University of Southern California Dissertations and Theses

A grain of sand: Yingzao fashi and the miniaturization of Chinese architecture

(USC Thesis Other)

A grain of sand: Yingzao fashi and the miniaturization of Chinese architecture

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

A Grain of Sand:
Yingzao Fashi and the Miniaturization of Chinese Architecture

Di Luo

A Dissertation
Presented to
Faculty of the USC Graduate School
University of Southern California
In partial fulfillment of the
Requirements for the degree
Doctor of Philosophy
(East Asian Languages and Cultures)

August 2016

Dissertation Committee
Professors Dominic Cheung (Chair), Sonya S. Lee, Bettine Birge
ii
Acknowledgements
My sincerest thanks go to the three distinguished scholars on my dissertation committee: Professors
Dominic Cheung, Sonya Lee, and Bettine Birge, who have directed and supervised the entire process
of my dissertation research and writing. Their remarkable scholarship in the fields of Chinese
literature, art history, and cultural history have been an inexhaustible wellspring of knowledge and
inspiration for me to tap over the years of my graduate study. They have guided and supported every
step of my academic journey with utmost patience and care.
I have received enormous help from many other professors in and outside USC. Professor
George Hayden has given me useful tips and suggestions for my translation of the Yingzao fashi and
my understanding of Chinese drama. Professor James Steele of the USC School of Architecture,
who was the advisor of my M.Arch. thesis, discussed my dissertation project with much euthusiam
and provided great insight into a lot of conceptual issues from his own professional perspective as
an architect and architectural historian. Professors Min Li, Richard von Glahn, and Katsuya Hirano
at UCLA have encouraged me to approach my study from the angles of a variety of disciplines
including landscape archaeology, history of religion, and popular culture. Professors Jeehee Hong
and Youn-mi Kim have generously shared with me their most recent studies on Liao architecture
and art, whereas Professor Stephen West has answered my questions about translating particular
terms in Northern Song miscellanea, for which I am grateful.
Professor Nancy Steinhardt has spent much of her own time reading and commenting on
my dissertation prospectus and chapters. Her deep knowledge of Chinese architectural history has
informed and influenced my own work significantly, and my conversations with her helped to shape
the overall theme and framework of the dissertation from its very inception.
My 2014 fieldwork in China--during which I was able to collect first-hand information and
data of the case studies presented in this dissertation--could not have been successful without the
iii
assistance of my academic advisor, Christine Shaw. Christine critiqued my research proposal and
agenda, helped me secure funding, and prepared necessary paperwork for my travel. Professors Li
Luke, Fang Xiaofeng, and Liu Chang at Tsinghua University helped me contact local authorities to
gain access to several restricted architectural sites.
This dissertation has received multiple fellowships and grants from the USC Graduate
School, the USC Department of EALC, the Metropolitan Center for Far Eastern Art Studies, and
the Harvard-Yenching Library. I am deeply thankful for their generous financial support.
My family has always been my strongest and most cherished source of courage, faith,
strength, and willpower. This dissertation is dedicated to them.

ii
Contents
List of Figures ........................................................................................................................... vi
Introduction ............................................................................................................................... 1
On Defining the Miniature: Philosophical, Religious, and Architectural Perspectives ....................... 3
Conceptual and Theoretical Framework: Deconstruction, Oneirism, and Simulation ....................... 8
Scholarship on the Miniature ..................................................................................................................... 12
Primary Sources, Digital Database, and 3D Modeling .......................................................................... 17
1. Miniatures in Texts .......................................................................................................... 21
Small-scale Woodworking (Xiaomuzuo) in the Yingzao Fashi ................................................................. 21
Types of miniature woodwork: shrines, repositories, and the “Heavenly Palace” ....................................... 23
Scaling and the cai-fen system .............................................................................................................................. 26
Models (Xiaoyang) and Ruled-line Paintings (Jiehua) ............................................................................... 29
Spiritual Vessels, Edible Architecture, Portable Shrines, Dollhouses, and Miniature Gardens ...... 32
Spiritual vessels ........................................................................................................................................................ 33
Edible architecture .................................................................................................................................................. 36
Shaluo shrines .......................................................................................................................................................... 37
Mohouluo dolls and dollhouses ............................................................................................................................ 39
Miniature gardens .................................................................................................................................................... 44
Puppets and the theatricality of miniatures ......................................................................................................... 45
Conclusion: Dreaming of Lilliput in Song China ................................................................................... 48

iii
2. Miniatures as Sacred Repositories, Part I: The Longxingsi Sutra Case ......................... 51
The Zhuanlun Jingzang (Wheel-turning Sutra Repository) at Longxingsi .............................................. 52
Dating the miniature: textual evidence ................................................................................................................ 54
Dating the miniature: a comparison with Yingzao fashi .................................................................................. 59
“Progressive miniaturization” in Chinese architectural history ....................................................................... 62
The Revolving Sutra Case in History ....................................................................................................... 64
Sixth century: legendary beginnings ..................................................................................................................... 66
Tang ........................................................................................................................................................................... 68
Northern and Southern Song ................................................................................................................................ 73
Yuan and later .......................................................................................................................................................... 78
Miniaturization as Deconstruction ........................................................................................................... 81
The octagon ............................................................................................................................................................. 83
The central pillar ..................................................................................................................................................... 86
The wheel ................................................................................................................................................................. 89
Conclusion: the Revolving World in a Nutshell ..................................................................................... 92
3. Miniatures as Sacred Repositories, Part II: The Huayansi Sutra Cabinets .................... 94
The Bizang (Wall Repository) at the Huayansi ........................................................................................ 95
Scale and form ......................................................................................................................................................... 97
Discovery, dating, and identification ................................................................................................................... 99
Redefining Liao architecture ............................................................................................................................... 105
The art historical perspective .............................................................................................................................. 107
Repositories, Shrines, Cabinets .............................................................................................................. 109
In worship halls ..................................................................................................................................................... 110
In monastic living quarters .................................................................................................................................. 111
iv
In houses ................................................................................................................................................................ 114
The Tamamushi Shrine: a distant echo from Japan ........................................................................................ 117
The Miniature and the Myriad ................................................................................................................ 119
The Flower Repository Universe ........................................................................................................................ 120
Indra’s Net ............................................................................................................................................................. 122
Sudhana’s epiphany in the Tower of Vairocana ............................................................................................... 124
Fazang’s mirror hall: the art of Huayan Buddhism ......................................................................................... 128
Conclusion ................................................................................................................................................. 130
4. Miniatures in the “Dome of Heaven” ........................................................................... 132
The Tiangong Louge Zaojing (Coffered Ceiling with Heavenly Palace Towers and Pavilions) at the
Jingtusi........................................................................................................................................................ 133
Tiangong louge, the “Heavenly Palace” ............................................................................................................ 134
Xiaodouba zaojing, the miniature octagonal ceiling coffer ............................................................................ 136
Jing, the magic square, and ceiling compartmentalization .............................................................................. 138
Miniature-making in Jurchen-Jin Material Culture .............................................................................. 140
Characteristics of Jin architecture: a revision.................................................................................................... 144
Miniature theaters ................................................................................................................................................. 147
Ruled-line painting ................................................................................................................................................ 149
The ethnic dimension ........................................................................................................................................... 152
Symbolism of the Chinese Dome .......................................................................................................... 155
Zaojing, the “water-weed well” .......................................................................................................................... 157
Wooden “domes of heaven” from the tenth century onward ....................................................................... 160
Ceiling Design and City Design ............................................................................................................. 165
The well-field and Neo-Confucianism .............................................................................................................. 166
v
The ideal city in miniature ................................................................................................................................... 172
Conclusion ................................................................................................................................................. 177
5. Miniatures, Models, Simulacra ...................................................................................... 179
The Model Pavilion at the Chongfusi ................................................................................................... 181
Dating the model: a conundrum......................................................................................................................... 183
A note on scale ...................................................................................................................................................... 185
The original and the copy .................................................................................................................................... 188
Modeling in Chinese History .................................................................................................................. 191
Modeling and drafting in the design process .................................................................................................... 192
Miniature pagodas and King Asoka’s 84,000 stupas ....................................................................................... 197
Armillary spheres and celestial globes: in simulation of heavenly images ................................................... 203
I Ching on the notion of simulation .................................................................................................................. 206
Conclusion ................................................................................................................................................. 208
Conclusion ............................................................................................................................. 210
Figures ................................................................................................................................... 215
Glossary ................................................................................................................................. 293
Bibliography .......................................................................................................................... 298

vi
List of Figures
1 Illustration of tiangong louge fodaozhang ................................................................................................. 215
2 Illustration of tiangong bizang ................................................................................................................ 215
3 Reconstructive drawings of douba zaojing, plan and section ............................................................ 216
4 Reconstructive drawings of xiaodouba zaojing, plan and section ..................................................... 217
5 Eight grades of cai in large-scale woodworking ................................................................................ 218
6 Six grades of cai in small-scale woodworking ................................................................................... 218
7 Guo Zhongshu, Summer Palace of Emperor Minghuang, detail ............................................................ 219
8 Tamamushi Shrine, detail of roof ....................................................................................................... 219
9 Illustration of huasheng .......................................................................................................................... 220
10 Line drawing of mural on east ceiling slope of Mogao Cave 31, showing a woman holding a
Mohouluo doll ....................................................................................................................................... 220
11 Wang Zhenpeng, Dragon Boat Regatta, partial .................................................................................... 221
12 Wang Zhenpeng, Dragon Boat Regatta, detail ..................................................................................... 221
13 Longxingsi sutra case, overview ......................................................................................................... 222
14 Sectional drawing of Zhuanlunzangdian at Longxingsi .................................................................. 223
15 Plan of Zhuanlunzangdian at Longxingsi ......................................................................................... 224
16 Bottom of pivot of Longxingsi sutra case ......................................................................................... 224
17 Master plan of Longxingsi ................................................................................................................... 225
18 Rhino 3D model of Longxingsi sutra case ........................................................................................ 226
19 Reconstructive drawings of zhuanlun jingzang, plans, elevation, section, and details of brackets227
20 Reconstructive drawing of zhuanlun jingzang, elevation and section .............................................. 228
21 Rotating core of Longxingsi sutra case .............................................................................................. 228
22 Longxingsi sutra case in the 1920s ..................................................................................................... 229
23 Longxingsi sutra case in the 1930s ..................................................................................................... 230
24 Longxingsi sutra case, detail of corner set ........................................................................................ 231
vii
25 Rhino 3D model of corner bracket sets of Longxingsi sutra case ................................................ 231
26 Longxingsi sutra case, detail of column-top and intercolumnar bracket sets .............................. 232
27 Zhuanlunzangdian at Longxingsi, detail of exterior bracket sets .................................................. 232
28 Diagram showing historical development of Chinese bracketing system .................................... 233
29 Cornice of Tianwangdian at Longxingsi, showing Qing bracket sets arrayed among Song
originals .................................................................................................................................................. 234
30 Daoxuan’s layout of ideal monastery, detail ..................................................................................... 234
31 Elevation of Yunyansi feitianzang ........................................................................................................ 235
32 Beishan Cave 136, interior ................................................................................................................... 236
33 Baodingshan Cave 14 ........................................................................................................................... 237
34 Drawing of Jinshansi revolving sutra case ........................................................................................ 238
35 Pingwusi revolving sutra case ............................................................................................................. 239
36 Gaolisi revolving sutra case ................................................................................................................. 240
37 Modern revolving sutra case installed by Tai Xiangzhou in a 2010 exhibition in Shanghai ...... 240
38 Yungang Cave 1, interior ..................................................................................................................... 241
39 Yungang Cave 2, detail of central pillar ............................................................................................. 241
40 Northern Wei miniature stupa from Gansu ..................................................................................... 242
41 Yingxian Wooden Pagoda ................................................................................................................... 242
42 Sanjie jiudi zhi tu 三界九地之圖 ......................................................................................................... 243
43 Su Song’s clock-tower .......................................................................................................................... 244
44 Wooden pagoda of Su Song’s clock-tower ....................................................................................... 244
45 Huayansi Bojia jiaozang, interior view ............................................................................................... 245
46 Sectional drawing of Huayansi Bojia jiaozang .................................................................................. 245
47 Huayansi Bojia jiaozang, ambulatory along the back of the central altar ..................................... 246
48 Huayansi Bojia jiaozang, ambulatory along the south wall ............................................................. 247
49 Huayansi sutra cabinets, detail of bracket sets ................................................................................. 248
viii
50 Arching bridge of Huayansi sutra case .............................................................................................. 249
51 Huayan Plaza in front of Huayansi .................................................................................................... 249
52 Arching bridge of Huayansi sutra case, detail ................................................................................... 250
53 Wooden miniature shrine in Binglingsi Cave 172 ............................................................................ 250
54 Yungang Cave 6, detail of central pillar ............................................................................................. 251
55 Erxianmiao miniature Daoist shrine .................................................................................................. 252
56 Huhuangmiao miniature shrine, detail of roof corner .................................................................... 252
57 Elevation of miniature shrine in Buddhist dormitory ..................................................................... 253
58 Plan of Jinshansi dormitory ................................................................................................................. 253
59 Drawings of Jingshansi miniature shrine ........................................................................................... 254
60 Plan of Jingshansi dormitory ............................................................................................................... 254
61 Line drawing of stone relief, north wall of east side-chamber, Dahuting Tomb 1 ..................... 255
62 Line drawing of stone relief, north wall of north side-chamber, Dahuting Tomb 1 .................. 255
63 Front elevation, section, and plan of shenchu, according to Lu Ban jing ......................................... 256
64 Yangshi Lei miniature shrine .............................................................................................................. 256
65 Reconstructive drawing of bizang, section ......................................................................................... 257
66 Diagram of the typology of Japanese zushi ....................................................................................... 258
67 Five Dynasty silk painting of “Seven Locations and Nine Assemblies” ...................................... 259
68 Bianxiang of Huayanjing, Mogao Cave 85, detail of Lotus Repository World ............................... 260
69 Bianxiang of Amitabha’s pure land, Mogao Cave 321...................................................................... 260
70 Compound eye of a fruit fly, detail .................................................................................................... 261
71 Mordern installation of Fazang’s mirror hall .................................................................................... 261
72 Main Hall of Jingtusi, west elevation ................................................................................................. 262
73 Scematic plan of Jingtusi ceiling ......................................................................................................... 262
74 Central coffer of Jingtusi ceiling, above the main Buddha ............................................................. 263
ix
75 Jingtusi Main Hall, interior view ......................................................................................................... 264
76 Miniature golden halls in central coffer of Jingtusi ceiling ............................................................. 264
77 Miniature Buddhas painted in central coffer of Jingtusi ceiling ..................................................... 265
78 Seven-tiered, fan-shaped bracket set at the southwest corner of Jingtusi ceiling ........................ 265
79 Double brackets in east coffer (Coffer E) of Jingtusi ceiling ......................................................... 266
80 West coffer of Jingtusi ceiling ............................................................................................................. 266
81 Partial view of Jingtusi ceiling, showing a combination of three different geometric shapes:
diamond, octagonal, hexagonal ........................................................................................................... 267
82 Baldachin roof above a painted Buddha at Jingtusi ......................................................................... 268
83 Sixteenth-century map of Yingzhou, showing location of Jingtusi ............................................... 269
84 Miniature bracket sets in the ceiling of Mituodian at Chongfusi ................................................... 269
85 Miniature theater in Houma Tomb 1 ................................................................................................ 270
86 Actor figures and a theater pavilion in Macun Tomb 4 .................................................................. 270
87 Line drawing of mural on the west wall of Manjusri Hall, Yanshansi .......................................... 271
88 Reconstructive plan of main building complex painted in Yanshansi murals ............................. 272
89 Coffered ceiling in Main Hall of Shanhuasi ...................................................................................... 272
90 Tiangong louge in ceiling coffer, Rear Hall of Fengshengsi................................................................ 273
91 Tiangong louge in ceiling, Offering Pavilion of Doudafuci ................................................................ 273
92 Tiangong louge in ceiling, Main Hall of Yong’ansi .............................................................................. 274
93 Tiangong louge in ceiling of Gongshutang ............................................................................................ 274
94 Zhihuasi ceiling coffer, now in the Philadelphia Museum of Art ................................................. 275
95 Zhihuasi ceiling coffer, detail of tianggong louge.................................................................................. 275
96 Tiangong louge in circular coffer, originally from Longfusi ............................................................... 276
97 Diagram of well-field system .............................................................................................................. 277
98 Diagram of Warring-states land-allocation system for administrative purpose, according to
Zhouli ....................................................................................................................................................... 277
x
99 Plan of ideal capital city, according to Kaogongji ................................................................................ 278
100 Plan of imperial palace, according to Kaogongji ................................................................................. 278
101 Reconstructive plan of Northern Song Dongjing ............................................................................ 279
102 Thirteenth-century map of Northern Song Dongjing .................................................................... 279
103 Plan of five-chambered mingtang ......................................................................................................... 280
104 Reconstructive plan of Zhou-dynasty mingtang ................................................................................. 280
105 Reconstructive elevation of Wu Zetian’s mingtang............................................................................ 281
106 Mandala city painted in ceiling, Yulin Cave 3 (d. Xi Xia) ............................................................... 281
107 Model pavilion at Chongfusi ............................................................................................................... 282
108 East-west cross section of Qianfoge .................................................................................................. 283
109 North-south cross section of Qianfoge ............................................................................................. 283
110 Chongfusi model, detail of triple and double brackets ................................................................... 284
111 Golden phoenix engraved between bracket-sets ............................................................................. 284
112 Qianfoge at Chongfusi, exterior ......................................................................................................... 285
113 A typical ceyang ....................................................................................................................................... 285
114 Model pagodas in Japan, Nara period ............................................................................................... 286
115 Model of Ming gatetower Qianlou, Huayansi Main Hall ................................................................ 287
116 Model of Qianlou, detail ...................................................................................................................... 288
117 Restored Qianlou in 2013 .................................................................................................................... 288
118 Yangshi Lei drawing and model of a building complex at Yuanmingyuan .................................. 289
119 Asokan stupa, excavated from Leifengta top chamber ................................................................... 290
120 Asokan stupa, excavated from Leifengta crypt ................................................................................ 290
121 Excavated bronze and iron Asokan stupas attributed to Qian Hongchu .................................... 291
122 Stone Asokan stupas ............................................................................................................................ 291
123 Pictorial reconstruction of Su Song’s clock-tower ........................................................................... 292
xi
124 Modern reconstruction of Su Song’s clock-tower at Taiwan National Museum of Natural
Science .................................................................................................................................................... 292
1
Introduction

Miniature architecture proliferated in China during 1000-1200 CE. Buddhist and Daoist icons were
sheltered by mini wooden pavilions, holy scriptures were stored in architectural-shaped bookcases
and cabinets, whereas the interior of a worship hall--especially the vaulted ceiling--was typically
ornamented with groups of tiny buildings to represent the “heavenly palace.” Portable relic shrines,
ceramic houses, container gardens, dollhouses, mini theaters, etc., became cherished items in social
life. Specifications of miniature-making have been written into the official building code, Yingzao
fashi 營造法式 (Building standards), promulgated in 1103 by the Northern Song imperial court.
However, even though a few of these miniatures have been discussed by scholars on separate
occasions, in general, miniature architecture has never received the systematic survey it deserves.
Miniaturization as a culturally significant form of artistic creation, too, has slipped past most
scholars’ attention. In fact, as this dissertation demonstrates, a critical understanding of miniatures
helps to positively reshape our premises and conclusions about architecture, art, and material
culture. The development of Chinese architecture from the eleventh century onward could be
described as a history of “progressive miniaturization”: as key structural members and ornamental
elements dwindled in size and scale over time, the overall form and structure of wooden architecture
also underwent drastic changes.
A major concern of this dissertation is the practical, spiritual, and aesthetic reasons behind
the fervor of miniature-making: what qualities made these small objects particularly appealing to
people? My study reveals that religious thought and practice, especially those associated with Huayan
Buddhism and Pure Land Buddhism, played a central role in the proliferation of the miniatures. I
argue that the Mahayana Buddhist worldview of the universe being a recursive, self-multiplying
system of “worlds-within-worlds,” a concept that resonates with William Blake’s poetic imagination
2
of “a world in a grain of sand,” has been translated into distinctive motifs of art by means of
miniaturization. The decrease in size allowed a much detailed display within a limited space; it
signaled the uncanny, the illusory, and the sublime, which helped to convey abstract Buddhist tenets
and assist one’s visualization of a transcendental realm beyond the everyday experience. Since
visualization was a key step of reaching Buddhahood, miniatures assumed liturgical as well as
soteriological functions. For both elites and commoners, they became all-important symbols of
spiritual power and “expedient means” of obtaining enlightenment and salvation.
Another important question is: what historical and social factors stimulated the flourishing
of miniature architecture? I propose that on the one hand, it was due to the high standardization of
Chinese architecture in the eleventh century. This standardization greatly facilitated miniature-
making, because carpenters only had to reduce the size of the standard timber material while the
same set of rules and formulas for large buildings would still apply. On the other hand, since
miniatures were never the main targets of sumptuary law, they granted carpenters much freedom to
execute their ideas and showcase their skills. With the installation of increasingly stricter statutes on
building activities, it was often safer and more economic to invest in miniatures than in large
structures to achieve similar levels of impressiveness and feelings of importance. The trend of
miniaturization was also observed in painting, sculpture, masonry, ceramics, and cabinetry; it became
a hallmark of the material culture of the eleventh- to twelfth-century China and endured well into
later centuries.

3
On Defining the Miniature: Philosophical, Religious, and Architectural Perspectives
The English word “miniature” hardly finds any equivalent in classical Chinese.
1
The concept of
being miniature, however, like in many other world civilizations, can be traced as far back as high
antiquity. Chinese myth tells how the cosmos was created by giant gods, compared with whom
humans are like dolls or little children. One account describes that the firstborn, Pangu, reached
ninety thousand leagues in height when fully grown. Upon death, his colossal body disintegrated,
generating the sun and the moon, mountains and streams, trees and rocks, and various other matters
between earth and heaven, whereas “[a]ll the mites on his body were touched by the wind and were
turned into the black-haired people” ( 身之諸蟲, 因風所感, 化為黎氓).
2
Another account
attributes the creation of human beings to the goddess Nuwa, who “kneaded yellow earth and
fashioned human beings” ( 摶黃土作人), perhaps following the image of herself.
3

It would seem that humans are born miniatures themselves. While the myth contrasts the
human body with gods and goddesses and with the immense world, the same contrasts is deeply
ingrained in the human consciousness between the trivial, fragile, and vulnerable self in front of
nature, of its formidable power and many unsolved mysteries. Coexisting with this consciousness,
however, is the intuitional drive to project the self to the surroundings, to see oneself as created in
the likeness of the pattern and structure of nature, and the unconscious to personify and

1
Modern Chinese equivalents of the word “miniature” include weixing 微型 and weisuo 微縮.
2
Yishi yin wuyun jinianji 繹史引五運歷年紀 (A Chronicle of the Five Cycles of Times, 3rd cent.), quoted in Anne Birrell,
1993, Chinese Mythology: An Introduction (Baltimore and London: The John Hopkins University Press), 33. On p. 25, Birrell
introduces the five main traditions of Chinese cosmologies expressed in six classical texts, the one on Pangu being the
latest. Birrell suggests that the legend of Pangu might have derived from certain Central Asian sources (30-31). See also
Bruce Lincoln, 1991, Death, War, and Sacrifice: Studies in Ideology and Practice (Chicago: University of Chicago Press).
3
Fengsu tongyi 風俗通義 (Explanation of Social Customs, late 2nd cent.), quoted in Birrell 1993, 35. Another famous
giant figure in Chinese myth is Kuafu.
4
anthropomorphize all forces, causes, and phenomena.
4
In other words, humans see themselves as
“miniatures” of the cosmos and the natural order personified.
The dialectic of microcosm-versus-macrocosm and small-versus-large has been
contemplated in early Chinese philosophical writings. In one of Zhuangzi’s most imaginative and
rhetoric-rich essays, “Autumn Floods” (qiushui 秋水), the small and minuscule is denoted as xiao 小,
jing 精, and wei 微. The Northern Sea, being the largest body of water on earth, reveals to the River
Lord that amid heaven and earth, he is “as a little (xiao) pebble or tiny (xiao) tree on a big mountain”
( 猶小石小木之在大山也).
5
Likewise, the expansive Middle Kingdom is but “a mustard seed in a
huge granary” ( 稊米之在大倉), whereas each person, in comparison with the myriad things, is like
“the tip of a downy hair on a horse’s body” ( 毫末之在於馬體).
6
The Northern Sea warns that one
must not belittle what appears small and tiny, since the capacity of things being forever smaller (wei)
and minute (jing) is limitless.
7
Small and large are relative but never absolute; one can freely “regard
the heaven and earth as a mustard seed and the tip of a downy hair as a mountain” ( 知天地之為稊
米也, 知毫末之為丘山也).
8

4
Lincoln 1991. The interrelations between the microcosm and the macrocosm is especially well illustrated in chap. 1--
even though Lincoln’s observations are mainly derived from Indo-European mythology, they are also applicable to this
case.
5
Victor H. Mair, trans., 1998, Wandering the Way: Early Taoist Tales and Parables of Chuang Tzu (Honolulu: University of
Hawaii Press), 153.
6
Mair 1998, 153-54.
7
Mair 1998, 155, “that which is minute is the smallest of the small” ( 精, 小之微也). Jing is roughly equal to
“minuteness” (also meaning refined) and wei is smaller than xiao. The term wei has been incorporated into weixing and
weisuo as modern terms for miniature. See n. 1.
8
Mair 1998, 155.
5
The same imagery is found in some Buddhist sutras claiming that the Sumeru mountain can
somehow be placed inside a mustard seed.
9
To be enlightened is to cross the boundaries of things
large and small, to transcend one’s mundane perceptions of the external world, and to see, or rather
envision, “a world in a lotus petal” ( 一葉一世界).
10
It is almost uncanny how this image resonates
with “a world in a grain of sand” and “eternity in an hour” in William Blake’s poem. To be sure,
such poetic images appear in Buddhist narratives to serve didactic and soteriological purposes.
Buddhism views the universe as not a singular but plural entities: the multiple worlds are all
miniatures as well as components of a grand network which makes up the entire buddhaksetra (Ch.
fotu 佛土), or Buddha land.
11
From this vantage point, the world we inhabit must look like a grain of
sand amid the countless stars in the cosmic river Ganga (which we now call the galaxies).
12
Time, like
space, is similarly composed of “miniatures” and need be measured and articulated in terms of a
fraction of a second.

9
“Placing something as high and wide as the Sumeru into a mustard seed, the size of the Sumeru does not increase or
decrease” ( 以須彌之高廣內芥子中, 無所增減). Vimalakirti Sutra (Ch. Weimojie suoshuojing 維摩詰所說經), T14.475:
546b, http://www.cbeta.org/result/normal/T14/0475_002.htm.
10
“[The Vairocana] lives in the Lotus Terrace Repository of the Sea of World. The terrace is covered by a thousand
leaves, each leaves being a world, and this makes a thousand worlds” ([ 盧舍那] 住蓮花臺藏世界海. 其臺周遍有千葉,
一葉一世界, 為千世界). Brahmajala Sutra (Ch. Fanwangjing 梵網經), T24.1484: 997c,
http://www.cbeta.org/result/normal/T24/1484_001.htm. The same trope is also used profusely in chap. 5 of the
Avatamsaka Sutra (Ch. Huayanjing 華嚴經).
11
Randy Kloetzli, 1983, Buddhist Cosmology: From Single World System to Pure Land: Science and Theology in the Images of Motion
and Light (Delhi: Motilal Banarsidass): ix. Four major models of the Buddhist cosmos are introduced: (1) the single world
system, (2) the “cosmology of thousands,” (3) the “cosmology of innumerables,” and (4) the cosmologies of the Pure
Land sects. The plurality of cosmologies pertaining to the Huayan School is explicated on pp. 52-54. An explanation of
the term buddhaksetra is in Stephen J. Laumakis, 2008, An Introduction to Buddhist Philosophy (Cambridge and New York:
Cambridge University Press): 207, 214-15.
12
Kloetzli 1983, 121, contains a very poetic contemplation on this matter: “Since the world is essentially a speck of sand
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.”
6
The atomic view emerged in the classical world and was adopted in proto-astronomy. A
good example is Archimedes’s calculation of the volume of the celestial space containing the sun
and the earth, a space so vast that it could be measured only by the infinitesimal. He reckoned that a
total of 10
63
grains of sand would be needed in order to fill this space.
13
In this sense, a grain of sand,
so humble and ordinary an object, offered the sage an approach to even the most abstruse subject
imaginable at the time, and yielded insight into the higher dimension. A certain similarity in the
physical form and structure of the measuring and measured cannot be dismissed: a grain of sand to
the earth, an atom to the universe, and a second to eternity.
In a similar light, in the works of Neo-Confucian thinkers, one finds propositions that the
principle (li 理, also translated as “coherence”) of heaven-and-earth can be sought in “a thing
smaller than a cricket, an ant, or a blade of grass.”
14
Contrary to the Buddhist view of the world and
the myriad things it contains being essentially empty and illusive, however, the li embodied by a
blade of grass forms the basis of learning for Confucian scholars, who were obliged to investigate
the principle of things so as to establish moral authority. The miniature discussed in this context,
therefore, is the epitome of a whole range of knowledge and wisdom essential for the cultivation of
the self.
What, then, is a miniature in an architectural sense? The rule of scale still applies in this case:
a miniature is several times smaller than a real structure, but the geometry of its basic form and
structure often remains unaltered. Here the contrast in scale is not as drastic (as in the case of

13
Kloetzli 1983, 115-17.
14
Quoted in Willard Peterson, 1986, “Another Look at Li,” The Bulletin of Sung-Yuan Studies (18): 15. Also, on p. 24,
“Coherence (li) is intelligible on all levels of integration, a blade of grass, a school of fish, the experience of a lifetime,
heaven-and-earth, the Great Ultimate.” Peterson’s article is an excellent exposition of the meaning of li and many of its
associated philosophical terms. Another great source concerning the concept of li in the context of the historical role of
Neo-Confucianism is Peter Bol, 2008, Neo-Confucianism in History (Cambridge: Harvard University Press). For a succinct
summary of this important work, see Bettine Birge’s review article in The American Historical Review 115 (2010.3): 822-23.
7
1:10
63
), and an architectural miniature usually ranges from 1/100 to 1/2 of the regular size.
Historical records in China refer to them using the prefix xiao-,
15
sometimes xiaoyang 小樣 (small-
scale prototype), which denotes architectural models. More specifically, in the Yingzao fashi, the term
xiaomuzuo 小木作 (small-scale woodworking) has been assigned to the category of non-structural
carpentry encompassing joinery (doors, windows, stairs, etc.), cabinetry, and miniature woodwork
that observes strict scaling principles.
Broadly speaking, miniature architecture could be a rather inclusive concept, denoting any
piece of architectural-shaped--but functionally non-architectural--object made using an identifiable
downscaling method. In this dissertation, for clarity and concentration, miniatures are discussed
within a narrower definition:
1. It has to be a three-dimensional, physical object. Architecture illustrated in some paintings,
although observing certain miniaturizing techniques, is not the main subject of this dissertation.
2. It has to be made using architectural (additive), not sculptural (subtractive) methods. This
confines our examination to wooden miniatures that applies the same post-and-lintel building
technique of Chinese timber-frame architecture. By contrast, miniatures carved of wood, cast in
bronze, or made of ceramic (such as the large number of Han pottery houses) adopt a disparate
structural logic and are excluded from this dissertation.
16

Miniatures pertaining to this definition largely correspond to several types of structures categorized
as the xiaomuzuo in the Yingzao fashi. Later in this introduction, I will explain further what these
structures are and what material evidence is available for the study of them.

15
The term xiao also refers to non-architectural miniatures, such as xiaoxiang; see Chapters 1 and 5.
16
This exclusion is due mainly to the purpose of this dissertation, which is to reveal the scaling method of miniature-
making without considering the change of building material or technique. Such a concentration allows me to expose the
exchanges between large- and small-scale woodworking, which further shed light on the historical development of
Chinese architecture.
8

Conceptual and Theoretical Framework: Deconstruction, Oneirism, and Simulation
Modern critcis of architecture often debate the interrelationship of a pair of concepts: form and
function. The form of a building, it has been argued, needs to follow the intended function,
sometimes to the extreme that any type of ornamentation--since it does not serve the “ultimate
purpose” of architecture, which is to provide shelter--is deemed superfluous, wasteful, and even
sinful.
17
To read architecture as a duality of form and function is to read it as a sign (or a collection
of signs) conveying certain meanings. The moment we cast our sight upon a building, the process of
reading and interpreting immediately takes place: this porch signifies an entrance, that belfry
indicates a church, etc. This process of interpretation, however, does not stay at the recognition of
the direct function of the building and its various parts, but always extends to the social and cultural
identities and values of architecture. Georges Bataille (1897-1962), for one, has compared the
museum to a “colossal mirror in which man finally contemplates himself in every aspect, finds
himself literally admirable, and abandons himself to the ecstasy expressed in all the art reviews;”
whereas public monuments, such as the Bastille, stand for the very existence of social order,
authority, and fear imposed on the multitudes.
18

Postmodernists and Poststructuralists, on the other hand, seek to provide an alternative to
the linear interpretation of form and function. Deconstruction, a term coined by Jacques Derrida
(1930-2004), has found many interesting intersections with architecture in this respect. The term
itself possesses a certain architectural underpinning in that it seems to reverse the usual process of

17
The most influential architect and theorist in history holding this view is perhaps Adolf Loos (1870-1933), whose
essay Ornament and Crime has received much criticism including that of Theodor W. Adorno (1903-1969). See Theodor
W. Adorno, “Functionalism Today,” in Rethinking Architecture: A Reader in Cultural Theory, ed. Neil Leach, 1997 (London
and New York: Routledge), 6-19.
18
Georges Bataille, “Architecture,” in Leach 1997, 21; “Museum,” in Leach 1997, 22-23.
9
construction by demolishing or dissembling the whole into parts. Derrida himself, however, clarifies
that the essence of deconstruction lies in neither architectural technology nor metaphor, but
architecture itself can be a deconstructive discourse, a way of thinking, and a form of writing.
19
His
theory has given rise to the so-called “deconstructivist” architecture, a postmodern style associated
with the works of many world-renowned architects including Frank Gehry, Daniel Libeskind, Rem
Koolhass, and Zaha Hadid, even though critics believe that the link between this particular style and
deconstruction is problematic and misleading.
20
Characteristic of the deconstructivist architecture is
the profusion of non-rectilinear forms, the accentuated asymmetry, irregularity, disorientation, and
ambiguity of space. Take, for example, the Walt Disney Concert Hall in downtown Los Angeles or
the CCTV Headquarters in Beijing: in a material discourse, deconstruction has been expressed as a
detachment of form from function, or indeed from any solid cultural contexts, allusions, or
innuendoes. It disrupts the linear reading of meanings and allows architecture to become a self-
referential existence.
Miniaturization, to some extent, can be analyzed as a mode of deconstruction. What
constitutes deconstructivist architecture finds many resonances here: the form is detached from the
assumed function (or rather, there is a general lack of architectural functionality); the usual
expectation for and experience of architecture is discontinued; articulation is replaced by ambiguity,
sometimes redundancy, which overwhelms us and prevents any straightforward interpretation. But
the distinctiveness of miniaturization is that the “deconstructing” force derives not from the
unfamiliar form but the unfamiliar scale. The twist of scale--even when the original form is largely

19
Jacques Derrida, “Architecture Where the Desire May Live,” in Leach 1997, 320-21. See also Leach 1997, 317-18;
Mark Wigley, 1993, The Architecture of Deconstruction: Derrida’s Haunt (Cambridge, Mass.: The MIT Press).
20
One of the first publication on deconstructivist architecture is Philip Johnson and Mark Wigley, 1988, Deconstructivist
Architecture (Museum of Modern Art).
10
maintained--bestows the architecture with a new identity and its observer a new spatial experience.
In essence, miniature architecture becomes non-architecture or anti-architecture in disguise; it
embraces strict rules of construction for the sake of “deconstructing” architecture and destabilizing
its very meaning. This “deconstructive” reading of miniature architecture shares many points in
common with other critical works on the topic of miniatures, as will be introduced later.
From the perspectives of reception theories and psychoanalysis, miniatures engender far
more profound impacts on the human psyche than what meets the eye. As Gaston Bachelard (1884-
1962) has proclaimed, “[v]alues become engulfed in miniature, and miniature causes men to
dream.”
21
Bachelard highlighted the oneirism, or dreamlike nature, of the minuscule: the toy-world
of our childhood, the botanist’s magnifying glass (to which we might add Alice’s looking glass), he
asserts, become the fountain of daydreams and memories, presenting to us a new world and a new
universe.
22
In the phenomenological inquiry, the miniature averts direct “reading” but evokes
powerful imaginations. A miniature garden, as Bachelard would argue, does not necessarily stand for
any real garden in the world, but it may bring to the mind reveries about the fairyland, the luxuriant
rainforest, the rolling hills, the tranquil countryside, or perhaps the distant isles in the sea. Herein lies
the poetics of the miniature--it opens up a vast space for the free wanderings of the mind and heart,
a world not restricted or predefined by any description, narration, or natural law.
This world created by the miniature might sound phantasmal and unreal. The miniature
itself, likewise, is sometimes regarded as a “fake,” a shadow of the past, a mere “copy” or
representation of something original and substantial. If, according to Plato, all art is but imitation of
a real object, which is itself the imitation of an ideal form created by God, miniatures could similarly

21
Gaston Bachelard, 1969, The Poetics of Space, trans. Maria Jolas (Boston: Beacon Press), 152.
22
Bachelard 1969, 149-50, 153-55, 157.
11
be categorized as “imitations of imitations” since they are twice-removed from the truth. Such a
hierarchical view of artistic creation, however, has become more and more problematic as we
entered the age of new media. Jean Baudrillard (1929-2007) has pointed out acutely the falsehood of
the presumed dichotomy between the “original” and the “copy,” between “reality” and
“simulation.” As much as an image imitates a basic reality, it also works to pervert reality or even
mask the absence of reality (think of, for instance, a painting of the Elysian), to an extent that the
image “bears no relation to any reality whatever: it is its own pure simulacrum.”
23
Baudrillard
depicted Disneyland as “a perfect model of all the entangled orders of simulation,” an imaginary,
“deep-frozen infantile world” full of illusions and phantasms.
24
Many of his observations still apply
as recent technologies of simulation--computer-aided design, video games, and virtual reality, to
name but a few--have been continuously reshaping and redefining modern ways of living and
thinking. These technologies have also encouraged new methods of approaching and preserving the
past: while the physical form of historical sites and artworks may perish, they may be archived and
forever stored in a digital form, i.e., as their own simulacra.
This dissertation engages with these three concepts in the discussion of miniature
architecture. I will reveal how miniature-making, within a particular historical and social context, can
be analyzed as 1) a deconstructive language of artistic creation as well as a deconstructive discourse
of Chinese architectural history; 2) a tool for inducing imaginations and daydreams, which lead to
insights; and 3) a method of simulation involved in the transmission of canonical forms, formulas,
and knowledge.

23
Jean Baudrillard, 2001, Selected Writings, ed. Mark Poster (Stanford: Stanford University Press), 173.
24
Baudrillard 2001, 174-75. Also see Leach 1997, 209, 221-22.
12
Scholarship on the Miniature
Scholars have approached the topic of miniature art from various perspectives. Rolf Stein is perhaps
the first to have commented extensively on Chinese miniatures from a cultural-historical point of
view. His celebrated work, The World in Miniature: Container Gardens and Dwellings in Far Eastern
Religious Thought (1990), focuses on the container garden, the traditional house, and the temple as
distinctive forms of miniature art in China as well as in Tibet, Mongolia, and Vietnam. He traces the
container garden back to the Han-dynasty incense burner (boshanlu 博山爐) which was often
fashioned into a miniature mountain; and this deep historical connection, as he demonstrates, has
been implicated by multiple literary references.
25
Lying behind the miniature mountain and miniature
garden was a set of “themes” associated with Daoist aesthetics and Buddhist motifs. The tiny
landscape, because of its altered size, transcended “from the level of imitative reality [into] the
domain of the only true reality: mythical space,” offering a retreat, a separate world, a world of magic
and imagination for the wandering soul.
26
A full-size house, on the other hand, was at once a
microcosm of the universe and a projection of the human body. Various parts of the primitive
house--the hearth, the skylight, and the central drainage--have carried rich symbolic meanings
ingrained in cultural-specific ritual practices.
As the name of the book indicates, Stein understands religious thought and aspiration to be
perhaps the strongest drive behind the creation of such miniature art. The Altar of Heaven in
Beijing, the legendary Hall of Light (mingtang 明堂) in history, and the many turning libraries used in
Buddhist monasteries, to name but a few, are interpreted as miniaturization based on cosmological

25
Rolf A. Stein, 1990, The World in Miniature: Container Gardens and Dwellings in Far Eastern Religious Thought, trans. Phyllis
Brooks (Stanford: Stanford University Press), 23-48, and chap. 2.
26
Stein 1990, 52-53, 77, 112. Stein proposes two sets of themes associated with miniature gardens. The Taoist themes
encompass ideas of cures, vital power, immortality, medicinal essences, horns of plenty, and retreats. The peasant themes
include cures (once again), vital power, longevity in its social context, continuation of life, fertility, and fecundity.
13
theories and cosmographical models such as the Kunlun mountain or its Buddhist equivalent, the
Sumeru. He accentuates that real architectural features have been rigorously incorporated to help
worshipers visualize the sacred landscapes or intended world order; the sutra libraries in Beijing and
Shanxi, for instance, can literally rotate, some carved in a way that they appear to be a miniature
Sumeru “emerging from the sea.”
27
However, notwithstanding these technical details, Stein’s
interpretation of the miniatures largely concentrates on their metaphorical and symbolic nature, but
specifications of the miniaturizing process and its socio-economic significance--such as will be
elaborated in this dissertation--have been left undiscussed.
A comparable approach is taken up by Susan Stewart in her On Longing: Narratives of the
Miniature, the Gigantic, the Souvenir, the Collection (1993). The miniatures analyzed in this work,
encompassing toys, dollhouses, and motifs in fairytales such as the Lilliputians, are examined as
metaphors and cultural products expressing the anxieties, triumphs, and desires of the bourgeois
society.
28
Though concerning a different historical context and intellectual tradition, Stewart agrees
with Stein that a chain of projections exist between the universe, the human body, and architecture:
the Vitruvian Man, therefore, besides being a golden formula for perfect proportions and
measurements, is a miniaturization and personification of the entire world with all its natural laws
and phenomena.
29
But the miniature is not entirely a reflection; it presents a theatrical stage where
we entertain ourselves with “a deliberately framed series of actions.”
30
On this stage, even the flow

27
Stein 1990, 254-56.
28
Susan Stewart, 1993, On Longing: Narratives of the Miniature, the Gigantic, the Souvenir, the Collection (Durham: Duke
University Press), xii.
29
Stewart 1993, 128-30.
30
Stewart 1993, 54.
14
of time is manipulated--quickened, slowed, or stopped, as in dreams--generating a total interiority
segregated from the external and the ordinary.
31

James Roy King’s Remaking the World: Modeling in Human Experience (1996) illuminates
miniature-making from an anthropological perspective. He emphasizes the modeling process as an
“experience” consisting of a broad spectrum of human activities: exploring dimensionality; world
making through realism, detail, and truth to prototype; solving technical problems; investigating
diverse materials; exchanging goods; collecting; exercising the muscles (kinesthetics); projecting
personal values and experiences; creating aesthetic experiences; and fantasizing.
32
This “activity set,”
as he calls it, aims to bring “pleasure and insight” for those involved in the experience.
33
On the one
hand is the display of the magic of verisimilitude and miniaturization, with the sole purpose of
telling stories, of stopping time, of escaping this world and venturing into the realm of imagination.
On the other hand, by contrast, is the passion for geometric precision, the attention to technical
details, and the rigorous practice of using models as analytic and pedagogical tools.
34
King
differentiates the two groups by referring to the former as miniatures and the latter as models.
35
In
fact, as this dissertation will unravel, the line between miniatures and models is at best a fuzzy and
shifting one. The apparent contradiction noted here only helps to expose the paradoxical nature of
the miniature--it has to be mimetic in order to be fantastic.

31
Stewart 1993, 65-66. See also 61, 68, on the interiority of the dollhouse.
32
James Roy King, 1996, Remaking the World: Modeling in Human Experience (Urbana and Chicago: University of Illinois
Press, 1996), 6.
33
King 1996, 3.
34
King 1996, 19, 148, 209. Models have been used since the Middle Ages, through High Renaissance, all the way to the
modern period.
35
King 1996, 3. “By model I mean a re-creation of some prototype or original, generally but not always smaller and
usually of materials different from those of the original.” And p. 19, “Miniatures...seem to be less interested in accuracy
and more devoted to a wide array of small, domesticated objects that look something like the real thing.”
15
In recent years, miniature art has started to receive increasing attention from art historians.
Lothar Ledderose’s Ten Thousand Things: Module and Mass Production in Chinese Art (2000), though
highlighting the concept of modularity rather than miniaturization, has in fact foregrounded many
important issues regarding miniature-making in traditional China. To this dissertation, the most
pertinent and illuminating point Ledderose has made is that both modular design and prefabrication
of wooden structures were enabled by the standardization of building components.
36
As he explains,
this high standardization was achieved by implementing a scaling principle: using cai 材 and fen 分 as
the primary and secondary measuring units, each structural member was proportioned to the
standard unit according to prescribed formulas.
37
It will be revealed further in this dissertation that
the same scaling principle was applied to create visually stunning miniature woodwork.
Besides wooden architecture aboveground, miniatures found underground have become a
special focus of interest. Archaeological finds have testified to the inclusion of miniature architecture
in tombs beginning no later than the late seventh century BCE, a practice which continued to
flourish well into later dynasties. Excavated architectural models are made of either ceramic, wood,
or bronze, their types ranging from simple granaries, stoves, and wells to houses, pigsties, and more
complex structures such as multistory towers, fortified courtyards, and paddy fields.
38
Generally
categorized as mingqi 明器 (spiritual vessels), these miniatures have been believed to accompany the
dead and ensure their wellbeing in the afterlife. Wu Hung, for one, has argued that these artifacts

36
Lothar Ledderose, 2000, Ten Thousand Things: Module and Mass Production in Chinese Art (Princeton: Princeton University
Press), intro. and chap. 5.
37
See Chapter 1 of this dissertation.
38
A table showing the earliest known architectural models from Chinese tombs is provided in Armin Selbitschka, 2005,
“Miniature Tomb Figurines and Models in Pre-imperial and Early Imperial China: Origins, Development and
Significance,” World Archaeology 47 (1): 26-27. Also see Qinghua Guo, 2010, The Mingqi Pottery Buildings of Han Dynasty
China: Architectural Representations and Represented Architecture (Sussex Academic Press), for a general introduction to mingqi
architecture.
16
were intentionally designed in diminutive forms based on the belief that the posthumous soul--a
duality of the ascending, heavenly-bound hun 魂 and a descending, earthly-bound po 魄--was
invisible and a miniature itself.
39
Not only the mingqi but tomb objects that symbolically or physically
contained the deceased--including the soul jars (hunping 魂瓶), the coffins, and the burial chambers--
were miniaturized structures serving the soul. Armin Selbitschka, on the other hand, points out that
miniatures similar in form to the mingqi have also been found at several residential sites, thus
problematizing the long-held opinion that they were prepared exclusively for burials.
40
Selbitschka
contends that architectural models in fact represented real land properties previously owned by the
deceased; they were placed in tombs with the good wish of the living that the land would remain a
source of income for the family.
41
Jeehee Hong’s recent study on miniature fruits and furniture
deposited in medieval tombs and pagoda crypts provides a more nuanced narrative. Entering the
tenth century, she argues, the mingqi were given new modes of representation which aimed to
function for both the dead and the living.
42
While most art historians emphasize the mimetic quality
of the miniatures, Nancy S. Steinhardt’s analysis of a Yuan-dynasty architectural-shaped wooden

39
Wu Hung, 2015, “The Invisible Miniature: Framing the Soul in Chinese Art and Architecture,” Art History 38 (2): 286-
303. See also Wu Hung, 2010, The Art of the Yellow Springs: Understanding Chinese Tombs (Honolulu: University of Hawaii
Press).
40
Selbitschka 2015, 20-44.
41
Selbitschka 2015, 39-40. According to the inscriptions and “land contracts” found in tombs, architectural models
represented private land ownership, and ensured the sources of income and tax for the living family members. Regarding
land contracts and ownership, an important work to be consulted is Valerie Hansen, 1995, Negotiating Daily Life in
Traditional China: How Ordinary People Used Contracts, 600-1400 (New Haven: Yale University Press).
42
Jeehee Hong, 2015, “Mechanism of Life for the Netherworld: Transformations of Mingqi in Middle-Period China,”
Journal of Chinese Religions 43 (2): 161-93. Hong’s main argument is that the burial chamber, which simulated a wooden
structure with downscaled wooden chairs and tables inside, transformed from a private section of the dead to a
“socialized,” public realm to be negotiated between the living and the dead.
17
coffin has led her to believe that a miniature could mimic and deviate from the real at the same time,
thus becoming a dual-purpose object which is both representational and decorative.
43

The existing scholarship has exposed many lingering issues: for funerary miniatures alone,
the debates go on as how their social functions and values should be understood and through what
artistic forms these functions became realized. A more general question, perhaps, is whether or not
miniatures were created to imitate and simulate, or else as mainly decorative, less symbolically
significant objects. But before satisfactory answers to these questions are sought, what seems still
lacking is a clear definition and classification of the miniatures. Also lacking is an explanation of the
specifications and procedures of miniature-making using proper nomenclature. While this
dissertation does not intend to solve these issues all at once, it does provide rudimentary conceptual
and technological foundations for the study of miniature architecture in China in a predefined
historical period. While in dialogue with former scholarly works on miniature, it focuses on scale
instead of form, style, or other physical attributes. I will demonstrate how scale and scaling
principles, when carefully observed, can shed light on questions not only about dating and
iconography but those concerning the historical background, social value, and religious significance
of the surveyed object.

Primary Sources, Digital Database, and 3D Modeling
The key text consulted in this dissertation is the Yingzao fashi, especially the part on terminology, the
scaling principles, and miniature woodworking. The historical background and significance of this
text will be briefly outlined in Chapter 1, where I will provide a summary of the technical details of
miniature-making so as to prepare the reader for the discussions in succeeding chapters. It has to be

43
Nancy S. Steinhardt, 2010, “The Architecture of Living and Dying,” in The World of Khubilai Khan, ed. James C. Y. Watt
(New Haven and London: Yale University Press), 65-73.
18
noted here that the Yingzao fashi has several limitations: first, it prescribes, instead of describes,
methods and procedures of building activities. The formulas and patterns listed in the text have been
associated mainly with official and canonical--and to some extent, ideal--forms, whereas in real
practice, we often observe vernacular and/or regional features differing from those exemplified in
the text. Second, the types of miniatures prescribed in the text are almost exclusively timber-frame
structures, meaning that “sculptural miniatures”--including the majority of mingqi--are not covered.
Moreover, the examples that have been included represented merely a selective few from all the
wooden miniatures available at the time. Despite such limitations, the Yingzao fashi has proven to be
the most indispensable and invaluable guideline and “grammar book” for the research conducted
here. A fuller picture of the variety of miniature architecture can be reconstructed by referring to
supplementary texts especially Song-dynasty miscellanea, which will also receive a detailed
examination in Chapter 1.
In terms of material evidence, four key examples will be examined, one for each of Chapters
2 to 5. These include a revolving sutra case at the Longxingsi 隆興寺 Buddhist monastery in
Zhengding, Hebei Province, a set of sutra cabinets at the Huayansi 華嚴寺 in Datong, a coffered
ceiling at the Jingtusi 淨土寺 in Yingxian, and a model pavilion at the Chongfusi 崇福寺 in
Shuoxian, the last three all in Shanxi Province. Chronologically, these four are dated to the period
from the eleventh to the twelfth centuries, and have been traditionally associated with the
architecture of the Northern Song, the Liao, and the Jin, respectively. Geographically, the group
represented the woodworking practice in China’s northern “borderlands”--an area under constant
territorial disputes and conflicts between the Han and non-Han regimes at the time, and one close to
the cultural center, i.e., modern-day Kaifeng in Henan, where the Yingzao fashi was compiled and
promulgated. The first two examples each display a remarkable conformity to the Yingzao fashi, the
third exposes both adherence to and deviation from the standard, whereas the fourth stands for a
19
type of miniatures completely omitted by the text. While the scaling method will be analyzed for
every example, the focus of each chapter is slightly different: the revolving sutra case is most ideal
for illustrating the interrelationship between the concepts of miniaturization and deconstruction; the
sutra cabinets and the ceiling will illuminate the kind of unusual visual experience and psychological
impact brought by miniatures; and the model pavilion will shed light on the issue of simulation.
The digital counterpart of this dissertation is an open-access online database called Project
CloudCastle,
44
which will eventually outgrow the scope of this dissertation and serve as a reference,
research guide, and digital preservation for Chinese architecture. Currently, for a selective examples
including the four key miniatures, it offers basic information on the date, geographical location,
layout, formal and structural features, ornamentation, brief history, and the inscriptions, textual
records, and secondary works discussing these examples. It will further include a glossary of
architectural terms and a digitized Yingzao fashi with my annotated English translation.
The most crucial feature of this website is the visualization of the dimensional data of the
miniatures I collected in fieldwork. I have developed several interactive photogrammetry and Rhino
3D models to achieve this, a process which has been instrumental to the writing of this dissertation.
My chapters use the data extracted from the models to identify and date the miniatures, to expose, in
quantitative terms, the downscaling formulas they adopted, and to unravel the technological
exchanges between Chinese architecture and its miniatures in history. In a greater sense, 3D models
not only enable a heightened awareness of, and a “tangible” way of critiquing the spatial and
dimensional attributes of the objects, but they also benefit and potentially transform processes of
architectural and art historical inquiry. Further, these models would constitute what I envision to be
a digital collection and exhibition of Chinese architecture accessible via new media and on different

44
https://chinesearchitecture.wordpress.com
20
platforms, to be used in classrooms, libraries, and museums. They not only encourage innovative
learning and research modes but would assist in physical as well as digital conservation projects.
21
1. Miniatures in Texts

Literary references to miniatures exist in abundance. They are, however, largely fragmentary in
nature, and they simply do not adhere to a set of specific terms when describing different
miniatures. For a start, it is necessary to narrow down the scope of research to targeted historical
period and genres of literature. In this dissertation, I take the Yingzao fashi as a point of departure
and expand my search to a selective few miscellanea (biji 筆記) written by Song-dynasty scholars.
The reason to focus on these texts is manifold. While the Yingzao fashi demonstrates the technical
details of miniature woodworking, the many eye-witness accounts in the miscellanea vividly depict
how the miniature world actually interacted with people in all aspects of life. While the former is a
legal text compiled by a court architect under imperial decree, the latter assume much more personal
tones and provide multiple lens for readers to inspect miniatures and their variegated roles in
different social venues. The former focuses on definition and technique, whereas the latter, practice
and significance.
The written evidence presented here is certainly not exhaustive but representative; in some
cases it may even appear convoluted. But it is my intention to expose the heterogeneous nature of
these records, and to showcase the types of narratives where traces of miniatures might be pursued.

Small-scale Woodworking (Xiaomuzuo) in the Yingzao Fashi
The Yingzao fashi, a canon often compared to the famous Ancient Roman architect Vitruvius’s (c. 80-
15 BCE) Ten Books on Architecture, has been internationally recognized as one of the most
fundamental works for the study of premodern Chinese architecture. Compiled by the court
architect Li Jie 李誡 (1035-1110) and promulgated state-wide in 1103, the Yingzao fashi not only
served as the Northern Song official building code, but has also remained the earliest surviving
22
treatise on Chinese architecture. Ying 營 means to conceive, to plan, zao 造 is to build and make. Put
together, yingzao is roughly equivalent to architecture in its broadest sense. The word fashi 法式
denotes law, rules, and standard forms and patterns to be followed; it excellently exposes the nature
of this text being a legal code written in a most serious and critical manner possible, one that had
been meticulously scrutinized by court officials before formal promulgation--a fact that has lent
much authority and credibility to this text even today.
1

The Yingzao fashi was “rediscovered” in 1919 by Zhu Qiqian 朱啟鈐 (1872-1964), who later
founded the Society for the Research in Chinese Architecture (Zhongguo yingzao xueshe 中國營造
學社) in 1930. Since then, the text has become an invaluable reference for modern architectural
historians, who constantly rely on the terminology, design methods, procedures of construction, and
other important information in this text in their efforts of exploring and interpreting Chinese
architecture.
2
A particularly interesting and engaging content of the text is the set of architectural
drawings attached at the end to illuminate what mere words can hardly convey. This makes the
Yingzao fashi a visual as well as a textual source.
3

1
Yingzao fashi (Building Standards), comp. Li Jie, vol.1 (Shanghai: Shangwu yinshuguan, 1954), 1-8, 15-18.
2
A brief history of the Society and their study of the Yingzao fashi is provided in Shiqiao Li, “Reconstituting Chinese
Building Tradition: The Yingzao fashi in the Early Twentieth Century,” Journal of the Society of Architectural Historians 62
(2003.4): 470-89, doi: 10.2307/3592498. Research on the Yingzao fashi have been conducted by scholars such as Liang
Sicheng 梁思成, Chen Mingda 陳明達, Pan Guxi 潘谷西, and Li Luke 李路珂 in China; Takeshima Takuichi 竹島卓
一 in Japan; Else Glahn, Qinghua Guo, and Jiren Feng in the West. See, for instance, Jiren Feng, Chinese Architecture and
Metaphor: Song Culture in the Yingzao Fashi Building Manual (Honolulu: University of Hawaii Press, 2012); Else Glahn,
“Some Chou and Han Architectural Terms,” Bulletin of the Museum of Far Eastern Antiquities 50 (1978): 105-25; Jiren Feng,
“Bracketing Likened to Flowers, Branches and Foliage: Architectural Metaphors and Conceptualization in Tenth to
Twelfth-century China as Reflected in the Yingzao Fashi,” T’oung Pao 93 (2007.4/5): 369-432. Other historical treatises on
Chinese architecture include the Gongcheng zuofa 工程做法 (Methods of architectural projects), Yingzao fayuan 營造法原
(Building standards and sources), Yuanye 園冶 (On gardening), Lu Ban jing 魯班經 (Carpenter’s classic), etc., but these
are much later works and far less influential than the Yingzao fashi.
3
The illustrations, however, needed be treated with caution, since the earliest illustrated edition is dated to the Ming
Yongle Period (1562-1567). For a discussion of the different historical and modern editions of the Yingzao fashi, see Li
Luke, “Chuanshi liang Song shiqi Yingzao fashi de canjuan, zhailu ji zhulu gouchen 傳世兩宋時期營造法式的殘卷,
23
The Yingzao fashi categorizes architectural design and construction into a total of thirteen
types of zuo 作, or works. These include stonework, large-scale woodwork (damuzuo 大木作), small-
scale woodwork (xiaomuzuo), woodcarving, bamboowork, plasterwork, brickwork, and so on. The
text is a tripartite system which first describes the rules and techniques for each work, second the
required labor force, and then the estimated amount of materials to be used. So far, the part on
large-scale woodworking is well-studied, whereas its counterpart--small-scale woodworking--receives
relatively less attention. This is mainly because the former deals with the loadbearing members of
the building, encompassing the columns, the beams, the brackets, and about everything integral to
the structural frame. The latter, on the other hand, covers architectural joinery and furnishing--the
doors, the windows, the stairs, the ceiling, the interior shrines and cabinets--which lack significant
structural functions but are more often the focal points of ornamentation.
4
It is, however, from the
part on small-scale woodworking that we learn about particular types of miniature architecture.

Types of miniature woodwork: shrines, repositories, and the “Heavenly Palace”
The ninth through eleventh chapters (juan 卷) of the Yingzao fashi introduce two major types of
miniatures:
1. Zhang 帳, canopy shrines, which include four subtypes:

摘錄及著錄鉤沉 (Transmitted manuscripts, excerpts, and copies of the Yingzao fashi in the Northern and Southern
Song period),” Zhongguo jianzhu shilun huikan 4 (2011): 31-46.
4
A full list of the types of small-scale woodworking is found in chaps. 6-11 of the Yingzao fashi, totaling forty-two
distinct types of works. These include partition walls, screens, fences, balustrades, sun-shades, plaques, etc. Else Glahn
has translated the xiaomuzuo as “lesser carpentry,” which is often followed by later scholars. But translating xiao as
“lesser” or “minor” could be misleading due to the implication of “lesser importance,” subordinate, and minor.
Counterintuitively, it is often the small-scale components that are more extravagantly embellished and elaborately crafted
in both secular and religious, urban and vernacular settings. My translation of “small-scale woodworking” was suggested
by Professor Nancy Steinhardt, who believes “small-scale” to be a more accurate and less misleading rendering of xiao.
“Small” not only indicates the generally small overall size of these objects, but also the smaller cai 材 (timber material)
they use in comparison with that of large-scale woodworking. See below.
24
1a. Fodaozhang 佛道帳 (Buddhist/Daoist shrines);
1b. Yajiaozhang 牙腳帳 (aproned shrines);
1c. Jiuji xiaozhang 九脊小帳 (nine-ridged small shrines);
1d. Bizhang 壁帳 (wall shrines).
2. Jingzang 經藏, sutra repositories, which include two subtypes:
2a. Zhuanlun jingzang 轉輪經藏 (wheel-turning sutra repositories);
2b. Bizang 壁藏 (wall repositories).
We are able to get a fairly clear idea of what these objects are and how they look thanks to the text
and the detailed drawings attached to it (figs. 1, 2).
5
The zhang appear to be wooden shrines where
religious icons are sheltered, and the jingzang are receptacles for the scriptures. Both types of
structures are installed in the interior as immobile fixtures of architecture, just like built-in furniture
pieces. The subtypes are functionally similar to each other but offer visually different schemes of
design. The four subtypes of zhang, for instance, involve different degrees of technical complexity
and would be selected based on need and budget.
Both zhang and jingzang can be identified as miniature architecture (in the sense defined in
this dissertation) for three reasons.
6
First, they are timber-framed, follow the same structural logic,
and adopt the same woodworking technique as large-scale buildings. The main physical difference
lies in the reduced scale of the wooden components. Second, they imitate, painstakingly and
convincingly, the form of real buildings, from the elevated platform to the colonnades, the
bracketing system, the suspended eaves and the sweep of roof, etc. Each corresponding part has
appropriated the name of its large-scale equivalent, i.e., a miniature hall is directly termed a “hall,”

5
For reconstructions of these shrines and repositories, see Takeshima Takuichi, Eizo hoshiki no kenkyu 營造法式の研究
(A Study on the Yingzao fashi), vol. 2 (Tokyo: Chuokoron bijutsu shuppan, 1971), 469-779; Pan Guxi and He Jianzhong
何建中, Yingzao fashi jiedu 營造法式解讀 (Interpreting the Yingzao fashi) (Nanjing: Southeast University Press, 2005),
137-52.
6
See Introduction of this dissertation for my definition of miniature architecture.
25
and a miniature rafter directly a “rafter.”
7
It would seem that each miniature is a replica of some
implied “prototype” or “original” found in real life--perhaps a pavilion, a tower, or a mixture of
several buildings. Third, specific downscaling formulas have been prescribed in the Yingzao fashi for
these structures, confirming that they are indeed to be created via a conscious and rigorous
miniaturizing process.
An overview of the downscaling formulas would illuminate how exactly miniatures were
produced back in time. But before we delve into the technical details, two more types of miniatures
recorded in the Yingzao fashi require our attention. One is the tiangong louge 天宮樓閣 (lit. heavenly
palace towers and pavilions), a set of wooden “buildings” to be placed at the top of the
Buddhist/Daoist shrine or sutra repository (see figs. 1, 2). These are even smaller in scale, and they
do not seem to have any functions except for ornamentation. The names assigned to particular parts
of the miniature group are quite intriguing--nine-ridged palatial halls, tea houses, corner towers,
galleries, etc.--do they suggest that the miniatures are supposed to mimic these specific building
types? The term tiangong 天宮 (heavenly palace) is even more problematic: does it imply that the
mini halls and towers are to symbolize a certain imagined, heavenly realm? These questions cannot
be answered based on the Yingzao fashi alone, yet it is remarkable that the zhang, jingzang, and tiangong
louge are all to be installed in a religious (Buddhist or Daoist) setting and are likely to be involved in
certain religious rituals or practices.

7
But this appropriation does not happen without certain modification; the tripartite structure of a typical Chinese
building--the terrace, the columns and bracketing (vertical supports), and the overhanging roof--are “transformed” into
the dais (zuo 坐), the body (shen 身, also including columns and bracketing), and the crown (tou 頭) of the miniature
shrines and repositories. A variety of architectural motifs such as the balustrade (goulan 鉤闌), the festive gate (huanmen
歡門), and the coffered ceiling (pingqi 平棊) do not always follow the rules of large-scale woodworking and are added
creatively to the miniatures. There are even elements commonly used for wooden couches, beds, and tables and chairs,
notably the apron (yajiao 牙腳) and the decorative pattern of the archway (kunmen 壼門).
26
The other additional type of miniatures is the douba zaojing 闘八藻井 (eight-ribbed vaulted
coffer) with its smaller form, the xiaodouba zaojing 小闘八藻井 (miniature eight-ribbed vaulted
coffer), both to be installed in the center of the ceiling (figs. 3, 4). The coffers do not have an
ostensible architectural appearance, but they do embrace elements of large-scale structures including
doors, windows, balustrades, and brackets.
8
Like the other three types mentioned above, they
observe the same set of scaling formulas.

Scaling and the cai-fen system
The primary measuring unit of a Chinese timber-frame structure is called cai 材 (lit. timber material).
A cai does not have an absolute numerical value but always keeps a height-to-width ratio of 3:2 in
cross section.
9
The cai is further measured by a secondary unit called fen 分 (lit. fraction), and one cai
equals 15 by 10 fen.
10
To design a building, the first step was to determine the dimension of the cai,
so that the majority of the building components could be proportioned and calculated accordingly;
this is often referred to as the cai-fen system (caifenzhi 材分制) by modern architectural historians.
The Yingzao fashi has proposed a total of eight grades of cai to be adopted for different ranks
of buildings (fig. 5):
11

Grade of cai Height of cai Width of cai Application
1 9 6 Nine- and eleven-bay palatial halls
2 8.25 5.5 Five- and seven-bay palatial halls

8
These coffers are discussed in Yingzao fashi, vol. 1, 165-69.
9
Yingzao fashi, vol. 1, 73-75.
10
A third unit, zhi 栔 (6 by 4 fen in cross section), is used especially for brackets and beams.
11
Ibid. The ranking of the timber material has been examined by many scholars. See, for instance, Liang Sicheng,
“Yingzao fashi zhushi 營造法式註釋 (Annotations on the Yingzao fashi),” in Liang Sicheng quanji 梁思成全集 (Complete
works of Liang Sicheng), vol.7 (Beijing: Zhongguo jianzhu gongye chubanshe, 2001), 79-80, 378.
27
3 7.5 5
Three- and five-bay palatial halls; seven-bay
ordinary halls
4 7.2 4.8 Three-bay palatial halls; five-bay ordinary halls
5 6.6 4.4
Small three-bay palatial halls; large three-bay
ordinary halls
6 6 4 Gazebos, small ordinary halls
7 5.25 3.5 Small palatial halls and gazebos
8 4.5 3
Vaulted ceiling coffers for palatial halls; brackets
for small gazebos
(The heights and widths of cai are expressed in cun, a Song-dynasty unit of length. 1 cun = 3.09-3.20
centimeters, or roughly 1.22-1.30 inches.)
From Grade 1 to Grade 8, as the cai decreases in size, the corresponding type of structures also
decreases in overall size. In most cases, the social rank of a Chinese building was manifested not by
its form but scale: an eleven-bay palatial hall with first-grade cai, for instance, would have been
reserved for imperial palace buildings and temples only. To select a specific grade of cai hence would
mean to choose an appropriate scale, which ought to match the intended social identity and function
of the building.
Interestingly, a similar scaling scheme exists for miniature woodworking (fig. 6):
12

Grade of cai Height of cai Width of cai Application
1 1.8 1.2 Ceiling coffers; Buddhist/Daoist shrines
2 1.5 1.0 Aproned shrines
3 1.2 0.8 Nine-ridged small shrines; wall shrines; well
gazebos
4 1.0 0.66 Sutra repositories
5 0.6 0.4 Miniature ceiling coffers; Heavenly Palace and
ceiling coffers for Buddhist/Daoist shrines
6 0.5 0.33 Heavenly Palace for sutra repositories
(unit: cun)

12
The dimensions of these cai, however, are not listed explicitly as those of the damuzuo in the Yingzao fashi. The first who
has noticed this scaling scheme is Chen Mingda; see his Yingzao fashi cijie 營造法式辭解 (Annotations on the glossary in
the Yingzao fashi), ed. Ding Yao 丁垚 et al (Tianjin: Tianjin University Press, 2010). The table includes only miniature
woodworks; two other types of small-scale structures regulated in the Yingzao fashi, jingtingzi 井亭子 and jingwuzi 井屋子,
are not considered in this dissertation.
28
Take the Buddhist/Daoist shrine for example: for the majority part of the shrine, the size of the cai
has to be decreased to 1.8 by 1.2 cun. This means that ideally, a Buddhist/Daoist shrine would be a
1:4 replica of a medium-size hall.
13
Moreover, when it comes to the tiangong louge ornaments and
ceiling coffers installed in the shrine, the cai has to be further downscaled to 0.6 by 0.4 cun--merely
one-third of that of the shrine--making them “miniatures within the miniature.”
14
This kind of
“double miniaturization” is also observed in sutra repositories.
15

The advantage of adhering to this scaling scheme and the overall cai-fen system was
immediate. Since the dimension of each building component was standardized, architects could
easily apply their usual woodworking knowledge to miniature-making. A floral bracket arm (huagong
華栱), for instance, was always one cai (15 by 10 fen) in cross section and 72 fen in length, miniature
or not.
16
Though the actual size might vary, this measurement was to be maintained regardless of
scale and building type.
The observation that a grander scale should signify a higher-rank building does not
necessarily apply to miniature woodworks. In fact, it might be possible for an intricately crafted
small structure to capture the same sense of imperial grandeur found in real palace buildings. The
dwindling scale, which has shrunken past a threshold that it now denies the intrusion of the human
body into the interior “architectural” space, seems to have bestowed the miniatures with some

13
By “medium-size” I mean a hall adopting the fourth-grade cai. This is an ideal situation; the shrine could never be an
exact replica due to its nature of being a receptacle. More on this issue in the following chapters.
14
Yingzao fashi, vol. 1, 199. Mainly it is the bracketing system, the balustrade, the ceiling coffers, and the roof frame that
observe this scheme. Other non-architectural elements, however, have their distinctive scaling formulas.
15
Ibid., vol. 2, 6, 26.
16
Ibid., vol. 1, 76-77.
29
magical attributes.
17
The logic is reversed: the smaller the architecture, the farther away it is
dislocated from this world, and the more transcendent it becomes. It is therefore no coincidence
that the “Heavenly Palace” has to be presented using the smallest scale possible.
Meanwhile, the sumptuary laws must have been rendered useless for miniatures. The
Northern Song statutes “Tianshengling 天聖令” promulgated in 1029, though strictly forbidding
extravagant building forms and schemes for all but a small privileged group, did not enforce any
restrictions on miniature architecture.
18
Miniature-making, therefore, could mean a greater freedom
for both patrons and miniaturists to execute their vision and maximize the potential of their project.
But miniatures were not at all thrifty objects; a complete set of tiangong louge for the Buddhist/Daoist
shrine, for one, would cost 1,525 workdays (gong 功)--amounting to some 12,200 hours--to make and
install.
19

Models (Xiaoyang) and Ruled-line Paintings (Jiehua)
Building practices that did not fall into the criteria of officially supervised and administered projects
are not addressed in the Yingzao fashi. Architectural models, for instance, have been omitted

17
One of the Daoist canonical texts, Daodejing 道德經, has a famous saying that it is the emptiness (interior space) of
architecture that can be used. This quote has been inspirational for many American architects and theorists including
Frank Lloyd Wright and Rudolf Arnheim.
18
The Tianshengling specifically forbids officials lower than the fifth rank to use the “double bracketing” scheme for
their residences, in addition to a number of other restrictions. See reprint in Tianyige cang mingchaoben tianshengling jiaozheng:
fu tangling fuyuan yanjiu 天一閣藏明鈔本天聖令校證: 附唐令復原研究 (Correction and examination of the Tiansheng
Statutes based on the Ming edition in the Tianyige collections), 2 vols. (Beijing: Zhonghua shuju, 2006). An especially
relevant essay in the book is Niu Laiying 牛來穎, “Tiansheng yingshanling fuyuan tangling yanjiu 天聖營繕令復原唐
令研究 (Study and reconstruction of the Tang statutes based on the Tiansheng Statutes on Building and Repairing),”
650-74. The sumptuary laws in pre-modern China never failed to expose the anxiety of the ruling elites about the
parallelism between the possession and exhibition of material wealth and that of fame, social rank, and political power.
An exemplary work addressing this issue is Craig Clunas, Superfluous Things: Material Culture and Social Status in Early
Modern China (Cambridge: Polity Press, 1991), in which Ming literature and material culture is the main focus.
19
Takeshima 1971, vol. 2, 562. The gong is a unit used in the Yingzao fashi to measure the length of the time of labor; one
gong is equivalent to eight hours on a regular workday. See Yingzao fashi, vol. 1, 22-23.
30
altogether. This omission is to some degree compensated by other contemporary written sources
such as Wen Ying’s 文瑩 Yuhu Qinghua 玉壺清話 (Pure talk in a jade pot, 1078), which recounts an
interesting incident involving a model pagoda:
Guo Zhongshu was good at painting multistory palace halls and pavilions, and when builders
compared his paintings [with real structures], the measurement [of the painted architecture] did not
err in even the slightest. Emperor Taizong (r. 976-997) heard of his fame and granted him the official
title of Jiancheng. At that time the pagoda of the Kaibaosi monastery was about to be built; its
architect, Yu Hao of the Zhe region, designed a thirteen-level structure. Guo made a xiaoyang [of the
pagoda] to calculate the size of each level from the bottom upward; and coming to the top level,
there was an excess of a one-chi-five-cun distance that could not be integrated into the smooth
curving profile of the pagoda. He informed Hao of this problem and said, “You had better inspect
it.” Thereafter, Hao spent a few sleepless nights reexamining the original design, and it was indeed as
Guo had warned. In the next morning, he knocked at Guo’s door and knelt in gratitude.
郭忠恕畫殿閣重複之狀, 梓人較之, 毫釐無差. 太宗聞其名, 詔授監丞. 將建開寶寺塔, 浙匠喻
皓料一十三層. 郭以所造小樣末底一級折而計之, 至上層餘一尺五寸, 殺收不得, 謂皓曰: “ 宜
審之.” 皓因數夕不寐, 以尺較之, 果如其言. 黎明, 叩其門, 長跪以謝.
20

The term xiaoyang, as briefly mentioned in the Introduction, indicates an architectural model.
Whereas miniature woodworks in the Yingzao fashi were made as part of architecture--as shrines and
repositories placed inside monasteries or as ceiling ornaments--a model was physically detached
from any nesting or sheltering structures. Another important difference lies in purpose: while shrines
and repositories were largely products to exhibit creativity, virtuosity, and the pursuit of aesthetic
forms and expressions, models had to be faithful to real buildings, to accurately replicate every
technical detail. If the former were meant to invoke powerful images and arouse the feeling of
religious solemnity and royal magnificence, the latter, on the other hand, were meant to serve as
practical tools for experts to exchange ideas, check measurements, detect problems, and make
necessary corrections before real constructions commenced. It appears that architectural models,
though miniaturized, were deprived of any possible symbolic meanings.

20
Yuhu qinghua 玉壺清話, by Wen Ying 文瑩, reprint in Xiangshan yelu, xulu, yuhu qinghua 湘山野錄, 續錄, 玉壺清話,
annotated by Zheng Shigang 鄭世剛 and Yang Liyang 楊立揚 (Beijing: Zhonghua shuju, 1997), 21.
31
Underlying all models and other miniature woodworks, however, was the basic scaling rules
to be followed. It is not clear how the model pagoda in the text was made, but the author does tell
us that it was used to “calculate (ji 計)” and “inspect (shen 審)” the dimensions of the real. In other
words, there must have existed a fixed numerical relationship between the two scales--one for the
model and the other for the planned pagoda--so that such calculations could be of any use. The
tradition of modeling in China could be traced back to around 100 BCE, and written records have
testified to the enduring appeal of models over a millennium.
21
Relating to modern-day experience, it
has been a common practice for architects today to develop a series of models at different stages of
design--from the most preliminary, concept models to study models and the final representational
models, from hand-made models to digital and 3D printed models--which serve different audiences
and purposes.
In the quoted text, however, the modeler is a painter. The type of paintings Guo Zhongshu
郭忠恕 (fl. 952-977) was famous for have been traditionally classified as the jiehua 界畫, or ruled-
line paintings, which depict architecture, bridges, boats and other structures using a set of drawing
tools including a ruler (fig. 7). Ruled-line paintings were critiqued by generations of Chinese literati
for the “indulgence” in minuscule details, which were believed to inhibit imagination and creativity,
but they were also admired by many for the breath-taking verisimilitude, and are now treasured as
invaluable visual evidence and an archive of Chinese wooden architecture.
Interestingly, Guo’s ability to produce lifelike paintings of “palaces and pavilions” was
primarily, if not solely, based on his mastery of scale. When builders came to compare his painted
work with real architecture, they did not seem to pay much attention to either form, style, or color,
but rather the represented dimension of the buildings, which was said to “not err in even the

21
This is a topic to be further investigated in Chapter 5 of the dissertation.
32
slightest.” Guo must have been quite familiar with the actual scaling scheme of real wooden
structures so that he was able to reproduce these structures in a miniature form--whether as painted
objects or as freestanding models.
Similarly remarkable is the implication from the text that Guo was able to apply his
knowledge across different types of media, from wood to silk, from a three-dimensional working
space to a two-dimensional, planar one.
22
To be sure, in comparison with modeling, the
“miniaturization” of buildings in painting would necessarily entail a rather complex procedure
involving certain degrees of abstraction and distortion. The two-dimensionality of silk or paper
determines that only a portion of the “data” of the original building could be preserved and
represented at a time--a six-sided box could, at best, be drawn showing a half (three sides) of it,
whereas the hidden sides could only be hinted at or imagined.
23

Spiritual Vessels, Edible Architecture, Portable Shrines, Dollhouses, and Miniature Gardens
A more holistic picture of the social life of miniatures in the eleventh- to thirteenth-century China is
represented by a series of Song-dynasty miscellanea, which describe the metropolitan life in the old
and new capitals, Dongjing 東京 (Eastern Capital, modern-day Kaifeng) and Wulin 武林 (modern-
day Hangzhou). Five such texts will be examined here, including:
Dongjing menghua lu 東京夢華錄 (The Eastern Capital: a dream of splendor, 1147) by Meng
Yuanlao 孟元老;

22
The idea and practice of the transfer of media is examined in Shih-shan Susan Huang, “Media Transfer and Modular
Construction: The Printing of Lotus Sutra Frontispieces in Song China,” Ars Orientalis 41 (2011): 135-63.
23
It is dubious if foreshortening--which makes things appear more “natural” to the observer’s eye--was ever considered
during this process of miniaturization. It would involve an extra complex system of calculation if the law of perspective
was to be applied. Even if a certain degree of foreshortening was adopted by some painters, it seems that no strict rules
or formulas were applied. Chapter 4 of the dissertation will continue this discussion.
33
Ducheng jisheng 都城紀勝 (Record of the famous sights in the capital, 1235) by Naideweng 耐
得翁;
Xihu Laoren fanshenglu 西湖老人繁盛錄 (Xihu Laoren’s record of prosperity, ca. 1253) by
Xihu Laoren 西湖老人 (Elder at the West Lake);
Menglianglu 夢粱錄 (Dream of the yellow millet, ca. 1275) by Wu Zimu 吳自牧;
Wulin jiushi 武林舊事 (Recollection of everyday life in Wulin, after 1280) by Zhou Mi 周
密.
24

The five texts can be discussed as a group because they overlap in terms of contents as well as
structure, and it is not uncommon for later texts to emulate--even reiterate whole sentences from--
former works on the same subject. In this group, the Dongjing menghualu has been traditionally
recognized as the “model;” it contains descriptions of a more transient, everyday type of miniatures
that have been rarely, if ever, preserved as material remains. In all five texts, miniatures are most
frequently associated with several major festivals of the year, including the Qingming 清明 Festival,
the Ghost Festival, the Duanwu 端午 Festival, the Buddha’s birthday, and the Qixi 七夕 Festival,
during which time they became highly treasured and beloved goods to be purchased, presented, and
exchanged among people, from the imperial family to commoners.

Spiritual vessels
Mingqi, a type of funerary objects found in burial chambers, encompasses a great variety of miniature
architecture including granaries, wells, kitchens, animal pens, houses, multistory towers, and
fortifications.
25
These objects might serve as surrogates of worldly possessions to accompany the

24
All five texts are in the reprint Meng Yuanlao 孟元老 et al., Dongjing menghualu 東京夢華錄 (Shanghai: Zhonghua
shuju, 1962). Studies and partial translations of these texts include: Stephen West, “The Interpretation of a Dream: The
Sources, Evaluation, and Influence of the Dongjing Meng Hua Lu,” T'oung Pao 71 (1985): 63-108; and Jacques Gernet,
Daily Life in China on the Eve of the Mongol Invasion: 1250-1276 (Stanford: Stanford University Press, 1970).
25
Many of these miniatures still retain their original shape upon excavation and have since provided much visual
information about Chinese wooden architecture back in the first and second centuries. See Introduction for a selective
list of scholarship.
34
dead, or as channels for the living to communicate and interact with their deceased relatives and
friends. While the earliest mingqi were often made of bronze and clay, the much more ephemeral
paper-made mingqi started to be widely adopted in the Song. According to Zhao Yanwei 趙彥衛 (fl.
1163-1206), “the ancient mingqi were [so named because they were] vessels for the spirits. Today,
people use paper to make such objects and call them ‘vessels of the netherworld’” ( 古之明器, 神明
之也. 今之以紙為之, 謂之冥器).
26
The paper mingqi were perhaps meant to be burned, rather than
buried, under the influence of the more and more frequent practice of cremation.
27

Two instances of the use of paper mingqi can be found in the Dongjing menghualu. The first is
during the Qingming Festival in spring, on the third day of the third month, when inhabitants of the
capital swarmed to the outskirts of the city to “sweep the tombs.” This was also a time for family
outings and picnicking under the blossoming trees. On this day,
All the paper-goods shops rolled and folded paper in the form of towers and pavilions and displayed
them along the street.
紙馬鋪皆於當街用紙袞疊成樓閣之狀.
28

These paper buildings would be bought as the gifts for the dead. To be portable they would certainly
have been miniatures instead of full-scale replicas; and they would not need to follow any strict
scaling rules as those prescribed in the Yingzao fashi but could be simplified wherever their makers
saw fit.

26
Yunli manchao 雲麓漫鈔, juan 5, http://ctext.org/wiki.pl?if=gb&chapter=856778.
27
For burial practices in the Song, see Patricia Ebrey, “Cremation in Sung China,” American Historical Review 95 (1990),
406-28; Hsueh-man Shen, “Shengsi yu niepan--Tang Song zhiji fojiao yu shisu muzang de jiaocuo lingyu 生死與涅槃--
唐宋之際佛教與世俗墓葬的交錯領域 (Where Secular Death and Buddhist Nirvana Intersect: Secular and Religious
Burials during the Tang-Song Transition)” (unpublished paper, 2012),
https://www.nyu.edu/gsas/dept/fineart/people/faculty/Shen_PDFs/Shengsi.pdf.
28
Dongjing menghualu, 39.
35
The second instance details the burning of paper mingqi during the Ghost Festival (fifteenth
day of the seventh month). Though not mentioning architecture in particular, it helps to illustrate
the social and religious setting wherein miniatures as surrogates were produced and used:
Several days before [the Ghost Festival], the types of mingqi sold in the marketplace included boots,
shoes, kerchiefs, caps, golden rhinoceros clasps, surrogate belts, and five-colored clothes. These were
displayed on paper-made racks and transported around to be sold. The Pan Tower and the east and
west wazi districts were [buzzing with excitement] just as in the Qixi Festival. Everywhere there were
vendors of cakes, potted seedlings, fruits and the like; some were selling woodblock-print Scripture of
the August Mulian. Still other vendors chopped the bamboo into a three-legged stand three to five chi
in height, and weaved its top into a bowl-like lantern base to represent the yulanpen (Sk. ullambana, the
upturned vessel for salvation in Buddhism), on which paper clothes and money were disposed to be
burned. Musicians and performers of the entertainment districts had started the show of “Mulian
Rescues His Mother” since past the Qixi and would continue until the fifteenth day; the spectators
multiplied.
先數日, 市井賣冥器靴鞋, 幞頭帽子, 金犀假帶, 五綵衣服. 以纸糊架子盤游出賣. 潘樓并州东
西瓦子亦如七夕. 耍鬧處亦卖果食種生花果之類, 及印賣尊勝目連經. 又以竹竿斫成三腳, 高
三五尺, 上織燈窩之狀, 謂之盂蘭盆, 掛搭衣服冥錢在上焚之. 构肆樂人, 自過七夕, 便般目連
救母雜剧, 直至十五日止, 觀者增倍.
29

This passage informs us of the variety of the paper mingqi available at the time: in addition to
miniature buildings, people also made paper clothes and accessories. A special container, yulanpan,
which was required for the ritual of burning the mingqi, served as a means of providing salvation for
the ghosts. Associated with this ritual was the popularization of the worship of Mulian, a paragon
filial son and pious Buddhist, whose heroic adventure to hell must have become a widely-circulated
story made readily accessible to the public by woodblock printing and theatrical performances. (The
scriptures mentioned in the text, however, was probably to be used as mingqi.) The adventurous and
courageous spirit exemplified in the story must have appealed to the multitudes, winning much
admiration. The multi-day show based on the story, too, must have been quite a spectacle featuring
horrific scenes of the burning hell and grotesque-looking ghosts, providing a feast for the senses and

29
Ibid., 49.
36
the imagination.
30
It seems that two factors--the development of paper and printing industry on the
one hand and the popularization of Buddhist rituals, stories, and related performances on the other--
have encouraged the spread of paper goods as the new form of mingqi.

Edible architecture
The Spring Fair was a time when hundreds of social groups, guilds, and clubs held meetings to
publicly showcase their activities and products in a certain competitive spirit, to attract customers
and perhaps also recruit new members. The Wulin jiushi records such groups to include clubs of
drama, kick-ball, singing, ci-lyrics, wrestling, music playing, archery, tattooing, martial arts,
storytelling, shadow plays, hairdressing, and tricks and magic, each having a unique and catchy name.
Among the most extraordinary displays on the fair was certain “edible architecture” from the cooks’
guild and the bakers’ shop (chuhang guoju 廚行果局):
Certain participating groups, flaunting the so-called ingenious design ideas of theirs, used the tongcao
plant [tetrapanax?] and silk gauze to sculpt and decorate [foodstuff] into towers, terraces, and various
dioramas. They embellished them with pearls and jade to attain an utmost exquisiteness. A dish [of
miniature architecture] like this could reach as much as tens of thousands of coins in value, even
though these were such wasteful and useless items created for nothing but momentary pleasure.
有所謂意思作者, 悉以通草羅帛, 雕飾為樓台故事之類, 飾以珠翠, 極其精緻, 一盤至直數萬,
然皆浮靡無用之物, 不過資一玩耳.
31

Obviously, the author of the text was most critical of such costly and unnecessary “art.” But for the
fair-goers, especially the culinary geniuses, the most rewarding aspect of the edible miniatures
perhaps lay not as much in the monetary value as in the enthralling experience of miniature-making,
which must have brought much delight, pride, and a feeling of self-fulfillment.

30
For an overview of Song plays and drama performances, see William Dolby, A History of Chinese Drama (New York:
Barnes & Noble, 1976), 14-20.
31
Dongjing menghualu, 377.
37
Architecture made of food was also produced on the Duanwu Festival (fifth day of the fifth
month), a midsummer’s day for people to row dragon boats and commemorate the great poet Qu
Yuan 屈原. According to the Wulin jiushi, for the imperial family and the court officials, a certain
food art was created using the zong 粽 (sticky rice dumplings), a traditional delicacy consumed on
this day:
The artfully made zong had many varieties; some were even connected and combined to form towers,
terraces, boats, and carriages.
巧棕之品不一, 至結為樓臺舫輅.
32

This observation resonates with the Xihu Laoren fanshenglu, which proudly claims that
[Wulin] is the only capital under heaven where people stack the zong together to make artful designs
such as towers, pavilions, gazebos, carriages, and the like.
天下惟有是都城將棕揍[ 湊] 成樓閣, 亭子, 車兒諸般巧樣.
33

We are not, however, provided with any further details as how such edible miniatures looked, and
there seems to be a general lack of visual evidence for this curious food art.
34

Shaluo shrines
A type of makeshift portable shrines, referred to as the shaluo 沙羅 shrines, was reportedly used for
sheltering Buddhist statues on special occasions. The Dongjing menghualu notes that on the eighth day
of the twelfth month, monks and nuns in Dongjing would parade in the streets and alleys of the city
in groups of three, four, and five. Chanting the Buddha’s name, they would
use a silver-covered bronze shaluo or some other basin of decent quality to contain a Buddha’s statue
made of gilt bronze or wood. They immersed [the bottom of] the statue into fragrant water and

32
Ibid., 379.
33
Ibid., 118.
34
Most of the zong Chinese people eat today are triangular or sometimes rectangular in shape. They do not come in any
immediately recognizable forms such as houses.
38
constantly sprinkled water over it with a sprig of willow tree. In this way they would go door to door
to preach [Buddhist teaching] and convert [the townsfolk].
以銀銅沙羅或好盆器, 坐一金銅或木佛像, 浸以香水, 楊枝洒浴, 排門教化.
35

According to the same text, the ritual of “bathing the Buddha” was also held in major monasteries
of the city and on the eighth day of the fourth month, the Buddha’s birthday.
Here the shaluo seems to have consisted of no more than a basin, but a sheltering structure
could be added on top to convert it into a proper miniature shrine suitable for its “occupant.” An
entry in the Xihu Laoren fanshenglu states that in Wulin, to celebrate the Buddhas’s birthday,
The nuns and monks of every monastery came to assemble flower pavilions and flower houses on
top of the sacrificial tables; in [each pavilion or house] they placed a golden Buddha in a shaluo basin
and filled the basin with fragrant water. The [entire assembly] was then carried to the market [for the
parade].
諸尼寺僧門卓上札花亭子並花屋, 內以沙羅盛金佛一尊, 坐於沙羅內香水中, 扛臺於市中.
36

A similar instance appears in the Wulin jiushi, also on the Buddha’s birthday:
Monks and nuns, in a competitive spirit, used small basins to contain bronze statues [of the Buddha],
immersed the statues in sweet water, and covered them with flower huts. Playing cymbals during
procession, they went to visit every great mansion and wealthy family in the city, where they used
small dippers to pour water [onto the statues] and asked for alms.
僧尼輩競以小盆貯銅像, 浸以糖水, 覆以花棚, 鐃鈸交迎, 徧往邸第富室, 以小杓澆灌, 以求施
利.
37

It is highly likely that the flower pavilions, houses and huts mentioned in these texts were
miniatures (since they were to be carried around the city) and could even be somewhat similar to the
shrines prescribed in the Yingzao fashi. A possible connection is suggested by the particular term hua
花 (flower). In classical Chinese this term is alternatively written as 華 (considered a more refined
written form) which carries the connotations of “decorated” and “splendid.” This latter form has

35
Dongjing menghualu, 61.
36
Ibid., 117-18.
37
Ibid., 378.
39
been adopted in the Yingzao fashi throughout to name many of the core components of the wooden
architecture, such as the huagong 華栱 (floral bracket arms) and the huaban 華版 (floral-patterned or
decorative panels).
38
In this light, the “flower pavilions” could mean either a structure made of real
flowers, or a structure that is heavily decorated, with or without distinctive floral motifs. Another
connection lies in the function of these miniatures. Like the examples in the Yingzao fashi, the
“pavilions,” “houses,” and “huts” were made to venerate and shelter the Buddha. Even though they
were not permanent structures installed in monasteries and temples, it might be possible that their
makers were more or less inspired by those immobile shrines during their creative activity.

Mohouluo dolls and dollhouses
On the Qixi Festival (the seventh night of the seventh month), a particular type of dolls called the
Mohouluo 摩侯羅 (alt. Mohele 磨喝樂; Sk. Rahula) became the most treasured goods for both
adults and children. According to the Dongjing menghualu, the Mohouluo dolls would be on sale
everywhere in the outer city during the festival. The dolls, though made of clay, were
placed on carved, colorfully embellished wooden daises surrounded by balustrades. Some were
sheltered by red silk gauze and blue-green envelopes; some were ornamented by golden pearls and
ivory jade. A pair of dolls could cost as much as several thousand coins.
悉以雕木彩裝欄座, 或用紅紗碧籠, 或飾以金珠牙翠, 有一對直數千者.
39

The dais, the balustrade, the sheltering gauze and the envelope seem to suggest a certain
architectural-like structure. As we encounter in the Yingzao fashi, the dais (zuo 座) and the balustrade
(lan 欄) are two distinctive components of the miniature woodworks. These forms were

38
More connotations of the term hua in relation to architecture are discussed in Feng 2012, 138-80. The use of hua as a
prefix could indeed suggest that flower-like forms are adopted as either structural supports or motifs of decoration of a
building, but it could also mean that something is exquisite and well embellished in general.
39
Dongjing menghualu, 48.
40
appropriated from real wooden buildings and miniaturized for the encasement of religious icons.
The silk gauze (sha 紗) and the envelope (long 籠), on the other hand, are more ambiguous terms.
The character long literally means a cage, a bamboo basket, or a trunk. Put together, shalong could
indicate a silk gauze canopy or something like a small tent for interior use.
40
Speaking of canopies,
we are reminded that the miniature shrines in the Yingzao fashi are termed as zhang, or canopy
shrines. Could this suggest that the shalong had a similar (though perhaps simpler) form?
The Xihu Laoren fanshenglu provides a brief description of how the Mohuoluo dolls looked:
“[They were] mostly dressed in crimson vests and blue-green silk skirts; some were wrapped in
swaddling clothes, and some wore hats” ( 多著乾紅背心, 繫青紗裙兒; 亦有著背兒, 戴帽兒者).
41

A more detailed account is given in the Menglianglu:
In the imperial court and the houses of wealthy families, Mohele, also known as the Mohouluo dolls,
were modeled and sold. [The dolls] were crafted using clay and wood; additionally, [craftsmen] made
colorfully embellished balustrades and daises, and enveloped the dolls with blue-green silk gauze.
Below, they supported the dolls with a table enclosed by dark green, gold-lined aprons, and those
decorated with gold, gems, pearls and jade were especially well-crafted…
Children in the marketplace, holding newly picked lotus leaves in their hands, mimicked the
appearance of Mohouluo. This custom was widespread in the Eastern Capital [i.e. Dongjing] and has
not changed ever since; no one knows from which textual source it was derived from.
內庭與貴宅皆塑賣磨喝樂, 又名摩㬋羅孩兒, 悉以土木雕塐, 更以造綵裝襴座, 用碧紗罩籠之,
下以桌面架之, 用青綠銷金桌衣圍護, 或以金玉珠翠裝飾尤佳...
市井兒童, 手執新荷葉, 效摩㬋羅之狀. 此東都流傳, 至今不改, 不知出何文記也.
42

A similar entry appears in the Wulin jiushi:
The clay dolls called the “Mohouluo” were sometimes extremely exquisite; decorated with gold and
pearls, their value could not be calculated…

40
Alternatively, it could mean a lantern made of silk gauze. Many Tang and Song texts speak of shalong or bishalong used
as lanterns or as envelopes/encasements for keeping the dust off works of calligraphy.
41
Dongjing menghualu, 120.
42
Ibid., 160.
41
Young boys and girls often wore short-sleeved coats made of lotus leaves and held lotus leaves in
their hands to mimic Mohuoluo. These were perhaps old customs of Central China. Prior to the
Qixi, as a tradition, the Department of the Construction and Repair of Imperial Buildings would
present to the court ten tables of Mohouluo dolls, each table containing thirty dolls. A large doll
could reach three chi in height; some were carved of ivory, some made of ambergris and foshouxiang
(lit. Buddha-hand fragrance), while engraved gold leaves, jewels and jade were used for all. Clothes,
caps, coins, hairpins, bracelets, jade pendants, pearl-strung curtains, hairs, and the toys held in their
hands were all made of the seven precious metal and gems, and each [doll] was encased in a five-
colored, silk gauze cabinet embellished with engraved gold leaves. Some regional military
commanders, dignitaries, and capital officials even ordered golden dolls to be cast and presented to
the court.
泥孩兒號摩㬋羅, 有極精巧, 飾以金珠者, 其直不貲...
小兒女多衣荷葉半臂, 手持荷葉, 效顰摩㬋羅. 大抵皆中原舊俗也. 七夕前, 脩內司例進摩㬋
羅十卓, 每卓三十枚, 大者至高三尺, 或用象牙雕鏤, 或用龍涎佛手香製造, 悉用鏤金珠翠. 衣
帽, 金錢, 釵鋜, 佩環, 真珠, 頭鬚及手中所執戲具, 皆七寶為之, 各護以五色鏤金紗廚. 制閫貴
臣及京府等處, 至有鑄金為貢者.
43

These sources confirm that a certain kind of highly decorative encasement was made for the
precious dolls. Such an encasement should include a dais (sometimes with a table underneath), the
balustrade, and a tent-like envelop. The last source gives it a specific name, shachu 紗廚, literally silk-
gauze cabinet, which could carry several layers of meanings. First, the shachu has been generally used
as a kind of interior partition in traditional Chinese architecture. Though the name suggests a certain
silk fabric as the primary material, in later historical developments, wooden partitions--which
constituted a wooden frame and lattice screens--have also been indiscriminately referred to as shachu.
Second, shachu is a term sometimes interchangeable with shazhang 紗帳, meaning silk
canopies. As we have read in the Yingzao fashi, a zhang did not necessarily involve the attachment of
any fabric. What lies in the heart of the zhang, then, is not the use of canopy, but the fact that it
defines an intimate, sheltered space for a single occupier. In this sense, the shachu of the Mohouluo

43
Ibid., 380-81.
42
dolls and the zhang in the Yingzao fashi shared a very similar function--to protect and enshrine.
44
This
point becomes even more prominent when we consider the meaning of the character chu 廚. The chu
denotes a cabinet or cabinet-like structure for storage, usually with openable door leaves. A number
of surviving historical miniature shrines have been named chu, such as the famous Tamamushi Zushi
玉蟲廚子, a miniaturized, single-story wooden hall elevated on a high plinth (fig. 8).
45
Intriguingly,
the name zushi 廚子 (Ch. chuzi), literally small cabinets, refers to not the Tamamushi alone but
almost all Buddhist miniature shrines in Japan. In light of the connections between shachu, zhang, and
zushi, it is possible that the Mohouluo dolls were sheltered by some kind of miniature shrines with or
without silk canopies.
Putting dolls in miniature architecture creates an interesting juxtaposition. While it makes
sense to encase religious icons in shrines for worship, it is, at first sight, a bit surprising that the
Mohouluo dolls would be enshrined in a similar manner, almost as idols to be admired by the
emperor and commoners alike. This myth can be dispelled considering the dual identity of
Mohouluo as both a human child and a Buddhist disciple. An earlier textual reference to the
Mohouluo dolls states that “During Qixi, as a custom, people would make wax models of infants
and float them on the surface of the water as a form of entertainment. Called huasheng, these were
considered auspicious dolls for women to give birth to boys. They originated from the Western
Regions where they were called Mohouluo” ( 七夕俗以蠟作嬰兒形, 浮水中以為戲, 為婦人宜子

44
More discussion on the function and meaning of zhang can be found in Neil Schmid, “The Material Culture of
Exegesis and Liturgy and a Change in the Artistic Representation in Dunhuang Caves, ca. 700-1000,” Asia Major 19
(2006): 171-210.
45
An overview of the shrine, its history, architectural features, and pictorial program can be found in Akiko Walley,
“Flowers of Compassion: The Tamamushi Shrine and the Nature of Devotion in Seventh-Century Japan,” Artibus Asiae
72 (2012.2): 265-322.
43
之祥, 謂之化生. 本出西域, 謂之摩喉羅).
46
The same huasheng 化生, a child-looking figure, is
included as one of the eight major woodcarving motifs in the Yingzao fashi (fig. 9). Scholars have
confirmed that Mohouluo was in fact none other than the historical Buddha’s only son Rahula (Ch.
Luohouluo 羅睺羅), who later converted to Buddhism and became one of the ten disciples of the
Buddha.
47
It seems that Mohouluo in Northern and Southern Song China was worshipped as the
divine son, a figure almost like the cherubs, who carried the dual significance of religious piety and
the mundane wish of the Chinese household to produce healthy boys as heirs of the family. Visual
and textual evidence of this long tradition of the Mohouluo worship has been found at the Mogao
cave temples in Dunhuang (fig. 10).
48

What can be added to this analysis is a note on the intersection between miniature
architecture and childhood. The shachu of Mohouluo dolls reminds us of the dollhouse. Of course,
unlike a dollhouse, which usually features a spacious, all-inclusive interior equipped with different
household articles, the “house” of a Mohouluo doll was as a much narrower container and would
not permit any imaginary activity of the doll. Nonetheless, in both cases there exists the intention to
create a miniature world suitable for the occupant; in return, this miniature world is meant to please
the eye and amuse those living in the “normal,” “real” world. What else could the Mohouluo dolls
bring to us aside from seasonal entertainment, good wishes, and religious inspirations? The two
dominant themes proposed by Susan Stewart for the dollhouse--wealth and nostalgia--also appear
true in this case. On the one hand, the Mohouluo “dollhouses,” embellished with precious metal and

46
Quoted in Guo Junye 郭俊葉, “Dunhuang bihua, wenxian zhongde ‘Mohouluo’ yu funu qizi fengsu 敦煌壁畫, 文獻
中的 ‘ 摩睺羅’ 與婦女乞子風俗 (The Mohouluo in Dunhuang murals and documents and its relationship with the
custom of ‘begging for sons’ of women),” Dunhuang yanjiu 142 (2013.6): 15.
47
Ibid., 13, 16-17.
48
Ibid., 13-17.
44
gems, were certainly “extravagant displays of upper-class ways of life.”
49
On the other hand, the
“dollhouses” would arouse a sense of nostalgia by presenting an encapsulated childhood or
infancy.
50

Miniature gardens
Another type of treasured items on the Qixi Festival were the miniature gardens, which seem to be a
kind of predecessors to today’s penjing 盆景 (container gardens or landscapes):
[Some vendors] applied a layer of earth to the surface of a small tray, upon which they planted grains
of millet and let them grow into seedlings. They then added small cottages, flowers and [dwarf] trees
[to the scene], and placed small figurines of peasants and farmers; all was in the likeness of a rural
village. This was called the “tray of grains.”
又以小板上傅土, 旋種粟令生苗, 置小茅屋花木, 作田舍家小人物, 皆村落之態, 謂之榖板.
51

While in this case the miniature world had an unmistakably agricultural, rural setting, there were also
mini landscapes cultivated in a similar fashion, as evidenced in many Tang and Song written sources.
The Yingzao fashi, for one, mensions the so-called “artificial mountains” (jiashan 假山) and
“container mountains (penshan 盆山), though it does not elaborate how they should be made.
52
Rolf
Stein’s research on the history of miniature gardens exposes excellently how the East Asian
fascination with miniature landscapes (and more generally, with the idea of the microcosm) can be
traced back to Han China.
53
It was during the Song, however, that the literature on the techniques of

49
Stewart 1993, 61-62.
50
Ibid., 44. As Stewart observes, such an intersection comes not only from the fact that “the child is in some physical
sense a miniature of the adult, but also because the world of childhood, limited in physical scope yet fantastic in its
content, presents in some ways a miniature and fictive chapter in each life history.”
51
Dongjing menghualu, 49.
52
Yingzao fashi, vol. 3, 82-83.
53
Stein 1990, 23-48.
45
gardening and other leisure activities as aspects of “elegant living” started to flourish and came to be
massively printed as manuals.
54
Stein speaks of two intertwining threads of themes found in these
miniatures: one is the Daoist aesthetic associated with longevity and immortality, and the other is the
peasant element emphasizing fertility and fecundity.
55
The “tray of grains,” displaying an idyllic scene
of the agricultural life, gives further support to Stein’s observation.

Puppets and the theatricality of miniatures
The miniatures in these examples had varied forms and functions, but they were all products of the
same social environment and cultural milieu. Techniques of miniature-making and mass production
spread widely across various crafting and manufacturing professions in the capitals of the Northern
and Southern Song. The types of trade goods made in miniature forms encompassed food, toys,
funerary objects, ritual artifacts, and home decor. What became miniaturized, of course, was not
confined to architecture alone but had extended to vehicles (such as imperial carriages and dragon
boats), animals (bulls, horses, elephants, lions, and fowls), and human figures of different ages and
occupations.
56
In many cases, these miniatures were not isolated objects but were actually put
together as part of an integral scene, a diorama, or a “stage” where certain “drama” was to be
enacted.
57
The shaluo shrines, the Mohouluo “dollhouses,” and the miniature gardens, for example,

54
Ibid. Stein points out the frustrating fact that none of these manuals, in addition to contemporary and later
encyclopedia, seems to have treated the miniature garden as a unique phenomenon under certain cultural criteria. He
tentatively traces the term penjing to xiezijing 些子景 (lit. a bit of landscape) referred to in a late-Yuan and early-Ming
source, where familiar elements such as dwarf trees, small balustrades, and figurines were used to create a miniature
world in the container. See also Clunas 1991.
55
Ibid., 112.
56
These miniatures are discussed in the five miscellanea consulted in this section.
57
A discussion on the theatricality of miniatures is in Stewart 1993, 54.
46
were each meant to present or evoke a spectacular “scene”--be it mythological or historical,
imaginary or nostalgic.
The magic of the miniatures in creating drama and stirring emotions and memories
originates from their innate ability to mirror and distort the real world. The result of miniaturization,
therefore, is the birth of a familiar-looking space-time nested in this world, which is at the same time
near and far, approachable and inaccessible. The miniature is essentially a paradox, a mimicry of
what is considered usual and ordinary, which turns out to be extraordinary and magical, creating a
clear departure from daily experience. In the Northern and Southern Song, there was an emerging
urban culture of drama taking the forms of shows, plays, storytelling, parades, acrobatic
performances, and festival extravaganzas, and it was no coincidence that miniatures rose quickly in
popularity with this new passion for theatricality. During the Ghost Festival, as briefly mentioned
earlier, the Buddhist story of Mulian saving his mother from hell was put on stage and became the
most spectacular show on this occasion not just for commemorating the dead but also for valorizing
and entertaining the living. This was not an isolated case. In fact, drama literature and performance
became so widely permeated in all echelons of the society that it gave birth to a series of professions,
organizations, and official departments specialized in different aspects of the performative art on
stage.
The Ducheng jisheng provides an excellent summary of who these professionals were. For
actors and actresses, there were five distinctive roles called moni 末泥, yinxi 引戲, fujing 副淨, fumo
副末, and zhuanggu 裝孤, each fulfilling specific purposes of the play.
58
Acrobatic performers
included wrestlers, pole climbers, sword dancers, bird tamers, archers, and those doing various

58
Dongjing menghualu, 96.
47
physical feats such as somersaults, walking on stilts, and playing fireworks.
59
Special performances
involving the use of miniatures were called “puppet plays (kuileixi 傀儡戲)” or “shadow plays (yingxi
影戲).” A puppet play was a three-dimensional, miniaturized version of a normal play; but instead of
being performed by adult actors and actresses, it featured the use of string puppets, rod puppets,
waterborne puppets, and most intriguingly “human puppets (roukuilei 肉傀儡),” which were actually
children or teenagers “manipulated” by the puppeteers in certain ways (don’t we see here, again, the
intersection between childhood and miniatures?).
60
On the other hand, a “shadow play” would be
unfolded on a two-dimensional “stage”--a single backdrop like the screen in today’s movie theaters--
on which the shadow of paper or leather puppets was cast and animated.
61
Puppet plays were a
beloved form of entertainment in the marketplace as well as on imperial feasts; the Dongjing
menghualu has detailed a waterborne extravaganza held annually at the imperial lake, the Jinmingchi
金明池 (Pond of Golden Light), located west of the Northern Song capital. During the feast, there
would be large and small dragon boats carrying all kinds of performers, musicians, wooden puppets,
kickball players, and spinning dancers.
62
The scene must have been so impressive and memorable
that it inspired a number of great artworks, most notably Wang Zhenpeng’s 王振鵬 (fl. 1280-1329)
painting illustrating the entire process of the dramatic performance (figs. 11, 12).
63
An intriguing
account in the Wulin jiushi mentions a curiously crafted “jade wine-boat” presented in 1179 by

59
Ibid., 97.
60
Ibid.
61
Ibid., 97-98.
62
Ibid., 40-41.
63
This painting is examined in Zhang Huazhi 張華芝, “Wanmin tongle: Yuan Wang Zhenpeng Longchi jingdutu 萬民
同樂: 元王振鵬龍池競渡圖 (Pleasure with the masses: the painting of the rowing competition in the dragon pond by
Wang Zhenpeng of the Yuan dynasty),” Gugong wenwu yuekan 361 (2013.4), 48-57.
48
Emperor Xiaozong 孝宗 (r. 1162-1189) of the Southern Song to his father, the former emperor:
“As the wine filled up the jade boat, most of the figurines on the boat were animated as if they were
alive” ( 酒滿玉船, 船中人物, 多能舉動如活).
64
Could this be a remote echo of the show on the
imperial lake?
In this culture of drama, spectators were also performers, and the world was a grand stage
where dramas of life unfolded. During the Qixi Festival, some households in Wulin “performed in
small towers using people [as puppets] for a gigantic shadow play” ( 戲于小樓, 以人為大影戲).
65

The theme of regarding the world as the ultimate stage and life as essentially theatrical and illusory
was not uncommon in many contemporary and later literary works.
66
It seems that the idea and
practice of miniaturization found a certain affiliation with this view of world and life, and in turn
engendered a surprisingly glamorous material culture of miniatures whereby the triviality and
transience of human life could be dramatized, experienced, and contemplated upon. Opposite to
this sense of humbleness was the unleashed human imagination, creativity, and perhaps a feeling of
self-importance; after all, we humans are the creator, collector, and manipulator of the miniature
world, of its architecture, its physical environment, and its “occupants” from dolls to puppets.

Conclusion: Dreaming of Lilliput in Song China
The miniatures introduced in this chapter served different purposes and were used for different
occasions and/or locations. Miniature shrines (including the zhang and the shaluo shrines), sutra

64
Dongjing menghualu, 471.
65
Ibid., 370. The xiaolou 小樓 might be a small or miniature tower--it is hard to determine.
66
Wilt Idema and Stephen H. West, Chinese Theater, 1100-1450: A Source Book (Wiesbaden: Steiner, 1982), 291. See also
Jeehee Hong, “Virtual Theater of the Dead: Actor Figurines and Their Stage in Houma Tomb No. 1, Shanxi Province,”
Artibus Asiae 71 (2011.1): 102-03.
49
repositories, and the “Heavenly Palace” were all created for “transcendent” or “sacred” causes,
especially religious rituals or practices. Paper towers and pavilions, in a similar manner, were
“spiritual” items to function in the world of the dead. On the other hand were miniatures designed
for overtly mundane purposes: architectural models were created to facilitate communication
between experts, to detect structural flaws, and to demonstrate the feasibility of a building project;
edible architecture, dollhouses, and miniature gardens were meant to please the eye, to flaunt wealth,
to arouse feelings of nostalgia, to induce imaginations, and to dramatize everyday experiences.
Though some of these miniatures were permanent structures while some were merely for overnight
entertainment, one important commonality was their magical ability to open up a new world by
downscaling and alienating the “real” world in front of us. They alluded to the familiar and the
ordinary, wove the daily elements together and represented them as half-real-half-fabricated stories,
as an impenetrable, self-contained, timeless universe. Admittedly, this world-making magic heavily
relied on the creative manipulation of the scale, on the miniaturizing process, whereby a mixed sense
of unfamiliarity and theatricality was engendered.
It is perhaps not too far-fetched to suggest that an Alice-in-Wonderland syndrome or
Lilliputian complex existed in the collective subconscious of Chinese literature. Gulliver’s encounter
with the tiny inhabitants of Lilliput is quite comparable in theme to several widely-read Tang tales on
uncanny dreams, notably the “Nanke Taishou zhuan 南柯太守傳 (Governor of the Southern
Tributary State)” by Li Gongzuo 李公佐 (fl. 766-818). The story tells of a certain desolate man,
Chunyu Fen, who dreamt of entering the “Kingdom of Huaian”--an anthill where he was to live for
more than twenty years.
67
Unlike Gulliver who was aware of the tiny kingdom he steps into, Chunyu

67
This tale became a classic in the Northern Song and has been incorporated into Taiping guangji 太平廣記, juan 475,
http://ctext.org/taiping-guangji/475/chunyufen/zh.
50
had no idea he was in a different world until the end of his dream, since his body was miniaturized
into the scale of an ant. Also miniaturized was the length of time--the twenty-odd years in the dream
actually lasted for only a few hours. In the beginning of those twenty-odd years, he married the
king’s daughter, became a governor, and enjoyed a wealthy life. But after suffering a major military
defeat, he was charged of treason and soon escorted back to the human world. Waking up from the
dream, Chunyu searched his backyard and found a little anthill beneath an old ash tree, where tiny
city walls, pavilions, towers, and swarms of ants could still be seen. The moral of this tale is to warn
against any worldly desire of amassing wealth and fame, because in the eyes of the wise, eventually,
they would turn out to be nothing but heaps of anthills. The sober and somewhat pessimistic tone
on the insignificance and impermanence of life carries a Daoist note on self-renunciation and the
Buddhist ideal of enlightenment. As will be elaborated in later chapters, such concepts and ideals
would become the intellectual underpinnings of the burgeoning of miniature art.
51
2. Miniatures as Sacred Repositories, Part I: The Longxingsi Sutra Case

This and the next chapters provide a close examination of the miniature woodwork known as
jingzang (sutra repositories) in the eleventh-century China. Corresponding to the categorization in the
Yingzao fashi, the chapters each focus on one of the two major subtypes--the zhuanlun jingzang (wheel-
turning sutra repositories) and the bizang (wall repositories)--by presenting and analyzing pertinent
architectural remains as well as revealing their dimensional characteristics in comparison with the
official templates.
The key example to be investigated in this chapter is the zhuanlun jingzang at the Longxingsi
in Zhengding, Hebei. Known as a masterpiece of Northern Song architecture, this particular
revolving sutra case has been regarded by scholars as an excellent reflection and representation of
the contemporary woodworking techniques. More often than not, it is looked upon as a certain
“model” or “replica” of large structures, whereas its distinctive nature as a miniature receives
relatively less consideration. The first and foremost question this chapter aims to answer, therefore,
concerns the identification of this sutra case: on what grounds can it be labeled as miniature
architecture? Does this identification change our view of the sutra case, and of the architectural
tradition it exemplifies? Equally important is the issue of dating--while the dating of a miniature
woodwork could be largely tentative, even speculative, observing the scaling scheme it follows might
shed new light.
Following identification and dating, this chapter turns to question how the Longxingsi sutra
case might have been used in history. It is interesting that the sutra case has been frequently
discussed in the discourse of Chinese architectural history--i.e. it has been recognized as a piece of
architecture more than anything else--whereas the case itself as a receptacle of scriptures involving
specific religious rituals and practices appears to be a much neglected matter. In fact, as will be
52
demonstrated in this chapter, the spiritual drive and materialistic concern lying behind the making of
sutra cases cannot be fully exposed without an inquiry into function. A survey of historical examples
preceding and following the Longxingsi sutra case will help us to better determine the application
and social significance of this type of miniatures.
The last section of this chapter, engaging with the concept of deconstruction, attempts to
read the miniature as a composite of several distinctive, iconic formal elements. The deconstructive
interpretation of the sutra case presents many new problems and intellectual challenges: was the
miniature made to consciously “copy” certain classical examples or prototypes? What forms have
been appropriated, altered, or reinvented? Does the change in scale induce a consequent change in
the meaning of form? I propose that the miniaturizing process can be compared to a deconstructing
process in which the assumed, long-established interrelationships between word and meaning, sign
and signified, and in this case, architectural forms and the significances they carry, are destabilized
and ruptured. Deconstructive reading therefore offers a means to analyze miniature architecture as
essentially non-architecture and anti-architecture, a dissolution of the established architectural
discourse.

The Zhuanlun Jingzang (Wheel-turning Sutra Repository) at Longxingsi
Standing some eight meters tall in the center of a two-story hall, the zhuanlun jingzang at the
Longxingxi is a fairly massive interior installment (fig. 13).
1
It would be, at first sight, awkward to call

1
My survey of this sutra case has been digitized and accessible at my online database,
https://chinesearchitecture.wordpress.com/2016/03/05/longxingsi/, which includes a Rhino 3D model and a
photogrammetry model. Studies on the Longxingsi and its sutra case include (in a chronological order): Liang Sicheng 梁
思成, “Zhengding diaocha jilue 正定調查紀略 (Brief report on the field survey in Zhengding),” Zhongguo yingzaoxueshe
huikan 4 (1933.2): 1-40; Laurence Sickman and Alexander Soper, The Art and Architecture of China (Penguin Books, 1984);
Nancy S. Steinhardt, Liao Architecture (Honolulu: University of Hawaii Press, 1997); Zhang Xiusheng 張秀生 et al. eds.,
Zhengding Longxingsi 正定隆興寺 (Beijing: Wenwu chubanshe, 2000); Liu Youheng 劉友恒 and Du Ping 杜平, “Woguo
xiancun zuizaode zhuanlunzang: Zhengding Longxingsi Songdai zhuanlunzang qianxi 我國現存最早的轉輪藏: 正定隆
53
this structure a “miniature,” a term which is usually associated with portable, hand-held items.
Indeed, by what standards can one identify something several times larger than the human body as
essentially small? Here I ask that we do not look upon this receptacle based on expectation or our
conventional judgment about size, but instead consider how it would compare with its surrounding
structure. One soon notices that the Zhuanlunzangdian 轉輪藏殿 (Hall of the wheel-turning sutra
repository, or library hall), reaching 23.05 meters tall and 13.98 by 13.3 meters across, is significantly
larger, in which the sutra case is nested like a fetus in the mother’s womb (fig. 14).
2
Such nesting
requires changes to be made to the wooden frame of the hall in order for the sutra case to fit in. As
scholars have observed, a major beam at the first level of the library is slightly curved and elevated at
one end to make room for the crown of the sutra case. Additionally, two interior columns have been
shifted outside the orthogonal column-grid, forming a hexagonal boundary (fig. 15).
3

The nesting of the sutra case inside the library hall demands more than structural
adjustments to be successful. The sutra case is fixed in a round pit on the ground floor by a single,
robust wooden pivot, the top end of which penetrates a small hole in the second floor (fig. 16). But
the link between the nest and the nested reaches far beyond physical contact; they are bonded also
through structural similarity and dimensional consistency. In terms of the scaling scheme, the sutra
case has adopted a cai of 4.5 by 3 centimeters, whereas the cai of the library hall ranges between (20-

興寺宋代轉輪藏淺析 (The earliest surviving revolving sutra repository in China: a brief analysis of the Northern Song
revolving sutra case at Longxingsi in Zhengding),” Wenwu chunqiu 59 (2001.3): 52-55.
2
The phenomenon of nesting (as in Russian dolls and Buddhist relic containers excavated from the Famensi) often
tellingly exposes the incremental change in scale.
3
Liang 1933, 153. Also see Steinhardt 1997, 198-199. The particular adjustments made are rarely found in contemporary
Buddhist buildings. One only has to compare it with its twin, the Cishige 慈氏閣 across the central avenue of the
Longxingsi, a pavilion which appeared exactly the same as the Zhuanglunzangdian from the outside but comes up with a
different interior structure.
54
22) by (15-18) centimeters.
4
This means that the smaller cai is roughly one-fifth of the larger cai.
Since the sutra case is made up of columns, beams, brackets, rafters, and eaves like all Chinese
wooden architecture, carpenters only had to produce miniature versions of these structural
components to create the desired outcome.
5

Dating the miniature: textual evidence
Though the nesting relationship suggests that the Longxingsi sutra case could be as old as the library
hall, its actual dating has been a highly debated issue. The earliest surviving stone stele in the
monastic precinct, the famous Longcangsi bei 龍藏寺碑 (Stele of the Monastery of the Hidden
Dragon), dates the foundation of the monastery to 586, the sixth year of the Kaihuang 開皇 Period
of Sui. The second earliest stele dates from 971 (the fourth year of the Kaibao 開寶 Period), when
the first Northern Song emperor, Zhao Kuangyin 趙匡胤 (r. 960-976), visited the monastery during
the war and ordered the colossal statue of Bodhisattva Guanyin to be recast and sheltered in a newly
built structure, the Dabeige 大悲閣 (Pavilion of Great Compassion).
6
The multistory Dabeige has
since become the dominant building of the monastery, which continued to flourish under imperial

4
Guo Daiheng 郭黛姮, ed., Song, Liao, Jin, Xi Xia jianzhu 宋, 遼, 金, 西夏建築 (Song, Liao, Jin, and Xi Xia
architecture), vol. 3 of Zhongguo gudai jianzhushi 中國古代建築史 (History of ancient Chinese architecture) (Beijing:
Zhongguo jianzhu gongye chubanshe), 382. This is roughly equal to the fourth-grade cai in the Yingzao fashi. Pan Guxi,
however, states that the pavilion uses a cai of 24 x 16.5 cm, which he equals to 7.5 x 5.16 chi, the third-grade cai in the
Yingzao fashi. See Pan and He 2005, 46. It is likely that the cross section of the timber material varies in dimension after
years of weathering, alteration, and rebuilding, and the various given sizes (21 x 15 cm and 24 x 16.5 cm) only show
rough average values.
5
Yingzao fashi, vol. 1, 206. “The curving of the roof and the proportioning of the brackets and other members all comply
with the design principles of the large-scale woodworking, while their sizes are decreased according to the cai to be used.
This is also true to the rounding, beveling, and segmentation of columns and flying rafters” ( 其屋蓋舉折及枓栱等分
數, 並準大木作制度, 隨材減之. 卷殺瓣柱及飛子亦如之).
6
A brief history of the monastery is recounted fairly comprehensively in Zhang et al. 2000, 1-2, 323-331.
55
auspices during the Yuan, the Ming, and the Qing dynasties. The Northern Song also seems to be
the earliest possible time any wooden structures on the premises could be traced back to; and
because of centuries of expansion, dilapidation, alteration, and repair (the most recent large-scale
restoration took place in 1997-1999), not a single building can be said to have remained a one-
hundred-percent, authentic Northern Song structure. Nonetheless, the overall layout of the
monastery and the predominant structural features of its architecture are generally believed to have
retained the Northern Song style (fig. 17).
7

When Liang Sicheng 梁思成 (1901-1972) made his field trip to the Longxingsi in 1933, he
proposed that the cruciform Monidian 摩尼殿 (Hall of the Mani Jewels)--then the most well
preserved building in the monastery--was an eleventh-century remain, based on his comparison
between this wooden structure and the prescriptions in the Yingzao fashi.
8
His judgment turned out
correct only posthumously in the 1977-1980 restoration of the hall, when inscriptions of “Da Song
huangyou sinian 大宋皇祐四年 (fourth year of the Huangyou Period of the Great Song, equivalent to
1052)” or simply “huangyou sinian” were found on the surfaces of multiple wooden components, thus
confirming that the hall was indeed a Northern Song original.
9
This discovery has led scholars to
date several other buildings--including the library hall--to the eleventh century, since they have
displayed a great consistency in form, style, and scaling scheme with the Monidian.

7
This conclusion is based on the scale of the timber unit and the woodworking techniques shown from the structures,
which have been discussed by many architectural historians to have generally followed the principles laid out in the
Yingzao fashi.
8
Liang 1933, 19-20.
9
Also found are inscriptions of “Ming Chenghua ershier nian” ( 明成化二十二年, 1486) and “Qing Daoguang ershisi
nian” ( 清道光二十四年, 1844), when major repairs or restoration works were carried out.
56
The dating of the sutra case, on the other hand, proves to be a lot more difficult. While
Sekino Tadashi 關野貞 (1868-1935) asserts that this is a Qing woodwork, Liang, however, considers
it to be a masterpiece of Northern Song carpentry, again based on his knowledge of the Yingzao
fashi.
10
Alexander Soper agrees with Liang and dates the sutra case to the eleventh century, a date
generally accepted by architectural historians today.
11
This is despite the fact that the earliest
inscription found on the sutra case indicates a year of 1365.
12
There have been other voices of
disagreement, too. A colleague of Liang, Liu Dunzhen 劉敦楨 (1897-1968), himself a highly
esteemed scholar of Chinese architecture, believes that “even though the wooden frame of the
library hall adopts the Song style, the sutra case seems to have come from the Yuan or the Ming…
In any event, the sutra case must have undergone considerable modifications during the late-Yuan
and early-Ming period.”
13

10
Liang judges the sutra case to be a Northern Song woodwork based on three distinctive structural features: 1) the use
of “genuine” ang 昂; 2) the concave contours (ao 䫜) of the lower part of the bearing blocks (qi 欹); 3) the use of
liaoyanfang 橑檐方, a roof purlin rectangular in cross section, which all comply with the Yingzao fashi regulations for large-
scale woodworking. See Liang 1933, 154-55. Surprisingly, Liang did not mention the zhuanlun jingzang in the Yingzao fashi,
which should be an obvious template for this woodwork. He seems to be mainly comparing the sutra case with full-scale
structures, which may potentially undermine his argument. It is understandable because the study on the Yingzao fashi at
the time merely just covered the damuzuo part and had not yet stepped into the xiaomuzuo. See Liang 1933, 23-24; Tokiwa
Daijo 常盤大定 and Sekino Tadashi 關野貞, Shina bunka shiseki 支那文化史蹟 (Historical remains of Chinese culture),
vol. 8 (Tokyo: Hozokan, 1940), 90-91.
11
Sickman and Soper 1984, 433. See also Qi Yingtao 祁英濤, “Monidian xinfaxian tiji de yanjiu 摩尼殿新發現題記的
研究 (A study on the newly discovered inscriptions in the Monidian),” in Qi Yingtao gujian lunwenji 祁英濤古建論文集
(Collected essays of Qi Yingtao on ancient architecture) (Beijing: Huaxia chubanshe, 1992), 106-13.
12
The twenty-fifth year of the Zhizheng 至正 Period of Yuan. This is briefly mentioned in Zhang et al 2000, 17. See
also Liu Youheng 劉友恒, Fan Ruiping 樊瑞平, and Du Ping 杜平, “Jin 50 nian Zhengding gujianzhu weixiu zhong
faxian de wenzi tiji chubu yanjiu 近 50 年正定古建築維修中發現的文字題記初步研究 (A preliminary study on the
textual inscriptions discovered during the recent fifty years of restoration of the ancient architecture in Zhengding),”
Wenwu chunqiu (2006.1): 44, which indicates that the inscriptions are found written on the hanging posts and beams and
were left by tourists in the Yuan.
13
“ 藏殿架構隨系宋式, 但轉輪藏則似元, 明間物… 此轉輪藏殆元末明初大經改作, 無可置疑.” Liu Dunzhen 劉
敦楨, “Hebei gujianzu diaocha biji 河北古建築調查筆記 (Survey notes on the ancient architecture in Hebei),” in Liu
Dunzhen quanji 劉敦楨全集 (Complete works of Liu Dunzhen), vol. 3 (Beijing: Zhongguo jianzhu gongye chubanshe,
1987), 17-18. Also see Zhang et al. 2000, 22-26, for a list of all steles on site.
57
Liu’s rebuttal is based on the inscription of a stele dated to 1259, titled “Dachao guoshi
Namodashi chongxiu Zhendingfu Da Longxingsi gongdeji bei 大朝國師南無大士重修真定府大
龍興寺功德記碑 (Stele recording the merit of the Namodashi, State Master of the Yuan, who
restored the Grand Longxingsi in Zhending Prefecture).” It mentions that in the year of yimao 乙卯
(1255), the Namodashi “ordered the jingzang to be repaired [or (re)built]” ( 隨令補修經藏), and that
after the restoration work, “the monastery was complete with a Buddhist zang, a sutra hall, and the
monks’ living quarters” ( 寺有佛藏有經堂有僧).
14
It is unclear if the jingzang mentioned in the
inscription is the one in situ today, but a certain form of repository must have stood in the
monastery by 1255.
There are two more steles with probable references to the sutra case. One is the “Shecai shi
yongyedi zhuan Dazangjing gongdeji bei 捨財施永業地轉大藏經功德記碑 (Stele recording the
merit of donating money and land for the turning of the Tripitaka),” erected in 1314, now in the
library hall. The phrasing of the title, “turning of the Tripitaka,” seems to refer to the ritual of
turning a sutra case where the Tripitaka was stored.
15
The character zhuan 轉 also appears in the
second stele, “Shengzhu benming changsheng zhuyan bei 聖主本命長生祝延碑 (Stele wishing for

14
Changshan zhenshizhi 常山貞石志, 15.21a, http://ctext.org/library.pl?if=gb&file=31654&page=42. The year 1259 is
the ninth year of Yuan Xianzong 憲宗 (Mongke 蒙哥). The term buxiu 補修 is frustratingly ambiguous; it could mean
“to fix and repair (in this sense it is the same as xiubu 修補)” or “to build so as to make whole.” My first impression
points to the first alternative. The inscription also records, “Since the time of war and commotion, [the monastery] had
become dilapidated while only the pavilion [of the Bodhisattva Guanyin] was intact 兵塵以來, 破落如是, 獨有此閣如
故,” implying that many buildings in the monastery had been damaged by then. About Namodashi 南無大士, see Ma
Xiaolin 馬曉林, “Dachao guoshi Namodashi chongxiu Zhendingfu Da Longxingsi gongde ji zhaji: jianlun Make Boluo
xingji de xiangguan lunshu 大朝國師南無大士重修真定府大龍興寺功德記劄記: 兼論馬可波羅行記的相關論述
(Notes on the record of the merit of the Namodashi, State Master of the Yuan, who restored the Grand Longxingsi in
Zhending Prefecture: with comments on pertinent contents in Marco Polo's travel logs),” Guoji hanxue yanjiu tongxun
(2012.6): 252-57.
15
This inscription has been much defaced and become largely illegible.
58
the longevity of the Sage Lord, d. 1317),” which records that “the Sutra for Humane Kings, in fifty
volumes and a hundred chapters, and the Sutra of the Medicine Buddha, in fifty chapters, were printed
and bestowed to this monastery, where rituals of reading, turning, reciting, and chanting these sutras
for the Sage Lord [Yuan Renzong (r. 1311-1320)] had to be performed under imperial decree” ( 印造
仁王護國般若波羅蜜經五十部計百卷, 藥師如來本願功德經五十卷, 施本寺, 欽為聖主本命
看轉誦讀).
16

According to these texts, it is safe to say that some sutra repository was built in the
Longxingsi no later than 1255, and that the ritual of “turning the Tripitaka” has been practiced on
site since at least the early fourteenth century. Can we wager for an even earlier date? In fact, since
the library hall can be fairly positively dated to the eleventh century for identifiable structural
characteristics, a certain revolving sutra case must have existed along with it considering the
exclusiveness of both the nest and the nested. It is highly likely that the two were originally designed
and built as a whole; it would be technically inefficient--if not impossible--to restructure the existing
hall and later install an individually designed sutra case in the interior.
17
This evaluation is further
supported by an in-depth analysis of dimensional and formal qualities.

16
Changshan zhenshizhi, 19.15b, http://ctext.org/library.pl?if=gb&file=31657&page=31. The calligraphy of the
inscription is by the famous calligrapher Zhao Mengfu 赵孟頫. The same inscription implies the ritual function of the
sutra case, which will be elaborated in the next section.
17
For the sutra case to fit in the pavilion perfectly well, its overall structure could not have been altered in any significant
way, though it must have undergone multiple repairs and renovations and may or may not be true to its original design
in every detail. The external form--especially the ornaments and decorative patterns--could have deviated remarkably
from one version to another throughout the maintenance and negligence in history.
59
Dating the miniature: a comparison with Yingzao fashi
The size of each part of the Longxingsi sutra case can be extracted from the 3D Rhino model I have
developed (fig. 18).
18
As mentioned earlier, the cai is 4.5 by 3 centimeters; this is actually larger than
the recommended value, 3.2 by 2.1 centimeters (1 by 0.66 cun), for the zhuanlun jingzang in the Yingzao
fashi.
19
The octagonal case measures 2.6 meters each side and 6.9 meters in diameter, also larger than
the values of 2.1 meters (66.6 cun) and 5.1 meters (160 cun) prescribed in the official standards.
20
This
“deviation” in size, in fact, is commonplace for surviving wooden structures from the same
historical period, and it has been usually interpreted as a result of regional building practices and
customs and the carpenter’s predilections in each individual project.
21

In spite of the enlargement in size, the sutra case displays certain similarities with the Yingzao
fashi template in terms of structure and form. As shown in the modern reconstructions of the
template (figs. 19, 20), the sutra case is supposed to consist of a rotating core (neicao 內槽) and an
immobile, pavilion-shaped outer structure (waicao 外槽).
22
From bottom to top, it should include a
dais (zhangzuo 帳坐), an octagonal case (zhangshen 帳身), a layer of skirting roofs (yaoyan 腰檐), a

18
This models is viewable at https://chinesearchitecture.wordpress.com/2016/03/05/longxingsi, together with a
photogrammetry model.
19
Yingzao fashi, vol. 2, 5, 11. This is smaller than the cai of the fodaozhang (1.8 by 1.2 cun).
20
Ibid., 1. Regarding the overall size of the jingzang, the Yingzao fashi regulates: “The principles of making revolving sutra
cases: the total height is 200 cun and the diameter 160 cun. Make in an octagonal prism with each side measuring 66.6 cun
wide” ( 造經藏之制: 共高二丈, 徑一丈六尺. 八棱, 每棱面廣六尺六寸六分).
21
Such a phenomenon has been noted by many architectural historians. For the most comprehensive survey, see Guo
2009.
22
Takeshima and Pan have each made their reconstruction of the structure of jingzang. The two proposals are somewhat
different. The most obvious discrepancy is Pan’s reconstruction of the pivoting mechanism, which comprises of a series
of diagonal braces spoking from the axle. Since the Yingzao fashi never indicates that the braces should be so placed,
which is only one of many possibilities, it is enticing to think that Pan could have closely studied the Longxingsi sutra
case (where diagonal braces are exposed) and determined that its inner structure could have reflected the Northern Song
standard. This is a reticent proclamation that the design of the Longxingsi rotating core is a Northern Song original.
60
layer of roof-top substructure (pingzuo 平坐), and the Heavenly Palace (tiangong louge) as
ornamentation of the crown (zhangtou 帳頭).
23
These components are clearly identifiable from the
Longxingsi sutra case, except that the latter has chosen not to include the Heavenly Palace, and has
combined the inner core and the outer structure into a single rotating entity.
In its current state of dilapidation, the sutra case is missing many elements: its interior has
been hollowed out, exposing the wooden pivot and the web of braces originally hidden behind the
shelves and drawers (fig. 21). Any scriptures once stored in the case, too, are completely gone.
24
In
one of Sekino’s photographs taken in the 1920s, a Buddhist icon is shown placed inside the sutra
case, directly facing the viewer (fig. 22).
25
Also shown are the ornately carved coiling dragons on the
columns, round mirrors hung on each side of the octagon, and two rings of miniature balustrades,
one on top and the other surrounding the bottom of the dais.
26
Most of these features had already
disappeared when Liang visited the Longxingsi in 1933 (fig. 23).
The bracket sets beneath the double-layered eaves, on the other hand, have been carefully
restored, even though their original paints are lost. Because of the small scale of the brackets, they
easily recall the type of bracketing used in Qing architecture--slender, tightly spaced, and
overwhelmingly decorative--features that have possibly led Sekino to attribute this the woodwork to

23
Yingzao fashi, vol. 2, 1-16.
24
The Yingzao fashi mentions that coffers (jingxia 經匣) should be used for the storage of scriptures inside the jingzang.
See Yingzao fashi, vol. 2, 15.
25
According to Sekino, there were in fact four bodhisattvas, one sitting at each cardinal direction of the sutra case. It is
not clear if scriptures were then kept behind the statues or totally lacking. See Tokiwa and Sekino 1940, 90-91.
26
Today, the balustrades have all gone, leaving only irregularly interspersed, rectangular holes on the stone-paved ground
along the perimeter of the sutra case. The restoration team obviously has not completed their work. During fieldwork, I
found stacks of dust-covered timber behind the sutra case in a restricted area of the library hall. Many of them appeared
to be broken or heavily weathered materials taken directly off from the sutra case, as they had mortises and tenons on
them; some arched ones appeared to fit with the curvature of the round sutra case. I took the risk of pushing the sutra
case to make it rotate, but it wouldn’t budge. The bottom of the pivot looked as though it was displaced slightly from
the center of the pit.
61
Qing carpenters.
27
Applying the scheme of “eight-tiered double-twig triple-arm counted double
bracket (bapuzuo shuangmiao sanxia’ang jixin chonggong 八鋪作雙杪三下昂計心重栱),” the brackets
appear grander than the six-tiered brackets prescribed in the Yingzao fashi. A greater number of tiers
usually signals a higher-rank building since it is more visually appealing and demands more time and
resources. In fact, according to the Yingzao fashi, eight is the maximum number of tiers a bracket set
may have; material evidence of such brackets has not been found elsewhere, making the Longxingsi
sutra case a singular example.
28

The complexity of the eight-tiered bracketing is especially well exposed at the eight corner
sets (zhuanjiao puzuo 轉角鋪作) (fig. 24). Their distinctive form is a result of their specific position at
the vertices of the octagon, where two adjacent sides intersect at 135 degrees. In this situation, all
tiers of bracket arms of the same set have to be “tripled” in essence: placed on top of the same
column, they grow into a cluster of three interconnected subsets--one parallel to the left side, one to
the right side, and one jutting out in the middle, with an angle of 22.5 degrees between each subset
(fig. 25). This method of tripling the corner set is in accordance with the instructions found in the
Yingzao fashi.
29

The column-top bracket sets (zhutou puzuo 柱頭鋪作) and intercolumnar bracket sets (bujian
puzuo 補間鋪作), on the other hand, are relatively simpler (fig. 26). A total of eight sets--including
two corner, two column-top, and four intercolumnar sets--have been evenly spaced for each side of
the octagon, spanning over one central bay and two end bays. The spacing and the arrangement, yet

27
Tokiwa and Sekino, 90-91; Sekino asserted that it should date between 1643 and 1661.
28
Guo 2009, 384. A case of nine-tiered bracketing is found in the Chongfusi; see Chapter 4 of this dissertation.
29
Yingzao fashi, vol.1, 79-80. It does differ considerably from the regular method of assembling corner sets (where
members are intersected orthogonally), but the angle formed by the octagonal shape of the sutra case has led to a slightly
different variation.
62
again, shows a striking conformity to the rules explained in the Yingzao fashi.
30
These brackets are
placed tightly to each other, leaving no space to breathe; whereas all existing Northern Song
buildings--such as the library hall--have fairly sparsely-spaced brackets (fig. 27). Such is the privilege
of miniature architecture: regulations for full-scale buildings concerning structural stability,
efficiency, and social rank often become invalid for miniatures.
31
No wonder why such a small-scale
woodwork could have had the most extravagant form of bracket sets in Chinese architectural
history, as its creators were free to exhaust their ingenuity and attempt the most complex design.
The Buddhist societies, too, would have loved to patronize the most attractive and “authoritative”
structure for the sheltering of sacred images and scriptures.

“Progressive miniaturization” in Chinese architectural history
The miniaturization of the Longxingsi sutra case was not an isolated phenomenon. It signaled a
trend of change, a critical historical moment when the ultimate source of architectural
impressiveness continued to transition from sheer mass to accumulated intricacy and redundancy.
As small-scale woodworking became standardized by the end of the eleventh century, the regular
components of large buildings also started to undergo a series of miniaturization. This is most
excellently exposed by one of Liang Sicheng’s hand-drawn diagrams (fig. 28), in which he reveals the
changing proportion between the bracket set and the column from the Tang to the Qing dynasties.

30
The “bay” is here the space between the columns and/or the hanging posts. According to the Yingzao fashi, the
number of intercolumnar sets should not exceed one per bay, at most two for the central bay. However, for the jingzang,
the same text also stipulates five intercolumnar sets to be used for the skirting eaves of the outer ring, the dais, and the
crown of the inner ring, and nine intercolumnar sets for the substructure of the Heavenly Palace. This would mean a
total number of seven or ten bracket-sets per side. Bracket spacing will be explained in Chapter 3 of this dissertation.
31
See Chapter 1 for the discussion on the Tianshengling. The contrast is particularly striking when comparing the sutra
case with the library hall. The latter, restricted by its rank, simply applies the “five-tiered counted single bracketing” for
the lower level and follows the regular spacing of intercolumnar sets--two in the central bay and one in all other bays.
63
Not only did the bracket set become increasingly miniaturized in comparison with the column, but
the size of the cai (manifested as the cross section of the bracket arm) also gradually decreased with
the passage of time.
32
This is evidenced by the Tianwangdian 天王殿 (Hall of the Heavenly Kings)
of the Longxingsi, where much smaller bracket sets have been added during a Qing repair to bolster
the original Northern Song structure, providing a salient contrast between the robust, medieval
brackets and their overtly decorative, diminutive descendants (fig. 29).
33

Accompanying this miniaturization of architectural components was the “degeneration” of
certain structural members. This is most clearly detected from the ang 昂, the slanting arms in the
bracketing system. Prior to the eleventh century, the ang had been used for strictly structural
purposes. Functioning like a lever, its outer end ought to follow the downward slope of the roof and
be suspended under the eaves, whereas the inner end should go all the way up to the roof frame to
provide extra support to the beams and purlins. In the Longxingsi sutra case, however, while the ang
all appear to be genuine slanting members from the outside, they have in fact lost their structural
integrity because the inner ends have been cut short and attached to a partitioning board. In other
words, they are functionally corrupt even though having partially preserved the original form. The
same phenomenon is also observed in some eleventh-century buildings as well as in the Yingzao
fashi.
34

32
Zhongguo gudai jianzhu jishushi 中國古代建築技術史 (History of the technology of ancient Chinese architecture)
(Beijing: Kexue chubanshe, 1985), 100-01.
33
Liang 1933, 28.
34
The famous Shengmudian 聖母殿 (Hall of the Sacred Mother, d. 1023-1032) at the Jinci 晉祠 in Taiyuan, Shanxi, for
instance, uses “fake” ang. In the Yingzao fashi, the ang of miniature shrines and repositories are essentially non-structural.
As far as large-scale woodworks are concerned, there is also a type of functionally corrupt ang termed cha’ang 插昂
(inserted lever arm), which coexists with functional ones like the basic ang and the tiaowo 挑斡 (cantilever). What Liang
regards as the “ancient-style genuine ang” in his 1933 report is in comparison with the “fake,” horizontally placed lever
arms that started to grow in fashion in the late eleventh century.
64
Coming to the Qing, even the bracket sets started to lose their original role and become
largely ornamental components. The implication is that carpenters could now make much smaller
brackets, since their size would not affect the validity of the structural frame. Indubitably, the
experience accumulated over centuries of miniature woodworking must have prepared for this
transformation, a process during which small and densely spaced brackets turned out to be one of
the most prominent features of Chinese architecture.
35

The Revolving Sutra Case in History
According to textural evidence, revolving bookcases came to be used in China as early as the sixth
century. Luther Carrington-Goodrich’s article, “The Revolving Book-case in China,” is perhaps the
first comprehensive study on this subject.
36
This article is primarily a literary survey; it explores
various forms of historical records encompassing stone inscriptions, gazetteers, building standards
(i.e., the Yingzao fashi), monks’ travel logs, and miscellanea from the Tang to the Qing dynasties.
Altogether, Carrington-Goodrich identifies some twenty-six sutra cases from the early tenth to the
mid-seventeenth century, but his major interest lies not in the specific architectural forms but rather
the curious revolving mechanism and its provenance. He argues that the notion of the revolving
bookcase had never really existed in either Confucian or Daoist tradition and should be mainly
considered a Buddhist inspiration, one that was perhaps derived from the “prayer cylinder” or
“prayer-wheel” used in Tibetan Buddhism.
37
Meanwhile, it is likely that the techniques of making

35
Generally speaking, the preference for the smaller scale might have been a choice based on economic and/or aesthetic
reasons. Economically, there was an increasing lack of large timber as time passed. Aesthetically, the slender and
intricately crafted miniatures could have intrigued later rulers include the Mongols, the Manchus, and the Chinese
themselves, who turned to seek a new architectural fashion and statement to suit their styles and needs.
36
Luther Carrington-Goodrich, “The Revolving Book-case in China,” Harvard Journal of Asiatic Studies 7 (1942.2): 130-61.
37
Ibid., 152-55.
65
revolving devices had come from local craftsmanship which gave birth to the chariot, the potter’s
wheel, the watermill, the wheelbarrow, the “taxicab (jili guche 記里鼓車),” and the like.
38
The reason
to make the bookcase turnable, he believes, was to assist translators and copyists of Buddhist sutras
to quickly locate books as well as to enact the ritual of circumambulation.
39

A more recent study is Helen Loveday’s “La Bibliotheque Tournante en Chine: Quelques
Remarques sur son Role et son Evolution.”
40
Based on Carrington-Goodrich’s work and other
previous studies, Loveday’s article has made several profound observations. She attributes the
popularity of revolving sutra cases in Tang and Song times to the culture of worshipping dharma
relics--i.e. Buddhist scriptures--and to the turning ability of the sutra case which enabled the ritual of
cakrapravartana (turning the wheel of law) to be performed.
41
She also comments extensively on
structure and iconography, explaining that the form of the sutra case displayed a great resemblance
to certain funerary and religious monuments, especially funerary stupas (muta 墓塔) and dharani
pillars (tuoluoni jingchuang 陀羅尼經幢), whereas the resemblance in form was further enhanced by
shared pictorial motifs symbolizing the Sumeru and other elements of the imagined cosmos.
42

While their works greatly inform the following discussion, what I aim to present below
include not only available textual sources but also material evidence of the sutra cases in history,
from the earliest time to the present. I will concentrate on the religious functions of the sutra cases
across time, the rituals they entailed in different settings, and their significances in the eyes of

38
Ibid., 156.
39
Ibid., 157-58.
40
Helen Loveday, “La Bibliotheque Tournante en Chine: Quelques Remarques sur son Role et son Evolution,” T'oung
Pao 86 (2000.4/5): 225-79.
41
Ibid., 239-46.
42
Ibid., 225-26, 246-58.
66
different audiences. This will contextualize the Longxingsi sutra case in the historical development
of this special device, leading to a better understanding of how it might have been used. Meanwhile,
it will give us a hint of how and why miniature-making came to be involved during the process.

Sixth century: legendary beginnings
Multiple textual sources have traced the invention of the revolving sutra case to the Liang dynasty
(502-557); some have traced more specifically to a historical figure, the monk Fu Xi 傅翕 (497-569),
who is believed to have invented this device. One exemplary text, Shanhui dashi lu 善慧大士錄
(Record of the Grand Master Shanhui), prefaced by a Tang scholar Lou Ying 樓穎 (fl. 744-?), gives
the following account:
When [Fu Xi] was at [the Shuanglinsi monastery], he often felt that the scriptures were too numerous
for people to read them all [in a lifetime]. Thus he built a large multilevel shrine (kan) by the
mountain, which consisted of a single column [as the pivot] and eight sides, and it was filled with
various scriptures. The shrine revolved without any hindrance and was called lunzang. [Fu Xi] then
made a vow: “Those who come to the gate of my zang shall never lose their human form in their lives
and afterlives [throughout the transmigration of souls].” He preached to the common folks: “Those
who seek enlightenment, endeavors sincerely and exhaust themselves in this pursuit will be able to
turn the wheel repository. Regardless of how many turns one might make, this person would achieve
the same amount of merit as those holding and reading the scriptures. All is contingent on one’s will
and mind. Everyone can be benefitted.” The lunzang built nowadays all feature the image of [Fu Xi];
and this is how it started in the beginning.
大士在日, 常以經目繁多, 人或不能遍閱, 乃就山中建大層龕一柱八面, 實以諸經, 運行不礙,
謂之輪藏. 仍有願言: “ 登吾藏門者, 生生世世不失人身. 從勸世人, 有發菩提心者, 志誠竭力,
能推輪藏. 不計轉數, 是人即與持誦諸經功德無異. 隨其願心, 皆獲饒益. 今天下所建輪藏皆
設大士像, 實始於此.
43

Carrington-Goodrich reminds us that the two earliest biographies of Fu Xi, one in the Xu
Gaosengzhuan 續高僧傳 (Extended biographies of eminent monks) by Daoxuan 道宣 (596-667) and

43
Xinzuan xu zangjing 新纂續藏經, X69.1335: 109c http://www.cbeta.org/result/normal/X69/1335_001.htm. This text
seems to be largely the same as in the Shishi jigu lue 釋氏稽古略 (ca. 1354?) 2.35a, which is quoted in Carrington-
Goodrich 1942, 132.
67
the other in the Jingde chuandelu 景德傳燈錄 (Transmission of the lamp) by Daoyuan 道原 (fl. 1004),
do not mention such an invention, which is why he cautiously considers Fu Xi’s revolving sutra case
a “legend.”
44
From the Northern Song onward, however, the account of Fu Xi’s legendary invention
became readily accepted and incorporated into both official and anecdotal discourses. In the
comprehensive Buddhist history Fozu tongji 佛祖統紀 (Complete chronicles of Buddhas and
patriarchs, 1260-1264?) compiled by the monk Zhipan 志磐 under the Southern Song, the revolving
sutra case is explained to have been first installed in the dharma field of the Shuanglinsi 雙林寺 by
Fu Xi, out of his “compassion for the laity who either did not have enough time for reading the
sutras or were simply illiterate” ( 愍世人多故不暇誦經及不識字).
45
The text contains some
overlaps with the excerpt from the Shanhui dashi lu quoted above; similarly, it records the vow of Fu
Xi that “one who can faithfully turn the sutra case in one complete circle receives the same merit as
reciting the sutras; one who can rotate the sutra case in countless turns receives the same merit as
reading and reciting the complete Tripitaka” ( 有能信心推之一匝。則與誦經其功正等。有能旋
轉不計數者。所獲功德即與讀誦一大藏經正等無異).
46

Legendary or not, it seems likely that in the beginning, the sutra case was invented as a
turnable device to enable the laity, especially those who were illiterate, to have an equal opportunity
to accumulate merit and achieve enlightenment. The sources do not detail the form of the lunzang,

44
Carrington-Goodrich 1942, 132-133.
45
T49.2035: 318c. The same entry mentions having the images of Fu Xi and the eight divine guardians by the side of the
sutra case.
46
Ibid. Additional textual evidence for Fu Xi’s invention comes from Qisong’s 契嵩 (1007-1072) Xinjin wenji 鐔津文集
14.7b-8b, Ye Mengde’s 葉夢得 (1077-1148) Jiankangji 建康集, 4.6ab, and the Shishi jigu lue. In more recent scholarly
works these have been incorporated into Nanjo Bunyu 南條文雄, A Catalogue of the Chinese Translation of the Buddhist
Tripitaka: the Sacred Canon of the Buddhists in China and Japan (Oxford Clarendon Press, 1883),
http://catalog.hathitrust.org/Record/010249920. Also see Carrington-Goodrich 1942, 131-33.
68
only that it was a single-pivot, eight-sided “shrine” (kan 龕), which was geometrically the same as the
Longxingsi sutra case. The employ of the term kan suggests that the sutra case might have been
modelled after some Buddhist halls or pavilions and was perhaps small in size like a niche. Clearly,
what the texts focus on is the function of the revolving device as a receptacle of the Buddhist
Tripitaka and a miraculous tool of salvation.

Tang
The seemingly humble beginning and charitable nature of the lunzang have been somewhat rewritten
soon afterward, when the ideal of offering equal accessibility was largely downplayed in certain cases
by an emerging emphasis on the exquisite and lifelike architecture. This is evidenced by a passage
from Daoxuan’s Zhong Tianzhu Sheweiguo Zhiyuansi tujing 中天竺舍衛國祇園寺圖經 (Illustrated
scripture of Jetavana vihara of Sravasti in Central India), a text providing a vivid picture of the ideal
monastery (fig. 30). It speaks of a certain lianhuazang 蓮花藏 (lotus repository; Sk. padmagarbha)
installed inside the Hall of the Grand Buddha (Dafodian 大佛殿):
[The repository] is one zhang three chi tall. Its form resembles a mingtang [Hall of Light]. Beneath the
dais are nine coiling dragons as the support; the dragon heads were made of purple-sheen polished
gold. Up above is a seven-jeweled lotus, in which there is a standing statue of the great divine
heavenly general Manibhadra. The statue is made of gold with silver engravings; it wears a seven-
jeweled necklace and a seven-jeweled cap. On top of the lotus is a seven-layered, silver dais. The dais
has eight sides; on each side there is a window above and a door with golden door leaves below. The
doors are locked with golden, lion-shaped locks; they open on their own but not by people. [The
repository] uses red crystal roof tiles. The tile-ends are all decorated with golden lions with golden
and silver bells in their mouths. The finial [of the repository] is similar to those of contemporary
pagodas; however, it is eight-sided, and on each side there is a golden chain where golden lions with
bells in their mouths are hung. The tongues of the lions are all made of eight-sided, rooster-shaped
king jewels. The jewel on top of the finial takes the form of the Garuda. On the back of the Garuda
is the Bodhisattva Puxian riding a white elephant and sheltered by a baldachin (the part from the
Garuda to the baldachin is made of a single jewel). Such is how the lotus repository looks.
高一丈三尺, 狀若此間明堂形. 臺下九龍盤結為腳, 紫磨金作龍頭. 上有七寶蓮花, 花中有摩
尼跋陀大神將立身, 用黃金作之, 白銀彫鏤, 項以七寶頭戴七寶. 蓮花之上以白銀為七層臺,
臺有八楞, 八面有窗, 窗下有門以金為扉, 有黃金鎖形如師子自然開, 開不以人功. 以紅頗梨
為瓦, 瓦頭皆有金師子, 師子口中皆銜金銀鈴. 臺上相輪如今塔上者, 然有八角, 角別金鎖具
69
之. 鎖上懸金師子如上銜鈴, 皆八楞珠王為舌, 珠王如雞. 相輪上珠如金翅鳥, 鳥上普賢菩薩
乘白象王, 覆以寶蓋( 從金翅鳥以上至蓋一珠所作). 上敘蓮花藏相.
47

It is not clear if the lotus repository is turnable, but like Fu Xi’s sutra case, it is similarly an
eight-sided, architectural-shaped receptacle. Specifically, Daoxuan has identified the seven-layered
dais (tai 臺), the windows and doors, the roof tiles (wa 瓦), and perhaps most importantly the finial
(xianglun 相輪), which strongly suggest the pagoda (ta 塔) to be the main source of the architectural
shape. Though Daoxuan’s text is allegedly based on the authentic Indian prototype, the author had
in fact never been to India, and scholars believe that his depictions were largely derived from his
personal observations of contemporary Buddhist and imperial buildings.
48
This could mean that by
the time the text was written, the lotus repository had already become a highly developed, popular
device with elaborate forms and distinctive iconography used in Chinese monasteries.
The same text informs us of the important role of the repository in Buddhist rituals. On
each of the six fasting days (liuzhairi 六齋日), the monks and nuns should come to worship the
repository. On such a day, the nine dragons would exhale the smoke of fragrant incense, the great
general Manibhadra would admonish the audience, whereas the golden lions and the bells would all
eulogize the virtue of upholding the Buddhist precepts.
49
Also involving the repository was the ritual
of ordination. As Daoxuan recounts, any monk wishing to be ordained should first come to pray to
the repository; in response, the Garuda and the Puxian would offer their sermon and
encouragement. Having heard their illuminating voices and cleared any remaining doubts in the
mind, the monk would then proceed to the altar of ordination. After the ordination was complete,

47
T45.1899: 887a-b.
48
Puay-peng Ho, “The Ideal Monastery: Daoxuan’s Descriptions of the Central Indian Jetavana Vihara,” East Asian
History 10 (1995): 1-7.
49
T45.1899: 887b.
70
he had to return to the repository; and if he had been ordained in a superior way, the doors of the
repository would open automatically for him, manifesting in front of him hundreds of thousands of
buddhas and a myriad of silver towers and pavilions of the Lotus Repository World (lianhuazang shijie
蓮花藏世界).
50

While Fu Xi’s sutra case was invented for the laity, Daoxuan’s lotus repository was to be
used in a strictly monastic setting and played a key role in the all-important ritual of ordination.
What mattered in the latter case was not the revolving mechanism but the vividly represented
Buddhist icons lavishly adorned with precious materials--gold, silver, crystal, and various other
jewels and gems. It is notable that architecture was an important part of this spectacle: the ordained
would be able to “see” numerous towers and pavilions inside the open repository, a vision perhaps
brought forth through certain forms of miniature architecture such as the tiangong louge.
51

The earliest known revolving sutra case which was actually built, however, came from 809
CE, in the Huayan 華嚴 court of the Daxingtangsi 大興唐寺 monastery outside Chang’an. We
know of its existence from the “Binguogong gongde ming 邠國公功德銘” inscribed on a stone
stele dated 823, which extols the deeds of Liang Shouqian 梁守謙 (779-827), a devout Buddhist and
powerful eunuch in the imperial court, who patronized the installation of this very sutra case and the
5,327 juan of scriptures it held.
52
According to the inscription,
[Liang Shouqian] has built a zhuanlun jingzang inside the [sutra] hall, where stone has been carved into
clouds and the ground dug open, from which [the repository] emerges. It is square in shape…
Countless floral dharani-pillars have been erected, making [the structure] comparable to the Tushita
Heaven; thousands of towers and pavilions have been built as if in the Mirage City. Resembling and

50
Ibid.
51
As will be further elaborated in Chapters 3 and 4, miniature architecture played an important role in the visualization
of the marvelous Buddhist world.
52
The date 809 is determined by Carrington-Goodrich by cross-referencing two gazetteers. See Carrington-Goodrich
1942, 133-34; also see Zhang Yong 張勇, Fu Dashi yanjiu 傅大士研究 (A research on Fu Dashi) (Dharma Drum
Publishing Corp, 1999), 434-35.
71
modeled after real objects, it is lifelike in appearance but transcendent in spirit. The phoenixes and
swans seem to be flying away, while the dragons appear to be sneaking out. Curved brackets and
bearing blocks, stacked one upon another, are intermingled with jewels and gems. Roof rafters, neatly
arranged in lines, are embellished with pearls and jades. [The repository] soars up in five stories and
has four doors opening to the four sides; the gems shine upon each other, their luster is reflected
back and forth… On the facade are drawn [carved?] portraits of the divine kings and various
monsters as extravagant ornaments. There are also images of Bodhisattvas and celestial beings
encircling the structure.
又於堂內造轉輪經藏一所, 刻石為雲, 鑿地而出, 方生結構, 遞[ ][ ] 緣. 立無數花幢, 竊比兜率;
造百千樓閣, 同彼化城. 狀物類本, 擬容奪真, 鵷鵠若飛而不飛, 虹螭似走而不走. 欒櫨櫛比,
雜之以琳琅. 榱桷駢羅, 飾之以珠翠. 淩雲五級, 方開四門, 璀錯相輝, 煥麗交映... 其外或圖寫
龍神鬼物之狀, 以為嚴飾; 或造菩薩天仙之類, 周匝其旁.
53

Clearly, in this case the miniatures have made their way into display: the entire repository was
shaped into to a five-story, tower-like structure adorned with “countless floral dharani-pillars” and
“thousands of towers and pavilions.” The text further alludes to the Tushita heaven (doushuai 兜率)
and the Mirage City (huacheng 化城), implying that these miniatures, indeed, were meant to evoke a
vision of the miraculous realm. The smaller the miniatures, the greater was the visual effect of the
world of the myriad they engendered;
54
but what has been stressed here is not just sheer number but
also verisimilitude. The architecture must have truthfully embodied reality in both form and spirit in
order to be visually persuasive. The miniature buildings, fully equipped with bracket sets and roof
rafters, might have also followed certain scaling rules to successfully achieve a sense of realness.
At least from the perspective of this inscription, the repository was first and foremost a
monument of Liang Shouqian’s personal merit and charitable deeds, which would serve as a
paradigm for other Buddhists to follow suit. It more or less stood as a conspicuous symbol of
wealth and power to awe and dazzle its audience. In this case, one did not have to turn the

53
Quan Tang wen 全唐文 998, http://ctext.org/wiki.pl?if=gb&chapter=158780&remap=gb; Guanzhong jinshiji 關中金石
集, 4.17b, http://ctext.org/library.pl?if=gb&file=27718&page=35.
54
This is inspired by a talk with Kun Yue in 2014, then an EE major doctoral student at University of Southern
California, who asserted that the logic of having to produce and operate on a smaller and smaller scale in today’s most
advanced technology (such as nanotechnology) was to allow more units (and a faster speed) in a limited, confined space.
72
repository to receive merit; but patronizing its installation was considered an even more admirable
act.
In most cases, the financing and building of the repository proved to be a communal effort.
The revolving repository in the Nanchanyuan 南禪院 monastery in the ninth-century Suzhou, for
instance, involved different social groups in the process of its installation, which became quite an
event at the time according to a record by the famous Tang poet Bai Juyi 白居易 (772-846). Bai Juyi
informs us that the project was initiated by none other than himself, then the prefect of Suzhou;
subsequently, two monks from the Shu region and three from the Wu region gathered building
materials, two patrons donated coins, and four monks at the monastery administered the
construction, which lasted from 829 to 836.
55
The total cost amounted to ten thousand strings for
the library hall and three thousand six hundred strings for the repository and the Tripitaka
combined. In the year following its completion, a new abbot, the Chan Master Yuansui, was invited
to the monastery. As a daily ritual, Master Yuansui would come to venerate the thousand Buddhas in
the library hall, open the repository, retrieve the scriptures, and lead the crowd to chant these sutras.
The chanting voice was so powerful that it immediately bestowed blessings to those who heard it,
and moved the hearts of the listeners who soon converted to Buddhism.
56

Bai Juyi’s repository was a nine-storied, eight-sided turnable structure hosting a thousand
Buddhas and 256 sutra coffers containing 5,058 juan of the Tripitaka.
57
To rationalize the installation

55
Bai Juyi 白居易, “Suzhou Nanchanyuan Qianfotang Zhuanlun jingzang shiji 蘇州南禪院千佛堂轉輪經藏石記
(Record of the revolving sutra case in the Qianfotang of the Nanchanyuan in Suzhou),” in Quan Tang wen 676,
http://ctext.org/wiki.pl?if=gb&chapter=398674.
56
Ibid.
57
Ibid. “Inside the hall, there is a baldachin above and a repository below. Between the [baldachin] and the repository
are nine layers of disks and a thousand Buddha niches painted with various colors and embellished with gold and jade.
Around the baldachin are sixty-two hanging mirrors. The repository is octagonal with two doors on each side; it is
painted with vermilion and reinforced with bronze fittings. Surrounding the repository are arranged sixty-four mats.
Inside, the repository is turned by a rotating core and stopped by a wooden block. There are a total of 256 sutra coffers
73
of such a costly device, he has dwelled on the belief that the dissemination of Buddhist teachings
among people, as would be facilitated by this repository, would naturally foster a humane society,
encouraging mutual help, generosity, and benevolence. To him, building the repository proved to be
a worthwhile endeavor, for it would “mark the path to enlightenment, lubricate the wheel of
dharma, entice the old man’s children out of the fire house through expedient means, and dispel
common people’s ignorance about life” ( 以表旌覺路也，脂轄法輪也，示火宅長者子之便門
也，開毛道凡夫生之大竇也).
58
Bai Juyi further stresses that the tripartite system of the sutra, the
repository, and the library hall served as an indispensable vehicle for the preservation of Buddhism.
The installation of the repository, therefore, was tantamount to the meritorious deed of continuing
the Buddhist law.
59

Northern and Southern Song
None of the Tang or earlier revolving sutra cases have survived. The one at the Longxingsi now
stands as the oldest example of this particular type of device. A slightly later example is the feitianzang
飛天藏 (celestial repository) at the Yunyansi 雲岩寺 in Jiangyou 江油, Sichuan (fig. 31). This
revolving repository, dated 1181, shares quite some similarities with the Longxingsi sutra case in
terms of overall form and dimension.
60
It takes the shape of a three-storied, octagonal wooden

and 5,058 scrolls of scriptures 堂之中上蓋下藏。[ 藏] 蓋之間輪九層，佛千龕，彩繪金碧以為飾，環蓋懸鏡六十
有二。藏八面，面二門，丹漆銅鍇以為固，環藏敷座六十有四。藏之內轉以輪，止以尼，經函二百五十有
六，經卷五千五十有八。”
58
Ibid.
59
Ibid. Additional textual sources on pre-Song sutra repositories are collected most comprehensively in Huang Min-chih
黃敏枝, “Zailun Songdai siyuan de zhuanlunzang 再論宋代寺院的轉輪藏 (A further discussion on the revolving sutra
cases in Song-dynasty monasteries),” Qinghua xuebao 26 (1996.2): 139-88; (1996.3): 265-96.
60
For studies on this repository, see Gu Qiyi 辜其一, “Jiangyouxian Ruishan Yunyansi feitianzang ji zangdian kancha
jilue 江油縣圌山雲岩寺飛天藏及藏殿勘察紀略 (A brief record of the survey of the celestial repository and the
74
pagoda with a central pivot anchored in a round pit 7.2 meters in diameter. The octagon is 5.2
meters across and roughly 2.11 meters wide per side. The cai measures 3 by 2 centimeters, smaller
than that of the Longxingsi sutra case but almost the same as the values prescribed in the Yingzao
fashi. The most notable feature of this repository is the three levels of tiangong louge on the upper part
of the pagoda, where the cai is further reduced to 2.3 by 1.3 centimeters. Inside these miniature
towers and pavilions were once held some two hundred wooden figurines of the Daoist pantheon,
such as the gods of the twenty-eight lunar mansions and those of the twenty-four solar terms.
61

Interestingly, no sutra or sutra coffers have ever been found inside the repository: unlike its
Buddhist counterpart, it seems that the feitianzang was designed to enshrine Daoist icons only.
62

Images of the revolving repository have also been captured by twelfth-century stone
sculptures. Two well-known examples come from Dazu 大足 in Sichuan. One is the Beishan 北山
Cave 136 (d. 1142-1146), the so-called “Zhuanlun jingzang ku 轉輪經藏窟 (cave of the wheel-
turning sutra repository),” where a four-meter-tall stone “repository” is sculpted at the center of the
cave (fig. 32).
63
The repository is hollowed inside and cannot be turned; but it has incorporated all

repository hall at Yunyansi in Ruishan, Xiangyou county),” Sichuan wenwu 14 (1986): 9-13; Huang Shilin 黃石林,
“Sichuan Jiangyou Douruishan Yunyansi feitianzang 四川江油窦圌山雲岩寺飛天藏 (The celestial repository at the
Yunyansi in Douruishan, Jiangyou, Sichuan),” Wenwu 文物 (1991.4): 20-33; Zuo Lala 左拉拉, “Yunyansi feitianzang jiqi
zongjiao beijing qianxi 雲岩寺飛天藏及其宗教背景淺析 (A preliminary analysis of the Feitianzang at Yunyansi and its
religious background),” Jianzhushi 21 (2005): 82-92; Guo 2009, 535-48; Shih-shan Susan Huang, Picturing the True Form:
Daoist Visual Culture in Traditional China (Cambridge: Harvard University Press, 2012): 108.
61
Gu 1986.
62
Zuo 2005. The feitianzang is considered a Daoist imitation and appropriation of the Buddhist lunzang. The revolving
ritual, however, is in this case equaled to the veneration of the images.
63
Studies on this particular repository include Hu Liangxue 胡良學, “Dazu Beishan Fowan shike Zhuanlun jingzang ku
zhi guanjian 大足北山佛灣石刻轉輪經藏窟之管見 (A glimpse at the Cave of the Wheel-turning Sutra Repository at
Fowan, Beishan, in Dazu),” Zhonghua wenhua luntan (2001.1): 112-16. A brief description and images are in Guo
Xiangying 郭相穎, Beishan shiku 北山石窟 (Beishan Caves), vol. 1 of Dazu shike diaosu quanji 大足石刻雕塑全集 (A
complete collection of the sculptural art of Dazu) (Chongqing: Chongqing chubanshe, 1999). Another source is Angela
Howard, Summit of Treasures: Buddhist Cave Art of Dazu, China (Trumbull, CT: Weatherhill, 2001).
75
essential formal and iconographical features appropriate for a wooden jingzang: a Sumeru dais, an
octagonal lotus throne (which appears in Daoxuan’s repository), eight columns carved with coiling
dragons, and an octagonal crown adorned with reliefs of single-, double-, and three-storied pavilions
and pagodas, reminding us of the tiangong louge.
The other is the Baodingshan 寶頂山 Cave 14 (d. 1174-1252), known as the “Pilu daochang
毘盧道場 (Dharma field of the Vairocana Buddha)” (fig. 33). The central hexagonal (octagonal?)
pillar has five niches hosting the images of the Vairocana, the Sakyamuni, the Amitabha, and two
six-storied dharani-pillars.
64
Though less conspicuous, the structure of the central pillar is
unmistakably that of a revolving repository. The Sumeru dais, the lotus throne, the columns with
coiling dragons, and the double-story miniature pavilions on top are all strikingly similar to the stone
repository in Beishan Cave 136. The skirting roof with its well-articulated tiles and rafters clearly
indicates that the repository was modeled after a wooden structure.
Further visual evidence for the revolving sutra cases comes from a set of drawings by
Japanese pilgrims to Southern Song Buddhist monasteries. The drawings are dated to 1248, and one
of them shows a particular wooden structure labeled as the “bajiao lunzang 八角輪藏 (octagonal
revolving repository)” found in the Jinshansi 金山寺 monastery in Zhenjiang 鎮江 (fig. 34).
65
The

64
The back of the repository is not hollowed out, leading Howard to the observation that it is five-sided. It could,
however, indicate an eight-sided structure; this is proposed in Hu Wenhe 胡文和, “Dazu Baoding Pilu daochang he
Yuanjue daochang tuxiang neirong yuanliu xin tansuo: poyi liudai zushi chuan miyin midi 大足寶頂毗盧道場和圓覺道
場圖像內容, 源流新探索: 破譯六代祖師傳密印謎底 (A new exploration of the contents and origination of the
images of the Field of Vairocana and Field of Perfect Enlightenment at Baoding in Dazu: deciphering the riddle of the
transmission of the esoteric mudra by the Sixth Patriarch),” Fagu foxue xuebao (2008.2): 247-310. Hu has attributed the
pictorial program of the cave to the representation of the Lotus Repository World. The connection to the Huayan
School and perhaps also esoteric Buddhism has been pointed out by Howard and other scholars. This connection with
Huayan visualization will be further discussed in Chapter 3 of this dissertation.
65
For a study of these drawings, their dating and transmission, see Zhang Shiqing 張十慶, Wushan shicha tu yu Nan Song
jiangnan chansi 五山十剎圖與南宋江南禪寺 (Drawings of the Five Buddhist Mountains and Ten Monasteries and their
connections with Southern Song Buddhist monasteries in the Jiangnan Region) (Nanjing: Southeast University Press,
2000). On p. 62, Zhang further supplies with information on Japanese repositories, which are believed to have been
76
section of the sutra case is excellently exposed: the entire structure appears to be hinged upon a
single pivot fixed in the center of a ground pit, whereas the other end of the pivot is inserted into
the mezzanine or roof frame. The rotating core consists of many diagonal braces, providing support
for the outer components. Like the examples mentioned above, the outer structure has a dais at the
bottom, the sutra cases (represented as lattices in the drawing) in the middle, and a crown at the top.
Other formal features--including the balustrade, the columns with carvings, and the tiangong louge--are
also recognizable.
According to the Northern Song scholar Ye Mengde’s 葉夢得 (1077-1148), while revolving
sutra cases had been rarely used when he was young, they soon became widespread across China
from metropolitan cities to poor villages, where six to seven out of ten Buddhist monasteries had
one installed.
66
Associated with these sutra cases were rituals in which horns would be blown and
drums would be beaten, bringing forth an audio-visual performance for the crowd, who would
swarm to the monastery and line up outside its walls carrying coins and bolts of silk on their backs.
67

Such ritual performances, in Ye Mengde’s opinion, were not able to convey Buddhist teachings or
inspire Buddhist followers as the devices were originally designed to do; instead, they became
corrupted tools for monasteries to seek profit, and for patrons to solicit blessings simply by paying

inspired by Chinese prototypes. Another work examining the relationship between Chinese and Japanese repositories is
Zhang Shiqing, “Zhongri fojiao zhuanlun jingzang de yuanliu yu xingzhi 中日佛教转轮经藏的源流与形制 (Sources
and types of Buddhist revolving sutra cases in China and Japan),” Jianzhu shilun wenji 11 (1999): 60-71.
66
Ye Mengde 葉夢得, “Jiankangfu Baoningsi lunzangji 建康府保寧寺輪藏記 (Record of the revolving repository at
the Baoningsi in Jiankang Prefecture),” in Jiankangji 建康集, 4.8a-10a,
http://ctext.org/library.pl?if=gb&file=10513&page=112 (image); http://ctext.org/wiki.pl?if=en&chapter=838905
(text).
67
“ 吹蠡伐鼓, 音聲相聞; 襁負金帛, 踵躡戶外, 可謂盛然.” Carrington-Goodrich renders “ 吹蠡伐鼓, 音聲相聞” as
“one can hear the sound of the wheels of the revolving cases turning (p. 137),” which was not necessarily the case. More
likely, this describes the musical instruments used or that some sutra-chanting rituals were performed when the wheel
was turned.
77
for the sutra case to be turned.
68
Huang Min-chih has demonstrated that Ye Mengde’s criticism is in
fact well-grounded: anyone who wanted to turn the sutra case was supposed to pay a formidable
amount of money, and their payments indeed remained a major source of income for Song-dynasty
monasteries.
69
The economic benefits stirred up construction: not only were revolving sutra cases
being built in a greater number, but monasteries often vied for producing the most grandiose work.
In some cases, the repository was built surrounded by four smaller ones, making a five-wheel
composite.
70

Others, however, have argued that the ultimate purpose for such extravagance was still to
make salvation available to everyone. The abbot of the Shengfasi 勝法寺 in Changshu 常熟, for
instance, has confided his opinion on this issue to Ye Mengde. The abbot asserted that the revolving
sutra case could work only by providing its beholder an extraordinary viewing experience. For
commoners, he argued, “to instruct them with words would confuse them, whereas to teach them
by books would tire them. The only means is to use the extravagant repository to convey religious
solemnity and grandeur through its spectacular, extremely elaborate carvings and colorful
embellishment. This will turn all those who come to seek blessings and repentance into faithful
followers” ( 與之言吾理則惑, 教以其書則怠. 惟轉輪藏侈, 極雕刻彩繪之觀, 以致其莊嚴之意.

68
Jiankangji 建康集, 4.9b, http://ctext.org/library.pl?if=gb&file=10513&page=115.
69
Min-chih Huang 1996, 272-73.
70
One such five-wheel repository is recorded in Huihong 惠洪 (1017-1128), “Tanzhou faifu zhuanlunzang lingyanji 潭
州開福轉輪藏靈驗記 (d. 1119)” in Shimen wenzi chan 石門文字禪, 21.2b-6a,
http://ctext.org/library.pl?if=gb&file=1929&page=54; also in J23.B135.21: 676b,
http://tripitaka.cbeta.org/J23nB135_021. It is said that “a large wheel was built on the crest of the mountain and
surrounded by four small wheels at the four corners, like five standing humans 建大輪山之顛, 而輔以小輪, 四棋布
峙, 立如人聚.” The miniature architecture is compared to the thirty-three heavens on the Sumeru (Daoli gongque 忉利宮
闕, Sk. Trayastrimsa), and other canonical Buddhist texts have been invoked to rationalize the symbolism of the five
wheels.
78
可使凡徽福悔過者, 一皆效誠於此).
71
In other words, it was no longer the contents but the
appearance of the container that really mattered.

Yuan and later
Back to the Longxingsi sutra case, can we assume that it was installed in a similar light, out of similar
motivations to preserve the Buddhist law, bestow blessings, convert the multitude, provide an
expedient means to enlightenment, and perhaps also to commemorate the deeds and virtue of a
certain patron? One has to admit that a single sutra case could have derived various different
meanings for different audiences, and that over the course of its installation, dilapidation,
restoration, and alteration, generations of users and patrons must have written and rewritten the
religious and social functions of this particular device.
The aforementioned inscription of “Shengzhu benming changsheng zhuyan bei” informs us
of how the Longxingsi sutra case was used in the early fourteenth century. It tells that Zhilihetai 執
禮和台 (Jirgu'atai, fl. 1339),
72
a Mongol official, donated a hundred chapters of the Sutra for Humane
Kings and fifty chapters of the Sutra of the Medicine Buddha to the monastery, together with a thousand
strings of coins from his own salary to set up a fund for the continued provision of food and
incense in the future.
73
As the inscription indicates, the monks were to perform rituals of “reading,
turning, reciting, and chanting [the sutras]” ( 看轉誦讀), which most likely involved the actual

71
“Changshuxian Shengfasi 常熟縣勝法寺,” in Wujunzhi 吳郡志, 35.6b-7b,
http://ctext.org/library.pl?if=gb&file=4805&page=122 (image), http://ctext.org/wiki.pl?if=gb&chapter=982353 (text).
72
Yuanren zhuanji ziliao suoyin 元人傳記資料索引 (Index to Yuan Biographical Materials), p. 16092, retrieved from the
China Biographical Database (CBDB), http://isites.harvard.edu/icb/icb.do?keyword=k16229&pageid=icb.page76535.
73
Changshan zhenshizhi 常山貞石志, 19.15b-16a, http://ctext.org/library.pl?if=gb&file=31657&page=29.
79
turning of the sutra case. What Zhilihetai asked for return was apparently the longevity of the
emperor and the prosperity of the dynasty: “[I] wish and pray that the rule of the emperor persists
for billions of ages, spanning over boundless territories, and that his royal bond with Buddhism lasts
thousands of years” ( 冀祝聖歷億萬載無疆之算, 結梵席千百年不斷之緣).
74
The same wish
certainly also served as a proclamation of Zhilihetai’s loyalty to the emperor.
75

In spite of Zhilihetai’s good wish, Emperor Renzong died prematurely at the age of thirty-
six, and the Yuan turned out a short-lived dynasty which collapsed in 1368, merely half a century
after the stele bearing this inscription was erected. The revolving sutra cases, however, managed to
endure the vicissitudes of time and continued to survive, even flourish, in later dynasties. A well-
preserved example is found in the Baoensi 报恩寺 in Pingwu 平武, Sichuan, dated 1446. Occupying
the eleven-meter-tall space inside the library hall, the sutra case takes the form of a three-story
octagonal wooden pagoda adorned with two levels of exquisite tiangong louge, showing a stunning
resemblance to the feitianzang at the Yunyansi (fig. 35).
76
By contrast, the sutra case at the Zhihuasi
智化寺 in Beijing (d. 1444) looks vastly different from all previous examples: it has a marble dais
and a rather bulky, octagonal body filled with densely arranged sutra coffers. The expression of
architectural elements has been kept to a minimum, and the structure cannot be turned. These new

74
Ibid.
75
Ibid. The inscription explicitly states toward the end, “Always contemplate the extreme devotedness of the
subordinate (Zhilihetai), [whose donation] will indeed provide for the long years to come” ( 永惟臣子至極之情, 寔為
歲月悠久之計).
76
This repository is introduced in Xiang Yuanmu 向遠木, “Sichuan Pingwu Ming Baoensi kancha baogao 四川平武明
報恩寺勘察報告 (Survey report of the Ming-dynasty Baoensi in Pingwu, Sichuan),” Wenwu (1991.4): 1-19.
http://www.nssd.org/articles/article_read.aspx?id=1002614790#; Li Xiankui 李先逵, “Pingwu baoensi 平武報恩寺,”
in Yuan-Ming jianzhu 元明建築 (Yuan-Ming architecture), ed. Pan Guxi, vol. 4 of Zhongguo gudai jianzhushi 中國古代建築
史 (History of Ancient Chinese Architecture) (Beijing: Zhongguo jianzhu gongye chubanshe, 1999), 323-29.
80
changes in form and style might have been due to the increasing popularity of Tibetan Buddhism at
the imperial court.
77

Even during the final years of the Ming, the miraculous power of the revolving sutra case to
bestow blessings and facilitate enlightenment was still firmly believed by many. The Dijing jingwu lue
帝京景物略 (Overview of the famous sights in the imperial capital, 1635), when commenting on
the sutra case at the Dalongfusi 大隆福寺 in Beijing, tells us that “people would chant sutras or
donate coins; and when their merit became equal to that of [reading] the entire Tripitaka, the
repository would be turned once” ( 人誦經檀施, 德福滿一藏為轉一輪).
78
The text seems to
suggest an almost reversal of the original function of the sutra case conceived by its inventor, Fu Xi:
in the beginning, one only had to turn the repository to obtain the same merit of reading the sutras;
in late Ming, however, one needed be either literate or wealthy.
79
But the story ends with an
interesting twist: “A poor girl, unable to either chant the sutras or donate coins, felt ashamed and
distressed at heart. She then placed a single coin on the wheel, and the sutra case started to turn by
itself incessantly” ( 一貧女不能誦經, 又不能施, 內愧自悲. 因置一錢輪上. 輪為轉轉不休).

77
Liu Dunzhen 劉敦楨, “Beiping zhihuasi rulaidian diaochaji 北平智化寺如來殿調查記 (Survey notes on the
Tathagata Hal of the Zhihuasi in Beijing),” Zhongguo yingzao xueshe huikan 3 (1932.3): 1-69, reprint in Liu Dunzhen quanji 劉
敦楨全集 (Complete works of Liu Dunzhen) (Beijing:Zhongguo jianzhu gongye chubanshe, 2007), vol. 1, 47-85; Yan
Xue 閆雪, “Beijing Zhihua chansi zhuanlunzang chutan: Mingdai Hanzang fojiao jiaoliu yili 北京智化禪寺轉輪藏初
探: 明代漢藏佛教交流一例 (A preliminary research on the revolving sutra case in the Zhihua Chan Monastery in
Beijing: a case of the interaction between the Han and Tibetan Buddhism in the Ming dynasty),” Zhongguo zangxue 85
(2009.1): 211-15.
78
“Da Longfusi 大隆福寺,” in Dijing jingwulue 帝京景物略, by Liu Tong 劉侗 (1593-1636) and Yu Yizheng 于奕正
(1597-1636), ed. Zhou Sun 周損, 1.78a-79b, http://ctext.org/library.pl?if=gb&file=24430&page=194&remap=gb
(image), http://ctext.org/wiki.pl?if=gb&chapter=666227&remap=gb (text).
79
Carrington-Goodrich’s translation goes: “People may read the canon or make donations. If one’s virtue or gift is equal
to one tsang 藏 [one of the three baskets of Buddhism] it is sufficient to make one turn of the wheel (p. 145).” To me,
however, the text still reads ambiguous. It could actually mean that one turn of the repository equaled reading the entire
Tripitaka, but not necessarily which one should go first or be the prerequisite for the other.
81
The construction and use of the revolving sutra case continued in the Qing and even today.
80

In 2007, the Song monastery Gaolisi 高麗寺 was reconstructed anew at its original location in
Hangzhou, where a library hall and an octagonal sutra case were also built. The reconstructed sutra
case is a fourteen-meter-tall, four-storied wooden pavilion believed to have faithfully embodied the
Song character and style (fig. 36). It weighs several tons and can be turned either manually or by the
mechanical transmission and hydraulic drive system at the bottom.
The century-old religious device has also become a peculiar cultural icon in contemporary
art. In April 2010, a modern version of the revolving sutra case--its form obviously inspired by the
Longxingsi example--was installed as the central piece of art at the exhibition of “Zhi de wenming
紙的文明 (All about paper)” in Shanghai (fig. 37).
81
The artist Tai Xiangzhou 泰祥洲 filled it with
the Siku quanshu 四庫全書 instead of the Buddhist Tripitaka, and explained that by turning the sutra
case, one was symbolically receiving all the wisdom recorded in the books. He called his installation
a “fangbian famen 方便法門 (expedient means),” a term derived from Buddhist literature, and
interpreted it to be a “cosmic model” or “cosmic mechanism” comparable to the particle colliders
devised by modern physicists.

Miniaturization as Deconstruction
As proposed in the Introduction, miniature architecture can be interpreted as deconstructive
architecture or anti-architecture because of the detachment of form from function (or in other
words, sign from signified) resulted from the deliberate manipulation of scale. The reduction in scale

80
Two well-known Qing repositories are in the Yonghegong 雍和宮 and the Summer Palace, both of imperial nature
and in Beijing. A probably late-Ming early-Qing example is at the Tayuansi 塔院寺 in Wutaishan.
81
A news report of this exhibition is at http://art.china.cn/zixun/2010-04/24/content_3480899.htm.
82
is apparently a numerical or “qualitative” change, but it entails a series of “quantitative”
transformations leading to structural redundancy and ambiguity. Such transformations are not only
closely related to the development of Chinese wooden structures in history, but they also
deconstruct and dissolve the very meaning of architecture, providing us a new perspective to
discourses on architectural history.
The Longxingsi revolving sutra case would be an excellent example to illustrate these points.
Being essentially a receptacle for religious scriptures and icons, the sutra case does not feature a
common-sense “architectural interior” to be occupied by humans. Instead, it denies any intrusion of
the human body as if to preserve a sacred, unsullied, and impenetrable space. This determines that it
no longer needs real entrances or exits, windows or doors, and yet the intention to keep the
distinctive form of a wooden pavilion or pagoda has generated miniature versions of these
architectural components which are now mainly decorative than functional. While imitation and
numerical precision are still the primary goal during miniature-making, alterations and simplifications
of design can nonetheless be freely embraced in the process as the miniaturist see fit.
The following is an analysis of what particular architectural icons and patterns--which I term
as “archetypes”--have been deconstructed in the making of the Longxingsi sutra case. Somewhat
similar to the Jungian archetypes, the archetypes discussed here are not pure geometrical forms and
shapes but ones fully imbued with religious and cultural significances, deeply ingrained in their
specific historical context, and evocative of certain feelings, memories, and notions. They are
reminders of the past, but in the deconstructive process of miniaturization, their images and
associated meanings have been forever destabilized and created anew.

83
The octagon
The most immediately perceptible archetype for the Longxingsi sutra case--and for almost all other
surviving sutra cases introduced in the preceding section--is the octagonal structure. More
specifically, it is the octagonal-based “pavilion-type pagoda (lougeshi ta 樓閣式塔)” which has been
generally accepted as the ultimate visual source. The tradition of building octagonal architecture in
China can be traced as far back as the Han tombs, but octagonal pagodas started to emerge not until
the first culmination of Buddhist art in the Northern and Southern Dynasties.
82

The earliest surviving Chinese pagoda was the twelve-sided stone-hewn Songyuesi 嵩岳寺
pagoda (d. 520) built under the Northern Wei. Most pagodas at this early stage, including the well-
known Simenta 四門塔 (d. 544) in Shandong, the stone pagodas found in the Yungang Caves 1 and
2 (figs. 38, 39), and the legendary Yongningsi 永寧寺 pagoda (d. 516, now in ruins), were generally
square in plan.
83
Interestingly, the octagonal design was perhaps first presented on a smaller scale by
a group of miniature stone stupas found in Northwest China. The fourteen stupas, dated to 426-436,
are collectively known as the “Beiliang shita 北凉石塔 (Northern Liang stone stupas)” with an
average height of forty centimeters (fig. 40).
84
Except but one, each stupa is composed of, from

82
Nancy S. Steinhardt, Chinese Architecture in an Age of Turmoil, 200-600 (Honolulu: University of Hawaii Press, 2014), 92-
93. A Han-dynasty octagonal burial chamber is found near Luoyang.
83
Liang Sicheng regards the early stage in the history of Chinese pagodas as “the Period of Simplicity, or the Period of
the Square Plan (ca. 500-900).” See Liang Ssu-ch’eng [Sicheng], Chinese Architecture: A Pictorial History, ed. Wilma Fairbank
(Cambridge: MIT Press, 1984), 124. A more nuanced discussion of the early period is in Nancy S. Steinhardt, “The Sixth
Century in East Asian Architecture,” Ars Orientalis 41 (2011): 27-71, which relates the earliest pagodas to multiple
sources including the mingtang and the Northern Liang miniature stupas.
84
An examination of the eight inscribed stupas is provided by Stanley Abe, who argues that they should not be identified
as miniatures of the stupas due to obvious differences in form, though they do display close visual connections with
Indian/ Central Asian architecture. See Stanley K. Abe, Ordinary Images (Chicago and London: University of Chicago
Press, 2002), 103-06, 123-66. Also see Alexander Soper, “Northern Liang and Northern Wei in Gansu,” Artibus Asiae 21
(1958.2): 131-64; Su Bai 宿白, “Liangzhou shiku yiji he Liangzhou moshi 涼州石窟遺跡和涼州模式 (Remains of
Liangzhou cave temples and the Liangzhou artistic mode),” Kaogu xuebao 83 (1986.4): 435-46; Yin Guangming 殷光明,
“Beiliang shita shulun 北涼石塔述論 (A discussion on the Northern Liang stone stupas),” Dunhuangxue jikan 敦煌学辑
刊 33 (1998.1): 87-107; Angela Howard, “Liang Patronage of Buddhist Art in the Gansu Corridor during the Fourth
84
bottom to top, an octagonal base, a cylindrical body, a shoulder in the shape of an inverted bowl, a
neck, a conical spire, and a hemispherical crown. The composition is largely “Indian or Central
Asian” as observed by scholars, but many iconographical elements, such as the cylindrical body and
the spire, later became incorporated into the Chinese pagoda.
85

The octagonal plan became widely adopted for large-scale pagodas in the eleventh century.
86

The eight-sided, nine-storied wooden pagoda (d. 1056) at the Fogongsi 佛宫寺 in Yingxian, Shanxi,
still stands as a supreme example of the pagodas of its time (fig. 41). Near the capital of the
Northern Song (modern-day Kaifeng), most wooden pagodas have been lost in time, but their
images have been largely captured by two contemporary brick pagodas--the Kaibaosi 開寶寺
pagoda (d. 1049, more commonly known as the “iron pagoda of Kaifeng”) and the Kaiyuansi 開元
寺 pagoda (d. 1055), both multistoried and octagonal in plan.
87
For miniature architecture, however,

Century and the Transformation of a Central Asian Style,” in Between Han and Tang, Religious Art and Archaeology of a
Transformative Period, ed. Wu Hung (Beijing: Cultural Relics Publishing House, 2000), 92-107.
85
Dietrich Seckel, “Stupa Elements Surviving in Eastern Asian Pagodas,” in The Stupa, Its Religious, Historical, and
Architectural Significance, ed. Anna Libera Dallapiccola and Stephanie Zingel-ave Lallement (Wiesbaden: Franz Steiner
Verlag, 1980). Seckel proposes two models of the development from stupa to pagoda. Generally speaking, one is the
elongation of the original stupa along the vertical axis, resulting in a multistory tower; the other is the miniaturization of
the hemispherical body and its spire to be placed on top of the pagoda, now as the latter’s finial. Many scholars discuss
the evolution of the Chinese pagoda, arguing that earlier prototypes include the que 阙 (gate towers or watchtowers) and
other similar multistory wooden pavilions or towers. See Liang 1984, 124; and Lothar Ledderose, “Chinese Prototypes
of the Pagoda” in Dallapiccola and Lallement 1980, 238-45, which considers the mingtang to be a likely prototype because
of its cosmological symbolism.
86
Liang 1984, 124. The Longxingsi repository falls in what Liang terms as “the Period of Elaboration, or the Period of
the Octagonal Plan (ca. 1000-1300),” the second stage of the development of Chinese pagodas.
87
See Steinhardt 1997, Conclusion. It includes an overview of the development of Chinese pagodas, focusing on Liao
and Song periods, the octagonal imagery, and the Womb World Mandala. Steinhardt claims the octagon to be the
distinctive feature and greatest legacy of Liao architecture. See also pp. 353, 363. The octagonal plan was also adopted
for burial chambers in the Liao and Xi Xia, sometimes accompanied by octagonal miniature pagodas as funerary objects
and an octagonal mound aboveground. The use of the octagon is again discusses in Steinhardt 2011, 56, this time
including the mingtang by Empress Wu and the Hall of Dream at the Horyuji. Also helpful is her lecture on the same
topic at the University of Southern California on February 19, 2014, where she argued that the purpose of the octagon
was to approach the circular stupa. The formal resemblance between the octagonal-based pagoda and the repository is
perhaps best showcased by a little known single-story wooden pagoda near Mogao named the Cishita 慈氏塔 (ca. 1000).
See Xiao Mo 蕭默, Dunhuang jianzhu yanjiu 敦煌建筑研究 (A study on Dunhuang architecture) (Beijing: Jixie gongye
85
the octagon might have been applied even earlier: Daoxuan’s lotus repository and Bai Juyi’s
revolving sutra case, as we have seen, were both eight-sided structures.
Why did most sutra cases adopt an octagonal plan? On the one hand, there must have been
technical considerations. The traditional square shape was replaced by the octagon because the latter
was closer to the shape of a circle but technically more practical than a circular structure. The
octagon was more exciting than the square; it demanded necessary breakthroughs of building
technology, bringing forth more complex and labor-intensive projects which at the same time
allowed the demonstration of advanced woodworking skills, virtuosity, and variety.
On the other hand were semantic considerations. It would seem only natural to make the
sutra cases octagonal since the Buddhist structure they emulated--i.e. the pagoda--were themselves
octagonal in plan. But one would lose sight of the bigger picture if one regards the sutra cases as
simple imitations and hence signifiers of the pagoda; what could be derived from the octagonal,
miniaturized form was in fact a hybrid of mixed images, iconographies, and symbolisms. As Eugene
Wang has observed, the octagon functioned as an architectural intermediary where Chinese cultural
codes and cosmological ideals (the Eight Trigrams, the shi 式 divinatory boards, etc.) sought to
“translate” and incorporate imported Buddhist terms and visions.
88
The outcome was often one of
ambiguity and confusion--a “Tower of Babel” effect--as not a single set of definitive references
could be pinpointed.

chubanshe, 2003), 387-91. Also to be considered is a type of pagodas known as the huata 華塔 (ornamented pagodas),
for instance the ones in Chengchengwan 成城灣, at the Guanghuisi 廣惠寺, and the Qinghuasi 慶華寺, which all
feature an octagonal base and a spire embellished by multiple miniature buildings reminiscent of the tiangong louge.
88
Eugene Wang, “What Do Trigrams Have to Do with Buddhas? The Northern Liang Stupas as a Hybrid Spatial
Model,” RES: Anthropology and Aesthetics 35 (1999): 70-91.
86
The central pillar
The second archetype for the Longxingsi sutra case is the “central pillar” in both physical and
metaphysical senses. Many early monasteries featured the pagoda as the single most important
monument occupying the center of the entire building complex. Scholars often trace the pagoda-
centered layout to the Yongningsi mentioned above, where the square-based pagoda dominated the
sky of the capital city of Luoyang, a spectacle which could be seen from one hundred li away.
89

Again, this layout was considered at the time as an Indian original: Daoxuan’s illustration (see fig.
30), for instance, shows a seven-story pagoda standing at the center of the ideal monastery.
90
Though
other spatial configurations gradually gained more popularity over time, in the eleventh century, the
pagoda-centered layout was still in use, such as at the Fogongsi. At the Longxingsi, even without a
pagoda, the idea of having a tall structure as the visual center and vantage point of the entire
monastery has nonetheless been realized by the 33-meter-tall Foxiangge 佛香閣 (Fragrance Pavilion
of the Buddha, est. 971, restored post-1950s).
For Buddhist cave temples, there is a certain type of “central-pillar caves (zhongxinzhu ku 中
心柱窟)” which typically feature a miniaturized, rock-cut pagoda at the center of the cave chamber

89
Yang Hongxun 楊鴻勛, “Beiwei Luoyang Yongningsi ta fuyuan yanjiu 北魏洛陽永寧寺塔復原研究
(Reconstruction of the Yongningsi pagoda in Northern Wei Luoyang),” in Jianzhu kaoguxue lunwenji 建築考古學論文集
(Collected essays on architectural archaeology) (Beijing: Tsinghua University Press, 2008), 328-341. Yang’s
reconstruction of the pagoda amounts to a total height of 147 meters. The earliest literary evidence for the pagoda-
centered layout is in Hou Hanshu 後漢書, which mentions Zuo Rong’s 笮融 temple in Xuzhou. See Marylin M. Rhie,
Early Buddhist Art of China and Central Asia, vol. 1 (Leiden: Brill, 1999): 20.
90
A detailed examination of Daoxuan’s illustration is given in Antonino Forte, Mingtang and Buddhist Utopias in the
History of the Astronomical Clock: The Tower, Statue and Armillary Sphere Constructed by Empress Wu (Paris: Roma
Istituto Italiano Per Il Medio Ed Estremo Oriente Ecole Francaise D’Extreme-Orient, 1988), 41-42, 46. The historical
account of the transmission and preservation of the illustration and its text is outlined in the appendix to the
Introductory Essay on pp. 51-52. See also Alexander Soper, The Evolution of Buddhist Architecture in Japan
(Princeton: Princeton University Press, 1942), 36-37. A similar pagoda-centered layout has been adopted by the
Shitennoji 四天王寺 (originally built in 593) in Japan. This is not the only layout adopted in Japan, however. Variations
include the pagoda-kondo pair (Horyuji) and the twin pagodas (Yakushiji and Todaiji).
87
serving as the main niches for Buddhist icons as well as the principal vertical support (pillar) in the
interior. The space around the central pillar then naturally provides the route for circumambulation.
Examples include the Yungang Caves 1 and 2 (see figs. 38, 39), the Beishan Cave 136 (see fig. 32),
and many others at Mogao, Xiangtangshan 響堂山, and Kizil.
91
Though the actual form of the
central pillar in each individual case varies, some apparently lacking identifiable architectural
elements, the central pillar could still be interpreted as a symbol of the stupa in most cases. This
symbolism is evidenced by an inscription (d. 698) found in Mogao Cave 332, where the patron, Li
Kerang 李克讓 (d. 880), ordered the craftsmen to “carve out a treasure stupa in the middle [of the
cave chamber], and leave the four sides open to form an ambulatory [around the stupa]” ( 中浮寶剎,
匝四面以環通).
92
Scholars often read the central-pillar plan as a Chinese translation of the Indian
chaitya, a rock-cut chamber of worship wherein a hemispherical stupa was usually placed at the center
of the apse.
93

From a large Buddhist monastery to a single hall or cave temple, the center seems to have
been generally reserved for a stupa or pagoda, if not for a Buddhist image alone. Whether full-scaled
or miniaturized, in wood or in stone, the stupa/pagoda functioned as an embodiment and

91
For studies on the central-pillar caves, see Li Chongfeng 李崇峰, Zhong-Yin fojiao shikusi bijiao yanjiu 中印佛教石窟寺
比較研究: 以塔廟窟為中心 (A comparative study on Chinese and Indian Buddhist cave temples: focusing on the
pagoda-temple caves) (Hsinchu: Chuefeng Buddhist Art Foundation, 2002); Andrew K. Y. Leung, “The Architecture of
Central-Pillar Cave in China and Central Asia: A Typological Study,” PhD diss., University of Pennsylvania, 2007.
92
Translation after Sonya S. Lee, Surviving Nirvana: Death of the Buddha in Chinese Visual Culture (Hong Kong: Hong Kong
University Press, 2010), 157. The full text of the inscription is included on pp. 278-81. The term baocha 寶剎 could mean
either a Buddhist monastery or a stupa, and the two are sometimes used interchangeably (the latter being a metonym of
the former). Cave 332 is identified by Lee as a “nirvana cave” based on its pictorial program and intended viewership.
For an analysis of its architectural and pictorial characteristics, see pp. 146-69, including an examination of the content
and background of the inscription.
93
Leung 2007, 62, 80.
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. tieweishan 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 monastery 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 it 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 an embodiment of the Buddhist ideal of
nirvana, as a relic container, and 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.”
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 characteristic 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 strongly 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 has 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 mainly 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 in 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 comments 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.
90
rotate constantly.
101
The “cakra” in the “cakravala” means “wheel;” this image and the momentum of
rotation 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 considered 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 might 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 revolving 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 systems. 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, commenting 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.
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 main 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 spectacular,
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 revolving 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 yixiang 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 Wang, 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: Made and Unmade, ed. Robert S. Nelson and Margaret Olin
(Chicago: University of Chicago Press, 2003), 107-32.
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 downscaling 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 fact, 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 woodworks, 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 sutra 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 ritual
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 spectacular 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 permitted 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
93
because it was no longer supposed to function 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, the
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,
unequivocal reading.
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 formal 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 previous 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 connections 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.
95
The intellectual interest of this chapter lies in the kind of visual experience brought forth by
miniature architecture. 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 it 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 large group of
statues in the middle of the library hall (figs. 45, 46).
1

A visitor can view the cabinets section after section by strolling along the nine-feet-wide
ambulatory between the central altar and 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”
96
ambulatory dramatizes this “paralyzed” viewing experience, yet it works perfectly well for creating
an atmosphere of an innermost shrine which is intimate, dark, 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 architecture 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,
nonetheless, 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.
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 observations 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 they 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, hence 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
theoretical value, is nonetheless comparable to the 32.57 centimeter (108.6 fen) observed in the

4
My own measurement 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 larger 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 the 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 designs.
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 value 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
architecture should shed some light 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.
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, continuous 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 by 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 alternative source of inspiration could have been the streetscape
of the Liao 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 surrounding
ancient-style buildings of which are claimed to have been modeled on none other than 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 spacing 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=MzA5OTgwNDAzMA==&mid=402361954&idx=1&sn=71736814947c62f2430fd
0239b6c8a02&scene=2&srcid=0229Wf1qAKQQ1daJoQRYodN7#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.
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
underneath 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 summarized 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 sources 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.
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 hundred 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 and 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 structure. 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 monastery 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 instance, 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.
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 correct 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 pavilion-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 category of miniature woodworks. 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 the 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 continue in Chapter 4.
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” which “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 increment 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 consciously 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 and 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 comparable 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 the 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. The 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).
103
of the cai for both Liao and Song architecture was generally kept at 3:2--in agreement 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 trend 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 traditional 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 architecture, 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 an 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 discuss 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
104
The modern “rediscovery” of the Bojia jiaozang and its cabinets was made by a group of
Japanese scholars.
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-important 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, they 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 the 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 Wooden
Pagoda at the Fogongsi 佛宮寺 and the Chongfusi 崇福寺. These expeditions greatly 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 Sekino’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 stele 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 Ibid., 83-84; 103, n. 1.
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 intellect 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 strictly 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 prototypes in Soper’s discussion.
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 highlight 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: not 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
unsurpassable 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 included political reasons, nationalism, and wars.
Many Chinese scholars today still shun Japanese materials because of “the political scars of Japanese occupation;” 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 architecture 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
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 chambers, 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 and 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 practices.
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 the 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 recounts 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 understands 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).
109
repositories might have been at the time.
47
The text describes “a circuit of wall repository, 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 bring 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 sutra 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 furniture 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 those 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=52170&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 instance, 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 each 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).”
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 repositories, 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
earliest 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 appears 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. The 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.
111
are many contemporaneous remains to illustrate this point. The Daoist shrine 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 monastic 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.
112
the Five Buddhist Mountains and Ten Monasteries), 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 portable 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 shanhua 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 plan 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 miniature
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 drawings 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. Zhang 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, and in the Yongping qinggui 永平清規, T82.2584. These
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 available.
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 monasteries.
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 Sumeru 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 the 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 dimensions 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, perhaps within certain receptacles. To 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 worshippers came to offer incense.
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 (figs. 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 on granaries and barns.
64
It is likely that by the early third century, architecture had
been adopted as a beloved 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 the 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
Eastern Han; see pp. 340-344.
64
Ibid., 26-27. Miniature granaries made of bronze have been found in Warring States burials; see Chapter 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 Bangtaizi 棒台子 in Liaoyang 遼陽, proposing that the particular
architectural shapes might have 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, authentic 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 Imperial China: A Study of the Fifteenth-century Carpenter's Manual Lu Ban
115
a certain portable shrine called shenchu 神廚 (spiritual cabinet) 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 Ruitenbeek 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, and 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 shangchu 上廚 (upper
shrine), denoting the miniature shrine on the upper level of the cabinet. Some of his interpretations are tentative but the
overall dimensions 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 framework using mortises and tenons.
Functionally, the legs of a chair or table is comparable to the columns, the stretchers 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 explorations. 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
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 other 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 as 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 scaling scheme
whereas the latter clearly lacks one, so far as the text suggests.

architecture brought changes to furniture 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 (Beijing:
Xianzhuang shuju, 2006), 125.
71
This measurement is extracted from my digital model and does not include the dais.
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 generally 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 between 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 Chinese 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 size
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 mountain”--this symbolism of the miniature architecture as the world pillar echoes my discussion in Chapter 2.
118
Ono Satoshi’s 大野敏 study of ancient and medieval Japanese zushi has demonstrated that
miniature shrines in East Asia were made 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 hall;
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 architecture 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, bookcases, 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 architectural 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 into 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 with
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 miniature 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.
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 of 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 view 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 dissertation, 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.
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 seas 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 treasure 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 boundless 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 sandalwood 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 Luoyang, 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: University of Washington Press,
2005), 133; also 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,
121
The description evokes a series 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 is 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, specks 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 seed” 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 contains 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, 1977), 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,
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 minute 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 ultimate 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 appear 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 that 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 concepts of interdependency and intercasuality.
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 individual 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 sitting 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 “entrance” 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 Teachings of Huayan), T45.1867.513a-
b. Translation by George Tanabe in William Theodore de Bary and Irene Bloom, Sources of Chinese Tradition, vol. 1, From
Earliest Times to 1600 (New York: Columbia 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 only 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 preaches 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 the 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
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 the entire universe.
90

Sudhana’s epiphany in the Tower of Vairocana
The theme of meditation is continued in an exposition by Fazang 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 Garland 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 Bodhisattva, 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 sudden, 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 consist 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 Cook 1977, 22.
92
T45.1876.640b; translation after Tanabe in de Bary and Bloom 1999, 474.
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 even 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 unendingly 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 provided 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.
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 the 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 10
63
grains of sand to be needed.
98
The revelation of Archimedes’ calculation is
that even the greatest infinite 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 his 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 remarked 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 constitute 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 sand 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. Kloetzli 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.
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 grain 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 held 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 from 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 and mystical implication of these 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 transformations, 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 metonym for Buddhist monasteries.
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 Garland 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 would 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 containing 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 wall 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 in

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 (pp. 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: Harrassowitz 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 Locations and Eight Assemblies.” Historical records indicates that transformation tableaux
of Avatamsaka also existed in the Tang capitals of Chang’an and Luoyang.
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 multiplication 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 compound 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 unfathomable 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 ten mirrors were used in Fazang’s demonstration: one at each of the eight cardinal and
ordinal points, one in the ceiling, 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 speaks 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 Daminggong 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.
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 images of each other, and the long ambulatory surrounding the
main altar is composed of an extended series of repeated 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 figs. 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 history. 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
miniature 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 technology, 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 structural 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 cabinets along the south and north walls. In general, the two parts are
symmetrical to each other.
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 traditional 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
significance 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 reanimate 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 stage 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.
132
4. Miniatures in the “Dome of Heaven”

The Jingtusi is located in the northeast of the Ying 應 County in northern Shanxi, 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 roads 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 Chinese architecture. Often 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 ceiling 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 decreased, 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 patterns 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
133
of Heaven,” a study of the symbolism of the dome 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 other 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 is 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 Buddhas 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/jingtusi/, with one Rhino 3D model and two photogrammetry
models.
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 identifications 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 towers 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 group 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, substructures, 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 penthouses (fujie 副階) and skirting roofs (yaoyan 腰檐).
135
corner towers (jiaolou 角樓), and six-tiered double brackets should be used for palace halls (dianshen
殿身).
5
All these appear to have tallied rather well with 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 an
ornament 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 more 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/parts 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 towers 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 (miniature 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, 1980), 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.
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 octagonal 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 bracketing belongs to
the eight-tiered double brackets 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 parts of the ceiling, it was
uniformity--and the underlying principle of modularization and standardization--that assured such a
sophisticated project to be ever accomplished with efficiency and quality. The Yingzao fashi
prescribes that ceiling coffers and tiangong louge can share the same cai. This is especially true for the
xiaodouba zaojing (miniature octagonal coffers), which resembles its larger counterparts but is also

10
See scaling schemes in Chapter 2 of this dissertation. 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.
137
different in certain ways. According to the Yingzao fashi, 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 Jingtusi, three octagonal coffers--located at the center (C), the east (E), and the west
(W) of the ceiling--are present (see fig. 73). Each is a superimposition of a diamond shape inside a
square, forming 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 arrayed 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 octagonal 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 small “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 same 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 woodworking: 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 from 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” than the eight-tiered brackets, the highest-rank
bracketing.
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 emphasized 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 magic 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 by three grid, or
a jing layout resembling the magic square. This specific layout corresponds to the three-bay-wide and
three-bay-deep Main Hall and is further accentuated by the miniature baldachin roofs along the
walls. Each coffer is distinctive: in addition 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 differentiation, by the
repetitive yet nuanced motifs and details integrated into these 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, officially 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 (Statutes on building and repairing activities, promulgated in 1029) decrees that the commoner’s house is not
139
interior space by a type of checkerboard ceiling called pingqi 平棊--a lattice structure with decorated
panels but without brackets 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 octagonal 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 reinforces the ambulatory space
around the central altar and encourages 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 aligned 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, 163-65. 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 functions are said to be “catching the dust (chengchen 承塵)” and concealing an unrefined roof frame. The three
types of ceiling are usually 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 mid-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).
140

Miniature-making in Jurchen-Jin Material Culture
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 Cultural
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 cloister 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 second 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 mid-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 first 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. See 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 (Gazetteer 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 unpaged 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.
141
The founding date of 1124 has been reiterated in later revisions and expansions of the local
gazetteer and incorporated into the 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 the 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 of Jingtusi is at
odds with all other monasteries, 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 later 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 titles 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 phoenixes
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 in 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 quality);
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.
142
of its owner.
24
Despite the deceiving modesty of the building from the outside, the ceiling design,
with its nine luxuriously ornamented coffers, could 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 whose 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 of 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 completion 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 long-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 solid 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 coffers, 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 development 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 Siyuan, 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.
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 around 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 demoralized 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 second 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 patron 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.
144
military commanders altered.
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 it 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 them 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 Amitabha 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 architecture in Shanxi),” Wenwu shijie (1994.1): 62.
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 virtuosity and imperial magnificence than the tiangong louge at Jingtusi, but
any illustrations or explanations of this woodwork are totally lacking in either the restoration 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 and 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 of 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 decorative 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: Wenwu 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 observation of the conformity and “distinctiveness” is made in Takeshima 1944, 15.
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 architecture: 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 (another 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 history 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 development in the expressiveness of miniature architecture after the experiments in the
Northern Song and Liao, and was produced with a proliferation of other miniature motifs 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 of enthusiasm for” monumentality and symbolism. See also
Steinhardt 1997, 236-37.
37
Steinhardt 2003, 80.
147

Miniature theaters
Miniatures not only appeared in wooden structures above ground but were also indispensable
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 miniature 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 of 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 Northern 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 themselves.
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 Death 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: tomb decor of the Jin dynasty from Shanxi) (Beijing:
Kexue chubanshe, 2012); Hong 2011, 75-114.
148
Underlying the emergence of miniature theaters in tombs was the popularity of drama during
the Northern Song and Jin times.
40
Theaters were places where historic, romantic, and religious plays
were performed and watched, in public as well as private settings. 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 and 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 ambiguous 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 theaters 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 Huizong’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.
149
Ruled-line painting
Miniaturization in the Jin further extended to the realm of painting. The best example to illustrate
this point is the murals 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 bridge 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 secular-looking background (fig.
87). Another discrepancy is that 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 comparable
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 1127, 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 Illustrating 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 preliminary 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.
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
complex 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 determine 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 even 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 reason, jiehua has
been historically evaluated as secondary to figure painting and landscape painting, which were

45
See Chapter 1 of this dissertation. Their accuracy 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 enclosure 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 more recent
study of jiehua is Anita Chung, Drawing Boundaries: Architectural Images in Qing China (Honolulu: University of Hawaii Press,
2004).
151
created by the free hand and with good artistic sense. Despite the generally unappreciative attitude
of the Chinese literati toward ruled-line painting, to modern historians, these 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 the 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 Northern 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.
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 helping 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 the 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 elaboration 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 encouraged 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 value 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 toward 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 and 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.
153
the Jin apart from any Northern Song or Liao antecedents.
53
Admittedly, to create a miniature world
was to be a master 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 determined 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 woodworking 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 material 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 techniques of miniature-making were 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 without a high-profile incorporation of Chinese political and cultural ideologies by the

53
Laing 1988-89, 119.
54
Hoyt Tillman, “An Overview 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.
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, legitimation 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 legitimation.
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 Song 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 relocation 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.
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 alienate 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 Dome
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 such 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.
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 “rationally descriptive” as well as
“emotionally visionary;” physical as well as transcendental, combining 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 Eastern traditions.
In his response article, Alexander Soper concurs that “celestial symbolism” was introduced
to Asia via direct borrowings 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 Hellenistic 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 was 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 diffusionist 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.
157
ceilings where painted star maps were often found. These, she argues, have greatly 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, could have envisioned and developed their systems of
domed ceilings 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 specifications 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 coffer 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
transformations 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 quoted 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 form 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.
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 and 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 Well
(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 why 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 Wang 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 been 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.”
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 China
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 physical 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 instructions 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 the 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 even 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,
160
Third, unlike a Western dome, which was a load-bearing member of the roof frame, the
zaojing was a self-sufficient structure which can be detached from the building. The jing structure
does not involve 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 beauty 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 architectural motif; it avoids suggesting a load-bearing roof structure as the term
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 surviving
example is the central coffer of the Guanyinge 觀音閣 (Avalokitesvara Pavilion, d. 984) at the

and many others) 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.
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 springing from an
octagonal base (here supported by miniature brackets) and converging at the top, forming three
beautiful cupolas. Similarly, other wooden coffers 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 where 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 mysterious miniatures in the ceiling of the Chongfusi 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 originally part of a certain “feitian louge 飛天樓閣 (towers

71
Evidence of zaojing and other related ceiling 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 Eastern 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 ceilings reappeared in a significant number of Buddhist cave temples at Dunhuang, Kizil, and
Bamiyan.
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 galleries, 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 legacy of the Jingtusi miniature endured well into the Yuan, Ming, and Qing dynasties.
An offering pavilion at the 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 type 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 unprecedented 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 multiple eaves, and the closely-spaced towers and corner towers
thrusting into the deep ceiling (fig. 93). The less rhapsodic, calmer example is a pair of almost

72
The date of this structure 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 (Collected illustrations of Chinese architecture) (Tianjin: Baihua wenyi
chubanshe, 2007), vol. 2, 526.
163
identical coffers from the Zhihuasi 智化寺 in Beijing (d. 1443), now in the collections of
Philadelphia Museum of Art and the Nelson-Atkins Museum 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 stands 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 an
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 Buddhist 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 different 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 partially 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 Longfusi mentioned
in Chapter 2 of the dissertation.
164
Miniaturization has been a beloved 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 not 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 square-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
buildings 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 realm? 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 Jolas (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.
165
buildings” suspended above Buddhist statues. Miniature architecture in the ceiling, therefore, would
generate an optical 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 consciousness of the investigable order of the minute things was part
of the Neo-Confucian worldview, which burgeoned in the eleventh century and continued as a
prolonged intellectual and political discourse into later history. While Neo-Confucian thought
exercised 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 anything
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
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 Confucians 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 with 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.
167
well in the center (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 divided 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 road 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 measuring 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) (Beijing: Zhongguo jianzhu gongye chubanshe, 1985): 40.
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 and 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 practice 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 society (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 revenues 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.
169
field), and was essential to the maintenance of a stable source of state income, providing a revenue
roughly equal to a ten percent 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 practiced). The purpose here is, of course, not to straighten
out all the 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 planning 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 though the
practicality of old social systems 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 state-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.”
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 was one of the many advocators who asserted that a revival of Confucian ethics and institutions
would heal the state of its “illnesses” 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 abolition 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 resources 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).
171
What was the charm of the jingtian, that after more than a thousand years, it still haunted
Chinese intellectuals? One cannot 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 found 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 brothers. 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 potentials to grow and proliferate.
93
See Chapter 3 of this dissertation
94
Wing-tsit Chan 1963, 406-08.
172
Neo-Confucians were more concerned about applying 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 exposition 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 thoroughfares, 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., 561-63. 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 historical 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 (not unlike the jing) to locate and describe the subject of investigation.
96
Wen 2008, 112-20.
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 the 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 guarded 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 permeated 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 since 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 reference 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.
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 less rigorous adaptation was the Dongjing of Northern Song, a metropolis later occupied
by the Jurchen emperors. The reconstructive plan of the city (fig. 101) shows 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 emulation 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 Chapter 3 includes a “giant” lotus at the bottom,
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 representation 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 painted
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.
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 display of the omnipresent Buddha; it also
shares 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 cycling 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 embraced 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 original 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 sinicization of an
esoteric pattern.
101
For an examination of the historical mingtang, see Ming-chorng Hwang, “Ming-tang: Cosmology, Political Order, and
Monuments in Early China,” PhD diss., Harvard University, 1996.
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 imprint of ideal
city planning. The central coffer 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 modeled 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, this is not the only case where miniature
“cities” were created in ceilings, but similar examples are found in contemporary as well as later
designs (fig. 106).
102

In this light, the symbolism of the tiangong louge and the “dome of heaven” needs 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 generate 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 convincing shape. As Chinese miniaturists 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
projection of the celestial and ideal. In essence, any designed environment, whether inhabited by

102
For instance, the ceilings of certain Yulin Caves 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 not 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 miniature 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).
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 been 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 commenced a golden age in
Chinese architectural history when miniature-making--including mini theaters and ruled-line
paintings--started to dominate ritual and religious spaces, 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 ceiling 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 from
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
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 visual representation of the “Heavenly Palace,” therefore, had to incorporate not
just stately buildings but also the most ideal layout.
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, the 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 viewed 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, especially 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 centuries) 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 miniaturization 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 peculiarities
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
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, between 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 copy” 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” and 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 manifested through
appearance or mechanism. By engaging with on-going discussions on the simulacrum, this chapter
proposes a new angle of viewing miniature architecture while providing 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 Chaps. 1-3 on concepts of representation, sign, and simulacrum.
181
The Model Pavilion at the Chongfusi
The architectural model 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 appears 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 elsewhere
5
--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 at
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 bay 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? Or 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 found 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.
182
used. Such an “aberrant” and simplified bracketing scheme 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 the
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 cornices 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 Dayanta 大雁塔, show no traces of ang. Hence, the lack of ang in
the model strongly suggests 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.
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 Chongfusi 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, the monastery first became
occupied by a certain court official surnamed Lin 林 as his personal residence, 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 apparently it had gone by the early Ming when a
major restoration project was 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 Linyasi. 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 to 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 sympathetic; to repair them, he ordered the stored grains and

11
Chai Zejun 1996, 3-10, gives an 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 decoration 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 Shuozhouzhi 乾隆朔州志.
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 Qianfoge 千佛閣 (Thousand-Buddha pavilion) (fig. 112) was
erected on the ruin of the original Tripitaka Library.
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 members 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 architecture (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 tucked into the narrow porch of the model.
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 that 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 forever, 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.
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 elevations 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 believed 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 known 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 especially 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 ornamental finial, etc. See Fu Xinian, “Riben Feiniao Nailiang shiqi
jianzhu zhong suo fanyingchu de Zhongguo 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 five stories and a thrusting spire, it looks fairly similar to
the Kairyooji pagoda.
19
Ibid., 158-65.
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, Baoguosi 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 wooden 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 yanjiu 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.
188
Mituodian 彌陀殿, Chongfusi 1143 7 26 Shuoxian, Shanxi
Sanshengdian 三聖殿, Shanhuasi 1143 5 26 Datong, Shanxi
Puxiange 普賢閣, Shanhuasi 1154 3 22.5
Gate, Shanhuasi 12th c. 5 25
Guanyindian 觀音殿, Chongfusi Jin 5 22

In three instances--the main halls of the Foguangsi and Huayansi, and the Qianfodian at the
Fengguosi--the height of cai amounts to 30 centimeters (29 in the third instance). Noticeably, all
three structures are the predominate buildings in their respective monasteries, and they belong to the
highest rank of buildings as indicated by the number of bays they each have (seven or nine).
21
On
the other hand, though quite a number of multistory wooden buildings from the tenth to fourteenth
centuries have survived, comparatively less has been done to retrieve their dimensional data, which
would have been more helpful for making comparisons with the model in question. Judging from
the data of the wooden halls alone, it is possible that the model pavilion is based on a 1/10 scale and
was intended to present the majesty of certain late-Tang and Liao masterpieces, such as its
precedent--the original library. The Tang vestiges rest not only in the robustness of the size but also
in the elegant simplicity of the overall design.
22

The original and the copy
It will hardly escape the notice of any observer that the Chongfusi model and the miniature pagodas
from Nara are all multistory structures. This is not a coincidence. As explained in Chapter 1, for

21
Yingzao fashi vol. 1, 74-75. The rank of cai to be adopted is determined by the number of bays (jian 間) of the building--
the more the bays, the greater the cai. This also determines the rank of the building.
22
Another scenario could be that the model was made on a 1/5 scale instead. While this possibility cannot be ruled out,
it would result in a much smaller (17-meter-tall) building, which would be less visually appealing. The proposal was
rejected perhaps because it was considered by the selection committee as too “grand” and expensive to be executed,
considering the growing scarcity of high-quality timbers as building material.
189
structurally complex buildings especially high-rises, it was always wise to build a model before
committing to real construction. When used as a pilot model, the miniature preceded the full-scale: it
was born directly out of a “mental image” and could in this case be regarded as the “original.” The
full-scale, on the other hand, would become a “copy”--though not always a faithful one, since
changes would have to be made--and it was precisely the aim of the modeler to detect structural
deficiencies, safety hazards, awkwardness in form and shape, and to respond promptly with
solutions and revisions of design.
The model served as an indispensable medium of communication between experts, since the
design of a three-dimensional object could find no other way to be fully articulated and discussed. In
most cases, however, a model was disposable, and usually more than one pilot model would be
needed at different stages of design. A pilot model would be ephemeral unless it was converted to a
demonstration model, a display for both experts and non-experts, designers and clients (patrons).
The reason for the preservation of the Chongfusi model is uncertain, but one might expect
monasteries to turn well-crafted models into exhibits and spiritual articles to be proudly kept and
even worshipped, as what might also have happened at the Kairyooji and Gangoji.
The same can be said for an architectural model preserved in the Main Hall of the Huayansi
(fig. 115). This is a Qing model of Qianlou 乾樓, the northwest gate-tower of the city-wall of
Datong built in the early Ming and demolished in the Qing.
23
The model sits on a wooden table
supported by four column-like legs. It is a four-story tower based on a five-by-five grid and a
cruciform plan, and is painted with gold, vermilion, blue, green, and white. Four-tiered double
bracketing has been used for the ground floor and five-tiered double bracketing for all upper levels,
and the intercolumnar sets appear only under the second layer of eaves (fig. 116). With regard to

23
I was informed of this history by Mr. Peng, a junior researcher working at the Office of Historic Preservation at the
Huayansi, who accommodated my fieldwork. No work has been published on this model yet.
190
proportion, the model is less elegant than the one at the Chongfusi considering its thick and clumsy
lintels, rafters, balustrades, and bracket arms. No written records as I know of tell us how the model
ended up in an imperially patronized Buddhist monastery, displayed in front of the Buddhas in the
main worship hall. But the effort of the long-time preservation has been well paid: it served as the
most important visual reference for the newly restored gate-tower of Datong--a gigantic “copy” of
the miniature completed in 2013 (fig. 117).
A few observations can be drawn from the discussions above. From a technological
perspective, architectural modeling diverged from small-scale woodworking; it was a process using
more straightforward and faithful downscaling, creating almost exact “copies” or doubles of large-
scale woodworks. The function also differed: a model could play multiple roles in different historical
moments, changing from a restorative design proposal (which combined elements of both
innovation and preservation) to a visual record of a historic building, to an exhibition, a display, or
an object of worship in a religious setting. A model became worshipped not because it was attached
to some holy images, relics, or scriptures, nor because it was believed to represent or symbolize
some holy places such as the Buddhist pure land. A model was self-referential; and its “holiness,” if
any, could only have derived from the architecture itself.
What seems to have tied modeling and small-scale woodworking together, on the other
hand, is a strong sense of uncanniness felt from the hyper-reality of the miniature form. The
Chongfusi model, for instance, was not an exact double of the Tang library, nor did it come to
fruition as a built project. It is forever a ghost existing in the imaginary as well as corporeal world of
miniatures, a world from which it cannot escape. Being an unrealized design scheme yet a fully
materialized mental image, it marks the absence of what common sense deems to be “real”--a full-
scale structure. According to Jean Baudrillard, there are four successive phases of any image:
1. It is the reflection of a basic reality.
191
2. It masks and perverts a basic reality.
3. It masks the absence of a basic reality.
4. It bears no relation to any reality whatever: it is its own pure simulacrum.
24

This model can be said to have adhered to some of these points: it reflects the styles and
characteristics of Tang architecture while distorting them by miniaturization and by adding a Ming
flavor with perhaps a personal touch. Moreover, its existence corresponds to the absence of the
“real,” as it stands as a self-referential simulacrum--a singular existence straddling the two worlds--
that of the palpable and of the imagined.

Modeling in Chinese History
The term yang could mean a template, a prototype, a pilot or demonstration model. Usually small in
size, a yang is often called a “xiaoyang,” whereas a “dayang 大樣” denotes a full-scale model. A yang is
not necessarily three-dimensional but could include drawings and sketches that similarly applied
miniaturizing principles. In the Yingzao fashi, a two-dimensional yang is termed “tuyang 圖樣,”
meaning “illustrated templates” or simply “illustrations.” In these cases, it seems that the yang was a
direct translation and visualization of the design scheme, one that preceded the manufacture of the
end product. This contrasts to what James King offers as a modern definition of the model--“a re ‐
creation of some prototype or original, generally but not always smaller and usually of materials
different from those of the original.”
25
When it comes to the issue of originality, shall we understand
modeling as a process of creation or re-creation? What do we mean by “originals” and “copies,” and
how do we tell them apart? Are all human creations re-creations? While the previous discussion

24
Jean Baudrillard, Selected Writings, ed. Mark Poster (Stanford: Stanford University Press, 2001), 173.
25
King 1996, 3.
192
already commented on this issue, this section will continue to explore and unravel the nature of
models and model-making through a number of historical texts and examples.
To King, modeling is a way by which the modeler interacts with the society, with the world,
and with him- or herself; it is a medium of communication and negotiation, a channel of self-
expression, and a projection of the ego on the external.
26
Models differ from miniatures because they
are “descriptive, analytic, and pedagogic tools... to illustrate concepts, to explain processes, and to
depict underlying structures,”
27
whereas for miniatures, “the concern is not as much accuracy or
authenticity as display, emotion, and creating an impression.”
28
King seems to be claiming that
miniatures are all about “pleasure” while models provide “insight.” In fact, as the following will
demonstrate, miniatures and models have overlapping identities and shared functionalities, especially
in the sense that both are simulacra of reality.

Modeling and drafting in the design process
The earliest recorded architectural model in Chinese history was a wooden model of mingtang (Hall
of Light) made by Yuwen Kai 宇文愷 (555-612), a court architect of the Sui dynasty. After an
imperial decree to build such a ritual architecture was announced in 593, Yuwen Kai, “following the
[instructions in the] text of Yueling, made a wooden yang of mingtang, which had double-eaves and two
levels and was divided into five chambers with four doors. The dimensions, scale, the round and the
square shapes were all based on certain textual grounds. He presented the model to the emperor”

26
Ibid., 8. King examines modeling as a human experience: “modeling itself may serve as an epitome of all the ways that
a variety of individuals make contact with the world... our interest in technology, our consumerism, our passion to
control things, our preoccupation with certain kinds of perfection, our supposedly abundant leisure, our need for escape,
our interest in cooperative activity…”
27
Ibid., 148.
28
Ibid., 19.
193
( 依月令文, 造明堂木樣, 重檐複廟. 五房四達, 丈尺規矩, 皆有準憑, 以獻).
29
The model surely
amazed the emperor, who was prepared to endorse the project; meanwhile, it also served as a
platform for other court officials to participate in the design process and contribute their own
opinions. Sometime between 605 and 618, Yuwen Kai presented another model with a memorial
explicating his proposal, but it never came to fruition.
30

The construction of the mingtang often involved model-making because the exact layout and
form of the building, being notoriously elusive due to the all-important ritual and political
significances on the one hand and the lack of surviving examples on the other, could be first
envisaged only on a reduced scale. In Yuwen Kai’s case, the model was a personal vision of the
legendary architecture, which he presented to his audience for inspection and examination.
Unfortunately, it turned out to be a target of criticism and even denial, leading to the failure of the
project. In Chinese history, efforts of building the mingtang often met with similar frustrations and
setbacks. Those which succeeded had invariably drawn inspirations from some instructive
“templates”: Han Wudi’s mingtang, for instance, was a structure based on a certain drawing allegedly
showing the primitive form of the mingtang in the time of the Yellow Emperor.
31
In the late

29
Suishu 隋書, 6.18b-19a. A description of the same model is found on 68.9b: “The model was made of wood. The
lower level was a square hall composed of five chambers; the upper level is a round ‘observing platform’ which had four
doors 其樣以木為之, 下為方堂, 堂有五室, 上為圓觀, 觀有四門.”
30
Ibid., 68.3b, http://ctext.org/wiki.pl?if=gb&chapter=417382#p6. Yuwen Kai’s memorial reveals not only what
specific scale he adopted for the model but also the historical tradition lying behind scaling: “In the past, when Zhang
Heng devised the armillary sphere and celestial globe, he adopted a 1/20 scale. When Pei Xiu mapped the land, he
adopted a 1/9000000 scale. The drawing I present here adopted a 1/100 scale 昔張衡渾象, 以三分為一度. 裴秀輿地,
以二寸為千里. 臣之此圖, 用一分為一尺.”
31
Shiji 史記, 12.27a-b, http://ctext.org/shiji/xiao-wu-ben-ji/zh. The drawing was presented to Han Wudi by a certain
Gongyu Dai 公玉帶, showing a structure with “no wall on the sides but a thatched roof on top and a water channel
around. Encircling the wall of the precinct were double-storied galleries with attics above. The entrance was located at
the southwest 四面無壁, 以茅蓋, 通水. 圜宮垣為複道, 上有樓. 從西南入.” Han Wudi soon ordered a mingtang to be
built following this drawing.
194
Northern Song, it was still a standard procedure to make a xiaoyang first, so that the design could be
discussed, developed, and modified by a group of participants including the emperor himself.
32

A xiaoyang, other than being a three-dimensional model, could denote a set of sketches or
drawings for artists to study and emulate. The Northern Song monograph on painting, Tuhua
jianwenzhi 圖畫見聞志 (Record of remarkable paintings), records an instance where original sketches
came in handy in a restoration project. The Grand Xianguosi 相國寺, a Buddhist monastery in
Dongjing, was severely flooded in 1065, when a number of murals executed by former master
painters were destroyed. These murals, often referred to as transformation tableaux (bianxiang),
featured complex, well-thought compositions of various pictorial elements ranging from highly
specific icons to narrative scenes deeply rooted in Buddhist literature. To make the restoration of
these masterpieces possible, famous painters of the day were summoned to “consult and imitate the
copies and xiaoyang collected in the imperial storehouse” ( 用內府所藏副本小樣重臨仿).
33
Where
did these xiaoyang come from? Most likely, they had been drafted prior to the painting of the original
murals; after the murals were completed, they were kept in the imperial archives for future reference
and consultation.
Drafting, like modeling, was an essential step in artistic creation; it preceded the final
execution, and its immediate product was the first-time materialization of a mental image (a
“realization” and “concretization” on a smaller scale), one that would serve as the “original” to be

32
Xu Zizhi tongjian changbian shibu 續資治通鑑長編拾補, 34.9b-10a,
http://ctext.org/library.pl?if=gb&file=18207&page=70. In 1115, under an imperial decree written by Huizong himself,
“the inner court presented a xiaoyang of the mingtang 內出明堂小樣” to all court officials, and Prime Minister Cai Jing
was appointed head of the building project responsible for organizing scholars and officials to discuss and refine the
design.
33
Tuhua jianwenzhi 圖畫見聞志, 6: 8a-b. It is worth noting that the xiaoyang were not replicated without adding personal
touches.
195
replicated and refined. In this sense, builders, modelers, and painters were all miniaturists adept at
manipulating scale and proportion; they must have shared certain knowledge and techniques to have
arrived at such a significant commonality in practice.
34

Drafts and sketches were also instrumental in building activities. An example to illustrate this
point is the construction of the Northern Song imperial palace Yuqing zhaoyinggong 玉清昭應宮.
In 1008, the emperor ordered the court painter Liu Wentong 劉文通 to “first establish a set of
xiaoyang drawings” ( 先立小樣圖) before any construction began.
35
In addition to imperial halls,
chambers, galleries, studies, and gardens, the palace was also to contain a pyramid-like structure
called the Yuluo xiaotai 郁羅蕭臺--a mythic Daoist “heavenly palace” as elusive (and religiously
significant) as the mingtang and the tiangong louge.
36
To turn the intangible into something material, Liu
Wentong was instructed to “copy and emulate the Yuluo xiaotai painted by the Daoist Lu Zhuo”
( 移寫道士呂拙郁羅蕭臺) and incorporate it into his drawings. The drawings were then given to
the carpenters who managed to build a magnificent structure. The painting of the Yuluo xiaotai was
described to be “exquisite, elaborate, and full of miniscule details” ( 精巧密細)--probably a ruled-line

34
Yingzao fashi quotes a Tang text, Liu Zongyuan’s 柳宗元 “Ziren zhuan 梓人傳 (Biography of a carpenter)” (ca. 801, in
Liu Hedong ji 柳河東集, 17.6a-b): “[The carpenter] drew a house on the wall; within an area less than one-chi square, his
drawing showed all details and specifications of the structure. When [his apprentices] build full-scale houses according to
the dimensions marked in this drawing, there was not a slightest error 畫宮於堵, 盈尺而曲盡其制, 計其毫釐而構大
廈, 無進退焉.” In addition to sections, plans were also important especially for large building complexes such as
palaces, cities, and necropolises. The earliest excavated material evidence in this sense is the master plan (zhaoyutu 兆域
圖) of the mausoleum of a king in the Warring States period, using a 1/500 scale. See Liu Keming 劉克明, Zhongguo
tuxue sixiangshi 中國圖學思想史 (An intellectual history of image-making in China) (Beijing: Kexue chubanshe, 2008),
124-31.
35
Shengchao minghuaping 聖朝名畫評 3.
36
Ibid. See also Huang 2012, 83-84, on the structure of the Yuluo xiaotai and its relation with the Daoist imagery of the
body and the universe.
196
painting.
37
Techniques of miniaturization aside, it must have captured so nicely the mood and feeling
of sublimity appropriate for a dream-like heavenly scene that it was admired by many, including the
emperor, and turned out to be the direct source of inspiration for the realized project.
Liu Wentong’s xiaoyang were drawn exclusively for the imperial palace, but part of the design
later became “recycled” and adopted during 1023-1032 for a Daoist monastery in Suzhou.
38
It is not
so surprising that a local monastery could have been modelled after an imperial palace--religious
buildings would often assume an appearance of imperial grandiosity and solemnity (as in the case of
the Jingtusi) while the imperial palace needed to embrace fantastic elements such as “distant islands”
and “heavenly realms” to enhance the persona of the emperor as both a spiritual leader (in this case
a Daoist celestial) as well as a universal ruler. This mutual “appropriation” was to some extent
facilitated by the xiaoyang, which contributed to the transplantation of canonical architectural styles
from the capital to other regions of the state.
Material evidence for the xiaoyang, unfortunately, has all gone but for several Qing-dynasty
examples, the most famous of which were made by the architects from the Lei family, who were
known as the Yangshi Lei 樣式雷 (lit. modelers Lei). Lei Fada 雷發達 (1619-1693) was the first
generation of the Yangshi Lei to be in charge of the yangfang 樣房--the imperial design studio--and
his sons and grandsons inherited the post consecutively in the next two hundred odd years. The
Yangshi Lei were actively involved in many of the Qing imperial architectural projects, including the
planning and construction of the Forbidden City, the Yuanmingyuan 圓明園, the Altar of Heaven,

37
Shengchao minghuaping 3.
38
Kuaijizhi 會稽志, 7.3a-b.
197
and the Qing mausoleums.
39
The models made by the Yangshi Lei were crafted on different scales
for different purposes: a 1/100 model (yifenyang 一分樣 or cunyang 寸樣) was to display the overall
layout of a building complex; a 1/20 or even larger model was to show structural details and/or
propose different schemes of interior design.
40
Many of these models have survived even though the
real buildings have met with their demise prematurely. The Yuanmingyuan, for example, still has
some of its “images” kept intact, on a smaller scale and in a fragmentary way, by a number of
Yangshi Lei models (fig. 118).
41
Here modeling functions as a type of historic preservation of
architecture, and the Yangshi Lei models have entered into libraries and museums as permanent
collections.
42

Miniature pagodas and King Asoka’s 84,000 stupas
Based on the examples of architectural models discussed so far, it seems that in most cases,
modeling was a means by which a certain kind of technical difficulty was overcome. The difficulty
might rise from the intrinsic complexity of the structure (towers, pagodas, the roof frame), from the

39
One of the earliest studies on the Lei family is Zhu Qiqian, “Yangshi Lei kao 樣式雷考 (An examination on the
lineage of Yangshi Lei),” in Yingzao xueshe huikan 4 (1933.1): 84-89.
40
Other scales used for modeling included 1/1000, 1/200, 1/50, 1/25, etc. See Liu 2008, 547; Huang Ximing 黃希明
and Tian Guisheng 田貴生, “Tantan Yangshi Lei tangyang 談談 ‘ 樣式雷’ 燙樣 (On the architectural models made by
Yangshi Lei),” Gugong bowuyuan yuankan 故宮博物院院刊 (1984.4): 93.
41
A large-scale restoration project of Yuanmingyuan was initiated in 1873 (twelfth year of Tongzhi 同治) after the
damage done by the war in 1860. A set of drawings and models were produced by Lei Siqi 雷思起 (1826-1876) and Lei
Tingchang 雷廷昌 (1845-1907) who worked on single pavilions (with openable roofs showing interior decor) as well as
large building complexes (including gardens). After being inspected by royal members and supervising officials,
calculations were made estimating how much the project was going to cost. Some of the buildings were never repaired
due to lack of good timber material, financial difficulties, and concerns for foreign invasions. They remained on paper
and as models kept in the palace or by the Lei family. See Liu Dunzhen, “Tongzhi chongxiu Yuanmingyuan shiliao 同治
重修圓明園史料 (Archival sources for the restoration projects of Yuanmingyuan in the Tongzhi Period),” Yingzao
xueshe huikan 4 (1933.2, 3/4), reprint in Liu Dunzhen quanji, vol. 1, 183-257. Also see Huang and Tian 1984, 93.
42
See, for instance, archives at the National Library of China (http://www.nlc.gov.cn/nmcb/gcjpdz/ysl/), and the
National Taiwan University Library (http://cdm.lib.ntu.edu.tw/cdm/landingpage/collection/ysl).
198
elusiveness of a legendary or mythic architecture (mingtang, Yuluo xiaotai), or from the carpenter’s
unfamiliarity with certain foreign forms (pagodas, transformation tableaux). Especially, to build a
pagoda was a highly intellectually challenging task which demanded both the techniques of building
high-rises and a certain level of familiarity with the Buddhist visual vocabulary. As demonstrated
earlier, a xiaoyang pagoda would come in handy in this situation--it could help to pinpoint potential
deficiencies of the proposed structure. In the following example, another important function of
these architectural models will be exposed--the xiaoyang pagodas became standards of production,
were mass produced by the imperial court, and were distributed and applied nationwide.
The first recorded mass production of the xiaoyang pagodas was during Sui Wendi’s (r. 581-
604) Buddhist relic campaigns. In 601, on his sixtieth birthday, the emperor issued a decree
proclaiming his devotion to Buddhism and initiated what would later be known as the first of his
three relic campaigns, during which the sarira recently acquired by the emperor were redistributed to
more than a hundred prefectures of the state where pagodas would be erected to enshrine the
relics.
43
The decree stated that there would be “relevant personnel making yang [of the pagodas] to be
sent to the prefectures” ( 所司造樣送往當州), probably to provide visual instructions to local
carpenters and to ensure quality and budget control.
44

43
Guanghong mingji 廣弘明集, T52.2103: 213a-213b, 217a. In 602, Sui Wendi ordered more relics to be sent to fifty-one
prefectures (fifty-two or fifty-three according to other sources). The Biography of Tan Qian 曇遷 in Xu Gaosengzhuan 續
高僧傳 (T50.2060: 573b-c) has recorded the third campaign in 604 when pagodas were to be erected in another thirty
prefectures. A detailed recount and analysis of Sui Wendi’s relic campaigns during 601-604 is in Du Doucheng 杜斗城
and Kong Lingmei 孔令梅, “Sui Wendi fen sheli jianta youguan wenti de zai tantao 隋文帝分舍利建塔有關問題的再
探討 (A further discussion on the issues concerning the relic distribution and pagoda building campaigns by Sui
Wendi),” in Journal of Lanzhou University (Social Sciences) 39 (2011.3): 21-33.
44
Guanghong mingji, T52.2103: 213a-213b.
199
The form of the relic enshrinement is said to be “following the example set up by King
Asoka in all aspects” ( 建軌制度一准育王).
45
No pagodas built during Sui Wendi’s time have been
preserved to this day to verify this statement, but according to some textual evidence, the pagodas
probably looked more Chinese than “foreign” and had little resemblance to any Indian or Central
Asian prototypes. One such pagoda built in modern-day Jixian 薊縣 is described as “a five-story
wooden structure decorated with gold and jewels, where the sarira was kept underneath [in the
crypt]” ( 五層大木塔, 飾以金碧, 扃舍利於其下).
46
Instead of mimicking a stupa, which was
characteristic of its round, hemispherical part of the body, the pagoda was more likely a magnified
duplicate of one of Sui Wendi’s xiaoyang sent from the capital.
47

During the Five Dynasties, Qian Hongchu 錢弘俶 (r. 948-978), the King of Wuyue 吳越,
was so inspired by Sui Wendi that he started his own mass production of relic stupas which allegedly
amounted to a total number of 84,000. This time, no models were made and sent to local
monasteries to monitor construction, but the miniatures themselves served as reliquaries to be
deposited in the crypts under the pagodas across the kingdom. The most exciting relic deposit
attributed to Qian Hongchu was discovered in 2000-2001, when two so-called “Asokan stupas
(Ayuwang ta 阿育王塔)” were excavated from the top chamber and the crypt of the Leifengta 雷峰
塔 in Hangzhou. The one from the top chamber (d. 972/976) (fig. 119) is 33.5 centimeters tall and

45
Xu Gaosengzhuan, T50.2060: 573b-c.
46
“Minzhongsi chongcang sheli ji 憫忠寺重藏舍利記 (Record of redepositing the sariras at Minzhongsi)” (d. 892), in
Jinshi cuibian 金石萃编, 118.4a.
47
The xiaoyang sent out by Sui Wendi could have been drawings instead of (or in addition to) architectural models. On
the other hand, we cannot be sure if all pagodas were truthful duplicates of the imperial template.
200
made of silver.
48
Upon a square base, on each side of the stupa is carved an archway framing a
certain scene from the Jatakas. The capitals of the four corner columns are shaped into four Garuda
figures. Directly above are four accentuated, slender, and erect acroteria (shanhua jiaoye) bearing more
narrative scenes and icons; at the center of the roof is a spire with five layers of dew-disks (lupan 露
盤) and an ornamental finial. The stupa from the crypt of the Leifengta, 35.6 centimeters in height,
looks basically the same (fig. 120).
49

A number of other Asokan stupas have been unearthed elsewhere in north and south China,
all sharing a great similarity in form (fig. 121). In the crypt of the Wanfota 萬佛塔 in Jinhua 金華
alone, archaeologists have recovered fifteen relic stupas, eleven of which are made of bronze and
bear the following inscription: “The King of Wuyue, Qian Hongchu, piously pledges to make eighty-
four thousand precious stupas; inscribed in the year of yimao” ( 吳越國王錢弘俶敬造八萬四千寶
塔乙卯歲記). The other four are cast iron and have a slightly different inscription: “The King of
Wuyue, Qian Hongchu, piously pledges to make precious stupas in the number of eighty-four
thousand to forever provide for [the relics]; inscribed in the year of yichou” ( 吳越國王俶敬造寶塔
八萬四千所永充供養時乙丑歲記).
50
The exact inscriptions with the two distinctive “time
stamps”--one equivalent to 955 (bronze) and the other to 965 (iron)--have been found on similar

48
This stupa contains a 4.4-cm-tall gold bottle with eleven pieces of relics. See He Qiuyu 何秋雨, “Zhejiangsheng
bowuguan cang Wudai Wuyueguo Ayuwangta 浙江省博物館藏五代吳越國阿育王塔 (Asokan stupas of the Wuyue
King of the Five Dynasties collected in the Provincial Museum of Zhejiang),” Shoucangjia (2011.3): 33.
49
This stupa contains a gold coffer in which a lock of the Buddha’s hair from the usnisa is believed to have been
interred. See Li Yuxin 黎毓馨, “Hangzhou Leifengta digong de qingli 杭州雷峰塔地宮的清理 (Excavated items from
the crypt of the Leifengta in Hangzhou),” Kaogu (2002.7): 19. A detailed summary of the crypt, its assemblage, history,
and miniature relic stupa is found in chap. 2 of Seunghye Lee, “Framing and Framed: Relics, Reliquaries, and Relic
Shrines in Chinese and Korean Buddhist Art from the Tenth to the Fourteenth Centuries,” PhD diss., University of
Chicago, 2013.
50
He 2011, 33-35.
201
stupas yielded from several other sites.
51
It seems likely that the stupas were first cast collectively in
the imperial foundry and then distributed to various monasteries to be deposited.
52

The excavations have well corroborated the textual record that Qian Hongchu, “admiring
King Asoka’s feat of erecting the stupas, ordered eighty-four thousand stupas to be cast in bronze
and iron, inside which were deposited precious sutra-cases and dharani scriptures in woodblock
prints. The stupas were distributed widely across the kingdom, and the undertaking took ten years to
accomplish” ( 慕阿育王造塔之事, 用金銅精鋼造八萬四千塔, 中藏寶篋印心咒經. 布散部內,
凡十年而訖功).
53
What has also been corroborated is the particular form of the “Asokan stupa” in
Chinese history--an archetype attributed to the Indian king but appeared perhaps first in China in
the late third century. Alexander Soper traced the origin of this type of stupa to a certain Ayuwangsi

51
Ibid., 34-37. Other relic stupas attributed to Qian Hongchu have been excavated in Zhejiang, Jiangxi, and Hebei.
52
Ibid., 35-37. Some of the stupas are inscribed with the character of a single word ( 全, 金, 人, 大, 了, 六, 尔, 万, 德,
保, 化, 安, 仁, 向, or 乙), and He suggests that the same character indicated the same batch of products from perhaps
the same module. The mass production of the relic stupas went hand-in-hand with sutra printing: an iron stupa (d. 965)
unearthed from the Dashanta 大善塔 in Shaoxing contains a woodblock-print sutra with a frontispiece which reads,
“The King of Wuyue, Qian Hongchu, piously pledges to make precious cases and print sutras in the number of eighty-
four thousand, to forever provide for [the dharma relics]; inscribed in the year of yichou 吳越國王錢俶敬造寶篋印經八
萬四千卷永充供養時乙丑歲記.” This, however, is the only example where a sutra was found inside the relic stupa,
and it was likely that not all 84,000 sutras were deposited in the 84,000 stupas. In Leifengta’s case, the dharani scripture
were deposited in the hollowed inside of the bricks, which similarly bore Qian Hongchu’s name and title and the
proclamation that they amounts to a number of 84,000. See Li Yuxin, “Hangzhou Leifengta yizhi kaogu fajue ji yiyi 杭州
雷峰塔遺址考古發掘及意義 (Significances of the archaeological excavation at the ruins of Leifengta in Hangzhou),”
Zhongguo lishi wenwu (2002.5): 7; Chen Ping 陳平, “Qian Hongchu zao bawan siqian baojia yin tuoluonijing--jiantan
Wuyue baojia yin tuoluonijing yu Ayuwangta de guanxi 錢弘俶造八萬四千寶篋印陀羅尼經-- 兼談吳越寶篋印陀羅
尼經與阿育王塔的關係 (The 84,000 precious sutra-cases and dharani scriptures made by Qian Hongchu, with a
discussion on the relationship between the precious sutra-cases, the dharani scriptures, and the Asokan stupas),” in
Rongbaozhai (2012.1): 36-47, (2012.2): 48-57.
53
Fozu tongji 佛祖統紀, T49.2035: 394c, http://www.cbeta.org/result/normal/T49/2035_043.htm. The number might
have been a hyperbole; it might not be exact but simply means “countless” and “myriad.”
202
阿育王寺 in Ningbo, where there once was a miniature stupa about twice as large as Qian
Hongchu’s bronze and iron stupas:
54

The stupa in Maoxian (in modern-day Ningbo) emerged from the ground in the second year of the
Taikang period of the Western Jin dynasty (281) after the monk Huida’s epiphany. It was one chi and
four cun tall and seven cun wide. It had five layers of dew-disks. The blue-green color looked like
stone but was not stone; the exterior was carved and engraved on the four sides, showing various
exotic images and scenes. Liang Wudi ordered a wooden pagoda to be built to shelter it.
越州東三百七十里, 鄮縣塔者, 西晉太康二年沙門慧達感從地出. 高一尺四寸, 廣七寸. 露盤
五層. 色青似石而非, 四外雕鏤, 異相百千. 梁武帝造木塔籠之.
55

A later text claimed that this was one of King Asoka’s eighty-four thousand stupas made in
the third century BCE and excavated in 265, that it was made of unearthly materials and carried
extraordinary decorations including four transformation tableaux each showing a specific scene from
the Jatakas.
56
Based on these descriptions, scholars believe that the stupa in Ningbo and Qian
Hongchu’s relic stupas should look fairly similar in terms of shape, structure, and perhaps even
pictorial configurations (fig. 122).
57
Does this mean that the latter might have been modelled after
the former, and ultimately after an authentic Asokan stupa?
Surviving stupas commissioned by King Asoka, such as the Great Stupa at Sanchi (d. third
century BCE), however, look vastly different from their Chinese “copies.” Soper believes that the
ultimate source of Chinese stupas was not the massive, hemispherical burial mound historically
associated with their Indian precedents, but instead the harmika--the small “pavilion” crowning the

54
Alexander Soper, “Japanese Evidence for the History of the Architecture and Iconography of Chinese Buddhism,”
Monumenta Serica 4 (1940.2): 645.
55
Guanghong mingji, T52.2103: 201b. Another textual record of the same stupa, also authored by Daoxuan, is in Ji shenzhou
sanbao gantonglu 集神州三寶感通錄, T52.2106: 404b, http://www.cbeta.org/result/normal/T52/2106_001.htm, which
describes the stupa to be “similar to those made in Khotan in the Western Regions 似西域于闐所造.” See Soper 1940,
642-43, for a translation of this text.
56
Tang daheshang dongzhengzhuan 唐大和上東征傳, quoted with translation in Soper 1940, 641-42.
57
Soper 1940; He 2011.
203
dome of the Indian stupa.
58
This is to say that the subject of emulation was in fact a miniature, a
fragment of the whole, a metonym of a powerful religious monument. The “Asokan stupas,”
therefore, was not so much of an imitation but a creative process in which literal transplantation was
replaced by signification--not the whole, but the most distinct and symbolic part of the original
structure was selected and reinterpreted on the Chinese soil to stand for the stupa.
The same process could be further understood as a process of simulation: it would be hard
to pinpoint any “authentic” Asokan stupas but each miniature stupa was a simulacrum, a model of
itself, a distortion and subversion of the historical image of the stupa, an “unreal” existence
conceived and materialized in the lack of a proper original. In an endless cycle, the simulacra
mimicked and copied themselves; together, they formed what must be likened to Indra’s net, where
the myriad jewels each contain and reflect the images of all other jewels and are forever entering and
penetrating one another.
59

Armillary spheres and celestial globes: in simulation of heavenly images
The concept of simulation is perhaps better explicated by investigating astronomical models and
implements, especially the hunyi 渾儀 and hunxiang 渾象 in Chinese history. Joseph Needham
translated the former as “armillary spheres” (for observing and measuring) and the latter as “celestial
globes” (for simulating and predicting the trajectories of the stars at any given moment, in addition
to time-keeping).
60
The two types of devices overlapped in function and were often combined as a
set known as the huntianyi 渾天儀 or yixiang 儀象, such as Su Song’s astronomical clock in the late

58
Soper 1940, 658
59
See Chapter 3 of this dissertation.
60
Needham, Wang, and Price 2008.
204
eleventh century.
61
Unlike architectural models, these were tools of studying the celestial bodies; but
like architectural models, they were miniatures and simulations of the intended “reality,” a reality
that has never been fully observable or comprehensible but could nonetheless be approached by
modeling.
While architectural models were static, the greatest challenge of building an astrological
model came from the design and construction of a machinery-core driving the entire device with an
automatic force so subtly rhythmic that it could simulate the movement of heaven. In this case,
simulation had little to do with formal resemblance but had to be achieved through a knowledge
about mechanism. And this was no easy task. Before Su Song started to build his astronomical clock,
a series of pilot models had to be made and tested. The very first model to be built was a “wooden
machinery-core model” (muyang jilun 木樣機輪) driven by the constant dripping of water and
monitored by an escape system (figs. 123, 124). In the fifth month of 1088, a small-scale wooden
pilot model of the clock was sent to the court to be examined, and a full-scale wooden model was
completed in the twelfth month of the same year for further testing.
62
Finally, a new bronze clock
was finished and installed in the imperial observatory in 1089. It is said to be “as large as the human
body; stepping inside, one felt as if entering into a cage. The apertures were opened according to the
[projections of the] stars on the globe, which was driven by spinning waterwheels. It simulated
correctly the movement of the central constellations and the timing of the dawn and dusk, which
could be told by observing through the apertures” ( 大如人體, 人居其中有如篝象. 因星鑿竅, 依
竅加星, 以備激輪旋轉之勢. 中星昏曉應時, 皆見於竅中).
63

61
Introduced in Chapters 2 and 3 of this dissertation.
62
Xin yixiang fayao, 1.2a-b, 3a. A full translation of Su Song’s memorial to the court (briefly describing the process of
making the device) is in Needham, Wang, and Price 2008, 20-21.
63
Quwei jiuwen 曲洧舊聞, 8.10a-b.
205
Su Song, however, was not the first person in Chinese history to invent such a device. His
forerunners were said to have possessed even greater skills and finesse, but records about their
knowledge and method had become largely fragmentary and incomprehensible while the devices
they made were passed down only in the form of xiaoyang. Su Song perhaps spent more time
studying the xiaoyang he obtained than observing the night sky, believing that the models preserved
certain fundamental truth and insight into the matter. On the other hand, simply replicating the
xiaoyang would not restore the old masters’ template back to a fully functional astronomical device,
but the designer had to solve a series of basic mathematical and mechanical problems and seek help
from experts who knew how to calculate the locations, trajectories, and velocities of the stars.
64

As imperfect as our astronomical knowledge has been, all simulations of heaven were
imperfect, “false” simulations. In China, armillary spheres and celestial globes often failed to
function after being used for an extended period, and needed be repaired, recalibrated, or remade
from time to time. During the eleventh century, before Su Song’s clock, there had been at least three
earlier astronomical devices, and the making of these devices all involved a modeling process. The
first was proposed in 1049, the second in 1074 (designed by Shen Kuo), and the third in 1082--less
than ten years before Su Song’s clock.
65
Soon after Su Song’s invention, which was generally believed
as a successful project arousing much admiration, the imperial court yet again felt the need to make a
new device in 1124.
66

64
Ibid.
65
Song huiyao jigao 宋會要輯稿, 296.
66
Ibid. After the Song dynasty relocated its capital in Lin’an (modern-day Hangzhou), the court had to start from scratch
since all Northern Song devices were looted by the Jurchens. The attempts were not always successful, and the products
were said to be less than half of the Northern Song ones in size (using 8,400 jin of bronze, comparing with 20,000 jin
used for Northern Song ones), and perhaps less adequate in terms of function. See Qidong yeyu 齊東野語, 15.5b-6a.
206
Rather than the normal wear and tear of the existing implements, or the alleged political
motives for proclaiming new eras, the principal reason for remaking the astronomical devices was
the inaccuracies and deficiencies of the old ones which became gradually exposed and exacerbated
with the passage of time. For example, the making of the 1074 device was rationalized as a result of
the dysfunction of the existing apparatus, which produced erroneous data due to original design
flaws.
67
It is understandable that astronomers at that time could not rely on telescopes or space
explorers like we do now, and that the models they created must have always had large or small
flaws which after a certain amount of time would be too noticeable to be ignored. It was an
impossible mission to generate a simulation that always tallied with the celestial “image;” even today,
our model of the universe is tentative and far from perfection. Hence, modelers needed to
continuously push themselves forward to calibrate and modify their devices while renovating
existing models and formulas.

I Ching on the notion of simulation
Historically, Chinese intellectuals argued that all man-made vessels and tools were made in
simulation of certain images, shapes, or phenomena found in nature. The earliest expression of such
an idea can be found in I Ching:
[The Changes has four (aspects of the sage’s way] in it: in terms of words [, it esteems its statements];
in terms of movements, it esteems its alternations; in terms of fashioning implements, it esteems its
images; and [in terms of divination,] it esteems its prognostications.
易有聖人之道四焉: 以言者尚其辭, 以動者尚其變, 以制器者尚其象, 以卜筮者尚其占.
68

67
Song huiyao jigao, 296.
68
Edward L. Shaughnessy, trans., I-Ching: The Classic of Changes (New York: Ballantine Books, 1998), 196-97. Original text
accessible at http://ctext.org/book-of-changes/xi-ci-shang.
207
The dynamics of qi and xiang has undergone sustained discussions by generations of Chinese
scholars, who understood qi to be basically any type of designed object (including architecture,
astronomical devices, and their models), and xiang not only images but also “imitations” and
“simulations.”
69
Hence, when fashioning implements, the way a sage pay respect to the Changes is by
simulating its figural or abstract images.
The same text claims that all inventions of the sage kings were inspired by the hexagrams--in
the beginning, the eight trigrams were made according to the observations of celestial and earthly
phenomena and mechanisms so as to “penetrate the virtue of spiritual brightness and to categorize
the real characteristics of the ten-thousand beings” ( 以通神明之德, 以類萬物之情).
70
Afterward,
the rope knot, the fishing net, the plow, the boat and oar, the carriage, the bow and arrow, the grave,
the inner and outer coffins--are all said to have been “taken from (quzhu 取諸)” certain hexagrams,
or rather, to have consciously simulated the “images” of the hexagrams and/or the natural principles
the hexagrams were believed to have encapsulated.
71
The primitive house made of ridgepoles and
roof slopes is said to be “taken from” the hexagram dazhuang 大壯, or the Grand Robustness (fig.
24), the literal meaning and divinatory connotation of which do not appear to have any relations
with architecture.
72
Rather, the text seems to imply that the earliest houses were built according to

69
Hence there is the seemingly outrageous claim that “the Changes is images. Images are imaged 易也者, 象. 象也者, 像
也.” See Shaughnessy 1998, 207; Liu 2008, 72-92.
70
Shaughnessy 1998, 205. Original text accessible at http://ctext.org/book-of-changes/xi-ci-xia.
71
Ibid., 204-07. Also see Liu 2008 for Chinese scholars’ debates on this issue.
72
“In high antiquity they dwelled in caves and located themselves in the wilds. The sages of later generations changed it
with palaces and chambers, with a ridgepole at the top and eaves below in order to attend to the wind and rain; they
probably took it from Dazhuang, ‘Great Maturity’ 上古穴居而野處, 後世聖人易之以宮室, 上棟下宇, 以待風雨, 蓋取
諸大壯.” Shaughnessy 1998, 204-07; also see pp. 88-89, for the divinatory statement of dazhuang.
208
some basic principle or mechanism of the universe that the ancient sages somehow “visualized” and
expressed through an abstract image--a particular hexagram.
Looking back to the architectural miniatures introduced in earlier chapters, especially the
tiangong louge, it is now easy to understand them as simulations in light of Baudrillard’s four phases of
the image. The “original” of the tiangong louge--real Buddhist heavens--were absent; they could not be
observed but only imagined, fantasized, conceived, and designed. (Iconoclasm, according to
Baudrillard’s theory, happens precisely out of the fear of how icons reveal the truth of the absence
of gods.) For miniatures as well as models, simulation is especially well facilitated and largely fulfilled
by miniaturization: it is exactly because of the lifelikeness and hyper-reality one perceives from these
small objects that turn them into simulacra of the “real” (the full-size) which may or may not exist in
a corporeal form. In a reverse way, one can also understand the full-size as a simulation of the small-
scale. The disparity between the two is not as great as we might have assumed, as neither is less real
or more illusory than the other, but both are crystallizations of our mental images, our instruments
of approaching the world and the self.

Conclusion
A preliminary study of architectural models in Chinese history has yielded several constructional
observations. When compared with miniatures, a model is more concerned with an accurate
(re)presentation of the exact form, structure, or mechanism. It is no longer designed for the contents
(sacred images or scriptures) or the foreground (religious statues on the altar); nor is it intended as a
receptacle or a pure ornament. It appears more technical but less rhetorical or evocative.
A model, therefore, stops to be overtly attached to (in a subordinate manner) an external
object but becomes a display of itself. Models are built for design, demonstration, problem-checking
and solving, communication, standardization, reproduction, and quality control. The modeling
209
experience, on the other hand, had profound epistemological connotations: it forces us to reconsider
the interrelationship between models and their “originals” or “copies,” between the small-scale
hyper-reality and the more familiar “real” world it appears to imitate. From this perspective, what
models and miniatures have in common is their nature of being the simulacra of the “mental
images” people visualize or conceptualize while observing and engaging with various natural images,
phenomena, and forces.
In this light, the model pavilion at the Chongfusi could be understood as a simulacrum of
the original Tang-dynasty Tripitaka Library. Even though its dating proves to be problematic and
would perhaps remain a conundrum, both the scale and style of the model bear the kind of
simplicity, grace, and vigor characteristic of Tang architecture. The Asokan stupas, on the other
hand, further complicate the issue of architectural simulation, translation and transplantation across
space and time. An “authentic” Asokan stupa might have never existed, but such an authenticity has
been transferred successfully through miniature forms and through the mass production and
distribution of the miniatures. One can claim that modeling, like miniature-making, largely dissolves
and deconstructs the dichotomy of the “original” and the “copy”; it is instead in a continuously
developing and changing discourse that images--mental or corporeal--are constantly defined and
redefined.
210
Conclusion
A close examination of the representative miniature architecture in 1000-1200 CE China--as
demonstrated in this dissertation--has several significant contributions to the field of Chinese
architectural and art history. While miniatures have been largely regarded and studied as mere
“copies” or “reflections” of full-size buildings in previous scholarship, I have exposed here the
importance of investigating them as distinctive artistic and cultural products which must be
understood in their own right. A history of miniature architecture thus serves as an antithesis to the
existing discourse in the field; it destabilizes the very meaning of the term “architecture” and leads to
much more comprehensive knowledge about, and a broadened view of, both miniatures and their
larger counterparts.
The study of miniaturization offers a different means of approach to architecture. Instead of
form, structure, and style, it concerns the scale of the object to be surveyed. As evidenced by many
examples shown in this dissertation, the change in size often entailed a series of modifications and
alterations of other physical attributes, which then became manifested as a change in taste.
Beginning in the twelfth century, the focus of architecture has shifted from building a massive,
grandiose exterior supported by robust structural members to the creation of a painstakingly
decorated interior full of minuscule details. Such an inward turn bestowed architecture at the time
with a remarkable “introverted” character.
Parallel to the emergence of this introvertedness was the progressive miniaturization of
buildings and their various components: as structural members became downscaled, they lost their
original functions and transformed into ornaments which could carry religious implications and
embody political ideologies in certain cases. This dissertation showcases that an inquiry into cultural
milieu in addition to woodworking technology is imperative in order to reveal the motivation for
miniature-making. Miniatures in a Buddhist worship hall, for instance, need be considered and
211
evaluated as signifiers of specific ideas and vehicles for specific rituals or practices. While religious
piety might have been the strongest drive for the fervent production of these miniatures, other
intellectual influences--especially the Neo-Confucian ideal of the perfect world order--also has to be
reckoned with.
The four case studies presented in this dissertation include a Northern Song revolving sutra
case (d. eleventh century), a set of Liao sutra cabinets (1038), a group of Heavenly Palace miniatures
in the ceiling of a Jin monastery (1124), and a model pavilion from a later period (ca. 1383). The
rationale behind selecting these four specific examples is manifold. First and foremost is my
intention to explore and demonstrate the correspondence between text and material evidence. The
first three examples each correspond to a particular type of miniature architecture regulated in the
Yingzao fashi in terms of form, structure, and scale. Such remarkable correspondence not only
testifies to the validity of the text as an indispensable guide book for investigating Chinese
miniatures, but it further highlights the necessity and benefits of textual research in art historical
inquiry.
Second, the first three examples all fall in the predefined timeframe of 1000-1200 CE, which
is contemporaneous with the Yingzao fashi (1103), and geographically they are located fairly close to
the Northern Song capital Dongjing where the text was compiled and promulgated. This has not
only enabled the exposure of the correspondence noted above, but also provided a platform where
the three miniatures can be discussed in a comparative light. One of the most important
observations I have made in this dissertation is that the three examples are based on a 1/5, 1/5, and
1/10 scale respectively. The numerical consistency between the first two attests to the intimate
cultural contact between the Northern Song and the Liao, whereas the change from 1/5 to 1/10
reflects the historical development in miniature-making in the twelfth century. Such a change is also
exemplary of the phenomenon of “progressive miniaturization” over time, when the scale of
212
architecture dwindled and the focus of architecture transitioned from sheer mass to ornamentation
and details.
On the other hand, the fourth example is presented in the last chapter of the dissertation to
alert readers of the potential limitations of the historical narrative of miniature architecture I have
projected based on the three case studies and their correspondence to the Yingzao fashi. Even though
architectural models existed in abundance in history, they are completely omitted from the Yingzao
fashi due to the legal nature of the text. Hence, to deliver a more comprehensive survey of Chinese
miniatures, one has to go beyond the text and the templates and formulas it contains. This can be
attempted by drawing evidence from additional material remains which may and may not have been
documented. However, one also has to stay mindful of how both documented and undocumented
examples might have adopted certain common--rather than vastly different--rules and customs of
miniature-making. This is especially true considering the fact that the fourth example examined in
this dissertation is based on a 1/10 scale just as the Jin miniature.
The limitations of the Yingzao fashi further lead us to deliberate several important issues in
future studies of Chinese architecture. Looking beyond 1200 CE and to a greater geographical range,
one has to determine if the majority of the building traditions and practices reflected in the text were
still generally followed in later dynasties, what new changes and trends need be taken into account,
whether or not regional styles sometimes took precedence over the official, canonical patterns, etc.
1

Another important issue concerns the carpenters themselves. How did they approach the text, being

1
The Yingzao fashi has been reprinted multiple times since 1103. The earliest surviving edition--now with only a few
pages left--dates from the Southern Song, while complete reprints are retrieved from Ming and Qing official
encyclopedic works such as the Yongle dadian 永樂大典 (Yongle encyclopedia) and the Siku quanshu 四庫全書
(Complete library in four sections). It appears that little effort has ever been made to edit the original text to
accommodate new building technologies and requirements. It was not until 1734 that the multi-volumed Gongcheng zuofa
(Methods of architectural projects)--the second oldest building code in Chinese history--was completed and published by
the Qing court. This text introduces a different system of terminology, classification, and scaling principles, which
displays both continuity and development since the promulgation of the Yingzao fashi.
213
largely illiterate? How has their experience and knowledge of woodworking been preserved and
transmitted orally or in written forms? As briefly brought up in this dissertation, an investigation
into Chinese miniaturists problematizes--if not breaks down--the assumed dichotomy between the
literati and illiterate artisans. Guo Zhongshu, for instance, received rigorous architectural training in
his early career before being appointed by the emperor as an imperial ruled-line painter. His
communication with the carpenter Yu Hao regarding the design of a multistory pagoda excellently
showcases the exchange of knowledge and skills among different professionals, from painters to
architects, and across different echelons of the society. Finally, while the limitations of texts cannot
be overlooked, those of the material evidence must not be underestimated either. Scholars still need
to grapple with the problem of authenticity since most of the existing wooden structures in China
have undergone multiple restorations and might not have preserved their original appearances.
Overall, the study of Chinese miniatures broadens the horizon of architectural and art
historical inquiry. Miniatures and miniaturization can be interrogated as an architectural motif, a
literary trope, a religious or philosophical concept, a historical trajectory, a cultural phenomenon, a
form of knowledge and practice, and many more. The image of “a grain of sand” invoked in the
very title of this dissertation echoes not just William Blake’s poem but also the atomic worldview of
Huayan Buddhism and Archimedes’s scientific experiment of measuring the universe. Paralleling
miniature architecture to a grain of sand is not a hyperbole but rather an allusion to such a rich array
of connotations and historical instances. It pinpoints one of the extraordinary abilities of miniatures,
that is, to evoke imaginations, arouse feelings, bring back memories, and induce contemplations and
Bachelardian reveries, which all make us dreamers, poets, and perhaps philosophers as well.
Future surveys of Chinese miniature architecture can be productive and meaningful if one
adopts a cross-disciplinary approach and keeps in mind the global context. As an idea and technique,
miniaturization is independent of specific cultures, geographical regions, and time periods.
214
Miniatures (architectural and non-architectural) can be found anywhere on earth, and miniature-
making has been part of the human delight and endeavor ever since the dawn of civilizations,
evolving from the crafting of simple idols to the manufacture of smaller and smaller electronic
devices measured in nanometers. Studying miniatures not only offers new insight into Chinese
architecture, art, and material culture, but miniaturization ought to be considered and investigated as
a recurrent and underlying theme which has persisted in other traditions of the world.
215
Figures

1 Illustration of tiangong louge fodaozhang
(Yingzao fashi, vol. 4, 95)

2 Illustration of tiangong bizang
(Yingzao fashi, vol. 4, 101)
216

3 Reconstructive drawings of douba zaojing, plan and section
(Liang Sicheng 2001, 214)
217

4 Reconstructive drawings of xiaodouba zaojing, plan and section
(Liang Sicheng 2001, 216)

218

5 Eight grades of cai in large-scale woodworking
(Model by author)

6 Six grades of cai in small-scale woodworking
(Model by author)
219

7 Guo Zhongshu, Summer Palace of Emperor Minghuang, detail
(Public domain, via Wikimedia Commons)

8 Tamamushi Shrine, detail of roof
(Public domain, via Wikimedia Commons)

220

9 Illustration of huasheng
(Yingzao fashi, vol. 4, 136)

10 Line drawing of mural on east ceiling slope of Mogao Cave 31, showing a woman holding a
Mohouluo doll
(Guo 2013, 14)
221

11 Wang Zhenpeng, Dragon Boat Regatta, partial
(http://theme.npm.edu.tw/exh102/dragonlake/ch/photo1.html)

12 Wang Zhenpeng, Dragon Boat Regatta, detail
222
(http://theme.npm.edu.tw/exh102/dragonlake/images/bphoto8.jpg)

13 Longxingsi sutra case, overview
(Photo by author)
223

14 Sectional drawing of Zhuanlunzangdian at Longxingsi
(Chen 2010, 54)
224

15 Plan of Zhuanlunzangdian at Longxingsi
(Liang 1933, 21)

16 Bottom of pivot of Longxingsi sutra case
(Photo by author)
225

17 Master plan of Longxingsi
(Liang 1933, 15)
226

18 Rhino 3D model of Longxingsi sutra case
(Model by author)
227

19 Reconstructive drawings of zhuanlun jingzang, plans, elevation, section, and details of brackets
(Takeshima 1971, 665)
228

20 Reconstructive drawing of zhuanlun jingzang, elevation and section
(Pan and He 2005, 145)

21 Rotating core of Longxingsi sutra case
(Photo by author)
229

22 Longxingsi sutra case in the 1920s
(Tokiwa and Sekino 1940, pl. 90)
230

23 Longxingsi sutra case in the 1930s
(Liang 1933, pl. 27)
231

24 Longxingsi sutra case, detail of corner set
(Photo by author)

25 Rhino 3D model of corner bracket sets of Longxingsi sutra case
(Model by author)
232

26 Longxingsi sutra case, detail of column-top and intercolumnar bracket sets
(Photo by author)

27 Zhuanlunzangdian at Longxingsi, detail of exterior bracket sets
(Photo by author)
233

28 Diagram showing historical development of Chinese bracketing system
(Liang 1984)
234

29 Cornice of Tianwangdian at Longxingsi, showing Qing bracket sets arrayed among Song originals
(Liang 1933, pl. 35)

30 Daoxuan’s layout of ideal monastery, detail
(Ho 1995, 2-3)
235

31 Elevation of Yunyansi feitianzang
(Guo 2009, 537)

236

32 Beishan Cave 136, interior
(Guo 1999, 84)

237

33 Baodingshan Cave 14
(Guo 1999, 49)

238

34 Drawing of Jinshansi revolving sutra case
(Zhang 2000, 122)

239

35 Pingwusi revolving sutra case
(Zhongguo wenwu dituji: Sichuan fence, 496)
240

36 Gaolisi revolving sutra case
(http://andonglaowang.lofter.com/post/2fe384_6a10cc9)

37 Modern revolving sutra case installed by Tai Xiangzhou in a 2010 exhibition in Shanghai
(Photo by luychen,
https://www.flickr.com/photos/34076990@N05/4561707565/sizes/o/in/photostream/)
241

38 Yungang Cave 1, interior
(Photo by author)

39 Yungang Cave 2, detail of central pillar
(Photo by author)
242

40 Northern Wei miniature stupa from Gansu
(Wang 1999, 70)

41 Yingxian Wooden Pagoda
(Photo by author)
243

42 Sanjie jiudi zhi tu 三界九地之圖
(P. 2824)
244

43 Su Song’s clock-tower
(http://ctext.org/library.pl?if=gb&file=86840&page=89)

44 Wooden pagoda of Su Song’s clock-tower
(http://ctext.org/library.pl?if=gb&file=86840&page=96)
245

45 Huayansi Bojia jiaozang, interior view
(Photo by author)

46 Sectional drawing of Huayansi Bojia jiaozang
(Liang and Liu 1933, pl. 5)
246

47 Huayansi Bojia jiaozang, ambulatory along the back of the central altar
(Photo by author)
247

48 Huayansi Bojia jiaozang, ambulatory along the south wall
(Photo by author)
248

49 Huayansi sutra cabinets, detail of bracket sets
(Photo by author)
249

50 Arching bridge of Huayansi sutra case
(Photo by author)

51 Huayan Plaza in front of Huayansi
(Photo by author)
250

52 Arching bridge of Huayansi sutra case, detail
(Photo by author)

53 Wooden miniature shrine in Binglingsi Cave 172
(Heireiji Sekkutsu, pl. 110)
251

54 Yungang Cave 6, detail of central pillar
(Photo by author)
252

55 Erxianmiao miniature Daoist shrine
(Zhongguo wenwu dituji: Shanxi fence, 449)

56 Huhuangmiao miniature shrine, detail of roof corner
(Zhongguo wenwu dituji: Shanxi fence, 467)
253

57 Elevation of miniature shrine in Buddhist dormitory
(Zhang 2000, 146)

58 Plan of Jinshansi dormitory
(Zhang 2000, 121)
254

59 Drawings of Jingshansi miniature shrine
(Zhang 2000, 142)

60 Plan of Jingshansi dormitory
(Zhang 2000, 118)
255

61 Line drawing of stone relief, north wall of east side-chamber, Dahuting Tomb 1
(Mixian Dahuting hanmu, 143)

62 Line drawing of stone relief, north wall of north side-chamber, Dahuting Tomb 1
(Mixian Dahuting hanmu, 181)
256

63 Front elevation, section, and plan of shenchu, according to Lu Ban jing
(Ruitenbeek, 201)

64 Yangshi Lei miniature shrine
(http://cdm.lib.ntu.edu.tw/cdm/singleitem/collection/ysl/id/5)
257

65 Reconstructive drawing of bizang, section
(Takeshima 1971, 734)
258

66 Diagram of the typology of Japanese zushi
(Ono 2002, 2)

259

67 Five Dynasty silk painting of “Seven Locations and Nine Assemblies”
(Duan and Fan 2005, 75)
260

68 Bianxiang of Huayanjing, Mogao Cave 85, detail of Lotus Repository World
(Duan and Fan 2001, 222)

69 Bianxiang of Amitabha’s pure land, Mogao Cave 321
(Wang 2005, pl. 9)
261

70 Compound eye of a fruit fly, detail
(Public domain, via Wikimedia Commons)

71 Mordern installation of Fazang’s mirror hall
(Wang 2005, 259)
262

72 Main Hall of Jingtusi, west elevation
(Photo by author)

73 Scematic plan of Jingtusi ceiling
(Sketch by author)
263

74 Central coffer of Jingtusi ceiling, above the main Buddha
(Photo by author)
264

75 Jingtusi Main Hall, interior view
(Photo by author)

76 Miniature golden halls in central coffer of Jingtusi ceiling
(Photo by author)
265

77 Miniature Buddhas painted in central coffer of Jingtusi ceiling
(Photo by author)

78 Seven-tiered, fan-shaped bracket set at the southwest corner of Jingtusi ceiling
(Photo by author)
266

79 Double brackets in east coffer (Coffer E) of Jingtusi ceiling
(Photo by author)

80 West coffer of Jingtusi ceiling
(Photo by author)
267

81 Partial view of Jingtusi ceiling, showing a combination three different geometric shapes: diamond,
octagonal, hexagonal
(Photo by author)
268

82 Baldachin roof above a painted Buddha at Jingtusi
(Photo by author)
269

83 Sixteenth-century map of Yingzhou, showing location of Jingtusi
(Yingzhouzhi, unpaged)

84 Miniature bracket sets in the ceiling of Mituodian at Chongfusi
(Photo by author)
270

85 Miniature theater in Houma Tomb 1
(Hong 2011, fig. 1)

86 Actor figures and a theater pavilion in Macun Tomb 4
(Zhongguo wenwu dituji: Shanxi fence, 424)
271

87 Line drawing of mural on the west wall of Manjusri Hall, Yanshansi
(Fu 1998, 290-91)
272

88 Reconstructive plan of main building complex painted in Yanshansi murals
(Fu 1998, 295)

89 Coffered ceiling in Main Hall of Shanhuasi
(Sickman and Soper 1984, 456)
273

90 Tiangong louge in ceiling coffer, Rear Hall of Fengshengsi
(Liang and Liu 2007, 526)

91 Tiangong louge in ceiling, Offering Pavilion of Doudafuci
(http://shanxi.abang.com/od/gujian/a/dc_p2.htm)
274

92 Tiangong louge in ceiling, Main Hall of Yong’ansi
(Zhongguo wenwu dituji: Shanxi fence, 498)

93 Tiangong louge in ceiling of Gongshutang
(Zhongguo wenwu dituji: Shanxi fence, 481)
275

94 Zhihuasi ceiling coffer, now in the Philadelphia Museum of Art
(Photo by Ruichuan Wu)

95 Zhihuasi ceiling coffer, detail of tianggong louge
(Photo by Ruichuan Wu)
276

96 Tiangong louge in circular coffer, originally from Longfusi
(Liang and Liu 2007, 528)
277

97 Diagram of well-field system
(Nongzheng quanshu, 4.21a)

98 Diagram of Warring-states land-allocation system for administrative purpose, according to Zhouli
(Nongzheng quanshu, 4.30a)
278

99 Plan of ideal capital city, according to Kaogongji
(Sanlitu jizhu, 4.6a)

100 Plan of imperial palace, according to Kaogongji
(He 1985, 97)
279

101 Reconstructive plan of Northern Song Dongjing
(Zhang 2003, 160)

102 Thirteenth-century map of Northern Song Dongjing
(Shilin guangji, 3.143)
280

103 Plan of five-chambered mingtang
(Sanlitu jizhu, 4.2a)

104 Reconstructive plan of Zhou-dynasty mingtang
(Yang 2008, 162)
281

105 Reconstructive elevation of Wu Zetian’s mingtang
(Yang 2008, 505)

106 Mandala city painted in ceiling, Yulin Cave 3 (d. Xi Xia)
(Anxi Yulinku, pl. 141)
282

107 Model pavilion at Chongfusi
(Photo by author)
283

108 East-west cross section of Qianfoge
(Chai 1996, 136)

109 North-south cross section of Qianfoge
(Chai 1996, 136)
284

110 Chongfusi model, detail of triple and double brackets
(Photo by author)

111 Golden phoenix engraved between bracket-sets
(Photo by author)
285

112 Qianfoge at Chongfusi, exterior
(Photo by author)

113 A typical ceyang
(Yingzao fashi, vol. 4, 6)
286

114 Model pagodas in Japan, Nara period
(Left: Model pagoda at Kairyooji; Right: Model pagoda at Gangoji)
(Fu 1992, figs. 13, 14)
287

115 Model of Ming gatetower Qianlou, Huayansi Main Hall
(Photo by author)

288

116 Model of Qianlou, detail
(Photo by author)

117 Restored Qianlou in 2013
(http://dtzk50.blog.163.com/blog/static/84375982013102453323314/)
289

118 Yangshi Lei drawing and model of a building complex at Yuanmingyuan
(Liu 1933, figs. 7, 8)
290

119 Asokan stupa, excavated from Leifengta top chamber
(Photo by Luo Jianqiang)

120 Asokan stupa, excavated from Leifengta crypt
(Photo by Zhou Qingling)
291

121 Excavated bronze and iron Asokan stupas attributed to Qian Hongchu
(He 2011, 37)

122 Stone Asokan stupas
Left: at Kaiyuansi in Quanzhou (d. 1145); Middle: on the Luoyang Bridge in Quanzhou (d. 1055); Right: at
Nanshansi in Zhangzhou (d. 13th century or later)
(Soper, 1958, pl. XXXIX)
292

123 Pictorial reconstruction of Su Song’s clock-tower
(Needham, Wang, and Price 2008, fig. 1)

124 Modern reconstruction of Su Song’s clock-tower at Taiwan National Museum of Natural Science
(http://www.nmns.edu.tw/nmns_eng/04exhibit/permanent/tower.htm)
293
Glossary
ang 昂
ao 䫜
Ayuwang ta 阿育王塔
bajiao lunzang 八角輪藏
bapuzuo shuangmiao sanxia’ang jixin chonggong 八鋪
作雙杪三下昂計心重栱
bianxiang 變相
biji 筆記
bizang 壁藏
bizhang 壁帳
boshanlu 博山爐
buddhaksetra (Ch. fotu 佛土)
bujian puzuo 補間鋪作
cai 材
caifenzhi 材分制
cakravala (Ch. tieweishan 鐵圍山)
ceyang 側樣
cha’ang 插昂
cheng 棖
chobo 帳房
chu 廚/橱
chuhang guoju 廚行果局
cunyang 寸樣
damuzuo 大木作
dayang 大樣
dazhuang 大壯
dharmacakra (Ch. falun 法輪)
dianshen 殿身
dizhong 地種
dongjing 東井
douba zaojing 闘八藻井
doushuai 兜率
dutuoju 度脫劇
fangbian famen 方便法門
fangjing 方井
fangtian junshui fa 方田均稅法
fashi 法式
feitian louge 飛天樓閣
feitianzang 飛天藏
fen 分
294
fodaozhang 佛道帳
fu 夫
fujing 副淨
fumo 副末
gan 干
ge 閣
gong 功
goulan 鉤闌
hua 花/華
huaban 華版
huacheng 化城
huagong 華栱
huanmen 歡門
huasheng 化生
huazang shijie 華藏世界
hun 魂
hunping 魂瓶
huntianyi 渾天儀
hunxiang 渾象
hunyi 渾儀
jiaolou 角樓
jiashan 假山
jiehua 界畫
jili guche 記里鼓車
jing 精
jing 井
jingdian 鏡殿
jingtian 井田
jingtingzi 井亭子
jingwuzi 井屋子
jingxia 經匣
jingzang 經藏
jiuji xiaozhang 九脊小帳
juan 卷
kan 龕
kuileixi 傀儡戲
kunmen 壼門
kyuden 宮殿
lan 欄
li 理
lianhuazang 蓮花藏
lianhuazang shijie 蓮花藏世界
295
liaoyanfang 橑檐方
liupuzuo dougong 六鋪作重栱
liuzhairi 六齋日
lixue 理學
long 籠
lougeshi ta 樓閣式塔
lupan 露盤
miao 杪
mingqi 明器
mingtang 明堂
moni 末泥
muta 墓塔
muyang jilun 木樣機輪
neicao 內槽
penjing 盆景
penshan 盆山
ping’an 平闇
pingqi 平棊
pingzuo 平坐
po 魄
puzuo zhongju 鋪作中距
qi 欹
qi 器
qichu jiuhui 七處九會
quan 泉
quzhu 取諸
ren 仁
renzheng 仁政
roukuilei 肉傀儡
sha 紗
shaluo 沙羅
shang’ang 上昂
shanhua jiaoye 山花蕉葉
shazhang 紗帳
shen 身
shenchu 神廚
shengseng gongdian 聖僧宮殿
shi 式
shierzhang 十二章
shijie zhong 世界地種
shizun 世尊
shuiyun yixiangtai 水運儀象臺
296
ta 塔
tai 臺
tengai 天蓋
tian 天
tiangong 天宮
tiangong bizang 天宮壁藏
tiangong louge 天宮樓閣
tiangong louge fodaozhang 天宮樓閣佛道帳
tiangong louge zaojing 天宮樓閣藻井
tianjing 天井
tiaowo 挑斡
tou 頭
tuoluoni jingchuang 陀羅尼經幢
tuyang 圖樣
wa 瓦
waicao 外槽
warabite 蕨手
wei 微
weisuo 微縮
weixing 微型
xiang 象
xianglun 相輪
xiao 小
xiaodouba zaojing 小闘八藻井
xiaomuzuo 小木作
xiaoyang 小樣
xiezijing 些子景
xinglang 行廊
xumizuo 須彌坐
ya 亞
yajiao 牙腳
yajiaozhang 牙腳帳
yang 樣
yangfang 樣房
yaochu 藥廚
yaoyan 腰檐
yazi 牙子
yichu 衣廚
yifenyang 一分樣
ying 營
yingxi 影戲
yinxi 引戲
297
yixiang 儀象
yuanquan 圜泉
zang 藏
zao 造
zao 藻
zaojing 藻井
zhang 帳
zhangshen 帳身
zhangtou 帳頭
zhangzuo 帳坐
zhi 栔
zhongliao 眾寮
zhongxinzhu ku 中心柱窟
zhuan 轉
zhuanggu 裝孤
zhuanjiao puzuo 轉角鋪作
zhuanjingtong 轉經筒
zhuanlun jingzang 轉輪經藏
zhuanlun xuange 轉輪懸閣
zhutou puzuo 柱頭鋪作
zong 粽
zuo 坐/座
zuo 作
zushi 廚子 (Ch. chuzi)

298
Bibliography
Abe, Stanley K. 2002. Ordinary Images. Chicago and London: University of Chicago Press.
Adorno, Theodor W. 1997. "Functionalism Today." In Rethinking Architecture: A Reader in Cultural
Theory, edited by Neil Leach, 6-19. London and New York: Routledge.
1997. Anxi Yulinku 安西榆林窟 (Yulin Caves in Anxi). Beijing: Wenwu chubanshe.
Bachelard, Gaston. 1969. The Poetics of Space. Translated by Maria Jolas. Boston: Beacon Press.
Bai, Yong. 2011. "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: 15-18.
Bataille, Georges. 1997. "Architecture." In Rethinking Architecture: A Reader in Cultural Theory, edited by
Neil Leach, 21. London and New York: Routledge.
Bataille, Georges. 1997. "Museum." In Rethinking Architecture: A Reader in Cultural Theory, edited by
Neil Leach, 22-23. London and New York: Routledge.
Baudrillard, Jean. 2001. Selected Writings. Edited by Mark Poster. Stanford: Stanford University Press.
Birge, Bettine. 2008. "Law of the Liao, Chin, and Yuan and its Impact on the Chinese Legal
Tradition." In New Perspectives on Chinese History, Legal History: The Formation and Transformation
of Traditional Chinese Legal Culture, 443-503. Taipei: Academia Sinica.
—. 2010. "Review of Peter K. Bol, Neo-Confucianism in History (Harvard East Asian Monographs,
number 307)." American Historical Review. Vol. 115. no. 3. 822-23. doi:10.1086.
—. 2002. Women, Property, and Confucian Reaction in Sung and Yuan China, 960-1368. Cambridge:
Cambridge University Press.
Birrell, Anne. 1993. Chinese Mythology: An Introduction. Baltimore and London: The John Hopkins
University Press.
Bol, Peter. 2008. Neo-Confucianism in History. Cambridge: Harvard University Press.
Carrington-Goodrich, Luther. 1942. "The Revolving Book-case in China." Harvard Journal of Asiatic
Studies 7 (2): 130-61.
Chai, Zejun, ed. 1996. Shuozhou Chongfusi 朔州崇福寺. Beijing: Wenwu chubanshe.
Chai, Zejun, and Zhengyun Li. 1993. 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: Wenwu chubanshe.
299
Chan, Hok-lam. 1984. Legitimation in Imperial China: Discussions under the Jurchen-Chin Dynasty (1115-
1234). Seattle and London: University of Washington Press.
Chan, Wing-tsit, trans. 1963. A Source Book in Chinese Philosophy. Princeton: Princeton University
Press.
Chang, K. C. 1983. Art, Myth, and Ritual: The Path to Political Authority in Ancient China. Cambridge:
Harvard University Press.
Chen, Jinhua. 2014. "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 (5): 120-24.
Chen, Mingda. 2010. Yingzao fashi cijie 營造法式辭解 (Annotations on the glossary in the Yingzao fashi).
Edited by Ding Yao 丁垚 et al. Tianjin: Tianjin University Press.
Chen, Ping. 2012. "Qian Hongchu zao bawan siqian baojia yin tuoluonijing--jiantan Wuyue baojia
yin tuoluonijing yu Ayuwangta de guanxi 錢弘俶造八萬四千寶篋印陀羅尼經-- 兼談吳越
寶篋印陀羅尼經與阿育王塔的關係 (The 84,000 precious sutra-cases and dharani
scriptures made by Qian Hongchu, with a discussion on the." Rong Baozhai (1, 2): 36-47, 48-
57.
Chen, Tao. 2011. "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: 244-51.
Cheung, Dominic. 2008. Yongrong siru 雍容似汝 (As grandiose as the ru wares: essays on Chinese bronze,
porcelain and paintings). Taipei: Yishujia.
—. 2014. Zhongguofeng: maoyi fengdong, qianfan donglai 中國風: 貿易風動, 千帆東來 (Chinoiserie and Its
Chinese counterparts). Taipei: Yishujia.
Chung, Anita. 2004. Drawing Boundaries: Architectural Images in Qing China. Honolulu: University of
Hawaii Press.
Cleary, Thomas F., trans. 1984. The Flower Ornament Scripture: Translation of the Avatamsaka Sutra.
Boulder: Shambhala Publications.
Clunas, Craig. 1991. Superfluous Things: Material Culture and Social Status in Early Modern China.
Cambridge: Polity Press.
Cook, Francis H. 1977. Hua-yen Buddhism: the Jewel Net of Indra. University Park: Pennsylvania State
University Press.
de Bary, William Theodore, and Irene Bloom. 1999. Sources of Chinese Tradition. Vol. 1. New York:
Columbia University Press.
300
Derrida, Jacques. 1997. "Architecture Where the Desire May Live." In Rethinking Architecture: A
Reader in Cultural Theory, edited by Neil Leach, 319-23. London and New York: Routledge.
Dolby, William. 1976. A History of Chinese Drama. New York: Barnes & Noble.
Du, Doucheng, and Lingmei Kong. 2011. "Sui Wendi fen sheli jianta youguan wenti de zai tantao 隋
文帝分舍利建塔有關問題的再探討 (A further discussion on the issues concerning the
relic distribution and pagoda building campaigns by Sui Wendi)." Journal of Lanzhou University
(Social Sciences) 39 (3): 21-33.
Duan, Wenjie, and Jinshi Fan, . 2005. Dunhuang shiku quanji 敦煌石窟全集 (Compendium of Dunhuang
Caves). Vol. 20. Hong Kong: Shangwu yingshuguan.
Duan, Wenjie, and Jinshi Fan, . 2001. Dunhuang shiku quanji 敦煌石窟全集 (Compendium of Dunhuang
Caves). Vol. 21. Hong Kong: Shangwu yinshuguan.
Ebrey, Patricia. 1990. "Cremation in Sung China." American Historical Review 95: 406-28.
Erickson, Susan N. 1992. "Boshanlu--Mountain Censers of the Western Han Period: A Typological
and Iconological Analysis." Archives of Asian Art 45: 6-28.
Feng, Jiren. 2007. "Architectural Metaphors and Conceptualization in Tenth to Twelfth-century
China as Reflected in the Yingzao Fashi." T'oung Pao 93 (4/5): 369-432.
—. 2012. Chinese Architecture and Metaphor: Song Culture in the Yingzao Fashi Building Manual. Honolulu:
University of Hawaii Press.
Forte, Antonino. 1988. Mingtang and Buddhist Utopias in the History of the Astronomical Clock: The Tower,
Statue and Armillary Sphere Constructed by Empress Wu. Paris: Roma Istituto Italiano Per Il
Medio Ed Estremo Oriente Ecole Francaise D’Extreme-Orient.
Fu, Xinian. 1998. Fu Xinian jianzhushi lunwenji 傅熹年建築史論文集 (Collected essays of Fu Xinian on
architectural history). Beijing: Wenwu chubanshe.
Fu, Xinian. 1992. "Riben Feiniao Nailiang shiqi jianzhu zhong suo fanyingchu de Zhongguo
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 Arch." Wenwu (10): 28-50.
Fu, Xinian. 1998. "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 preliminary
analysis of the architecture painted in the Jin-dynasty murals of Yanshansi in the Fanzhi
County of Shanxi Province)." In Fu Xinian jianzhushi lunwenji, 282-313. Beijing: Wenwu
chubanshe.
Fu, Xinian. 2016. "Zhongguo gudai mugou jianzhu sheji 中国古代木构建筑设计 (Design of
ancient Chinese wooden architecture)." Sheji yu yanjiu (39).
http://mp.weixin.qq.com/s?__biz=MzA5OTgwNDAzMA==&mid=402361954&idx=1&s
301
n=71736814947c62f2430fd0239b6c8a02&scene=2&srcid=0229Wf1qAKQQ1daJoQRYod
N7#rd.
Gernet, Jacques. 1970. Daily Life in China on the Eve of the Mongol Invasion: 1250-1276. Stanford:
Stanford University Press.
Glahn, Else. 1978. "Some Chou and Han Architectural Terms." Bulletin of the Museum of Far Eastern
Antiquities 50: 105-25.
Gu, Qiyi. 1986. "Jiangyouxian Ruishan Yunyansi feitianzang ji zangdian kancha jilue 江油縣圌山雲
岩寺飛天藏及藏殿勘察紀略 (A brief record of the survey of the celestial repository and
the repository hall at Yunyansi in Ruishan, Xiangyou county)." Sichuan wenwu 14: 9-13.
Guo, Daiheng. 2009. Song, Liao, Jin, Xi Xia jianzhu 宋, 遼, 金, 西夏建築 (Song, Liao, Jin, and Xi Xia
architecture). Vol. 3, in Zhongguo gudai jianzhushi 中國古代建築史 (History of ancient Chinese
architecture). Beijing: Zhongguo jianzhu gongye chubanshe.
Guo, Junye. 2013. "Dunhuang bihua, wenxian zhongde ‘Mohouluo’ yu funu qizi fengsu 敦煌壁畫,
文獻中的 ‘ 摩睺羅’ 與婦女乞子風俗 (The Mohouluo in Dunhuang murals and documents
and its relationship with the custom of ‘begging for sons’ of women)." Dunhuang yanjiu 142
(6): 13-17.
Guo, Qinghua. 2010. The Mingqi Pottery Buildings of Han Dynasty China: Architectural Representations and
Represented Architecture. Sussex Academic Press.
Guo, Xiangying. 1999. Beishan shiku (Beishan Caves). Vol. 1, in Dazu shike diaosu quanji 大足石刻雕塑
全集 (A Complete Collection of the Sculptural Art of Dazu). Chongqing: Chongqing chubanshe.
Hansen, Valerie. 1995. Negotiating Daily Life in Traditional China: How Ordinary People Used Contracts,
600-1400. New Haven: Yale University Press.
He, Qiuyu. 2011. "Zhejiangsheng bowuguan cang Wudai Wuyueguo Ayuwangta 浙江省博物館藏
五代吳越國阿育王塔 (Asokan stupas of the Wuyue King of the Five Dynasties collected in
the Provincial Museum of Zhejiang)." Shoucangjia (3): 33-38.
He, Yeju. 1985. Kaogongji yingguo zhidu yanjiu 考工記營國制度研究 (A study on the city-planning methods
recorded in the Kaogongji). Beijing: Zhongguo jianzhu gongye chubanshe.
1986. Heireiji sekkutsu 炳霊寺石窟 (Binglingsi Caves). Tokyo: Heibonsha.
Ho, Puay-peng. 1995. "The Ideal Monastery: Daoxuan’s Descriptions of the Central Indian Jetavana
Vihara." East Asian History 10: 1-18.
Hong, Jeehee. 2016. "Crafting Boundaries of the Unseeable World: The Ontology of the Bhagavat
Sutra Repository." Art History (forthcoming).
302
Hong, Jeehee. 2015. "Mechanism of Life for the Netherworld: Transformations of Mingqi in
Middle-Period China." Journal of Chinese Religions 43 (2): 161-93.
Hong, Jeehee. 2011. "Virtual Theater of the Dead: Actor Figurines and Their Stage in Houma Tomb
No. 1, Shanxi Province." Artibus Asiae 71 (1): 75-114.
Howard, Angela. 2000. "Liang Patronage of Buddhist Art in the Gansu Corridor during the Fourth
Century and the Transformation of a Central Asian Style." In Between Han and Tang: Religious
Art and Archaeology of a Transformative Period, edited by Hung Wu, 92-107. Beijing: Cultural
Relics Publishing House.
—. 2001. Summit of Treasures: Buddhist Cave Art of Dazu, China. Trumbull, CT: Weatherhill.
Hu, Liangxue. 2001. "Dazu Beishan Fowan shike Zhuanlun jingzang ku zhi guanjian 大足北山佛灣
石刻轉輪經藏窟之管見 (A glimpse at the Cave of the Wheel-turning Sutra Repository at
Fowan, Beishan, in Dazu)." Zhonghua wenhua luntan (1): 112-16.
Hu, Wenhe. 2008. "Dazu Baoding Pilu daochang he Yuanjue daochang tuxiang neirong yuanliu xin
tansuo 大足寶頂毗盧道場和圓覺道場圖像內容, 源流新探索 (A new exploration of the
contents and origination of the images of the Field of Vairocana and Field of Perfect
Enlightenment at Baoding in Dazu)." Fagu foxue xuebao (2): 247-310.
Huang, Min-chih. 1996. "Zailun Songdai siyuan de zhuanlunzang 再論宋代寺院的轉輪藏 (A
further discussion on the revolving sutra cases in Song-dynasty monasteries)." Qinghua xuebao
26: 139-88, 265-96.
Huang, Shih-shan Susan. 2011. "Media Transfer and Modular Construction: The Printing of Lotus
Sutra Frontispieces in Song China." Ars Orientalis 41: 135-63.
—. 2012. Picturing the True Form: Daoist Visual Culture in Traditional China. Cambridge: Harvard
University Press.
Huang, Shilin. 1991. "Sichuan Jiangyou Douruishan Yunyansi feitianzang 四川江油窦圌山雲岩寺
飛天藏 (The celestial repository at the Yunyansi in Douruishan, Jiangyou, Sichuan." Wenwu
(4): 20-33.
Huang, Ximing, and Guisheng Tian. 1984. "Tantan Yangshi Lei tangyang 談談 ‘ 樣式雷’ 燙樣 (On
the architectural models made by Yangshi Lei)." Gugong bowuyuan yuankan (4): 91-94.
Hwang, Ming-chorng. 1996. Ming-tang: Cosmology, Political Order, and Monuments in Early China. PhD
diss., Harvard University.
Idema, Wilt, and Stephen H. West. 1982. Chinese Theater, 1100-1450: A Source Book. Wiesbaden:
Steiner.
Johnson, Philip, and Mark Wigley. 1988. Deconstructivist Architecture. Museum of Modern Art.
303
Karetzky, Patricia. 1980. "The Recently Discovered Chin Dynasty Murals Illustrating the Life of the
Buddha at Yen-shang-ssu, Shansi." Artibus Asiae 42 (4): 245-60.
King, James Roy. 1996. Remaking the World: Modeling in Human Experience. Urbana and Chicago:
University of Illinois Press.
Kloetzli, Randy. 1983. Buddhist Cosmology: From Single World System to Pure Land: Science and Theology in
the Images of Motion and Light. Delhi: Motilal Banarsidass.
Kuhn, Dieter. 2000. "Liao Architecture: Qidan Innovations and Han-Chinese Traditions?" T'oung
Pao 86 (4/5): 325-62.
Laing, Ellen Johnston. 1988-89. "Chin ‘Tartar’ Dynasty (1115-1234) Material Culture." Artibus Asiae
49 (1/2): 73-126.
Laumakis, Stephen J. 2008. An Introduction to Buddhist Philosophy. Cambridge and New York:
Cambridge University Press.
Leach, Neil, ed. 1997. Rethinking Architecture: A Reader in Cultural Theory. London; New York:
Routledge.
Ledderose, Lothar. 1980. "Chinese Prototypes of the Pagoda." In The Stupa, Its Religious, Historical,
and Architectural Significance, edited by Anna Libera Dallapiccola and Stephanie Zingel-ave
Lallement, 238-45. Wiesbaden: Franz Steiner Verlag.
—. 2000. Ten Thousand Things: Module and Mass Production in Chinese Art. Princeton: Princeton
University Press.
Lee, Seunghye. 2013. Framing and Framed: Relics, Reliquaries, and Relic Shrines in Chinese and Korean
Buddhist Art from the Tenth to the Fourteenth Centuries. PhD diss., University of Chicago.
Lee, Sonya S. 2012. "Repository of Ingenuity: Cave 61 and Artistic Appropriation in Tenth-Century
Dunhuang." Art Bulletin 94 (2): 199-225.
—. 2010. Surviving Nirvana: Death of the Buddha in Chinese Visual Culture. Hong Kong: Hong Kong
University Press.
Lehmann, Karl. 1945. "Dome of Heaven." Art Bulletin 27 (1): 1-27.
Leung, Andrew K. Y. 2007. The Architecture of Central-Pillar Cave in China and Central Asia: A Typological
Study. PhD diss., University of Pennsylvania.
Li, Chongfeng. 2002. Zhong-Yin fojiao shikusi bijiao yanjiu 中印佛教石窟寺比較研究: 以塔廟窟為中
心 (A comparative study on Chinese and Indian Buddhist cave temples: focusing on the pagoda-temple
caves). Hsinchu: Chuefeng Buddhist Art Foundation.
Li, Luke. 2011. "Chuanshi liang Song shiqi Yingzao fashi de canjuan, zhailu ji zhulu gouchen
(Transmitted manuscripts, excerpts, and copies of the Yingzao fashi in the Northern and
Southern Song period)." Zhongguo jianzhu shilun huikan 4: 31-46.
304
Li, Shiqiao. 2003. "Reconstituting Chinese Building Tradition: The Yingzao fashi in the Early
Twentieth Century." Journal of the Society of Architectural Historians 62 (4): 470-89.
doi:10.2307/3592498.
Li, Yuxin. 2002. "Hangzhou Leifengta digong de qingli 杭州雷峰塔地宮的清理 (Excavated items
from the crypt of the Leifengta in Hangzhou)." Kaogu (7): 18-20.
Li, Yuxin. 2002. "Hangzhou Leifengta yizhi kaogu fajue ji yiyi 杭州雷峰塔遺址考古發掘及意義
(Significances of the archaeological excavation at the ruins of Leifengta in Hangzhou)."
Zhongguo lishi wenwu (5): 4-12.
Li, Zongshan. 2001. Zhongguo jiajushi tushuo 中国家具史图说 (A pictorial history of Chinese furniture).
Wuhan: Hubei meishu chubanshe.
Liang, Sicheng. 2001. Yingzao fashi zhushi 營造法式註釋 (Annotations on the Yingzao fashi). Vol. 7, in
Liang Sicheng quanji 梁思成全集 (Complete works of Liang Sicheng). Beijing: Zhongguo jianzhu
gongye chubanshe.
Liang, Sicheng. 1933. "Zhengding diaocha jilue 正定調查紀略 (Brief report on the field survey in
Zhengding)." Zhongguo yingzao xueshe huikan 4 (2): 1-40.
—. 2012. Zhongguo gujianzhu diaocha baogao 中國古建築調查報告 (Survey reports on ancient Chinese
architecture). Vol. 1. Beijing: Sanlian shudian.
Liang, Sicheng, and Dunzhen Liu. 1933. "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 (3/4):
1-168.
Liang, Sicheng, and Zhiping Liu. 2007. Zhongguo jianzhu yishu tuji 中國建築藝術圖集 (Collected
illustrations of Chinese architecture). Vol. 2. Tianjin: Baihua wenyi chubanshe.
Liang, Ssu-ch'eng [Sicheng]. 1984. Chinese Architecture: A Pictorial History. Edited by Wilma Fairbank.
Cambridge: MIT Press.
Lin, Wei-cheng. 2011. "Underground Wooden Architecture in Brick: A Changed Perspective from
Life to Death in 10th- through 13th-Century Northern China." Archives of Asian Art 61: 3-36.
Lincoln, Bruce. 1991. Death, War, and Sacrifice: Studies in Ideology and Practice. Chicago: University of
Chicago Press.
Liu, Dunzhen. 1932. "Beiping zhihuasi rulaidian diaochaji 北平智化寺如來殿調查記 (Survey notes
on the Tathagata Hal of the Zhihuasi in Beijing)." Zhongguo yingzao xueshe huikan 3 (3): 1-69.
Liu, Dunzhen. 1987. Hebei gujianzu diaocha biji 河北古建築調查筆記 (Survey notes on the ancient
architecture in Hebei). Vol. 3, in Liu Dunzhen quanji 劉敦楨全集 (Complete works of Liu Dunzhen),
1-23. Beijing: Jianzhu gongye chubanshe.
305
—. 2007. Liu Dunzhen quanji 劉敦楨全集 (Complete works of Liu Dunzhen). Vol. 1. Beijing: Zhongguo
jianzhu gongye chubanshe.
Liu, Dunzhen. 1933. "Tongzhi chongxiu Yuanmingyuan shiliao 同治重修圓明園史料 (Archival
sources for the restoration projects of Yuanmingyuan in the Tongzhi Period)." Yingzao xueshe
huikan 4 (2, 3/4).
—. 1980. Zhongguo gudai jianzhushi 中國古代建築史 (History of premodern Chinese architecture). Beijing:
Zhongguo jianzhu gongye chubanshe.
Liu, Keming. 2008. Zhongguo tuxue sixiangshi 中國圖學思想史 (An intellectual history of image-making in
China). Beijing: Kexue chubanshe.
Liu, Youheng, and Ping Du. 2001. "Woguo xiancun zuizaode zhuanlunzang: Zhengding Longxingsi
Songdai zhuanlunzang qianxi 我國現存最早的轉輪藏: 正定隆興寺宋代轉輪藏淺析
(The earliest surviving revolving sutra repository in China: a brief analysis of the Northern
Song revolving sutra case at Zhengding Longxingsi)." Wenwu chunqiu 59 (3): 52-55.
Liu, Youheng, Ruiping Fan, and Ping Du. 2006. "Jin 50 nian Zhengding gujianzhu weixiu zhong
faxian de wenzi tiji chubu yanjiu 近 50 年正定古建築維修中發現的文字題記初步研究
(A preliminary study on the textual inscriptions discovered during the recent fifty years of
restoration of the ancient architecture in Zhengding)." Wenwu chunqiu (1): 43-54.
Loveday, Helen. 2000. "La Bibliotheque Tournante en Chine: Quelques Remarques sur son Role et
son Evolution." T'oung Pao 86 (4/5): 225-79.
Lu, Jimin. 2006. 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. Beijing: Xianzhuang
shuju.
Ma, Liang, ed. 1992. Yingxianzhi 應縣志 (Gazetteer of the Ying County). Taiyuan: Shanxi renmin
chubanshe.
Ma, Xiao. 2007. Zhongguo gudai mulouge 中國古代木樓閣 (Wooden towers and pavilions in ancient China).
Beijing: Zhonghua shuju.
Ma, Xiaolin. 2012. "Dachao guoshi Namodashi chongxiu Zhendingfu Da Longxingsi gongde ji zhaji
大朝國師南無大士重修真定府大龍興寺功德記劄記 (Notes on the record of the merit
of the Namodashi, State Master of the Yuan, who restored the Grand Longxingsi in
Zhending Prefecture)." Guoji hanxue yaniu tongxun (6): 252-57.
Maeda, Robert. 1975. "Chieh-hua: Ruled-line Painting in China." Ars Orientalis 10: 123-41.
Maeda, Robert. 1979. "Sung, Chin, Yuan Representation of Actors." Artibus Asiae 41 (2/3): 132-56.
Mair, Victor H., trans. 1994. Wandering the Way: Early Taoist Tales and Parables of Chuang Tzu.
Honolulu: University of Hawaii Press.
306
1993. Mixian Dahuting Hanmu 密縣打虎亭漢墓 (Han-dynasty tombs at Dahuting in Mixian). Beijing:
Wenwu chubanshe.
Muller, A. Charles. n.d. Digital Dictionary of Buddhism. http://buddhism-dict.net/ddb/index.html.
Nanjo, Bunyu. 1883. A Catalogue of the Chinese Translation of the Buddhist Tripitaka: the Sacred Canon of the
Buddhists in China and Japan. Oxford Clarendon Press.
http://catalog.hathitrust.org/Record/010249920.
Needham, Joseph, Ling Wang, and Derek J. De Solla Price. 2008. Heavenly Clockwork: The Great
Astronomical Clocks of Medieval China. Cambridge: Cambridge University Press.
Nelson, Robert S., and Richard Shiff, . 2003. Critical terms for Art History. Chicago and London:
University of Chicago Press.
Niu, Laiying. 2006. "Tiansheng yingshanling fuyuan tangling yanjiu 天聖營繕令復原唐令研究
(Study and reconstruction of the Tang statues based on the Tiansheng Statue on Builidng
and Repairing)." In Tianyige cang mingchaoben tianshengling jiaozheng: fu tangling fuyuan yanjiu 天一
閣藏明鈔本天聖令校證: 附唐令復原研究 (Correction and examination of the Tiansheng Statutes
based on the Ming edition in the Tianyige collections), 650-74. Beijing: Zhonghua shuju.
Norberg-Shulz, Christian. 1984. Genius Loci: Towards a Phenomenology of Architecture. New York: Rizzoli.
Ono, Satoshi. 1998. "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 miniature shrines in Medieval Japan)." Nihon kenchiku gakkai keigakukei
ronbunshu 日本建築学会計画系論文集 505: 191-98.
—. 2002. 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 architectural style
of palace-type miniature shrines during the middle to late Muromachi Period). Shikaban.
Pan, Guxi. 1999. Yuan-Ming jianzhu 元明建築 (Yuan-Ming architecture). Vol. 4, in Zhongguo gudai
jianzhushi 中國古代建築史 (History of Ancient Chinese Architecture). Beijing: Zhongguo jianzhu
gongye chubanshe.
Pan, Guxi, and Jianzhong He. 2005. Yingzao fashi jiedu 營造法式解讀 (Interpreting the Yingzao fashi).
Nanjing: Southeast University Press.
Peterson, Willard. 1986. "Another Look at Li." The Bulletin of Sung-Yuan Studies (18): 13-31.
Qi, Yingtao. 1992. "Monidian xinfaxian tiji de yanjiu 摩尼殿新發現題記的研究 (A study on the
newly discovered inscriptions in the Monidian)." In Qi Yingtao gujian lunwenji 祁英濤古建論
文集 (Collected essays of Qi Yingtao on ancient architecture), 106-13. Beijing: Huaxia chubanshe.
Rhie, Marylin M. 1999. Early Buddhist Art of China and Central Asia. Vol. 1. Leiden: Brill.
307
Ruitenbeek, Klass. 1996. Carpentry and Building in Late Imperial China: A Study of the Fifteenth-century
Carpenter's Manual Lu Ban Jing. E. J. Brill.
Schmid, Neil. 2006. "The Material Culture of Exegesis and Liturgy and a Change in the Artistic
Representation in Dunhuang Caves, ca. 700-1000." Asia Major 19 (1/2): 171-210.
Seckel, Dietrich. 1980. "Stupa Elements Surviving in Eastern Asian Pagodas." In The Stupa, Its
Religious, Historical, and Architectural Significance, edited by Anna Libera Dallapiccola and
Stephanie Zingel-ave Lallement. Wiesbaden: Franz Steiner Verlag.
Sekino, Tadashi, and Takuichi Takeshima. 1934. 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.
Selbitschka, Armin. 2015. "Miniature Tomb Figurines and Models in Pre-imperial and Early Imperial
China: Origins, Development and Significance." World Archaeology 47 (1): 20-44.
Shao, Xiaofeng. 2010. 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.
Shaughnessy, Edward L., trans. 1998. I-Ching: The Classic of Changes. New York: Ballantine Books.
Shen, Hsueh-man. 2012. "Shengsi yu niepan--Tang Song zhiji fojiao yu shisu muzang de jiaocuo
lingyu 生死與涅槃-- 唐宋之際佛教與世俗墓葬的交錯領域 (Where Secular Death and
Buddhist Nirvana Intersect: Secular and Religious Burials during the Tang-Song
Transition)."
https://www.nyu.edu/gsas/dept/fineart/people/faculty/Shen_PDFs/Shengsi.pdf.
Shi, Jinming, and Weilan Hai, . 2012. 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: tomb decor of the Jin dynasty from Shanxi).
Beijing: Kexue chubanshe.
Sickman, Laurence, and Alexander Soper. 1984. The Art and Architecture of China . Penguin Books.
Soper, Alexander. 1940. "Japanese Evidence for the History of the Architecture and Iconography of
Chinese Buddhism." Monumenta Serica (2): 638-79.
Soper, Alexander. 1958. "Northern Liang and Northern Wei in Gansu." Artibus Asiae 21: 131-64.
Soper, Alexander. 1947. "The ‘Dome of Heaven’ in Asia." Art Bulletin 29: 225-48.
—. 1942. The Evolution of Buddhist Architecture in Japan. Princeton: Princeton University Press.
Standen, Naomi. 2007. Unbounded Loyalty: Frontier Crossings in Liao China. Honolulu: University of
Hawaii Press.
308
Stein, Rolf A. 1990. The World in Miniature: Container Gardens and Dwellings in Far Eastern Religious
Thought. Translated by Phyllis Brooks. Stanford, California: Stanford University Press.
Steinhardt, Nancy S. 2003. "A Jin Hall at Jingtusi: Architecture in Search of Identity." Ars Orientalis
33: 76-119.
—. 2014. Chinese Architecture in an Age of Turmoil, 200-600. Honolulu: University of Hawaii Press.
—. 1997. Liao Architecture. Honolulu: University of Hawaii Press.
Steinhardt, Nancy S. 2010. "The Architecture of Living and Dying." In The World of Khubilai Khan,
edited by James C. Y. Watt, 65-73. New Haven and London: Yale University Press.
Steinhardt, Nancy S. 2011. "The Sixth Century in East Asian Architecture." Ars Orientalis 41: 27-71.
Stewart, Susan. 1993. On Longing: Narratives of the Miniature, the Gigantic, the Souvenir, the Collection.
Durham: Duke University Press.
Su, Bai. 1986. "Liangzhou shiku yiji he Liangzhou moshi 涼州石窟遺跡和涼州模式 (Remains of
Liangzhou cave temples and the Liangzhou artistic mode)." Kaogu xuebao 83 (4): 435-46.
Takeshima, Takuichi. 1971. Eizo hoshiki no kenkyu 營造法式の研究 (A Study on the Yingzao fashi).
Vol. 2. Tokyo: Chuokoron bijutsu shuppan.
—. 1944. 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.
Tanaka, Tan. 1989. Chugoku kenchikushi no kenkyu 中國建築史の研究 (Research on Chinese architectural
history). Tokyo: Kobundo.
2006. Tianyige cang mingchaoben tianshengling jiaozheng: fu tangling fuyuan yanjiu 天一閣藏明鈔本天聖令
校證: 附唐令復原研究 (Correction and examination of the Tiansheng Statutes based on the Ming
edition in the Tianyige collections). 2 vols. Beijing: Zhonghua shuju.
Tillman, Hoyt. 1995. "An Overview of Chin History and Institutions." In China Under Jurchen Rule:
Essays on Chin Intellectual and Cultural History, edited by Hoyt Cleveland Tillman and Stephen
H. West. Albany: SUNY Press.
Tokiwa, Daijo, and Tadashi Sekino. 1939-1940. Shina bunka shiseki 支那文化史蹟 (Historical remains
of Chinese culture). Vols. 1, 8. Tokyo: Hozokan.
Twitchett, Denis C., Herbert Franke, and John King Fairbank. 1994. The Cambridge History of China.
Vol. 6. Cambridge: Cambridge University Press.
Walley, Akiko. 2012. "Flowers of Compassion: The Tamamushi Shrine and the Nature of Devotion
in Seventh-Century Japan." Artibus Asiae 72 (2): 265-322.
309
Wang, Eugene. 2005. Shaping the Lotus Sutra: Buddhist Visual Culture in Medieval China. Seattle:
University of Washington Press.
Wang, Eugene. 1999. "What Do Trigrams Have to Do with Buddhas? The Northern Liang Stupas
as a Hybrid Spatial Model." RES: Anthropology and Aesthetics 35: 70-91.
Wang, Guixiang. 2011. "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 Zhongguo gudai mugou jianzhu bili yu chidu
yanjiu 中國古代木構建築比例與尺度研究 (Studies on the proportion and scale of historical
Chiense wooden architecture), edited by Guixiang Wang, Chang Liu and Zhijun Duan, 150-97.
Beijing: Zhongguo jianzhu gongye chubanshe.
Wang, Long. 1979. "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 (3):
72-77.
Watson, Burton, trans. 1993. The Lotus Sutra. New York: Columbia University Press.
Wen, Renjun. 2008. Kaogongji yizhu 考工記譯註 (The Kaogongji interpreted and annotated). Shanghai:
Shanghai guji chubanshe.
Wigley, Mark. 1993. The Architecture of Deconstruction: Derrida’s Haunt. Cambridge, Massachusetts: The
MIT Press.
Wong, Dorothy C. 2012. "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, edited by Robert Gimello, Frederic Girard and Imre Hamar,
223-60. Wiesbaden: Harrassowitz Verlag.
Wu, Hung. 2003. "Monumentality of Time." In Monuments and Memory: Made and Unmade, edited by
Robert S. Nelson and Margaret Olin, 107-32. Chicago: University of Chicago Press.
—. 2010. The Art of the Yellow Springs: Understanding Chinese Tombs . Honolulu: University of Hawaii
Press.
Wu, Hung. 2015. "The Invisible Miniature: Framing the Soul in Chinese Art and Architecture." Art
History 38 (2): 286-303.
Xiang, Yuanmu. 1991. "Sichuan Pingwu Ming Baoensi kancha baogao 四川平武明報恩寺勘察報
告 (Survey report of the Ming-dynasty Baoensi in Pingwu, Sichuan)." Wenwu (4): 1-19.
http://www.nssd.org/articles/article_read.aspx?id=1002614790#.
Xiao, Mo. 2003. Dunhuang jianzhu yanjiu 敦煌建筑研究 (A study on Dunhuang architecture). Beijing: Jixie
gongye chubanshe.
310
Xie, Yubao. 2009. "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 (4): 35-37.
Yan, Xue. 2009. "Beijing Zhihua chansi zhuanlunzang chutan: Mingdai Hanzang fojiao jiaoliu yili 北
京智化禪寺轉輪藏初探: 明代漢藏佛教交流一例 (A preliminary research on the
revolving sutra case in the Zhihua Chan Monastery in Beijing)." Zhongguo zangxue 85 (1): 211-
25.
Yang, Hongxun. 2008. "Beiwei Luoyang Yongningsi ta fuyuan yanjiu 北魏洛陽永寧寺塔復原研究
(Reconstruction of the Yongningsi pagoda in Northern Wei Luoyang)." In Jianzhu kaoguxue
lunwenji 建築考古學論文集 (Collected essays on architectural archaeology), by Hongxun Yang, 328-
41. Beijing: Tsinghua University Press.
—. 2008. Yang Hongxun Jianzhu kaoguxue lunwenji 楊鴻勛建築考古學論文集 (Collected essays of Yang
Hongxun on architectural archaeology). Beijing: Tsinghua University Press.
Yang, Zirong. 1994. "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 architecture in Shanxi)." Wenwu shijie (1):
62-67.
Yin, Guangming. 1998. "Beiliang shita shulun 北涼石塔述論 (A discussion on the Northern Liang
stone stupas)." Dunhuangxue jikan 33 (1): 87-107.
Yin, Zhenxing. 2014. "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 (2): 31-34.
Zhang, Huazhi. 2013. "Wanmin tongle: Yuan Wang Zhenpeng Longchi jingdutu 萬民同樂: 元王振
鵬龍池競渡圖 (Pleasure with the masses: the painting of the rowing competition in the
dragon pond by Wang Zhenpeng of the Yuan dynasty)." Gugong wenwu yuekan 361 (4): 48-57.
Zhang, Shiqing. 2000. Wushan shicha tu yu Nan Song jiangnan chansi 五山十剎圖與南宋江南禪寺
(Drawings of the Five Buddhist Mountains and Ten Monasteries and their connections with Southern Song
Buddhist monasteries in the Jiangnan Region). Nanjing: Southeast University Press.
Zhang, Shiqing. 1999. "Zhongri fojiao zhuanlun jingzang de yuanliu yu xingzhi 中日佛教转轮经藏
的源流与形制 (Sources and types of Buddhist revolving sutra cases in China and Japan)."
Jianzhu shilun wenji 11: 60-71.
Zhang, Xiusheng, ed. 2000. Zhengding Longxingsi 正定隆興寺. Beijing: Wenwu chubanshe.
Zhang, Yong. 1999. Fu Dashi yanjiu 傅大士研究 (A research on Fu Dashi). Dharma Drum Publishing
Corp.
311
Zhang, Yuhuan. 2003. Zhongguo chengchi shi 中國城池史 (History of Chinese cities). Tianjin: Baihua
wenyi chubanshe.
1985. Zhongguo gudai jianzhu jishushi 中國古代建築技術史 (History of the technology of ancient Chinese
architecture) . Beijing: Kexue chubanshe.
2006. 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.
2009. Zhongguo wenwu dituji: Sichuan fence 中國文物地圖集: 四川分冊 (Atlas of Chinese cultural relics:
Sichuan). Vol. 1. Beijing: Zhongguo ditu chubanshe.
Zhu, Qiqian. 1933. "Yangshi Lei kao 樣式雷考 (An examination on the lineage of Yangshi Lei)."
Yingzao xueshe huikan 4 (1): 84-89.
Zuo, Lala. 2005. "Yunyansi feitianzang jiqi zongjiao beijing qianxi 雲岩寺飛天藏及其宗教背景淺
析 (A preliminary analysis of the Feitianzang at Yunyansi and its religious background)."
Jianzhushi 21: 82-92.

Abstract (if available)

Abstract Miniature architecture proliferated in China during 1000-1200 CE. Buddhist and Daoist icons were sheltered by mini wooden pavilions, holy scriptures were stored in architectural-shaped bookcases and cabinets, whereas the interior of a worship hall—especially the vaulted ceiling—was typically ornamented with groups of tiny buildings to represent the “heavenly palace.” Specifications of miniature-making have been written into the official building standards, Yingzao fashi, promulgated in 1103 by the Northern Song imperial court. However, even though a few of these miniatures have been discussed by scholars on separate occasions, in general, miniature architecture has never received the systematic survey it deserves. Miniaturization as a culturally significant form of artistic creation, too, appears to have slipped most scholars’ attention. In fact, as my dissertation demonstrates, a critical understanding of miniatures helps to positively reshape our premises and conclusions about architecture, art, and material culture. The development of Chinese architecture from the tenth century onward could be described as a history of “progressive miniaturization”: as key structural members and ornamental elements dwindled in size and scale over time, the overall form and structure of wooden architecture also underwent drastic changes. ❧ A major concern of this dissertation is the practical, spiritual, and aesthetic reasons behind the fervor of miniature-making: what qualities made these small objects particularly appealing to people? My study reveals that religion, especially Mahayana Buddhism, played a central role in the proliferation of Chinese miniatures. I argue that the Buddhist worldview of the universe being a recursive, self-multiplying system of “worlds-within-worlds,” a concept that resonates with William Blake’s poetic imagination of “a world in a grain of sand,” has been translated into distinctive motifs of art by means of miniaturization. The decrease in size allowed a much detailed display within a limited space

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The magic mirror: representations of monsters in Chinese classical tales

PDF

The Buddhist world in modern Russian culture (1873-1919): Literature and visual arts

PDF

"East is red": a musical rarometer for cultural revolution politics and culture

PDF

Buddhist literature and gender in Korea: a preliminary survey and science fictional explorations

PDF

The many voices of the Lanting poems: society and seclusion in the Eastern Jin

PDF

"Double" in traditional Chinese fantastic fictions

PDF

Confrontation and compromise: The worlds of the supernatural and humans in Pu Song-Ling's "Liaozhai Zhiyi"

PDF

Banqueting for power: imperial women and banquets in the Tang Dynasty

PDF

Yamashiro, imagined home and the aesthetics of Hollywood Japanism: memory contained in architectural space

PDF

Family matters: women's negotiation with Confucian family ethics in Qing and Republican China

PDF

Translating Chinese poetry: Bei Dao as a case study

PDF

Factory 798: an everchanging microcosm of contemporary Chinese culture

PDF

Reconfiguring Chinese modernism: the poetics of temporality in 1940s fiction and poetry

PDF

Zhang Ailing's novellette: "Xiao'ai"

PDF

Dominant language influence in acquisition and attrition of the Chinese reflexive ziji By Chinese-English bilinguals

PDF

Romance rapscallions on the cusp of modern: male courtship in Eileen Chang stories

PDF

Making the modern scholar-priest: Buddhist universities and clerical education reform in Meiji Japan

PDF

The social life of nations: a comparative study of turn-of-the-century American and Chinese utopian fiction, 1888-1906

PDF

Mandate and manipulate: constructing political legitimacy during the Yuan-Ming transition, 1351-1370

PDF

Wong Kar-wai

Asset Metadata

Creator Luo, Di (author)

Core Title A grain of sand: Yingzao fashi and the miniaturization of Chinese architecture

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program East Asian Languages and Cultures

Publication Date 06/17/2016

Defense Date 05/06/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag 3D modeling,Buddhist art,Chinese architecture,Chinese art,digital art history,miniature,OAI-PMH Harvest,Yingzao fashi

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cheung, Dominic (committee chair), Birge, Bettine (committee member), Lee, Sonya S. (committee member)

Creator Email dluo@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-252739

Unique identifier UC11280558

Identifier etd-LuoDi-4442.pdf (filename),usctheses-c40-252739 (legacy record id)

Legacy Identifier etd-LuoDi-4442.pdf

Dmrecord 252739

Document Type Dissertation

Format application/pdf (imt)

Rights Luo, Di

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA