound eyes of a bug (fig. 70). The same optical effect could be alternatively experienced by setting up multiple mirrors in a room, where they generate infinite reflections of the objects placed inside. This was attempted by Fazang, whose installation of the octagonal “mirror hall ” (jingdian 鏡殿 ) displayed a honeycomb of reflections and an unfath omable depth of space (fig. 71). 108 In Eugene Wang's interpretation, the mirroring effect has been characteristic of the “pictorial illusionism” created by the central zone of the transformation tableaux in Medieval China, which usually show the frontal image of the Buddha and his entourage against a map-like background of landscapes, cities, and spiritual beings. 109 The example from the Library Cave, on the other hand, does not seek such a contrast between the mirroring and mapping effects, but is generated by a kind of “self-reflection” into a nine-fold matrix--or indeed a mandala. Does this not 107 This tableau is thought to be a representation of Amitabha's Pure Land; see Wang 2005, 235. 108 T50. 2061: 732a-b. A total of t en mirrors were used in Fazang' s demonstration : one at each of the eight cardinal and ordinal points, one in the ce iling, and one on the ground. Other sutras mentioning the “mirror hall” or “mirror wall (jingbi 鏡壁 )” include the Zhengfa nianchu jing 正法念處經, T17. 721: 178a-184a, which speak s of Indra's “ piliuli (Sk. vaidurya ) bi 毘琉璃壁 (wall of lapis lazuli)” as mirrors reflecting one's karma and retributions; and the Shoulengyan jing 首楞嚴經 (Sk. Suramgama Sutra ), T19. 945: 133b-c, which details how a dharma-field ( Ch. daochang 道場 ) should include the installation of eight round mirrors on the sides and eight more suspended in the air. According to Wang 2005, 256-59, Sui Yangdi's Tower of Labyrinth ( milou 迷樓 ) might have included a hall of mirrors, which preceded Wu Zetian's mirror hall at the Damin ggong 大明宮 and the one at the Jianfusi 薦福寺 in Tang Chang'an. Today, at the Todaiji 東大寺, in the Lotus Flower Hall, mirrors are still hung over the altar (Wang 2005, 264), which reminds us of Bai Juyi's repository discussed in Chapter 2 of this dissertation. 109 Wang 2005, chaps. 4 and 5. Mapping and mirroring are interpreted as two contrasting features of pictorial presentation that have been brought together and juxtaposed in a single transformation tableau.	di-luo-a-grain-of-sand-full-draft-3.pdf
130 echo the endless reflections among the jewels of Indra's Net? At the Huayansi, a vista of the mirroring world is similarly brought forth by repetition and self-reflection: the north-and south-wall cabinets are almost exact mirror i mages of each other, and the long ambulatory surrounding the main altar is composed of an extended series of repe ated miniature architectural units. 110 The quintessential architectural manifestation of the mirroring world is found in the Yingzao fashi templates of the fodaozhang and bizang (see fig s. 1, 2), where miniature towers and pavilions are multiplied to set up a theater of the myriad. Conclusion The examination of the Huayansi sutra cabinets excellently problematizes--and deconstructs--current discourses on Chinese architectural hist ory. The notion that miniatures are often accurate “models” of full-scale building falls short when the particularities of miniature-making are to be investigated. Instead, one should also study miniature architecture in relation to other forms of miniatur e art (including tomb art) and the material culture of a certain historical period, ethnic group, or dynastic regime. The dual identity of the Huayansi cabinets--straddling the realms of both architecture and furniture--indicates that in terms of technolog y, miniature-making and cabinetry mutually informed and influenced each other. Such a mutual relationship engendered a common repertory of numerical conventions and decorative motifs for both realms, but it also produced hybrid woodworks whereby the struct ural integrity of both architecture and furniture is dissolved and redefined. The Huayansi example, while displaying high numerical and structural consistencies with the Longxingsi sutra case, also echoes many existing Buddhist and Daoist miniature shrines and repositories from the eleventh- 110 Only minor differences are found between the cabin ets along the south and north walls. In general, the two parts are symmetrical to each other.	di-luo-a-grain-of-sand-full-draft-3.pdf
131 to thirteenth-century China. One can trace the origin of these religious receptacles to early household cabinets and cupboards, a connection which is further supported by the Lu Ban jing and a typological study of tradi tional Japanese zushi. The size and complexity of the miniature, on the other hand, depended on the nature of its setting--public, semi-public, or private. The aesthetic value of the Huayansi cabinets has to be revealed by considering the religious signifi cance of miniature-making. The Flower Garland Sutra depicts a self-multiplying, recursive world system (Flower Repository Universe) which is often conveyed through literary tropes of Indra's Net and Sudhana's revelation inside the Tower of Vairocana. To re animate such a vision, Chinese carpenters have invented a three-dimensional visual language whereby the world of the myriad is recreated through miniaturization, multiplication, and mirroring. The Huayansi miniature provided precisely such a theatrical sta ge or backdrop around the main altar, and similar cases are found in murals and silk paintings representing the Huayan worldview. The ultimate goal was to assist visualization (an essential component of Buddhist meditation) by evoking a series of reveries and imaginations--this concerns the phenomenological aspect of miniature architecture, a topic to be further elaborated in the next chapter.	di-luo-a-grain-of-sand-full-draft-3.pdf
132 4. Miniatures in the “Dome of Heaven” The Jingtusi is located in the northeast of the Ying 應 County in northern Shan xi, about five hundred meters east of the famous Liao Wooden Pagoda. A first-time visitor would have some difficulties finding the monastery, since it lies deep in the midst of many single-story, tiled-roof traditional houses, where a network of bumpy road s and alleys spreads out rather irregularly. No street signs help to point the direction, and the entrance has the most inconspicuous appearance. Behind the gate, the Main Hall ( 大雄寶殿 Daxiong baodian) stands as the only survivor of the original monastery (fig. 72). The Main Hall looks modest from the exterior, but it features one of the most awe-inspiring ceiling in the entire history of C hinese architecture. O ften referred to as the “tiangong louge zaojing 天宮樓閣藻井 (coffered ceiling with Heavenly Palace towers and pavilions ),” this ceiling consists of a group of exquisitely crafted miniature architecture. An examination of the Jingtusi ceili ng will not only highlight the complexity of dealing with Song-Liao-Jin architecture but further illuminate the nature and role of miniature-making. The tiangong louge, though recorded in the Yingzao fashi but appeared at neither the Longxingsi nor the Huayansi, finally made its debut here at the Jingtusi, lending us the opportunity to investigate it at a close distance. Comparing the Jingtusi ceiling with the Northern Song and Liao examples, one notices that not only did the size and scale of miniatures further decrease d, but the location where miniatures were installed also shifted from furniture pieces to the ceiling. These changes dictated that while certain elements of the earlier projects could be recycled, new forms and patt erns also needed to be generated. The “dome of heaven” is an important notion and phenomenon exposed in Alexander Soper's 1947 article “The 'Dome of Heaven' in Asia”--itself a response to Karl Lehmann's “Dome	di-luo-a-grain-of-sand-full-draft-3.pdf
133 of Heaven,” a study of the symbolism of the dom e in Western architecture. 1 This term is invoked here not only as a way of engaging with the existing scholarship, but more importantly as a platform for exploring the connections between Chinese ceilings and their Central Asian--even Western--parallels. The “dome” is here interpreted as an archetype which is to be deconstructed by miniature-making and experienced phenomenologically. In addition to probable Western sources, other intellectual and technological fountains of the Chinese ceiling design should be considered, especially the pattern of jing 井 (nine-square layout) as a powerful icon and ideology in the Confucian tradition. The Tiangong Louge Zaojing (Coffered Ceiling with Heavenly Palace Towers and Pavilions) at the Jingtusi The Jingtusi ceiling consists of several groups of miniatures in or around a total of nine coffers (fig. 73). 2 The most extraordinary are the golden miniature wooden halls installed in the central octagonal coffer above the main Buddha Shakyamuni (fig. 74), while the othe r two octagonal coffers are above the east and west Buddhas (Ksitigarbha and Amitabha). The rest of the nine coffers are either hexagonal or diamond in shape, and together they form a three by three grid--a jing layout. Along the perimeter of the ceiling i s a continuous course of miniature gallery roofs covering the east, west, and north walls, forming a U-shape enclosure. Projecting from these “roofs” are eight more impressive-looking, hip-and-gable roofs which are like baldachins sheltering the eight Budd has painted on the walls (fig. 75). 1 Alexander Soper, “The 'Dome of Heaven' in Asia,” Art Bulletin 29 (1947): 225-48; Karl Lehmann, “Dome of Heaven,” Art Bulletin 27 (1945. 1): 1-27. The same topic is picked up in Steinhardt 2014, 277-81. 2 My survey of this unique ceiling has been digitized and accessible online at https://chinesearchitecture. wordpress. com/2016/02/01/jingt usi/, with one Rhino 3D model and two photogrammetry models.	di-luo-a-grain-of-sand-full-draft-3.pdf
134 Tiangong louge, the “Heavenly Palace” Modern scholarship on Chinese architecture uses the term tiangong louge almost indiscriminately to refer to any surviving example of miniature architecture, often without careful reasoning. The term is first mentioned in the Yingzao fashi, and in the 1930s it started to be identified with certain small-scale woodworks such as the Huayansi sutra cabinets, even though such identif ications are often problematic. 3 In the case of the Jingtusi, the same issue lingers: on what grounds can one identify the miniatures as the tiangong louge ? Are they similar to, or different from, the falsely-labeled tiangong louge at the Huayansi? The golden miniature halls in the central coffer are built on a bracketed substructure (pingzuo ) and encircled by red-and-green openwork balustrades (figs. 76, 77). Between the four halls are galleries with four corner tower s signified by the tips of their elevated, outstretching eaves. Each hall faces a cardinal direction, and only the one facing south comes with two side chambers and a suspended platform at the front. Six-tiered double bracketing ( liupuzuo dougong 六鋪作重栱 ) have been adopted for the halls and the substructure, and five-tiered bracketing for the galleries. Miniature Buddhas are painted inside each bay of the halls and galleries, as if to accentuate the “heavenliness” of the golden palace. The entire gro up measures about 3. 70 meters long and wide. Comparing the miniatures with the Yingzao fashi template, one discerns many significant commonalities between the two. In the text, the tiangong louge is prescribed to feature hip-and-gable roofs, substr uctures, and balustrades. 4 The complex should include galleries ( xinglang 行廊 ) and 3 See Chapter 3. 4 Yingzao fashi, vol. 1, 199-200. The tiangong louge is prescribed to be used for three types of small-scale woodworks--fodaozhang, zhuanlun jingzang, and bizang. The way the tiangong loug should be made in each case is principally the same: it should be a group of two-story buildings ranging between 50 to 72 cun (160 to 230 centimeters) in height, with additional structural features including penthouse s (fujie 副階 ) and skirting roofs ( yaoyan 腰檐 ).	di-luo-a-grain-of-sand-full-draft-3.pdf
135 corner towers ( jiaolou 角樓 ), and six-tiered double b rackets should be used for palace halls ( dianshen 殿身 ). 5 All these appear to have tallied rather well wit h the Jingtusi miniatures. However, the discrepancies are not to be neglected. In a strict sense, louge means “towers and pavilions,” that is, multistory structures, but the Jingtusi miniatures are single storied. 6 The tiangong louge is said to be a n ornam ent on top of wooden shrines and repositories, but here they are fixed in the ceiling instead. 7 Do these discrepancies rule out the Jingtusi miniatures as a type of tiangong louge ? To solve this issue, one needs to consider not just formal features but mor e importantly the scale of miniaturization. My survey indicates that the cai of the miniatures (as well as the entire ceiling) is approximately 2. 78 by 1. 85 centimeters. 8 This is larger than the theoretical value in the Yingzao fashi. 9 Note, however, that this deviation is merely numerical, but in terms of scale, the miniatures have been proportioned to a degree that they share the same cai with the ceiling coffers, just as proposed 5 Ibid. The tiangong louge has never been defined in the Yingzao fashi since perhaps such a definition was thought irrelevant in a technical manual. Nonetheless, six building types/par ts are said to be included: except dianshen, xinglang, and jiaolou, there are also tea houses ( chalou 茶樓 ), wings ( jiawu 挾屋 ), and gabled porches ( guitou 龜頭, lit. tortoise head). The brackets they apply range from four to six tiers. 6 In fact, structures on the pingzuo could be recognized as ge, which basically means any building “suspended” or elevated from the ground. See Ma Xiao 馬曉, Zhongguo gudai mulouge 中國古代木樓閣 (Wooden towers and pavilions in ancient China) (Beijing: Zhonghua shuju, 2007). The corner tow ers of the Jingtusi tiangong louge are mostly hidden from view but they should indeed indicate multistory buildings. 7 Miniature doors and balustrades should also be used for a type of ceiling coffer known as the xiaodouba zaojing 小闘八藻井 (miniatur e eight-ribbed vaulted coffer), which might have been a precursor for the later installment of the tiangong louge in ceiling coffers. See Yingzao fashi, vol. 1, 168-69. 8 The value of fen is calculated by measuring a wooden bearing block ( jiaohudou 交互斗 ) found behind the west Buddha, probably fallen from the west or northwest coffer. It is not necessarily the exact value adopted in the ceiling but should have fairly accurately reflected the designed value. This calculated value has also been corroborated with the architectural drawings in Liu Dunzhen, ed., Zhongguo gudai jianzhushi 中國古代建築史 (History of premodern Chinese architecture) (Beijing: Zhongguo jianzhu gongye chubanshe, 198 0), 248. 9 The value of fen translates into a cai of 0. 87 by 0. 58 cun, about one and a half times larger than the theoretical values in the Yingzao fashi--0. 6 by 0. 4 cun for xiaodouba zaojing and fodaozhang, and 0. 5 by 0. 33 cun for jingzang.	di-luo-a-grain-of-sand-full-draft-3.pdf
136 in the Yingzao fashi. 10 The scale of the miniatures not only qualifies them as tiangong louge, but also testifies to scholars' speculation that they date from 1124 just as the Main Hall, an issue to be elaborated later in this chapter. Xiaodouba zaojing, the miniature octagona l ceiling coffer As noted above, the miniature halls, gallery roofs, and ceiling coffers all share the same cai. 11 This can be perceived from the surprising uniformity of some one thousand bracket sets installed in the ceiling. Such uniformity does not mean that each bracket set looks exactly the same; instead, a controlled diversity has been achieved by switching between various schemes of bracketing. For instance, while the gallery roofs are mostly supported by six-tiered bracket sets, a unique, seven-tiered and fan-shaped set is installed in the southwest corner of the ceiling (fig. 78). Brackets used for substructures differ from those under the eaves, and the highest rank of bracket ing belongs to the eight-tiered double bracket s inside the east coffer, featuring double twig arms ( miao 杪) and triple uplifting lever arms ( shang'ang 上昂 ) (fig. 79), a scheme never found in surviving wooden buildings. While diversity allowed miniaturists to highlight certain p arts of the ceiling, it was uniformity--and the underlying principle of modulari zation and standardization--that assured such a sophisticated project to be ever accomplished with efficiency and quality. The Yingzao fashi prescrib es that ceiling coffers and tiangong louge can share the same cai. This is e specially true for the xiaodouba zaojing (miniature octagonal coffer s), which resembles its larger counterparts but is also 10 See scaling schemes in Chapter 2 of this dissert ation. More specifically, the cai for both the tiangong louge of miniature shrines and the miniature octagonal coffer should be the same ; this fact is not explicitly pointed out in the Yingzao fashi but stated separately (vol. 1, 168-69, 199). Carpenters during the eleventh and twelfth centuries seem to have adhered to some “hierarchy” of small-scale woodworking: in this hierarchy, the tiangong louge was of the smallest scales among all miniatures. See also Chen 2010, 186-87. 11 This needs to be further verified by measurement, which has not being done (by me or others) and would have to involve electronic surveying equipment because of the small scale.	di-luo-a-grain-of-sand-full-draft-3.pdf
137 different in certain ways. According to the Yingzao fash i, a regular ceiling coffer is 256 centimeters square, while a miniature coffer is 154 centimeters square. 12 Equipped with an octagonal well and a cupola, it is often inserted into a penthouse ceiling or inside a miniature shrine. Most interestingly, one ought to “attach miniature doors, windows, and balustrades” to the side panels of the miniature coffer (see fig. 4), a practice which certainly foreshadowed the full flowering of miniature-making in the ceiling. 13 At the Jingtu si, three octagonal coffers--located at the center (C), the east (E), and the west (W) of the ce iling--are present (see fig. 73). Each is a superimposition of a diamond shape inside a square, formin g four triangles at the corners (fig. 80). Inscribed in the diamond is an octagon creating four additional, smaller triangles. Within the octagon is a circle surrounded by a pair of writhing dragons. Along the squares, diamonds, triangles, and octagons are array ed densely-arrayed bracket sets (the scheme of which varies from coffer to coffer), except for the triangles where golden dragons and phoenixes are engraved on the panels. Coffer C is the only one that comes with an additional level of tiangong louge circling the edge of the coffer, as if to maximize its centrality and importance. 14 12 Yingzao fashi, vol. 1, 165-69. The regular octagona l coffer (douba zaojing 闘八藻井 ), usually installed inside a palatial hall and in front of screens and partitioning walls, is composed of three parts from bottom to top: a square well ( fangjing 方井), an octagonal well ( bajiaojing 八角井 ), and an eight-ribbed smal l “dome” or cupola known as the douba 闘八 (lit. converging the eight ribs). As small-scale woodworks, the regular coffer should use a cai measuring 1. 8 by 1. 2 cun, and the miniature coffer 0. 6 by 0. 4 cun. When making a fodaozhang, one ought to apply the sam e cai to both the tiangong louge and the octagonal ceiling coffer(s) it has; this is also the cai assigned to all miniature octagonal coffers. See Yingzao fashi, vol. 1, 194-95, 199. The zaojing is also mentioned in large-scale w oodworking : in Yingzao fashi, vol. 1, 75, the zaojing used in the interior of palatial halls is required to adopt a cai of 4. 5 by 3 cun, without giving specifications regarding the structure. This zaojing is probably different fr om the other two and might have still retained some structural functions. See below. 13 Ibid., vol. 1, 168-69. This introduction of miniature elements has been noted in Soper 1947, 246. 14 This seems to suggest that miniature buildings are even “grander” tha n the eight-tiered brackets, the highest-rank bracketing.	di-luo-a-grain-of-sand-full-draft-3.pdf
138 The three coffers, considering their cai, are much closer to xiaodouba than regular octagonal coffers, even though their overall sizes are significantly larger than what has been prescribed. While the octagonal element is all the more prevalent, we also see the interplay between the square and the diamond, which forms what Soper refers to as the “square-and-diamond” pattern, a point to be further explored later. Here it is to be e mphasized that the xiaodouba opened the gate for miniature architecture to be introduced and incorporated almost effortlessly into the ceiling structure, allowing the most extravagant display even in a limited, moderate-size interior space. Jing, the magi c square, and ceiling compartmentalization As the schematic plan (see fig. 73) shows, the squarish ceiling is divided by beams, joists, and panels into nine coffers (compartments). Such a compartmentalization has created a three b y three grid, or a jing layout resembling the magic square. This specific layout corresponds to the thr ee-bay-wide and three-bay-deep Main H all and is further accentuated by the miniature baldachin roofs along the walls. Each coffer is distinctive: in addi tion to the three octagonal ones (C, E, W), two more geometric shapes--the hexagon and the diamond--have been incorporated into the side and corner coffers (fig. 81). A sense of rhythm and control is aroused by the subtle differen tiation, by the repetitive yet nuanced motifs and details integrated into the se coffers. The way the ceiling was compartmentalized at the Jingtusi was unprecedented and perhaps to this day remains a singular case where a total of nine coffers are present. Wooden halls of the eleventh-to thirteenth-century China usually came with one ceiling coffer (at most three, in a few cases), while the use of coffers was altogether banned for residences of commoners and low-rank officials. 15 The much more moderate, offi cially approved treatment of the ceiling was to cover the 15 In real practice, the application of zaojing was strictly moderated by sumptuary law. The Northern Song Yingshanling 營繕令 (Statute s on building and repairing activities, promulgated in 1029) decree s that the commoner's house is not	di-luo-a-grain-of-sand-full-draft-3.pdf
139 interior space by a type of checkerboard ceiling called pingqi 平棊--a lattice structure with decorated panels but without bracket s or domes--or an even simpler checkerboard, ping'an 平闇, which was devoid of any ornamentation whatsoever. 16 The ceiling design at the Jingtusi would have been a serious violation of sumptuary law had it not been endorsed, or more likely patronized, by the imperial court itself. Clearly, the ceiling was not designed alone but together with the sculptures and murals inside the hall. While the three main statues are “sheltered” under the three octagona l coffers, the murals are covered by the encircling gallery roofs. 17 Such a configuration reminds us of the U-shaped enclosure in the Huayansi library hall ; here, in a similar light, it reinforce s the ambulatory space around the central altar and encourage s circumambulation around the Buddhist triad. Moreover, eight unidentified, painted Buddhas--three on the east wall, three on the west, and two on the north--have been perfectly al igned with the eight baldachin roofs above, as if to suggest that the roofs were also part of the painted scenes (fig. 82). Viewed from below, the coffers appear to be “floating” above the gallery roofs. The miniature halls, the interlaced ceiling joists, and layers and layers of tiny brackets all add up to increase depth of the ceiling. allowed to have double bracketing or zaojing. This was modelled after the Tang sumptuary law and had very likely been adopted in the Jin. See Tianyige cang Mingchaoben Tianshengling jiaozheng, 2006. 16 Yingzao fashi, vol. 1, 1 63-65. Pingqi as a checkerboard ceiling ( qi 棊 literally means checkers or checkerboard) is similar to a paneled ceiling today--its segmentation is always rectilinear, without special geometric shapes such as diamonds, circles, or octagons, and the panels have a shallower recess which is structurally simplistic and less ornamental. Its original function s are said to be “catching the dust (chengchen 承塵 )” and concealing an unrefined roof frame. The three types of ceiling are usua lly used in combination, with zaojing occupying the center and the pingqi /ping'an along the periphery, such as in the Sutra Library at the Huayansi. The Jingtusi is a rare case where some of the side coffers and the borders between different coffers apply arrays of decorated ceiling panels identifiable as pingqi. 17 The particular structure of the gallery roofs deserves some further explanation. What makes them different from the tiangong louge in the central coffer is the fact that they are suspended in mi d-air: with no columns below, the roofs are simply projected from the vertical walls, forming a flat, paneled soffit underneath. The soffit, somewhat comparable to a checkerboard ceiling in spite of its linearity, again exemplifies the principle of modular design. A large panel of the soffit is twice as long and wide as a small panel (64 as opposed to 32 centimeters per side ), and the entire ceiling corresponds to forty-five small panels lengthwise and thirty small panels crosswise (7,800 by 5,200 fen).	di-luo-a-grain-of-sand-full-draft-3.pdf
140 Miniature-making in Jurchen-Jin Material C ulture The dating of the Jingtusi is primarily based on the oldest surviving gazetteer of the Ying County:18 The Jingtusi is located northeast to the administrative headquarter of the prefecture. It was built in the second year of the Tianhui Period of the Jin, by the monk Shanxiang according to an imperial decree. In the twenty-fourth year of the Dading Period (1184), it was repaired by the monk Shansong. 淨土寺 : 在州治東北. 金天會二年, 僧善祥奉敕創建. 大定二十四年, 僧善聳重修. 19 A map from a much later source (fig. 83) shows that, before the iconoclasm in the Cult ural Revolution, the monastery was composed of two moderate-size cloisters. The west cloister included the gate, a relic stupa, the Hall of Heavenly Guardians, the bell and drum towers, the east and west side halls, and the Main Hall, while the east cloist er had a meditation hall, utility rooms, an image hall, and a library. 20 In 1969, all buildings but the Main Hall were destroyed. 21 18 In my fieldwork I did not find any on-site inscriptions verifying the 1124 date. There are, however, three Ming inscriptions attached to the underside of the ceiling joists in the Main Hall. The first inscription goes, “Repaired on the twenty-ninth day, gengxu, of the fourth month, jisi, in the fifth year, jiaxu, of the Jingtai Period of the Great Ming dynasty (1454), by Tang Jian, Commander-in-chief stationing at Yingzhou, and the abbot of this monastery 維大明景泰伍年歲次甲戌四月己巳二十九日庚戌守備應州都指揮僉事唐鑒本寺住持... 重修. ” The secon d inscription: “Beautified and repaired on the twenty-sixth day, wuwu, of the third month, bingchen, in the nineteenth year, guimao, of the Chenghua Period of the Great Ming dynasty (1483), by Zong Yue, monk in charge of repair works of this monastery 維大明成化十九年歲次癸卯三月丙辰二十六日戊午本司修造僧宗鉞... 粧修. ” The third inscription: “Repair work initiated with a fund raising on the twelfth day of the fourth month in the seventh year, jiaxu, of the Chongzhen Period of the Great Ming dynasty (1634) and completed on an auspicious mi d-summer's day in the ninth year, bingzi (1636), financed by Buddhist believers of the entire prefecture 維大明崇禎七年歲次甲戌四月十二日募緣興建至九年歲次丙子仲夏吉旦合州眾善施財重修. ” According to Steinhardt 2003, 79, the earliest inscription carries the date of 1184, when the hall had its fi rst major repair. I have not found this inscription perhaps due to its location. Additionally, the plinth of a broken stone relic stupa in the monastic courtyard bears the date of 1040. The hall was also repaired in the Qing. S ee Ma Liang 馬良 et al, eds., Yingxianzhi 應縣志 (Gazetteer of the Ying County) (Taiyuan: Shanxi renmin chubanshe, 1992), 557. 19 Yingzhouzhi 應州志 (Gazett eer of the Ying Prefecture), eds. Tian Hui 田蕙 and Wang Yourong 王有容, first published in 1599, reprint in 1984, 60. 20 Ma 1992, 557. The unpa ged map inserted into the book might have been based on earlier records. 21 Ibid., 556, 738. In 1966, the statues, scriptures, and scroll-paintings from the Wooden Pagoda and the Jingtusi were confiscated and burned in an open field by student organizations.	di-luo-a-grain-of-sand-full-draft-3.pdf
141 The founding date of 1124 has been reiterated in later revisions and expansions of the local gazetteer and incorporated into t he Comprehensive Gazetteers of Shanxi. 22 Among the many reiterations, Wu Bing's 吳炳 compilation, published in 1769, provides us a personal observation of the architecture and an intriguing legend about the origin of the monastery : Under the roof rafters of t he Buddha hall, there are wooden panels carved with dragons and phoenixes installed in the space between [the ceiling joists?]; the luster of their gold and jade-blue colors illumines the interior, as the paint has not yet started flaking off. The design o f Jingtusi is at odds with all other mo nasteries, and senior townsfolk say that the hall was formerly the worshipping hall of Emperor Mingzong's (Li Siyuan 李嗣源, r. 926-33) ancestral temple. The rear hall of the original temple now lies outside the northern city walls in dilapidation due to a la ter moving of the city walls [southward] which cuts across the site of the temple. I checked the History of the Five Dynasties and realized that in the twelfth month of the second year of the Tiancheng Period (927), Emperor Mingzong bestowed posthumous tit les of emperors and empresses upon his progenitors of the past four generations and established a temple in Yingzhou. Hence there must have been such a temple in Yingzhou, and the rumor I heard might not have been groundless after all. 佛殿榱桷之下, 以木板雕鏤龍鳳, 嵌置其間, 金碧照耀, 尚未剝落. 其制異於他寺, 故老傳系明宗祖廟前室. 寢殿在今北城外, 後移建城垣, 隔斷故址, 遂廢. 考五代史, 明宗天成二年十二月追尊四代祖考皆為皇帝, 妣為皇后, 立廟應州, 勢必實有其地, 所言或非訛傳. 23 The Jingtusi Main Hall was considered a unique design even as early as the eighteenth century, as it was “at odds with” traditional monasteries, and its “oddity” was largely perceived from the ornamental panels below the roof, i. e., the ceiling. Wu Bing suggests that the ceiling might have preserved some features of the Later Tang temple--the lavishly painted dragons and p hoenixes seemed to be proudly reminding the onlookers of the past glories of this building and the eminence 22 Shanxi tongzhi 山西通志, 169. 50a. The same information is reiterated in Guangxu Shanxi tongzhi 光緒山西通志, eds. Zeng Guoquan 曾國荃 et al, 57. 43b. See Ma 1992, 769-71, for a list of the different versions of Yingzhouzhi : One edited by Xue Jingzhi 薛敬之, published i n 1488 (only preface remains); One edited by Tian Hui and Wang Yourong, published in 1599 (the oldest surviving version); One edited by Xiao Gang 肖綱, published in 1726; One edited by Wu Bing 吳炳 and examined by Dai Zhen 戴震, published in 1769 (highest qualit y); One edited by Tang Xuezhi 湯學治, published in 1879. The earliest gazetteers were compiled by some Song scholars and though now lost, they might have been available for a few Ming editors. 23 Yingzhou xuzhi 應州續志, ed. Wu Bing, 4. 5b.	di-luo-a-grain-of-sand-full-draft-3.pdf
142 of its owner. 24 Despite the deceiving modesty of the building from the outside, the ceiling design, with its nine luxuriously ornamented coffers, cou ld not have been executed without imperial sanction. 25 The two monks associated with the Jingtusi in the gazetteers, Shanxiang and Shansong, are not listed in major hagiographies. With the lack of information, it becomes difficult to know exactly under whos e decree the monastery was founded, and whether or not it had anything to do with a specific Jin emperor. The given date, “the second year of the Tianhui Period of the Jin,” falls in the reign of Emperor Taizong (Wanyan sheng 完顏晟, r. 1123-35), a younger brother of the founding emperor Aguda 阿骨打. Was Taizong the one who ordered Jingtusi to be built? If he was, the imperial patron o f the Jingtusi would be the same person who obliterated the Northern Song forces, plundered Dongjing, and kidnapped Huizong and his son--the last two emperors of the Northern Song--to the Jin capital, all of which to take place three years after the comple tion of the monastery. However, the political landscape of the year 1124 was a lot more complicated: Aguda had been dead only for months and the Jin had not yet conquered north China; the Northern Song was still negotiating with the Jin to retrieve its lon g-lost northern territories, historically known as the Sixteen Prefectures of the Yan-Yun 燕雲 Region, including Yingzhou; and the last emperor of the Liao, 24 Despite the lack of so lid evidence, scholars have determined that the ceiling is actually a Jin design. “According to records [the names of which have not been specified in the text], the ceiling coffers of the Main Hall were Jin originals... The tiangong louge of the ceiling cof fers, except for the group of towers and pavilions on the southwest corner and some bracket sets on the northwest corner of the galleries (which underwent later repair and replacement), have generally kept their original forms during the Jin dynasty. ” See Ma 1992, 557-58. This judgment, as it turns out, corroborates well with my earlier argument that the tiangong louge miniatures were contemporaneous with the Yingzao fashi. This dating is also consistent with the historical develo pment of miniature woodworking in the twelfth century; see below. 25 The connections between Yingzhou and the Li family of the Later Tang were reflected by local anecdotes. For instance, Shanxi tongzhi, 165. 41b, records that Li Keyong, the father of Li Siyu an, was born a “divine boy clad in a golden armor” out of the walls of the Wenchangci 文昌祠 in Yingzhou, where his mother prayed to the gods. See also Ma 1992, 646, for an anecdote about Li Keyong and the Wooden Pagoda.	di-luo-a-grain-of-sand-full-draft-3.pdf
143 Tianzuodi 天祚帝 (r. 1101-25), still had hopes to retaliate the Jin armies for their intrusions and keep his dynasty alive. From 1123 to 1127, Yingzhou was not in the firm grasp of any regimes. 26 An inquiry into the official histories of the Northern Song, Liao, and Jin unfolds incredibly rich narratives of the convoluting and tumultuous events occurred aroun d Yingzhou during this period. The Songshi 宋史 records that in 1123, “the Khitan general, Su Jing, surrendered Yingzhou to the Northern Song court,” and the Jin launched an immediate attack against the town, though we are not told if the siege was successful. 27 In the same year, Tianzuodi and his dem oralized ten-thousand-men army were fleeing from the tightening pursuit of the Jin army, who chased them from today's Inner Mongolia to Yingzhou and managed to capture the majority of Liao princes, princesses, and other imperial family members at the Liao camp. 28 From the perspective of the Jinshi 金史, the final stage of the struggles between the Jin and the Liao was in Yingzhou, where Tianzuodi constantly sought refuge to restore the strength of his forces, a stronghold that the Jin failed to grasp after multiple attempts. 29 It was not until the sec ond month of 1125 that Tianzuodi, in his last escape “sixty li east of the new town of Yingzhou ” (應州新城東六十里 ), fell into the hands of the Jurchens, an event signifying the demise of the Liao. 30 It is therefore hard to pinpoint the identity of the imperial pat ron of the Jingtusi. It will be equally hard to label Yingzhou as an undisputable territory of any state in 1124, when the borders became highly fluctuating and could be easily crossed and recrossed overnight as the loyalty of the 26 Ma 1992, 456, 717. 27 Songshi, 90. 19b. 28 Jinshi, 2. 28a, 74. 12b-13a; Liaoshi 遼史, 29. 7b-8a. 29 Jinshi, 68. 7a, 74. 13a. 30 Liaoshi, 30. 1b; Jinshi, 3. 7a, 76. 18a.	di-luo-a-grain-of-sand-full-draft-3.pdf
144 military commanders alter ed. 31 In this light, can one still identify the Jingtusi as a Jin structure? The statement given in the gazetteers was clearly written in hindsight--the Jingtusi was listed as a Jin monastery because Yingzhou later came under the Jin's control even though i t had remained a highly contested area in 1123 and the years immediately afterward. It is more accurate to say that the Jingtusi was built at a particular historical moment when the three regimes were in a total clash for political and military superiority. Regarding architectural style and technique, Jingtusi and its ceiling were largely created as a product of the local culture and tradition rather than an overt expression of the ambition or vision of any particular dynasty. Still, as will be unraveled below, the local woodworking tradition was never free of the influences of imperial ideologies, and it would not have taken long for local traditions to constitute, and become identified with, a dynastic culture full of distinctive, exciting characteristics. Characteristics of Jin architecture: a revision More than sixty wooden structures dated to the Jin dynasty now still stand in Shanxi, the rest few existing in the provinces of Hebei, Henan, Shaanxi, and Shandong, though we are not certain how many of the m still carry miniature woodworks that are comparable to our case. 32 This focus on full-size structures can be seen from the restoration projects of another Jin Buddhist monastery in northern Shanxi--the Chongfusi 崇福寺 in the Shuo 朔 County, of which the Amit abha Hall 31 Also, during 1122-1124, several members of the Yelu clan claimed to be emperors at different time planning to overtake Tianzuodi's place ; their regimes are known as the Northern Liao and Western Liao in history. For the issue of border crossing and the loyalty of military commanders in the Liao, see Naomi Standen, Unbounded Loyalty: Frontier Crossings in Liao China (Honolulu: University of Hawaii Press, 2007). 32 A number of representative Jin wooden structures is listed in Steinhardt 2003, 114, n. 22. Also see Yang Zirong 楊子榮, “Lun Shanxi Yuandai yiqian mugou jianzhu de baohu 論山西元代以前木構建築的保護 (On the reservation of pre-Yuan wooden architectur e in Shanxi),” Wenwu shijie (1994. 1): 62.	di-luo-a-grain-of-sand-full-draft-3.pdf
145 (Mituodian 彌陀殿, d. 1143) is often regarded as the epitome of Jin architecture. 33 The miniature woodwork in the ceiling of the Amitabha Hall (fig. 84), though only a small fraction of it has remained, suggests no less virt uosity and imperial magnificence than the tiangong louge at Jingtusi, but any illustrations or explanations of this woodwork are totally lacking in either the restor ation reports or scholarly works. Regarding the “origin” and sources of Jin architecture, scholars stress that they expressed a strong tie and affinity to Han-Chinese architecture and culture, especially to the “degraded” culture of the late Northern Song and the Southern Song, whereas little of the Jin's own culture had any impact on building activities. 34 This affinity was most clearly exposed by structure and technique, which showed a great degree of conformity to the Yingzao fashi ; the same conformity was sometimes mixed with an uncertainty and confusion of style an d scale, leading to a sort of unidentifiable yet palpable “distinctiveness” of Jin architecture. 35 In terms of scale, the size of Jin buildings were modest in general (the larger ones were often rebuilt from Liao originals), and the carpenters' good sense o f proportion and scale was failing as they swayed between Liao and Song traditions. There were no enthusiasms for monumentality. Instead, a Jin carpenter turned inward and was more sensitive to details. His works have been criticized as being more decorati ve than symbolic, more 33 Chai Zejun 柴澤俊 and Li Zhengyun 李正雲, Shuoxian Chongfusi Mituodian xiushan gongcheng baogao 朔縣崇福寺彌陀殿修繕工程報告 (Reports on the restoration projects of the Amitabha Hall of the Chongfusi in the Shuo County) (Beijing: We nwu chubanshe, 1993). The restoration projects started with systematic measuring and numbering of each wooden members. The small-scale woodwork inside the hall, however, was not included. 34 Takeshima 1944, 8-9; Sickman and Soper 1984, 458-59. 35 The observa tion of the conformity and “distinctiveness” is made in Takeshima 1944, 15.	di-luo-a-grain-of-sand-full-draft-3.pdf
146 conservative than innovative, focused more on entertainment than on imperial or religious visions. They were the outcome of complacency, not revolution. 36 These evaluations have exposed the seemingly contradictory qualities of Jin arc hitecture: the wooden structures, previously understood as “so lacking in architectural challenge, creativity, inspiration, or symbolism,” have presented to us one of the most stunning ceilings in history. 37 Was Jin architecture Chinese or non-Chinese (anot her way to put it is “Jurchen or non-Jurchen”)? Was it creative or non-creative? Why was there a contrast between the boring exterior and the exciting interior? The same questions drive us to ponder upon the potential issues of the existing evaluations: seeking the expression of the Jurchen identity in Jin architecture can be frustrating, while focusing on large buildings alone often leads to a biased assessment of the architectural achievements of the Jin. The Jingtusi ceiling was a crucial link in the his tory of Chinese (including Jin) architecture. It was the epitome of an era marked by bolder relinquishing of monumentality, the diminishing of cai, the advance to even smaller scales, and the exploration of the depth of interior spaces. It was a further de velopment in the expressiveness of miniature architecture after the experiments in the Northern Song and Liao, and was produced with a proliferation of other miniature motif s and objects in literature, masonry, paintings, and ceramics. One can justifiably regard China under Jurchen rule as an age of “introverted” architecture, when the center of focus shifted from a boasting facade (as in the Northern Song and Liao) to the extremely sophisticated details--or indeed miniatures--of the interior. Such a shift of focus was due to the nature of the building material and technology available as well as the conquerors' self-consciousness of their Jurchen ethnicity. 36 Steinhardt 2003, 86-110. Steinhardt stresses that the Jin were not innovators of the increasingly exquisite interior design, and that the detailed ornamentation bespoke a “lack o f enthusiasm for” monumentality and symbolism. See also Steinhardt 1997, 236-37. 37 Steinhardt 2003, 80.	di-luo-a-grain-of-sand-full-draft-3.pdf
147 Miniature theaters Miniatures not only appeared in wooden structures above ground but were also ind ispensable elements in Jin tombs. The burial chamber of a Jin tomb was typically sculpted in simulation of a wooden residential hall or courtyard, with lifelike but downsized brackets, paneled doors, openwork balustrades, roof tiles, and most curiously min iature theater stages, all carved out of stone. 38 Household furnishings such as chairs, tables, foldable screens, basin stands, and so on were similarly made in miniature forms, some in high reliefs, and some as free-standing, three-dimensional models. The mini theaters underground have long intrigued scholars. A conscious procedure of proportioning and downscaling was certainly embraced, and the shocking resemblance between the overall burial chamber and real wooden structures could have only derived from a rigorous application of woodworking formulas. What is puzzling, rather, is the purpose of miniaturization: why did people place theaters in tombs and fill them with figurines of actors and musicians? In Houma 侯馬 Tomb 1 (d. 1210), a mini theater with five actor figurines is placed rather awkwardly on the roof that shelters the image o f the deceased couple (fig. 85). In one of the Macun 馬村 Tombs (ca. 1100), a theater stage is implied by the sculpted balustrades ( goulan, a term also denoting theaters in Nort hern Song literature ) (fig. 86). Scholars argue that these theaters were meant to entertain the dead in the afterlife, to allow the continuation of the pleasure they had found in drama, and even to transform the dead into actors and actresses themsel ves. 39 38 A recent study of the architecture of these tombs is Wei-cheng Lin, “Underground Wooden Architecture in Brick: A Changed Perspective from Life to Deat h in 10th-through 13th-Century Northern China,” Archives of Asian Art 61 (2011): 3-36. 39 See, for instance, Shi Jinming 石金鳴 and Hai Weilan 海蔚藍, eds., Shengsi tongle: Shanxi Jindai xiqu zhuandiao yishu 生死同樂 : 山西金代戲曲磚雕藝術 (Theater, life, and the afterlife: to mb decor of the Jin dynasty from Shanxi) (Beijing: Kexue chubanshe, 2012); Hong 2011, 75-114.	di-luo-a-grain-of-sand-full-draft-3.pdf
148 Underlying the emergence of miniature theaters in tombs was the popularity of drama during the Northern Song and Jin t imes. 40 Theaters were places where historic, romantic, and religious plays were performed and watched, in pu blic as well as private s ettings. They invited the audience to a virtual world where real-life anxieties and ambitions were temporarily cast aside so that the joys and pains of an imagined life could be savored in an almost dream-like state. This virtual realm was where theaters a nd miniatures became connected: they both prompted contemplations of matters of death and dream. The viewing experience they provided was illusory, dramatic, and oneiric; they were more evocative than representational, and the languages (visual or verbal) they used were meant to be a mbiguous and suggestive. In this light, miniature theaters were never direct depictions of real wooden stages. On the one hand, miniaturists had to pay careful attention to form and size in order to achieve a certain degree of “ realism. ”41 On the other hand, miniature th eaters were placed in tombs to destabilize any sense of reality. Often found alongside pictorial representations of Daoist immortals, Confucian paragons, and Buddhist icons, the theater was installed as if to enact a “deliverance play ” (dutuoju 度脫劇 ) in the underground to emancipate the soul of the deceased. 42 In this sense, the particular form of the miniature mattered little as long as it reminded us of the theatrical, illusory nature of life and death; and such an epiphany was to be bestowed by the miniaturized details of an enclosed interior--rather than an exterior--space. 40 See Chapter 1 of this dissertation. 41 Robert Maeda, “Sung, Chin, Yuan Representation of Actors,” Artibus Asiae 41 (1979. 2/3): 148. This realism is referred to by Maeda as “a kind of institutionalized realism (less breathtaking than at Pai-sha) that may have been a product of both a conventional popular taste and a deeper realism once popularized by Hu izong's court artists. ” 42 Ellen Johnston Laing, “Chin 'Tartar' Dynasty (1115-1234) Material Culture,” Artibus Asiae 49 (1988-89. 1/2): 81-82, 117. A pithy exposition of the “deliverance plays” is in Idema and West 1982, 305-08. See also Shen 2012.	di-luo-a-grain-of-sand-full-draft-3.pdf
149 Ruled-line painting Miniaturization in the Jin further extended to the realm of painting. The best example to illustrate this point is the mur als of the Manjusuri Hall (d. 1167) at the Yanshansi 巖山寺 in Fanzhi 繁峙, Shanxi. The murals have been compared by scholars to the famous scroll painting of the Qingming shanghe tu 清明上河圖 (Along the river during the Qingming Festival), as both include scenes of the imperial palace, the wine shop, the b ridge over water, the watermill, and ramparts with parapets, showing the activities in and around a bustling metropolis. 43 Unlike the Qingming scroll, however, the Yanshansi murals have taken on an overt Buddhist theme as indicated by the cartouches and many Buddhist motifs--the Buddha and his holy attendants, the wafting clouds and mists, the haloes and radiating light, the jumping flames of fire, etc.--inserted into a secu lar-looking background (fig. 87). Another discrepancy is th at the murals have been “unfolded” along the walls of an interior space, which grants an immersive, three-dimensional viewing experience, and is itself a “backdrop” of the main altar. In fact, in terms of theme and function, the murals are in many ways com parable to the Jingtusi ceiling. The head painter in charge of the murals was a certain Wang Kui 王逵 (1100-?), a former imperial painter at the Jin court. Wang Kui must have known or studied the masterpieces of Northern Song paintings before the fall of the capital Dongjing to the Jin in 1 127, since his murals display a strong stylistic affinity to the works produced in Huizong's imperial painting academy. 44 Indeed, the legacy of Northern Song paintings, especially ruled-line paintings ( jiehua ), has been 43 Patricia Karetzky, “The Recently Discovered Chin Dynasty Murals Illust rating the Life of the Buddha at Yen-shang-ssu, Shansi,” Artibus Asiae 42 (1980. 4): 245-60; Laing 1988-89, 76. 44 Fu Xinian 傅熹年, “Shanxisheng Fanzhixian Yanshansi Nandian Jindai Bihua zhong suohui jianzhu de chubu fenxi 山西省繁峙縣巖山寺金代壁畫中所繪建築的初步分析 (A preliminar y analysis of the architecture painted in the Jin-dynasty murals of Yanshansi in the Fanzhi County of Shanxi Province),” in Fu Xinian jianzhushi lunwenji 傅熹年建築史論文集 (Beijing: Wenwu chubanshe, 1998 ): 307-11.	di-luo-a-grain-of-sand-full-draft-3.pdf
150 warmly embraced and inherited in this case. 45 The buildings in the murals were meticulously drawn, showing every structural and ornamental details possible. They were executed in such a precise manner that the architectural historian Fu Xinian has been able to “reconstruct” the building comple x on the west wall, which he argues to have been modeled after the Jin--and ultimately Northern Song--imperial palace (fig. 88). 46 When painting architecture, calculation was an important and inevitable task, as it was needed to de termine the correct proportioning of each part of the building. It is not surprising that many excellent ruled-line painters received some architectural training or participated in building activities themselves, such as Guo Zhongshu (see Chapter 1 ) and Li Song 李嵩 (fl. 1190-1230), who applied their knowledge of the carpenter's line and ink-mark to the painter's brush. Like woodworking, ruled-line painting was an activity that “required discipline and infinite patience:” brush-strokes had to be executed in e ven widths to avoid ambiguity,47 not unlike in modern architectural drawings where a system of lineweight control is enforced. Arguably, the rigorous method applied in ruled-line painting led to the rigidity of artistic creation. Art collectors and critics in history noted that a ruled-line painting could appear “lacking” in talent, ingenuity, and vitality, because of too much borrowing from practices of craftsmanship which went against the ideal of a free-spirited, intelligent literati-painter. For this rea son, jiehua has been historically evaluated as secondary to figure painting and landscape painting, which were 45 See Chapter 1 of this dissertation. Their accura cy was due to the particular drawing tools--jiechi 界尺 (a type of ruler, hence the name jiehua, or jiechihua ), the compass, and the square--used by painters to assist necessary measuring. 46 Fu 1998, 294-98. The plan Fu reconstructed retains some elements of the ideal city plan in terms of the overall shape, the centrality, symmetry, the orthogonal avenues, and the location and number of the gates, but it also in many ways differs from the ideal model--lacking a conspicuous jing layout but being a small enclo sure in a larger one. See below. 47 Robert Maeda, “Chieh-hua: Ruled-line Painting in China,” Ars Orientalis 10 (1975): 125-26, where Guo Zhongshu is mentioned. Also see pp. 129-30, 134, regarding the high demand in personal skill while painting jiehua. A mo re recent study of jiehua is Anita Chung, Drawing Boundaries: Architectural Images in Qing China (Honolulu: University of Hawaii Press, 2004).	di-luo-a-grain-of-sand-full-draft-3.pdf
151 created by the free hand and with good artistic sense. Despite the generally unappreciative attitude of the Chinese literati toward ruled-line pa inting, to modern historians, the se paintings are of immense value precisely because the rigorous “language” they applied have turned them into fairly reliable visual archives. However, it is also to be noted, as much as miniatures were not replicas of full-size structures, a painter of wooden buildings, confronted by the hurdle to represent a three-dimensional object on a two-dimensional surface as comprehensively as possible, had to incorporate the use of diagonal lines, the change of perspectives, and th e techniques of shading, foreshortening, and distortion. Scholars refer to the Yanshansi murals as the “miniaturist style” because on the one hand, the minute details of architecture--the multiple tiers of bracket arms and bearing blocks, the layers of round and square rafters, the mouldings of balustrades, and so on--have been captured and represented most painstakingly. 48 On the other hand, a specific scale has been used throughout to downsize buildings and other structures. Though we do not know the exact value of the scale used for the Yanshansi murals, according to the Norther n Song scholar Li Zhi 李廌 (1059-1109), a scale of 1:10 was generally adopted by painters as a rule. 49 This is to say that a painted building ought to be a one-tenth miniature of a real one. In the Yingzao fashi, the tiangong louge is also theoretically a 48 Karetzky 1980, 251. 49 This is recorded in Li Zhi's Deyuzhai huapin 德隅齋畫品 (Deyuzhai connoisseurship of paintings ), http://ctext. org/wiki. pl?if=gb&chapter=638091 : “[Guo Zhongshu] replaced cun with hao, chi with cun, and zhang with chi [when calculating dimensions]. If one magnified and multiplied [the size of the architecture he painted] to build a large structure, he would find that the structure fitted well with the woodworking rules, never with a slightest discrepancy. This could not have been accomplished had the painting not followed the rules in every detail 以毫計寸, 以分計尺, 以尺計丈. 増而倍之, 以作大宇, 皆中規度, 曽無小差, 非至詳至悉委曲於法度之內者不能也. ” See also Maeda 1975, 126.	di-luo-a-grain-of-sand-full-draft-3.pdf
152 one-tenth miniature of a full-size, medium-rank wooden hall. 50 It seems almost certain that ruled-line painters and small-scale woodworkers were aware of each other's formulas of miniaturization and were actively learning from and he lping to increase each other's expertise. The aesthetics of the minutiae of the interior space was shared between architecture and painting, by carpenters and artists, in the Northern Song, Liao, and Jin alike. As observed by Robert Maeda, “the subjects of these small works invite perusal of their tiniest details, provoking the kind of viewer delight and involvement often produced by miniature paintings. ”51 At the Jingtusi as well as Yanshansi, such a delight at heart was mixed with feelings of awe toward th e Buddhist heaven and its imperial grandeur, with perhaps also an urge to seek one's own place and identity in relation to the miniature world. The ethnic dimension Major critiques of Jurchen-Jin material culture consider it mainly a preservation and elab oration of Northern Song forms and was disappointingly conservative, complacent, passive, and non-innovative. 52 Now, viewed in light of miniaturization, it is hard to totally agree with such criticism. What appeared mediocre from the outside actually encour aged an introspection and an exploration of the inside, marked by “a distinct taste for the ornate, the dense, and the multilayered,” which set 50 A medium-rank hall could use the sixth-grade cai (6 by 4 cun), which is ten times greater than that of the tiangong louge (0. 6 by 0. 4 cun). Of course, the valu e of cai was not a constant, meaning that a miniature was not always a 1/10 “replica” but its size could fluctuate between 1/15 and 1/7. 5 of real buildings. 51 Maeda 1975, 138. 52 Laing 1988-89, 119, stresses that the influence from Southern Song culture tow ard the Jin was mainly in the realm of landscape painting, whereas the majority of Jin decorative arts and furniture was “a preservation of Northern Sung forms and an elaboration of them. ” Herbert Franke argues that the Jin visual arts were “conservative a nd traditional” and remained to be a continuation of Tang and early Song styles; the Jurchens, on the other hand, did not contribute too much as far as culture is concerned, and “the Jurchens' acceptance of Chinese culture was eager but more passive than active. ” See Denis C. Twitchett, Herbert Franke, and John King Fairbank, The Cambridge History of China, vol. 6, Alien Regimes and Border States, 907-1368 ( Cambridge: Cambridge University Press, 1994): 310, 313.	di-luo-a-grain-of-sand-full-draft-3.pdf
153 the Jin apart from any Northern Song or Liao antecedents. 53 Admittedly, to create a miniature world was to be a m aster of the formulas and rules, but such an apparent “docility” to standardization did not destroy originality or diversity. Creativity found another way in: instead of inventing new forms, the focus shifted to producing new scales and proportions, and to reigniting excitement and imagination through the theatricality and oneirism of miniaturization. An inevitable question to be asked is : “what was the role of the Jurchen ethnicity here?” It has been observed that during the Jin, ethnicity mainly determine d issues of socio-political status and legal rights and responsibilities;54 whereas in the realm of artistic creation, Jurchen ethnicity did not become overtly expressed, but the consciousness and passion of absorbing Han-Chinese culture and art kept the wo odworking tradition alive and ongoing. In other words, Jurchen architecture and art seemed to be wearing a conspicuous “Chinese” mantle, as if fearing they were not Chinese-looking enough. One of the major contributions of the Jurchen rule to Chinese mater ial culture, therefore, can be understood as a seamless continuation and development of the established tradition. There were no major breaks or a total overhaul. As early as the beginning years of the dynasty, when the Jingtusi ceiling was built, known te chniques of miniature-making wer e openly and fervently embraced without necessarily incurring any identity issue. Miniaturizing techniques managed to flourish and advance under the Jin and later into Ming and Qing times, and this could not have happened wi thout a high-profile incorporation of Chinese political and cultural ideologies by the 53 Laing 1988-89, 119. 54 Hoyt Tillman, “An Ove rview of Chin History and Institutions,” in Hoyt Cleveland Tillman and Stephen H. West, China Under Jurchen Rule: Essays on Chin Intellectual and Cultural History (Albany: SUNY Press, 1995): 24.	di-luo-a-grain-of-sand-full-draft-3.pdf
154 Jurchen rulers, most notably Hailingwang 海陵王 (r. 1150-1161), Zhangzong 章宗 (r. 1189-1208), and Xuanzong 宣宗 (r. 1213-1224). As all other conquest dynasties on the Chinese soil, the Jurchens were faced with the challenging task of legitimizing their rule, which meant that the policies they adopted and the art they patronized needed to serve the rule and serve the purpose of legitimation. 55 According to Hok-lam Chan, legitimat ion became a major concern for the court especially during the reigns of Zhangzong and Xuanzong, who called for two court assemblies where officials debated the “Five-agent” theory of dynastic successions and proposed their own solutions to Jurchen legitim ation. 56 The first assembly under Zhangzong proposed that the Jin was inheriting the Northern Song whereas Liao was deemed largely irrelevant. As a result, the court promulgated the Jin code of law known as the Taihe luyi 泰和律議 (based on Tang and Northern So ng laws), canonized Ouyang Xiu's New History of the Five Dynasties (Xin Wudaishi 新五代史 ), and suspended the compilation of the Liao history. The second assembly under Xuanzong, while further disqualifying the Liao as a legitimate regime, also led to the relo cation of the imperial capital to Nanjing (Dongjing in the Northern Song) in 1214, a move which would later prove devastating to the survival of the dynasty. 57 Such open incorporations of Chinese ideologies and political system certainly raised anxieties among the Jurchens. The fear for the disasters of total sinicization and annihilation of the Jurchen identity loomed larger and larger as the Jurchens continued their rule and relocated the capital 55 Ibid., 37-38. 56 This is the main topic in Hok-lam Chan, Legitimation in Imperial China: Discussions under the Jurchen-Chin Dynasty (1115-1234) (Seattle and London: University of Washington Press, 1984). 57 The relocation of the capital made the Jin more vulnerable to the attacks of Mongol riders. During the final years of the Jin, issues of legitimation started to occupy the center of state agenda as the rulers deemed them to be a means to tackle domestic crisis as well as the Mongol invasion. See Twitchett, Franke, and Fairbank 1994, 245-64.	di-luo-a-grain-of-sand-full-draft-3.pdf
155 further south into the heartlands of China. Countermeasures such as education were taken by the court in the attempt to restore Jurchen traditions, but to no avail. The same identity crisis, however, did not seem to have stopped the overwhelming tendency to embrace the legacies of Chinese material culture. In fact, it was perhaps due to such anxieties about legitimation and identity that drove the Jurchen conquerors to seek more strenuously solutions from the wisdoms and experience of their precedents. The point, therefore, was not to “represent” or to some extent a lienate the Jurchens as non-Chinese and new conquerors, but to prove them as legitimate successors as well as capable innovators of the same, unbroken culture. A culture of miniatures was created under such dynastic consciousness. Miniaturization invited a journey to the interior, and it is hard to believe that such a focus on interiority had not been, in the slightest, stimulated by the inward-looking, and fundamentally self-reflective, state ideologies. A seeking for the self was simultaneously a seeking for universal truth; as cosmological theories were hotly debated at court to locate the Jin in the long lineage of Chinese dynasties, cosmological patterns were adopted and explored in the process of miniature-making to convey the vision of a world of the myriad. Symbolism of the Chinese D ome The “dome of heaven,” in Karl Lehmann's exposition, was a Christian vision of Heaven painted in the domes or vaulted structures of early Christian and Byzantine architecture. 58 Represented by world-renowned examples su ch as the east dome of San Marco in Venice, Lehmann's domes focus on the visual representations of heaven projected onto the domed ceiling, which was often decorated with images of divine figures (with Jesus as Pantokrator, or ruler of universe, at the 58 Lehmann 1945, 1.	di-luo-a-grain-of-sand-full-draft-3.pdf
156 center), anthropomorphized planets (such as Jupiter), evangelists, mythic figures (Sirens, giants, grotesque animals), baldachins, and the zodiac against a starry, floral background, sometimes encircled by inscriptions. 59 The depiction of heaven could be “rati onally descriptive” as well as “emotionally visionary;” physical as well as transcendental, combin ing the knowledge from astronomy and theology at the time. 60 The probable origin, however, was traced back to pagan motifs and astrological practices in Near E astern traditions. In his response article, Alexander Soper concurs that “celestial symbolism” was introduced to Asia via direct borrowing s facilitated by military conquests such as the ones by Alexander and Great and the Arabs, but the predominant motifs were altered to serve the Buddhist worldview and teachings, especially those of the much Westernized Mahayana Buddhism. 61 Hence, domes in Asia were not necessarily comparable to their Western counterparts in structure; but in decoration, they combined Helle nistic as well as early Christian visual elements and symbolism with Eastern religious traditions. One would assume that such a route of transmission of art forms in many degrees overlapped with the Silk Road, especially the land routes by which Buddhism w as introduced to China. Indeed Soper seems to suggest such a trajectory in his organization of the materials, which cover the areas of Ajanta, Kashmir, Bamiyan, Khotan, Kucha (Kizil and Kumtura), Dunhuang, Yungang, and finally, Korea and Japan. This diffus ionist view by Lehmann and Soper has been criticized by more recent scholarship, which argues that the dome in each culture could have sources of its own. Nancy Steinhardt, for instance, has proposed that a more likely source for Asian domes were Han tombs with vaulted 59 Ibid., 2. 60 Ibid., 4, 27. 61 Soper 1947, 226.	di-luo-a-grain-of-sand-full-draft-3.pdf
157 ceilings where painted star maps were often found. These, she argues, have gr eatly influenced the construction of later cave temples at Kumtura, even the residences at Penjikent. 62 The two ends of the Silk Road, Han China and ancient Rome, cou ld have envisioned and developed their systems of domed cei lings independently. While the attention of these scholars is paid mainly to the pictorial representation and religious symbolism of “heaven,” this section returns to the “dome” and the specificati ons of dome-building in Asia. My interest lies in how the Jingtusi ceiling may be understood in light of the very concept of the “dome of heaven ”: is tiangong louge a representation of heaven, as its name suggests? Can we consider the Chinese ceiling coffe r a dome? How do we define a Chinese dome? Are Chinese domes miniatures? The goal here is not to clarify the similarities and discrepancies between the domes in China and other parts of the world, but to seek the historical manifestations and transformatio ns of a particular archetype, which eventually found its way into the Jingtusi. Zaojing, the “water-weed well” A Chinese ceiling coffer was typically topped with an eight-sided domical structure which resembled a cupola. The Yingzao fashi informs us that the coffer had three alternative and more ancient names: zaojing 藻井 (lit. water-weed well), yuanquan 圜泉 (round fountain), and fangjing 方井 (square well). A total of four instances where these names appear in early literature are quot ed and annotated in the text: 1. “Western Metropolis Rhapsody:” Rooted the inverted lotus stalks inside the coffer, enshrouded in red flowers that joined one to another. (The coffer is installed in the center of the ridgepole: the timbers crisscross to for m a well-like structure, paint it with water-weed patterns, and decorate it with lotus stalks. Connect the roots of the lotuses to the well and let the flowers dangle upside-down, hence they become “inverted. ”) 62 Steinhardt 2014, 277, 281.	di-luo-a-grain-of-sand-full-draft-3.pdf
158 西京賦 : 蔕倒茄於藻井, 披紅葩之狎獵. (藻井當棟中, 交木如井, 畫以藻文, 飾以蓮莖, 綴其根於井中, 其華下垂, 故云倒也. ) 2. “Rhapsody on the Hall of Numinous in Lu:” In the round pools and the square wells, invertedly planted are the lotus flowers. (Make a square well, and illustrate inside it a round pool a nd a lotus. The petals of the lotus flower are upside-down, hence it is said to be “invertedly planted”. ) 魯靈光殿賦 : 圜淵方井, 反植荷蕖. (為方井, 圖以圜淵及芙蓉. 華葉向下, 故云反植. ) 3. Comprehensive Interpretation of the Customs : The [coffer of the] palace hall imitates the Eastern We ll (eight stars in the constellation of Gemini) and is carved into shapes of lotuses and water caltrops. Water caltrops are aquatic plants and are used to subjugate fire. 風俗通義 : 殿堂象東井形, 刻作荷蔆. 蔆, 水物也, 所以厭火. 4. Shen Yue, History of the Liu-Song: The reason wh y round fountains and square wells are installed in the ceilings of palatial halls and decorated with lotuses flowers is to subjugate fire. (Today, a coffer made into a square shape is called dousi. ) 沈約宋書 : 殿屋之為圜泉方井兼荷華者, 以厭火祥. (今以四方造者謂之鬬四. )63 Several points can be made based on these quotations. First, the basic structure of a ceiling coffer was the “well,” a pattern found in the character jing, after which the coffer was named. A jing structure was composed of two pairs of orthogonally intersected timbers. 64 The inclusion of a jing in the ceiling was meant to signify certain celestial bodies, for instance the Eastern Well ( dongjing 東井 ), 63 Yingzao fashi, vol. 1, 35-36; the texts in parentheses are Li Jie's annotations. The first two entries come from the works of Zhang Heng 張衡 (78-139) and W ang Yanshou 王延壽 (ca. 140-165), respectively. They belong to the literary genre of fu 賦, rhapsody, which historically applied a highly stylistic language and has been often criticized as hyperbolic. Nevertheless, some technical terms and information have be en conveyed and passed down by these rhapsodies and investigated as reliable textual evidence by scholars. The dates of the four quoted texts span from the first to the fifth century, more than five hundred years before the Yingzao fashi. Despite the antiquity of the texts, one can, as the compiler Li Jie did, rely on the observations of these highly esteemed writers to get a glimpse of the ancient Chinese dome. 64 It could have been an extension of the jinggan 井干, one of the most primitive and widely adopted building techniques for log-cabins, vertical wells into the ground, and certain burial types such as the huangchang ticou 黃腸題湊. An annotation by Xue Zong 薛綜 (d. 243) on the “Western Metropolis Rhapsody” explains the zaojing to be “made by intersecting timbers into a square as if making a jinggan-structure 交木方為之, 如井干也. ” According to the Qing etymologist Duan Yucai 段玉裁 (1735-1815), “ Gan were the wooden railings of a well; the shape of gan could be either quadrilateral or octagonal 干, 井上木闌也, 其形四角或八角. ”	di-luo-a-grain-of-sand-full-draft-3.pdf
159 one of the twenty-eight Chinese lunar mansions, which leaves one to ponder if this astrological association suggests any specific religious or ritual messages. 65 Second, while Western domes largely displayed visions of heaven, the ceiling coffer in Ch ina has had, since the Eastern Han, unmistakable connotations of the element of water. A jing is where fresh water could be obtained; a quan 泉 is a fountain or spring of flowing water. 66 There were also multiple types of aquatic plants--the zao 藻 (water-weed), the lotus, and the water caltrop--which were either painted or carved in the ceiling. Though the lotus was more often associated with Buddhist iconography in post-Han Chinese architecture, in the quoted texts, rather, the purpose was to subjugate fire--the greatest enemy of wooden buildings. 67 Records of using the image of zao as a decoration of the roof frame were found in the Analects and Liji. 68 The zao soon became a term connoting “embellishment” and “extravagance,” to describe a lavishly made physica l object or a rhetorical language of literature. During the Northern Song and the Liao, the zao was one of the “twelve imperial insignias ( shierzhang 十二章 ),” an institution inherited from the Tang. 69 65 Additional connections between early Chinese architecture and practices of astronomical observation can be found in the section “Orientation ( quzheng 取正 )” in juan 2 of the Yingzao fashi, which includes inst ructions of determining the north and south by using a gnomon (for measuring shadow cast by the sun) and a viewing scope (for observing the Polar Star). A few Han rhapsodies inform us of certain ritual functions of the imperial halls and towers, such as th e well-known Jinggantai 井干臺 by Han Wudi, where the emperor was supposed to perform self-retrospection, watch the activities of the multitudes in his realm, learn lessons from the examples of good and evil historical/legendary figures, foster virtue, and ev en communicate with heaven. These were to be enacted by a fully immersive and evocative interior enhanced by wall paintings, the burning of incense, and the playing of music. Zaojing might have very well been part of such settings. 66 The connection between zaojing and water is noted in Takeshima 1971, 386. 67 Soper 1947, 238. 68 The phrase “ shanjie zaozhuo 山节藻棁 ” literally means mountain-shaped brackets and water-weed-patterned ceiling posts; it has been used to describe a luxuriously ornamented architectural interior appropriate only for the Son of Heaven. Anyone else who occupied such an interior would be condemned by the society as a transgressor of the ritual code. The Analect, for instance, disapproves of Zang Wenzhong 臧文仲 for residing in a hall decorated with shanjie zaozhuo. 69 Sanlitu jizhu 三禮圖集注 (Annotated compendium of the Illustrated Three Rites), compiled by Nie Chongyi 聶崇義 (fl. 10th century), 1. 5b. The zao-pattern was listed among the twelve patterns (the sun, the moon, dragons, mountains,	di-luo-a-grain-of-sand-full-draft-3.pdf
160 Third, unlike a Western dome, which was a load-bearing mem ber of the roof frame, the zaojing was a self-sufficient structure which can be detached from the building. The jing structure does not involv e any method of vaulting, and circular elements such as the “round fountain” was simply drawn on a flat surface. A zaojing was never visible from the outside: the sweeping slopes of a Chinese roof make a sharp contrast to the exposed domes of Pantheon, Hagia Sophia, and Taj Mahal. Rather, the zaojing was an interface between the interior space and the roof, and its be auty lies solely on its interiority. In English scholarship, the term zaojing has been translated alternatively as caisson, cupola, laternendecke (lantern ceiling), and coffered ceiling. 70 These translations do suggest that the zaojing is in certain ways comparable to the Western dome, vault, or coffering, whether or not this resemblance lies in structure or symbolism, or both. I use “ceiling coffer” or simply “coffer” to denote zaojing in this dissertation, hoping to grasp the essence of the jing--a sunken (but not always vaulted) space in the ceiling. The word “coffer” already carries the connotation of “decorated” and “embellished” pertaining to the term zao. A coffer means it is not necessarily the entirety of the ceiling, but is more often a repetitive a rchitectural motif; it avoids suggesting a load-bearing roof structure as the te rm dome might lead us to think. Wooden “domes of heaven” from the tenth century onward Wooden remains of zaojing before the tenth century are totally lacking. The earliest su rviving example is the central coffer of the Guanyinge 觀音閣 (Avalokitesvara Pavilion, d. 984) at the and many o thers) of the imperial gown since the time of the Sage Emperor Shun 舜. Also see Dieter Kuhn, “Liao Architecture: Qidan Innovations and Han-Chinese Traditions?” T'oung Pao 86 (2000. 4/5): 341. 70 Steinhardt 2014, 271.	di-luo-a-grain-of-sand-full-draft-3.pdf
161 Dulesi 獨樂寺 in Tianjin, a pure, concise example devoid of any brackets, whereas all its forerunners have to be sought in non-wooden or non-Chinese structures. 71 The coffers in the Main Hall of the Baoguosi 保國寺 (d. 1013) in Ningbo include eight arching ribs spring ing from an octagon al base (here supported by miniature brackets) and converging at the top, forming three beautiful cupola s. Similarly, other wooden c offers of the eleventh century, such as those in the Huayansi library hall and the Yingxian Wooden Pagoda, all seem to be fairly concise and lacking the tiangong louge in general. This means that the Jingtusi Main Hall is the earliest surviving example wh ere the tiangong louge entered the ceiling. Its unique design must have soon become popular, as it has been emulated at the Shanhuasi 善化寺 (d. 1128) in Datong, where the central ceiling of the Main Hall consists of one octagonal, one square-and-(double-)diamond, and three square coffers (fig. 89). The small-scale, densely-arrayed brackets look quite familiar. Though no miniature halls or galleries are included, the miniature Buddha images painted along the periphery of the group of coffers, above the substructure and separated by small posts, are largely reminiscent of the Jingtusi tiangong louge. The my sterious miniatures in the ceiling of the Chongf usi Amitabha Hall (see fig. 84), as mentioned earlier, probably belonged to a larger group of tiangong louge which is now lost. According to the staff of the monastery, they were ori ginally part of a certain “ feitian louge 飛天樓閣 (towers 71 Evidence of zaojing and other related c eiling types or archetypes--oculus, laternendecke, truncated pyramid, corbelling, and coffering--has been found in many examples across Asia. The earliest is the Dahuting Tomb 2 (d. late Eastern Han) in Mixian, near Luoyang, where diamond-and-square motifs (representing laternendecke ceilings) are painted on the barrel vaults of the burial chambers. Other early examples include the Jinguyuan 金谷園 tomb (d. 9-23 CE) in Luoyang, Henan, the Yinan 沂南 Tomb 1 (d. late second century) in Shandong, and multiple Easte rn Han cliff tombs in Qijiang 郪江, Sichuan. A few Korean tombs, notably the Anak 安岳 Tomb 3 (d. 357, better known as Dong Shou's 冬壽 tomb), the Daeanri 大安里 Tomb 1 (ca. first half of the fifth century), the Tomb of the Celestial Kings and Earthly Spirits ( Cheonwangjisinchong 天王地神冢, ca. fifth to sixth century), showcase the techniques of lantenendecke in its early, primitive form. Similar ceiling s reappeared in a significant number of Buddhist cave temples at Dunhuang, Kizil, and Bamiyan.	di-luo-a-grain-of-sand-full-draft-3.pdf
162 and pavilions in the air)” installed at the four corners of the ceiling. A more revealing example is the Rear Hall of the Fengshengsi 奉聖寺 at Tianlongshan in Taiyuan, where a band of miniature wooden gal leries, supported by eight-tiered bracketing, encircles the central octagonal coffer and accommodates tiny Buddha figurines underneath (fig. 90). 72 Unlike the Jingtusi ceiling, the brackets supporting the substructure have not been miniaturized, whereas the brackets of the galleries are significantly smaller than those of the cupola above. By contrast, at Jingtusi, the substructures, the tiangong louge, and the cupolas are always of the same scale. The l egacy of the Jingtusi miniatu re endured well into the Yuan, Ming, and Qing dynasties. An offering pavilion at t he Temple to Minister Dou (Doudafuci 竇大夫祠, repaired in 1267) in Taiyuan features a tiangong louge ceiling (fig. 91). Notably, the pavilion itself is braced by four diagonal beams, which resonate with the multiple layers of diagonal members in the octagonal cupola. Four miniature halls stand between the substructure below and the cupola above, connected by encircling galleries. A slightly different typ e of tiangong louge is found in the main hall of the Yong'ansi 永安寺 (rebuilt in 1315 based on a Jin original) in Hunyuan, Shanxi, where two rows of miniature buildings are installed on top of the central beams (fig. 92). Coming to the Ming, the tiangong louge developed an unpreceden ted level of intricacy, perhaps also of overwhelming superfluity in certain cases. The ceiling of the Gongshutang 公輸堂 in Huxian, Shaanxi (built between 1403-1424) is covered by an overflow of miniatures--the luxuriant, interlocked bracket sets, the multipl e eaves, and the closely-spaced towers and corner towers thrustin g into the deep ceiling (fig. 93). The less rhapsodic, calmer example is a pair of almost 72 The date of this stru cture is unknown but stylistically it is believed to be traced back to the Jin. Further information regarding this hall is nowhere to be found, but an old image is included in Liang Sicheng and Liu Zhiping 劉致平, Zhongguo jianzhu yishu tuji 中國建築藝術圖集 (Collect ed illustrations of Chinese architecture) (Tianjin: Baihua wenyi chubanshe, 2007), vol. 2, 526.	di-luo-a-grain-of-sand-full-draft-3.pdf
163 identical coffers from the Zhihuasi 智化寺 in Beijing (d. 1443), now in the co llections of Philadelphia Museum of Art and the Nelson-Atkins Museu m of Art (figs. 94, 95). 73 Patronized by Wang Zhen 王振 (?-1449), one of the most powerful and notorious eunuchs in early Ming, the Buddhist temple has assumed an undeniable imperial character: a profusion of dragons in high-relief appear in the ceiling coffer, which is encircled by miniature double-story towers. The number of towers amounts to seven per side (not including corner towers), and each st ands on a cluster of clouds. This rhythmic facade, overall, is visually similar to the type of tiangong louge illustrated in the Yingzao fashi. The zaojing, too, underwent certain changes: it has proliferated into an interlaced network of squares, diamonds, octagons, and triangles. Such exquisite miniature-making only escalated with the passage of time. The coffer from the Longfusi 隆福寺 in Beijing (d. 1452) is a n even more spectacular “well” with multiple layers of miniature towers, pavilions, and figurines of celestial deities (fig. 96). 74 These examples suggest that tiangong louge was not exclusively used for Budd hist halls but also applicable to ancestral temples and other ritual or religious space. In this sense, the miniature buildings were not necessarily representations of Buddhist heavens such as the Pure Land, but could be granted a different meaning at a di fferent location for a different purpose. Broadly speaking, being an idealized projection of the worldly architecture, they were signifiers of an ambiguous, loosely defined utopian and spiritual realm--the tian 天 (heaven). 73 A complete survey of this temple is in Liu 1932. See also Sickman and Soper 1984, 461-63, for a general analysis of the style. 74 The original monastery was pa rtially burnt in an accident in 1901 and the remaining structures were dismantled in 1976 after the disastrous Tangshan earthquake. The coffer mentioned here was restored in 1994 and now exhibited in the Beijing Museum of Ancient Architecture. According to Dijing jingwulue, “The form of the ceiling coffer in the hall originally came from the West; it contains the Eight Classes of Celestial Beings and Demigods and an entire Lotus Repository World in full display 殿中藻井, 制本西來, 八部天龍, 一華藏界具. ” This is the same Lon gfusi mentioned in Chapter 2 of the dissertation.	di-luo-a-grain-of-sand-full-draft-3.pdf
164 Miniaturization has been a belove d technique of Chinese carpenters to bring such a realm of imagination into full display. On the level of human perception, the alteration of scale almost always indicates a change in structure and nature. This is because in the physical world, scale is no t a mere number but an attribute inherent to any object. The dwindling scale and increasing intricacy of the Chinese dome over the history eventually disqualified its original role as a loadbearing component and converted it into something different--a squ are-and-diamond, and later octagonal, motif in the ceiling, which has since shifted its function from structural stability to emotional evocation. As the ceiling was being alienated from the roof frame, the introduction of miniaturized brackets and buildin gs helped to accentuate its new symbolic and decorative nature, adding to both depth and meaning of the interior space. At the Jingtusi, specifically, how did the tiangong louge miniatures help the viewers see heaven, or in this case a Buddhist heavenly re alm? Aside from the resemblance of the buildings to those depicted in Buddhist paintings and literary descriptions, it is perhaps miniaturization that has proved to be the foremost force of inducing the mind into the dream-like, rhapsodic state capable of seeing the transcendental. Gaston Bachelard alleges that “[v]alues become engulfed in miniature, and miniature causes men to dream. ”75 It is precisely the nature of being miniatures--the size has been greatly reduced but the basic geometry stays the same--that deconstructs architecture and sharpens senses. 76 The distorted scale arouses an uncanny and illusory feeling as one faces tiny “palace 75 Gaston Bachelard, The Poetics of Space, trans. Maria Jol as (Boston: Beacon Press, 1969), 152. Bachelard's examination of miniatures is mainly based on literary analysis. One of his major arguments is that poetic qualities of a space--including miniaturization--create psychological impacts, and that these impacts are transmitted through seeing, hearing, smelling, touching, and through other activities in and around the space. Importantly, what we experience in space makes us to imagine, or to “daydream,” and Bachelard asserts that these daydreams and reveries are themselves “images” of reality, which reveal just as much about us as our dreams and subconscious. 76 Stein 1990, 52. “In fact, the more altered in size the representation is from the natural object, the more it takes on a magical or mythic quality. ” Also see Stewart 1993, 65-66, for the observation of how reduced scales distort space and time of the everyday world.	di-luo-a-grain-of-sand-full-draft-3.pdf
165 buildings” suspended above Buddhist statues. Miniature architecture in the ceiling, therefore, would generate an optic al illusion, a sense of remoteness, depth, and perhaps also of multiplicity and infinity. It facilitated the creation of an immersive, gravitational field overhead which was at once limited and expansive, full of miniscule yet extravagant details. Ceiling Design and City Design A Neo-Confucian of the twelfth century would argue that it was within not just things of grand scales but also the miniscule--as trivial as a grain of sand or a blade of grass--that principles of heaven was encapsulated. This consci ousness of the investigable order of the minute things was part of the Neo-Confucian worldview, which burgeon ed in the eleventh century and continued as a prolonged intellectual and political discourse into later histo ry. While Neo-Confucian thought exerci sed profound impacts on the handling of “human affairs,” i. e. the political and social facets of life, its influence actually extended to the creation of material culture, including miniature-making. Such influences, at first sight, could be hard to detect--did the Jingtusi ceiling have anyt hing to do with aspects of the Neo-Confucian ideology ? Rather than affecting the exact form of an art object, the impact was probably an epistemological one--how the design of the ceiling was conceived, what practice, knowledge, and spiritual drive led to this particular design, and how it was received and interpreted by viewers, etc. As far as the Jingtusi is concerned, Neo-Confucianism provided a worldview which, not unlike Buddhist or Daoist worldviews, can be used to justify the symbolism of the “dome of heaven;” more importantly, Neo-Confucianism was a perspective to history, an intellectual tradition to be reckoned with. This section focuses on the archetype of jing, especially its manifestations in the realms of architecture and urban planning along the Confucian (and Neo-Confucian) tradition. In its simplest and most abstract form, jing is a plane divided by two horizontal and two vertical lines into nine	di-luo-a-grain-of-sand-full-draft-3.pdf
166 distinctive sectors : the four cardinal sectors (N, S, E, W), the four ordinal sectors (NE, NW, SE, SW), and the center. This is perhaps the most easily conceived reference of orientation, and similar coordinate systems have been applied to cartography since ancient times. 77 The way we understand space determines the way we design it; jing in Chinese history remained to be an ideal layout for natural as well as man-made environment, from farmlands to capital cities, palaces, ritual and symbolic structures (most notably mingtang, the Hall of Light), and ceilings. To Con fucians as well as Neo-Confucians, the spatial order was often a direct projection of the social structure. It is hence not surprising that a seemingly purely geometric layout was soon imbued with highly political and ethical significances. The well-field and Neo-Confucianism The earliest graphs of jing found on oracle bones and bronze vessels are written almost invariably as two horizontal lines intersecting two vertical lines, sometimes wit h a dot in the center. The Shuowen jiezi 說文解字 (Explanation of graphs and characters) interprets it to be a pictograph resembling the form of gan 干, which, explained the Qing annotator Duan Yucai 段玉裁 (1735-1815), indicates a quadrilateral or octagonal wellhead made of timbers. 78 The first application of the jing layout to spatial planning was the jingtian 井田, the well-field, which was basically a square piece of farmland divided into nine equal portions, sometimes with a 77 In addition to cartography, the jing-pattern appeared in numerology ( Luoshu 洛書, or the Luo Writ), divination (Eight Trigrams), astrology, and military theories regarding battle formation and the division of units. 78 See n. 64.	di-luo-a-grain-of-sand-full-draft-3.pdf
167 well in the ce nter (fig. 97). 79 The well-field system was advocated by Mencius : “a field one li square makes a jing, which is nine hundred mu, and at the center is the public field. The eight families [of the same jing] each hold a hundred mu as their private lands, and collaboratively they farm the public field” (方里而井, 井九百亩, 其中为公田. 八家皆私百亩, 同养公田 ). 80 The rather nondescript layout reminds us of the checkerboard, and the charm of this pattern lies in its ability to be repeated and expanded, conveniently and infinitely, to a larger and larger framework, or to be further segmented and div ided by adding orthogonally intersected lines. Indeed, it can forever grow inward or outward into a recursive grid. As demonstrated in the Kaogongji 考工記, builders and civil engineers of the Warring States period designed a special irrigation network and ro ad system by expanding the basic layout of jing:81 Name of ditch Location Dimension (width by depth) Name of road quan within a fu 1 by 1 chi N/A sui between fu 2 by 2 chi jing gou between jing (9 fu) 4 by 4 chi zhen xu between cheng (100 fu) 8 by 8 chi tu kuai between tong (1,000 fu) 2 xun by 2 ren dao chuan between ji (10,000 fu) lu 79 Chunqiu Guliangzhuan zhushu 春秋穀梁傳註疏, 5. 212b. “In ancient times, three hundred bu equaled one li, and [a field one by one li] was called jingtian. A jingtian was nine hundred mu and the public unit of it occupied one hundred mu 古者三百步为里, 名曰井田. 井田者, 九百亩, 公田居一. ” Bu, li, mu were all measuri ng units. 80 Mengzi zhushu 孟子註疏 5(1). 12a. 81 This system focused on the intersecting “lines,” not the field units. Fu was the smallest unit of farmland (tilled by one man), whereas jing was a basic form of configuration/grouping. The pertinent text is found in Wen Renjun 聞人軍, Kaogongji yizhu 考工記譯註 (Kaogongji, interpreted and annotated) (Shanghai: Shanghai guji chubanshe, 2008), 120-23; He Yeju 賀業矩, Kaogongji yingguo zhidu yanjiu 考工記營國制度研究 (A study on the city-planning methods recorded in the Kaogongji ) (Beij ing: Zhongguo jianzhu gongye chubanshe, 1985): 40.	di-luo-a-grain-of-sand-full-draft-3.pdf
168 Internally, the jing was divided into nine units called fu 夫 (area of field tilled by one man); externally, it was multiplied by ten at a time to form larger “checkerboards,” with wider and deeper water channels for irrigation, which were magnified incrementally that they would finally flow into natural rivers an d streams ( chuan 川). Alternatively, according to the Zhouli 周禮 (Zhou institutions), “nine fu makes a jing, four jing makes a yi, four yi makes a qiu, four qiu makes a dian, four dian makes a xian, four xian makes a du” (九夫為井, 四井為邑, 四邑為丘, 四丘為甸, 四甸為縣, 四縣為都 ). 82 The multiplier here is four, different from that of the irrigation grid probably because the system is administrative rather than agrarian, growing from private farmlands to villages, towns, counties, and states (fig. 98). This was the type of land allocation that facilitated the enfeoffment system of the Zhou, providing a simple geometric (and geographical ) solution for the aristocrats--from the Son of Heaven to the princes and their sons, brothers, and other male relatives--to divide their lands, population, and other natural resources in the process of lineage segmentation. 83 Ironically, lying in the heart of such a hierarchical system was jing--a pattern bearing the ideals of egalitarianism and public responsibility. The practi ce of the well-field system is allegedly traced back to the Xia and Shang dynasties, when an efficient method of land distribution and irrigation would prove vital to the survival of a newly emerging agrarian culture. It was the basic structure of the soci ety (eight families as the smallest social unit), the basic form of civil obligations (collaborative labor on the public unit of 82 Zhouli zhushu 周禮註疏 11. 6b-7a. The lands were grouped in this way so that “official posts of land administration could be assigned and tributes and taxes could be collected; these all concerned state reven ues 以任地事而令贡赋, 凡税敛之事. ” 83 He 1985, 26. See also K. C. Chang, Art, Myth, and Ritual: The Path to Political Authority in Ancient China (Cambridge: Harvard University Press, 1983), 16.	di-luo-a-grain-of-sand-full-draft-3.pdf
169 field), and was essential to the maintenance of a stable source of state income, providing a revenue roughly equal to a ten per cent tax collected in the form of harvested crops. 84 The well-field underwent significant changes and was interpreted differently over the history (some argue that it was never actually practi ced). The purpose here is, of course, not to straighten out all t he subtle or radical changes and developments in any given historical period or to prove that it was actually practiced at a time, but rather to expose the endurance of this abstract, seemingly simplistic pattern as an ideological model for later urban pla nning and architectural design. Its tenacity is especially due to the ethical values associated with this particular pattern by some Confucian thinkers. Already in the Warring States, Mencius advocated that a humane government (renzheng 仁政 ) should start with the drawing of correct lines and borders to ensure rightful land distributions. 85 He envisioned a utopia, a somewhat egalitarian society where people of the same community would befriend and help each other and always prioritize the collaborative work on the public lands. Ironically, Mencius's time was when ancient institutions and the virtue of the sage kings were gradually forgotten or abandoned under drastic social and political changes. Even t hough the practicality of old social sy stems such as the well-field in a new era often became questionable, they were remembered as admirable feats of the past and embraced by later Confucians as a hallmark for good government. 84 The 10% tax has been mentioned by many historical documents; see He 1985, 117. 85 Mengzi zhushu, 5(1). 10b-12a. “A humane government has to start with drawing lines and making boundaries. An uncorrected border leads to unequal divisions ( jing) of farmlands and uneven disbursement of grains and salaries. This is why tyrannical sta te-lords and corrupt officials always tend to ignore the correction of lines and boundaries. Only after the lines and boundaries are corrected can one settle down the distribution of lands and the disbursement of salaries 夫仁政, 必自經界始. 經界不正, 井地不鈞, 谷祿不平. 是故暴君污吏必慢其經界. 經界既正, 分田制祿可坐而定也. ”	di-luo-a-grain-of-sand-full-draft-3.pdf
170 Even into the Northern Song, the revitalization of the well-field became the center of the debate on the economic reform, especially concerning the problem of land distribution. 86 Ouyang Xiu w as one of the many advocators who asserted that a revival of Confucian ethics and institutions would heal the state of its “illnesse s” and corruption. 87 Cheng Hao 程顥 (1032-1085) presented a memorial to the emperor and stressed, “The boundaries of the land had to be defined correctly, and the well-fields had to be equally distributed--these are the great fundaments of government. ”88 The urge to put the jingtian in statewide application was similarly echoed by other highly influential scholar-officials of the time. 89 Wang Anshi's 王安石 (1021-1086) Xinfa 新法 (New measures) promulgated and enforced during the period of 1068-1076 introduced a square-field system ( fangtian junshui fa 方田均稅法 ), which, though not copying the ancient well-field, was allegedly devised to achieve the same goal and effect. 90 86 de Bary and Irene Bloom 1999, 596-98. The problem of land distribution was at the center of Fan Zhongyan's 范仲淹 (989-1052) reforms during the reign of Renzong 仁宗 (r. 1022-1063) and the later political struggles between the two “parties” of officials at court. 87 Ibid., 590-95, translation by Robert Hymes and Burton Watson. This is in Ouyang Xiu's “Benlun 本論 (Essay on fundamentals)” (written in 1040s), an essay suggests that the corruption of the Qin dynasty started with the abolit ion of the well-field system. 88 Ibid., 601-02, translation by William de Bary. Cheng Hao argued that the revitalization of the well-fields was not only to restore and maintain “the order of things” but also to reclaim the official control over natural reso urces including hills and streams. With much confidence, he asserted that the laws and institutions of the Three Dynasties “can definitely be put into practice. ” 89 Another notable advocator was Zhang Zai 張載, who pursued the ideal of egalitarianism embedded in the well-field system: “The land of the empire should be laid out in squares and apportioned, with each man receiving one square. ” See de Bary and Bloom 1999, 605-06. There were, of course, voices of caution and objection. For instance, Sun Xun 蘇洵 (1009-1066) analyzed the potential difficulties in practicing the ancient irrigation grid used in the well-field system and proposed alternative means and solutions (pp. 606-09). 90 Ibid., 610-11. To legitimize his square-field system, Wang Anshi even authored a Zhouguan xinyi 周官新義 (New interpretations of Zhouguan, Zhou Institutions).	di-luo-a-grain-of-sand-full-draft-3.pdf
171 What was the charm of the jingtian, that after more than a thousand years, it still haunted Chinese intellectuals? One c annot neglect the fact that several advocators of the revitalization program were also founders of lixue 理學, or Neo-Confucianism, including Zhang Zai 張載 (1020-1077) and Cheng Hao. Perhaps to many of the great philosophers (who were often idealists) of the time, the well-field was not just a practical solution to social problems but more essentially part of the universal truth or principle ( li) manifested in a geometric form, which, being universal, was not restricted to any specific dynasty or cause. Neo-Confucians claimed that “Principle is one but its manifestations are many,” forever pursuing an omnipresent, all-embracing li which was the source of all things and phenomena. 91 It would not surprise us if a Neo-Confucian asserted that an age-old institution could be revitalized in a different time and under different social conditions based on the immutable “principle” found in that ancient pattern. This immutable principle to be fou nd in the well-field was an ethical one--ren 仁, or humanness--the epitome of human virtue which was demonstrated graphically in the egalitarian layout of the pattern. 92 The tendency to eliminate the differences between historical periods, and to seek a certain immutability and indestructibility in the nature of things reminds us of the Huayan Buddhism of its grand, indiscriminative, “one-in-all” view of the universe. 93 Indeed, the Huayan School had influenced many Neo-Confucian thinkers,94 but a major disparity between the two schools was that 91 The dynamics between the “one” and the “many” has been exposed by the works of many Neo-Confucian thinkers, including Zhou Dunyi 周敦頤 (1017-1073), Zhang Zai, and the Cheng brother s. See Bol 2008, chs. 5 and 6; Wing-tsit Chan, trans., A Source Book in Chinese Philosophy (Princeton, New Jersey: Princeton University Press, 1963), 474, 499, 519, 544. 92 Wing-tsit Chan 1963, 499, 560, 571. The conscious effort to equalize ren to li is especially evident in Zhang Zai's and Cheng Yi's works. Most interestingly, Cheng Yi compared the mind-and-heart ( xin 心) to a seed of grain ( ren) having unlimited potentia ls to grow and proliferate. 93 See Chapter 3 of this dissertation 94 Wing-tsit Chan 1963, 406-08.	di-luo-a-grain-of-sand-full-draft-3.pdf
172 Neo-Confucians were more concerned about a pplying metaphysical understandings to human affairs, and lying at the center of their focus was still the idea of “humanness”--a focus since Confucius 's time. Parallel to this “one-in-all” view was the conviction that “size does not matter” in the exposit ion of universal truth. Echoing the Buddhist image of the “grains of sand of the Ganges,” Neo-Confucians claimed that li could be found in things as small and inconspicuous as “a blade of grass. ”95 Hence the li, like dharma, must be non-physical and able to permeate all material objects; it existed in all sizes and on all scales, from particles to stars. While this view might sound largely meditative, it helps to explain why the pattern of jing remained to be embraced in the eleventh-century and later China as a fundamental blueprint for almost all kinds of spatial design, which surfaced and resurfaced in land distribution, urban planning, and architectural design. The ideal city in miniature The jing layout had been adopted in the planning of capital cities by the late Warring States period. A guideline is provided in the Kaogongji : The builder builds the capital. It should be nine li square, having three gates on each side. Inside the capital there should be nine longitudinal and nine latitudinal thoroughfa res, each nine gauges wide. To the left [of the imperial palace] there should be the ancestral temple; to the right, the altar to the earth; at the front, the outer court; and at the rear, the market. The court and the market each occupies one fu. 匠人營國. 方九里, 旁三門. 國中九經九緯, 經涂九軌. 左祖右社, 面朝後市, 市朝一夫. 96 95 Ibid., 56 1-63. Cheng Yi wrote, “... every blade of grass and every tree possesses principle and should be examined. ” The important Neo-Confucian notion of gewu 格物 (the investigation of things) has a total of seventy-two interpretations in histori cal documents, whereas the ultimate goal was believed to help establish correct human relationships. According to the fourteenth-century Yupian 玉篇, ge 格 is “a model or measure;” it might be understood as a grid, a frame of reference, or a measuring tool (n ot unlike the jing) to locate and describe the subject of investigation. 96 Wen 2008, 112-20.	di-luo-a-grain-of-sand-full-draft-3.pdf
173 The overall layout of the capital (fig. 99) shares many commonalities with the jing: it is a grid of nine equal-size, square blocks, with longitudinal and latitudinal roads and drains. At t he center is the imperial palace, the seat of power, which is the terminus of the inflows of all revenues and the institution of the enforcement of collective labor. Contrary to the accessible, public field in a jingtian, however, the palace is heavily gua rded and secluded from the rest of the capital. Moreover, the palace itself could also be based on the jing layout (fig. 100). 97 Nine is a recurring number in the text. The capital measures nine li square, equaling to eighty-one jing in size. The outer court and the market each measure one fu, or one-ninth of a jing. While the jing mainly serves as a module for scale control (somewhat comparable to the neighborhood blocks in today's urban planning), the number nine is clearly permea ted with ritual connotations. It was a number appropriate only for the one and true Son of Heaven, whereas seven and five were numbers assigned to his vassals and lower ranks of aristocrats and officers. Though born out of an agrarian model, the jing has s ince become an imperial symbol, an archetype to be pursued by later Chinese cities. 98 Early capitals, including the Han Chang'an, never appeared to have truthfully followed, or realized, the ideal plan in every aspect. Rather, the jing functioned as a refer ence to be consulted, a perfection and a legacy to be honored. The closest, archaeologically excavated example is the Tang Chang'an, which was a square city with a checkerboard of roads, neighborhoods, and markets, though the imperial palace was placed not in the center but to the far north. The grid was stubbornly 97 Ibid. “[The palace should include] nine chambers in the inner court, where the nine imperial concubines reside, and nine halls in the outer court, where the nine ministers took office. The land of the capital should be divided into nine districts to be administered by the nine ministers 內有九室, 九嬪居之 ; 外有九室, 九卿朝焉. 九分其國, 以為九分, 九卿治之. ” 98 The development of Chinese imperial cities and its relation to city-planning methods in the Kaogongji is the main subject of He 1985 ; see, for instance, p. 141.	di-luo-a-grain-of-sand-full-draft-3.pdf
174 applied to the Tang Luoyang where it was superimposed onto the unruly waterways. These are the most well-known and exemplary cases of city planning with far-reaching influences in history. A les s rigorous adaptation was the Dongjing of Northern Song, a metropolis later occupied by the Jurchen emperors. The reconstruc tive plan of the city (fig. 101) show s a group of three nested squares where the imperial palace occupies the center. In fact, the ideal plan could not have been followed strictly since the topography, especially the network of waterways, needed to be taken into account. Yet residents and visitors of Dongjing (who lacked any satellite images or aerial views of the city ) might have very well perceived it to be closely adhering to the model, as indicated by a thirteenth-century map (fig. 102). The Central Capital (Zhongdu 中都, modern-day Beijing) of the Jin was built in e mulation of Dongjing. It was a square roughly 35. 52 li long and wide, with three gates per side, and an imperial palace in the middle. 99 The same pattern was generally followed in the construction of Beijing in the Yuan, Ming, and Qing dynasties. Miniature cities as a distinctive artistic motif emerged in the tenth century. The silk painting from the Dunhuang Library Cave discussed in C hapter 3 includes a “giant” lotus at the bo ttom, which floats above a cloud and contains a checkerboard of tiny enclosures each appearing to be a residential courtyard (see fig. 67). The checkerboard, fairly comparable to the layout of Tang Chang'an, is believed to be a re presentation of the Buddhist vision of a universe made of multiple world-systems--a vision which the painter saw fit to be conveyed through depicting a miniature metropolis. Curiously, the main body of the painting features a jing layout, framing the paint ed surface into nine sections each devoted to one episode of the Buddha teaching at the Nine Assemblies. Here the nine sections do not necessarily denote geographical discrepancies (the Nine Assemblies are said to be held in seven or eight, instead of nine, locations), but each labels a 99 He 1985, 5-6.	di-luo-a-grain-of-sand-full-draft-3.pdf
175 different time. In other words, a spatial pattern is in this case projected on a temporal one, or more precisely a spatio-temporal one. The spatio-temporal framework (what Eugene Wang refers to as the “chronotope”) becomes a visual formula fitful for the di splay of the omnipresent Buddha; it also share s certain commonalities with the mandala, which had been merged into the jing layout by the Tang. 100 The adding of the temporal dimension to the jing was first attempted by pre-Qin scholars in planning the legendary architecture mingtang. The actual form and structure of the mingtang are forever elusive: scholars over the history have argued for either a five-or nine-chambered layout, which resembled the character ya 亞 or the jing (fig. 103). The “Yueling 月令 (Monthly ordinances)” in the Liji 禮記 (Classic of rites) prescribes that the Son of Heaven should change his chamber of residence within the mingtang according to the passing of the twelve months, the c ycling of the four seasons, and the rising and descending of the heavenly and earthly qi (fig. 104). 101 The jing, therefore, was at once a temporal layout and a cosmic clock. It was as much ideal as the utopian city model and was em braced with as much imperial fervor: Wang Mang 王莽 (r. 8-23) and Wu Zetian 武則天 (r. 690-705), for instance, were avid patrons of the mingtang and left enough material remains for scholars to reconstruct their personal visions of this cosmic mansion (fig. 105). Was the design of the Jingtusi ceiling inspired by the ideal city, the mandala, and/or the mingtang ? The shared spatial configuration--jing--is an important clue to this question. The squarish 100 Chen Jinhua 陳金華, “Yixing yu Jiugong: yige Yindu sixiang Zhongguohua de li'an 一行與九宮 : 一個印度思想中國化的例案 (Yixiang and Jiugong: A case of the sinicization of Indian ideas ),” Journal of Shenzhen University (Humanities & Social Sciences) 31 (2014. 5): 122-23. Regarding the visual connection between jing and the mandala, Chen points out that the nine-quarter mandala was introduced by monk Yixing 一行 (673-727), who revised the o riginal six-quarter white sandalwood mandala in the Darijing 大日經 (Great Sun Sutra) and intentionally incorporated it into the magic square recorded in the Luoshu to bring in the latter's directional and numerological significances; such was the sinicizat ion of an esoteric pattern. 101 For a n examination of the historical mingtang, see Ming-chorng Hwang, “Ming-tang: Cosmology, Political Order, and Monuments in Early China,” Ph D diss., Harvard University, 1996.	di-luo-a-grain-of-sand-full-draft-3.pdf
176 plan of the Jingtusi Main Hall and its ceiling, the nine orderly arranged coffers, the nearly unbroken wooden galleries along the walls, and the baldachin roofs over the murals--three on the east and three on the west, similar in location to the city-gates--all suggest an unmistakable impr int of ideal city planning. The central cof fer is the climax of the design: with the four dazzling, golden miniature halls, it is the “imperial palace” within the “miniature city. ” This is not to say that carpenters who made this very ceiling consciously m odeled their work after Northern Song or Jin capitals and palaces, but ideal city models must have deeply influenced their way of thinking and designing space--be it exterior or interior, large or small. The ceiling, while not necessarily intended so, can be experienced as an imperial capital in miniature. Of course, t his is not the only case where miniature “cities” were created in ceiling s, but similar examples are found in contemporary as well as later designs (fig. 106). 102 In th is light, the symbolism of the tiangong louge and the “dome of heaven” need s be reexamined. The word tiangong literally means “heavenly palace,” which is supposed to be evoked by a group of miniature architecture. However, architecture alone cannot generat e a “palace”--a meaningful place or genius loci--but it has to be organized in certain ways to form a proper enclosure. 103 It was only after the tiangong louge entered the ceiling that the “palace” started to take a more conv incing shape. As Chinese miniatur ists have shown us, even heavenly cities and realms have to observe a certain spatial order and hierarchy, while the everyday world is in fact an imperfect project ion of the celestial and ideal. In essence, any designed environment, whether inhabited by 102 For instance, the ceilings of certain Yulin Cave s excavated during the Xi Xia period feature a Buddhist mandala which looks like a walled city sometimes drawn in a (recursive) jing-layout. 103 In the illustrations in the Yingzao fashi, these miniature buildings are organized no t according to any specific plans but simply in a linear fashion, forming a straight-line facade or an octagonal or U-shaped enclosure. The jing layout cannot be identified anywhere in these illustrations because it is inherently planar, whereas the miniat ure woodwork attached to furniture pieces would invariably form vertical, instead of horizontal, interfaces of the interior. For the concept of genius loci (spirit of place) and its application in architectural phenomenology, see Christian Norberg-Shulz, Genius Loci: Towards a Phenomenology of Architecture (New York: Rizzoli, 1984).	di-luo-a-grain-of-sand-full-draft-3.pdf
177 humans or gods, could neither be conceived nor perceived without an epistemological, spatial frame of reference. The vision of heaven expressed at the Jingtusi, therefore, is a religious as well as an imperial one. Buddhist icons and pictorial motifs have b een freely combined with the image of the imperial city, an ideal pattern sanctified over the history and passed down as part of the Confucian intellectual tradition. The representation of a world of myriad worlds had to be aided by miniature-making, which was to involve a neatly arranged group of many buildings based on a rhythmic, forever expandable and multipliable pattern of the jing. Conclusion Official and local histories have revealed that the Jingtusi cannot to be simply labeled as Jin architecture, but it was built during the tumultuous years around 1124 when the Northern Song, Liao, and Jin were having their final struggles for superiority and survival. National boundaries and ethnic differences aside, it manifested historical continuations rather than gaps, highlighted advancement rather than stagnation in artistic creation. The Jingtusi commenc ed a golden age in Chinese archit ectural history when miniature-making--including mini theaters and ruled-line paintings--started to dominate ritual and religious space s, presenting the audience an ever-expanding interior which brought much pleasure and insight when gazed upon. The beauty of the Jingtusi ceiling lies not only in the tiangong louge--the miniature towers and pavilions--but also in the nine cei ling coffers and their overall layout. The coffers are themselves miniaturized domes tracing back to the primitive laternendecke ceiling. By the third century, the Chinese dome had become a dysfunctional part of the roof frame, and its focus had shifted fr om maintaining structural stability to the symbolism of heaven, which would be brought to full strength in the twelfth century by the dream-inducing tiangong louge. The nine-square configuration of the	di-luo-a-grain-of-sand-full-draft-3.pdf
178 ceiling, on the other hand, is derived from the jing, a geometric pattern tracing back to ancient land distribution and administration systems. The jing soon became an archetype for Chinese cities and palaces because of the connotations of egalitarianism and humanness associated with it by Confucians. The vis ual representation of the “ Heavenly Palace,” therefore, had to incorporate not just stately buildings but also the most ideal layout.	di-luo-a-grain-of-sand-full-draft-3.pdf
179 5. Miniatures, Models, Simulacra At the Chongfusi, a Jin-dynasty Buddhist monastery in Shuoxian 朔縣, Shanxi, the art of miniature architecture is exemplified by a free-standing architectural model of a three-story pavilion (fig. 107). This model, unlike the three key specimens examined in the previous chapters, does not fall into the Yingzao fashi category of small-scale woodworking, and is in many ways different from the miniature shrines and repositories discussed earlier. In fact, t he subject of this chapter--architectural models at the Chongfusi and elsewhere--is never discussed in the official manual. Understandably, a model is not usually considered a part of architecture; it is more individual and rarely raises public or state concerns about budget, resource, labor, or law. A model is often vie wed not as an “end product” but a work in progress or an intermediate, existing mainly in the transitory, fast-changing design process. A model may also be kept as a collectible, a copy, or even as some visual “ record ” of historical images and memories, es pecially when it remains as the sole survivor and witness to the art and architecture of its time. As the final chapter of this dissertation, here I intentionally cross over geographical boundaries and the pre-defined time-frame (eleventh to thirteenth cen turies) and extend my discussion to a longer, continuous tradition of miniature-making in Chinese history. The aim is not to focus on one example but rather to go beyond particularities and return to more general and essential issues considering miniaturiz ation as an epistemological process. This does not mean that geographical and historical contexts become irrelevant, but in this particular case, the intrinsic attributes of all architectural models have allowed me to shift the center of focus from peculia rities to generalities. As introduced in Chapter 1, a model is commonly referred to as xiaoyang, or simply yang 樣, in Chinese written records. Here I further delve into the nature and roles of models in history--how	di-luo-a-grain-of-sand-full-draft-3.pdf
180 did models differ from other types of miniatures in terms of form, function, and technique? Were models copies of their large-scale equivalents, or vice versa? A greater interest lies in how models negotiated with the “real”--the indicated and signified. Were models imitations of pre-existing objects, or visualizations of conceived, yet-to-be-built, or never-meant-to-be-built projects (i. e. mental images) ? Contemplations of the relationship between the “original” and the “copy” have given rise to a great many critical discourses on the issues of mimesis, representation, and semiotics from the time of Plato and Aristotle onward. In our age of mass media, of industrial production and reproduction, and of computer-simulated virtual reality (movies, video games, computer-aided design, etc. ), the clash between “originals” and “copies” has resurfaced as a deep entanglement of the real and the unreal/hyper-real, b etween the corporeal and the imagined/fantastic. 1 The entanglement of reality and virtual reality is best exposed and addressed by the concepts of “simulation” and “simulacrum,” two terms historically fraught with the negative meanings of “fake” and “bad c opy” and have in recent decades been reinterpreted in a new light by philosophers such as Jean Baudrillard (1929-2007). In China, the ideas of mimesis and simulation can be traced to the I Ching 易經 (Book of Changes), where the dialectic of the “original” a nd the “copies” lies in how the qi 器 (vessels, apparatus, implements)--the designed--is in resemblance to the xiang 象 (images, forms, phenomena)--the natural, whether or not such a resemblance is manifest ed through appearance or mechanism. By engaging with on-going discussions on the simulacrum, this chapter proposes a new angle of viewing miniature architecture while provid ing insight into the interrelationship between miniaturization and simulation. 1 An inspirational source for theories on the dynamics between “reality” and “imagery,” the sign and signified, the original and the copy is Robert S. Nelson and Richard Shiff, eds., Critical Terms for Art History (Chicago and London: University of Chicago Press, 2003), especially Ch aps. 1-3 on concepts of representation, sign, and simulacrum.	di-luo-a-grain-of-sand-full-draft-3.pdf
181 The Model Pavilion at the Chongfusi The architectural m odel in question is a 3. 4-meter-tall wooden pavilion ( ge 閣) sitting on top of a one-meter-tall altar table (figs. 108, 109). 2 It features a typical hip-and-gable roof, with two additional layers of skirting eaves on the first and second levels. 3 The pavilion measures 140 by 102 centimeters across. The breadth of each bay appear s to have been well controlled, but the major columns--each thrusting to more than one meter in height--are so thin that they would have posed a threat to the overall structural integrity in real cases. 4 All three levels feature six-tiered bracketing which alternates between triple and double brackets (fig. 110)--a scheme not seen elsewhere5--and most bracket-sets simply end with a floral bracket arm at the top, where the suspended arm reaches out and supports the eaves-board immediately without cushions or longitudinal members, which is highly unusual. What also makes the model unusual is the total missing of the levering member in a bracket-set, which matured no later than the middle of the ninth century and prevailed in the Northern Song, Liao, Jin, and later dynasties. 6 The bracketing for the balcony is even simpler: only a scarce number of floral arms are 2 The dimensional data of this model come from my fieldwork (digitized and accessible a t https://sites. google. com/site/sourcesofchinesearchitecture/shanxi/chongfusi ), as well as Chai Zejun, ed., Shuozhou Chongfusi 朔州崇福寺 (Beijing: Wenwu chubanshe, 1996 ), 16. 3 Despite the three layers of roofs, it is sometimes referred to as a two-story building because the “attic” between the middle and topmost roofs has no balcony or windows. 4 Yingzao fashi regulates that the height of a b ay should always be smaller than its breadth. This rule has been generally followed in practice, with the exceptions of certain types of towers--such as bell/drum towers--which seem to have extra-long columns. Examples of these towers are seen in Dunhuang murals and some book illustrations. 5 Though the double bracket was a common feature of Chinese architecture from the tenth century onward, a triple bracket was an extremely rare, if not singular, case--could it be an archaic, and later abandoned scheme? O r was it an innovation by the modeler? More unusual about this woodwork is the fact that the double bracket is suspended from the centerline of the columns whereas the triple bracket sits at the center of a bracket-set. 6 The oldest examples of ang are fou nd in the Foguangsi 佛光寺 main hall (d. 857) in Wutaishan, showing a certain level of development. Earlier structures, including the oldest wooden structure in China--the Nanchansi 南禪寺 main hall (d.	di-luo-a-grain-of-sand-full-draft-3.pdf
182 used. Such an “aberrant” and simplified bracketing sche me implies a somewhat archaic style preceding the ninth century. 7 The cai of the model measures roughly 3 by 2 centimeters across,8 almost equal to the Jingtusi miniature, and about two-thirds of those of the Longxingsi and Huayansi repositories. What clearly differentiates the model from the other three, however, is the significant sparsity of brackets: the spacing of the bracket-sets supporting the lowest layer of eaves is 28 centimeters, equal to 140 fen (1 fen = 0. 2 centimeters), in contrast to the 100-120 fen adopted by the other miniatures. 9 As a result, the model only has some fifty bracket-sets in total. The sparsity of bracket-sets was characteristic of full-size buildings, for which economy and structural efficiency were usually prioritized over th e many luxuries enjoyed by miniature architecture. Ornamentation, on the other hand, has been applied in a similarly sparing way. Golden phoenixes and dragons, now much faded, are engraved on the surface of c ornices and columns (fig. 111). 10 Overall, it seems that the pavilion was not meant to impress its viewer by overflowing structural details, but it was to be appreciated because of the pure, well-articulated form--a form that appears “simplistic” in comparison with Northern Song, Liao, and Jin architecture but largely reminiscent of the robustness and forcefulness of the Tang. 782)--as well as one engraved on one of the lintels of Daya nta 大雁塔, show no traces of ang. Hence, the lack of ang in the model strongly suggest s an archaic scheme. 7 Alternatively, the modeler might have attempted to capture and revive such a style, or he had other concerns such as keeping with the budget and the deadline, which was less likely the reason. 8 This value was extracted from one of the bracket-sets on the first level of the model. This is an approximate as I was not able to perform a systematic measurement. 9 See Chapter 3 of this dissertation. 10 Tour guides to the monastery claim that this is a motif of “phoenixes above/ascending and dragons below/descending ( fengshang longxia 鳳上龍下 ),” and attribute the original sutra library which the model is said to be based upon to the period of Wu Zetian's reign.	di-luo-a-grain-of-sand-full-draft-3.pdf
183 Dating the model: a conundrum The exact date of the model pavilion is shrouded in the mists of history. According to official records, the origin of the C hongfusi is traced back to 665, when a Dazangjingge 大藏經閣 (Tripitaka Library) was established on site under imperial decree, and additional Buddhist halls were added surrounding the library to form a monastic precinct. 11 In the Liao, t he monastery first became occupied by a certain court official surnamed Lin 林 as his personal resi dence, but it was soon restored after spotting of miraculous light on the site were reported in 983-1012. In 1143, the Jurchen general Zhai Zhaodu 翟昭度 built the Mituodian, now the main hall of the monastery, and the Chongfusi started to assume its modern-day shape and layout. We are not told if the original library was still in use during the Liao or Jin, but apparentl y it had gone by the early Ming when a major r estoration project w as carried out. A record of this restoration is found on a silk dharani banner (d. 1383) discovered in 1955 above a lintel of the Mituodian : This monastery was converted from the Khitan Grand Master Linya's residence, which is why it is also called the Lin yasi. It was repaired in the Liao, Jin, Song, and Yuan dynasties. Coming to the Ming, during wartime, it was used as a storage for grains and supplies, while the monks scattered to the four wilderness. Holy images and murals were all destroyed, there were no bricks or stone left on the site, and only one dormitory survived. Later, in the fifteenth year of the Hongwu Period (1382), thanks to the imperial edict, offices of Buddhist and Daoist monks were established under heaven. I, monk Lixiang, was elected t o take the entrance exam administered by the Department of Rituals, and [passing the exam,] I received the official title of Sengzheng (Head of monks) of this prefecture. In the middle of my reconstruction of the local monastic community, on the fourteenth day, fifth month of the sixteenth year of the Hongwu Period (1383), I came across Grand Minister Xie Cheng (1339-1394), who was traveling in this prefect and paying a visit to the Chongfusi. He saw dismantled halls and dilapidated statues and was sympathe tic; to repair them, he ordered the stored grains and 11 Chai Zejun 1996, 3-10, gives a n outline of the history of the monastery. The primary sources include: nine Jin-dynasty inscriptions found in the Mituodian stating that the building project of the hall started in 1143 whereas the paintings and interior dec oration were completed in 1153. Other on-site inscriptions and steles date between 1354 and 1884, the contents of which can be found on pp. 391-401. Textual sources include Shanxi tongzhi, Huanyu tongzhi 寰宇通志, Yongzheng Shuozhouzhi 雍正朔州志, and Qianlong Shuo zhouzhi 乾隆朔州志.	di-luo-a-grain-of-sand-full-draft-3.pdf
184 supplies to be removed completely and the carpenters to start their work immediately. He appointed me and other officials as supervisors of the restoration project. 此寺契丹國臨衙太師改宅為寺, 因立異號臨衙寺. 後遼金宋元歷代重修. 至大明兵興, 設為倉所, 屯放糧儲, 僧各散於四野 ; 聖容壁飾具摧, 基址亦無磚石, 僧舍惟存壹廈. 後洪武十五年壬戌歲, 欽蒙勑旨, 天下開設僧道衙門, 選舉僧立祥前赴禮部發僧録司考試, 得參究禪學. 除授本州僧正司僧正. 整致院門間, 於洪武十六年五月十四日忽遇大臣永平侯謝大人出巡到此, 謁見本寺. 殿宇崩摧, 聖像損壞, 哀憐古寺, 以可重修, 將原囤糧儲即時般運一空, 隨命諸匠即日興工, 令本衛指揮孫等官監修重造, 以為記耳. It was during this time that the Qian foge 千佛閣 (Tho usand-Buddha pavilion) (fig. 112) was erected on the ruin of the original Tripitaka L ibrary. 12 The Tripitaka has been lost together with the library ; what is left to this day is just the model pavilion, sitting at the center of the Qianfoge as the main object of worship. Due to the lack of written records, we are not sure if the model was made during the 1383 restoration or a woodwork from an earlier (or later) date. One theory claims it to be one of the several design proposals made by Ming carpenters to restore the Tang library. 13 Indeed, dating becomes an especially difficult task here. Nonetheless, one can infer from an analysis of how the miniature is downscaled and stylized, and how it is placed in its immediate physical environment. As observed earlier, the design of the model pavilion still retains an identifiable Tang vigor: the bracketing scheme appears fairly archaic and atypical, the spacing of bracket-sets is unusually large, and several advanced structural m embers are missing. In terms of its intended function, the pavilion was obviously not designed to shelter any icons or scriptures. 14 Unlike shrines and repositories which are essentially receptacles, the miniature in this case remains to be a piece of archi tecture (and art), which steps from the background into the spotlight and becomes 12 Ibid., 15. The dating of this structure is relied on historical records and architectural style. 13 Ibid., 16. See also p. 367, description of pl. 25, claiming the model to be “one of several design proposals of the Qianfoge during the Ming restoration project. ” We do not know how much it resembles the Tang library. 14 A moderate-size statue of the Big-belly Maitreya Buddha cannot fit in but has to be placed in front of the model ; on the north side is a smaller figurine of Weituo 韋陀 barely tu cked into the narrow porch of the model.	di-luo-a-grain-of-sand-full-draft-3.pdf
185 the “protagonist ” of drama. It is the appearance--rather than the content--that derives the sense of sacredness of the religious space. Therefore, it is a likely scenario tha t the pavilion is a fourteenth century work modeled, to some extent, after the Tang library. It would make sense that after the loss of the Tripitaka, the model was made and placed in the most prominent place of the new building to “stand for” the original library and serve as a visual link to, and a commemoration of, the Tang legacy. This hypothesis is supported by the fact that the model and the Qianfoge share a similar plan--a three-by-three column grid. When the Qianfoge was built on top of the ruin, it is very likely that builders chose to utilize what had been left on the site--especially the foundation and the stone plinths of the columns--which was perhaps also a three-by-three grid. It is interesting to think that even though the original is gone fo rever, part of it might have been rematerialized--in a however distorted and fragmentary way--by the juxtaposition of the miniature and the full-size. A note on scale The cai of the model pavilion is 3 by 2 centimeters--how does this help dating ? As noted above, this value is close to the cai of the Jingtusi tiangong louge, which is presumably based on a 1/10 scale. 15 The 1/10 scale is not only observed in miniature woodworking and ruled-line painting, but it was adopted in architectural drawing. According to the Yingzao fashi, before putting up the roof frame of a building, carpenters needed to draw sketches on the wall to figure out the correct structure and profile of the roof. To build a roof could easily go wrong; it was such a demanding task that the ceyang 側樣, a sectional drawing of the building (fig. 113), was required for necessary calculation and clarification: 15 See Chapter 4 of this dissertation.	di-luo-a-grain-of-sand-full-draft-3.pdf
186 Set up a [1/10] scale so that one chi equals one zhang, one cun equals one chi, one fen equals one cun, one li equals on fen, and one hao equals one li. Draw a cross section of the roof to be built on the flat surface of the wall to help determine the steepness of the slopes and the roundness of the curves. Then one can examine the elevation s of the beams and columns, and the distance between the openings (mortises) of different structural members. 先以尺為丈, 以寸為尺, 以分為寸, 以厘為分, 以毫為厘. 側畫所建之屋於平正壁上, 定其舉之峻慢, 折之圜和. 然後可見屋內梁柱之高下, 卯眼之遠近. 16 Did the 1/10 scale also apply to architectural models? While not many models have survived in Medieval China, two miniature pagodas from Nara-period Japan might shed some light on this issue. One of the pagodas (d. 710-750), a five-story, 4. 1-meter-tall wooden structure kept at the Kairyooji 海龍王寺 in Nara, is belie ved to be a 1/10 model (fig. 114). 17 The other one (d. 751-794), also a 1/10 model, is at the Gangoji 元興寺 in Nara ( see fig. 114). 18 Similar to the Chongfusi model, neither of the two pagodas have any apparent “function” of containing or preserving Buddhist images or scriptures; the reason why they were made was little know n except that they were excellent displays of themselves. In terms of technique and style, they have expressed a striking consistency with full-size pagodas built in eighth-century Nara such as the Yakushiji 藥師寺 east pagoda and the Muroji 室生寺 pagoda. 19 This means that the miniatures might have been made as 16 Yingzao fashi vol. 1, 34. This is the method of juzhe 舉折 (lit. raising and bending [the roof members]). 17 Similar to the Chongfusi model, it is espe cially well-articulated in the details of architecture--the sizes and exact locations of each structural members, the proportions between different parts, the gradual tapering of the body toward the top, the spire with multi-layered disks and the ornamenta l finial, etc. See F u Xinian, “Riben Feiniao Nailiang shiqi jianzhu zhong suo fanyingchu de Z hongguo Nanbeichao Sui Tang jianzhu tedian 日本飛鳥奈良時期建築中所反映出的中國南北朝隋唐建築特點 (Characteristics of Chinese architecture of the Six Dynasties, the Sui, and the Tang periods as reflected by Japanese Architecture of the Asuka and Nara periods),” Wenwu (1992. 10): 28-50, reprint in Fu Xinian, Fu Xinian jianzhushi lunwenji (Beijing: Wenwu chubanshe, 1998), 147-67. 18 Fu 1998, 161-62. Reaching to a height of 5. 5 meters with its fiv e stories and a thrusting spire, it looks fairly similar to the Kairyooji pagoda. 19 Ibid., 158-65.	di-luo-a-grain-of-sand-full-draft-3.pdf
187 pilot models prior to the construction of real pagodas, or as post-construction models of exemplary pagodas to be studied and emulated by later projects. Assuming that the Chongfusi model has also adopted a 1/10 scale, this would mean that it was intended to propose (or emulate) a 34-meter-tall structure using a cai of 30 by 20 centimeters. While it is no longer possible to compare the model with the Tang library, one can nevertheless compare it with other surviving wooden buildings. Below, a list of representative wooden structures from the eighth to twelfth centuries shows the typical range of cai in relation to the rank of the building (indicated by the number of bays lengthwise):20 Name of Structure Date Rank Cai (in height) Location Main Hall, Nanchansi 南禪寺 782 3-bay 24 cm Wutaishan, Shanxi Main Hall, Foguangsi 佛光寺 857 7 30 Wutaishan, Shanxi Main Hall, Geyuansi 閣院寺 966 3 26 Laiyuan, Hebei Gate, Dulesi 獨樂寺 984 3 24. 5 Jixian, Tianjin Guanyinge 觀音閣, Dulesi 984 5 24 Main Hall, Baogu osi 保國寺 1013 3 21. 5 Ningbo, Zhejiang Qianfodian 千佛殿, Fengguosi 奉國寺 1020 9 29 Yixian, Liaoning Shengmudian 聖母殿, Jinci 晉祠 1020s 7 21 Taiyuan, Shanxi Main Hall, Kaishansi 開善寺 1033 5 23. 5 Gaobeidian, Hebei Bojia jiaozang 薄伽教藏, Huayansi 1038 5 23. 5 Datong, Shanxi Wooden Pagoda, Fogongsi 佛宮寺 1056 3 25. 5 Yingxian, Shanxi Xiandian 獻殿, Jinci 1068 3 21 Main Hall, Shanhuasi 善化寺 11th c. 7 26 Datong, Shanxi Wenshudian 文殊殿, Foguangsi 1137 7 22. 5 Main Hall, Huayansi 1140 9 30 20 The list has been compiled based on two main sources: Chai Zejun 1996, 61-62, which includes a chart comparing the cai of a number of Song, Liao, and Jin w ooden buildings; and Wang Guixiang 王貴祥, “Fujian Fuzhou Hualinsi Dadian yanjiu 福建福州華林寺大殿研究 (A study on the Main Hall of Hualinsi in Fuzhou, Fujian),” in Wang Guixiang, Liu Chang 劉暢, and Duan Zhijun 段智鈞, eds., Zhongguo gudai mugou jianzhu bili yu chidu yanji u 中國古代木構建築比例與尺度研究 (Studies on the proportion and scale of historical Chinese wooden architecture) (Beijing: Zhongguo jianzhu gongye chubanshe, 2011), 183-84.	di-luo-a-grain-of-sand-full-draft-3.pdf