The Tang Architectural Icon and the Politics of Chinese
Architectural History
Nancy Shatzman Steinhardt
The discovery in June 1937 of a wooden building, the east
hall of Foguang Monastery, reliably dated to the year 857 by
two inscriptions, one on the building itself and the other on
a small, octagonal, commemorative pillar with Buddhist im-
agery, was unquestionably the crowning moment in the mod-
ern search for China's ancient architecture (Fig. 1). It was
made by China's premier architect and architectural histo-
rian, Liang Sicheng (1901-1972), his wife and research part-
ner, Lin Huiyin (1904-1955), and two other architects and
architectural historians, Mo Zongjiang (1916-1999) and Ji
Yutang (1902-ca. 1960s), all sponsored by the Beijing-based
Society (later renamed Institute) for Research in Chinese
Architecture. The final twelve-mile ascent on mule to the
lower reaches of the sacred Buddhist mountain Wutai, in
Shanxi Province, where the monastery was found, in the
seventh and last year of the society's quest for old buildings
(conditions of war made it impossible to continue their
search except in small, demilitarized pockets of China after
1937), is recorded in moving detail in Liang's personal notes
from the journey.1 The research notes and drawings made
during the seven days at the site were printed in handwritten
form in the society's last volume, distributed when Beijing
was underJapanese attack and China's major universities and
research institutes had moved to Sichuan and Hunan Prov-
inces. The formal publication appeared in 1953 and has been
reprinted in Liang Sicheng's collected works.2
The drama of the moment and the times has escaped no
one who knows the history of the study of Chinese architec-
ture or the man who, between the late 1920s and his death in
1972, was almost entirely responsible for transforming it from
a discipline grounded in careful reading and explication of
classical texts that referred to buildings and described ideal-
ized structures to the study of buildings themselves. The chief
proponent of this transformation, Liang Sicheng, was the
eldest son of one of China's most outspoken intellectuals of
the late nineteenth and early twentieth centuries, Liang
Qichao (1873-1929).3 Sicheng's birthright accorded him the
best of classical Chinese and contemporary Western educa-
tion, culminating in a Master's Degree in architecture at the
University of Pennsylvania in 1927. He went on to found
several of China's premier departments of architecture; write
thousands of pages of Chinese architectural history, most
accompanied by his own drawings; train the first generation
of China-educated modern architects and architectural his-
torians; and represent China on international design com-
missions and task forces, including the design for the United
Nations Plaza. Liang Sicheng came under strong attack for
his traditionalist views during the last years of his life, dying a
disgraced citizen of the People's Republic in 1972, just as the
Cultural Revolution was drawing to a close. Restoration of his
reputation began almost immediately afterward, eventually
resulting in his elevation to near demigod status.4
During his six years of searching, Liang "discovered" dozens
of China's pre-fifteenth-century buildings, several of them argu-
ably as pivotal in understanding Chinese architecture as the hall
at Foguang Monastery.5 The latter's aura was in part due to its
date: when found, the east hall (Fig. 1) was the only known
wooden building of the Tang dynasty (618-907), predating the
Guanyin Pavilion at Dule Monastery, found by Liang in 1931, by
127 years. Yet within a year of publication of the complete article
and photographs of Foguangsi (Monastery), a building three-
quarters of a century older was discovered on the other side of
the same mountain range, at Nanchan Monastery (Fig. 2).6 In
Liang Sicheng's lifetime, two more Tang wooden buildings, at
Tiantai Hermitage and Five Dragons Temple, were discovered
in the same province by men who had searched with Liang
during the years when the Society for Research in Chinese
Architecture had been active (Figs. 3, 4).7 By the beginning of
the twenty-first century, two additional buildings and parts of
two others had been dated to the Tang period.8
Why a building discovered sixty-seven years ago that is
neither the only Tang wooden building nor even the oldest
continues to captivate Chinese and Western attention is just
one of the issues addressed here. The impact of this exclusive
and unchallenged focus on Foguangsi East Hall is also ex-
plored. It will be argued that the spotlight on Foguangsi East
Hall and the resulting limited discussion of other Tang build-
ings is due to the fact that it was found by Liang Sicheng and
that it was precisely the kind of building he hoped to find.
Almost immediately after its discovery it became, it will be
suggested, an architectural icon, a sacrosanct structure whose
constituent parts and total image had been anticipated by
pictures of Buddhist buildings of the Tang period, by classical
writings,9 and by a building in Japan at a time when archi-
tecture earlier than the Ming and Qing dynasties (1368-
1911) was little known in China and essentially unknown to
those who aspired to write China's architectural history. Fi-
nally, the symbolic status of the east hall will be shown to be
so interwoven with the man who found it that even today
scholars cling to the myth of Foguangsi East Hall almost as
exclusively as did Liang Sicheng.
Chinese Architectural History before the Discovery of
Foguangsi
In 1932, Liang Sicheng published his first modem study of early
Chinese architecture, an article entitled "Buddhist Monasteries
and Palaces of the Tang Period We Know at This Time."1' Much
of the evidence was drawn from paintings of Buddhist paradises
that had been made known to the West and to the Chinese
scholarly community only several decades earlier through in-
trepid missions to (some might say raids of) Buddhist caves near
the modem town of Dunhuang by Sir Aurel Stein (1862-1943),
Paul Pelliot (1878-1945), and others.1" Then as now, these
THE POLITICS OF CHINESE ARCHITECTURAL HISTORY 99()
1 Foguang Monastery East Hall, Mt.
Wutai, Shanxi Province, 857 (from
Nancy Steinhardt, ed., Chinese Archi-
tecture [New Haven: Yale University
Press, 2002], fig. 4.21, courtesy of
Yale University Press; photo: unless
otherwise noted, all photographs from
books were taken by Constance Mood,
Visual Resources Collection, University
of Pennsylvania)
2 Nanchan Monastery Main Hall,
Mt. Wutai, Shanxi, 782 (photo:
N. Steinhardt)
paintings were dated to the Tang period (Fig. 5). In 1932, they
were the most tangible links to Tang architecture.
Until Liang and his colleagues organized, fieldwork was
rarely conducted at Chinese buildings. Nearly two millennia
of scholars had spent their research lives at major urban
institutions, the most prestigious in the capital, for the pre-
vious five centuries, Beijing, and its secondary counterpart,
Nanjing. The all-pervasive image of Chinese architecture con-
sisted of buildings of the Forbidden City. Those who had
ventured outside Beijing knew that the models for its impe-
rial architecture lay in fourteenth-century imperial palaces,
altars, and tombs in Nanjing and that Beijing's fifteenth-
century imperial architecture had been cloned in seven-
teenth-century palatial, sacrificial, and funerary architecture
in capitals of the last dynasty, Qing (1644-1911), in Shenyang
(Mukden), Liaoning Province, and Chengde (Jehol), Hebei
Province. In addition, it was assumed that the palace style
preserved in Beijing and the other capitals was little changed
from what one would have seen in the eleventh-century Song
capital in the city today called Kaifeng or the eighth-century
Tang capital Chang'an (today Xi'an).12 The few nonimperial
buildings outside the capitals known to scholars, such as the
Confucian temples in Qufu, Shandong Province, were of the
same architectural style as the Ming and Qing palaces.13
Other buildings were largely unknown to the educated Chi-
nese community because Chinese literati of imperial times
had considered it a hardship to be stationed anywhere less
urbane than a provincial capital, and intelligentsia of the
early twentieth century, Liang's father included, believed it
foolhardy and beneath one's dignity to ride mules through
the countryside, sleep in barns, and seek information from
peasants in pursuit of old buildings.14
Three-quarters of a century of study has passed since the
establishment of the Society for Research in Chinese Architec-
ture, but long-standing notions of Chinese architecture-that it
is in essence ahistorical and that it is a codified, formulaic system
;;- .'c.--rr r?-. rrrs- -..?.?-- -- ---c-.'.r
r-?? ? ? - -? -?.r-
\tr
r5 t; , ?Is
i 'p'V Y'JIIYV
230 AR BU I.I.ETIN JUNE 2'00 VO LUME I.XXXVI NUMBER 2
3 Tiantai Hermitage Main Hall,
Pingshun County, Shanxi, Tang
period (photo: N. Steinhardt)
4 Five Dragons Temple (Prince
Guangren Temple) Main Hall,
Ruicheng County, Shanxi, 831
(photo: N. Steinhardt)
of iconic archetypes-have not faded. These notions are con-
sistent with a perennial cultural construct of Chinese civilization
as one with supreme reverence for its past and that defines itself
according to descriptions in classical writings.15
Painting and artifacts have fit into this construct. Paintings of
court ladies in the manner of Zhang Xuan (act. 714-42) and
Zhou Fang (ca. 730-ca. 800) and horse painting of the Tang
dynasty, the period in which Foguangsi East Hall was con-
structed, experienced revivals in each subsequent Chinese dy-
nasty and in Japan, Central Asia, and sometimes even West
Asia.'1 An inscription on the back of the earliest Chinese site
plan, a bronze plate dated 323-315 B.C.E. excavated in Hebei
Province in the mid-1970s, informs us that two identical designs
were cast, one for burial in the tomb of King Cuo and the
second to be preserved in the palace so that future generations
would know how to construct a royal necropolis.17 This kind of
inscription not only reinforces the idea that later architecture
follows earlier models, it also encourages Chinese scholars to
reconstruct the architectural past based on pictures when other
physical evidence does not exist.
Supported by these kinds of examples and, of course, the
literary record, archaeology also has been used to further per-
ceptions that Chinese architecture changes little over time. Ex-
cavation in 1976 of a site near Fengchu, Shaanxi Province, dated
to the last centuries of the second millennium B.C.E., for exam-
ple, yielded a pillar-supported building complex, each of whose
main structures was elevated on a pounded-earth platform (Fig.
6).18 It is impossible to miss the similarities between this com-
plex and the Three Great Halls of the Forbidden City in Beijing
(Fig. 7), each a pillar-supported structure elevated on a shared,
-,"* - LI,!
. $ 1 a 'W.Ajf ijaC' -
THE POLITICS OF CHINESE ARCHITECTURAI. HISTORY 231
5 Buddhist paradise, painting from
the north wall of Mogao Cave 172,
Dunhuang, Gansu Province
(published with permission of the
Foreign Affairs Office, Beijing)
I-shaped platform, aligned to the four directions, and fronted by
a courtyard and gate. Even knowing that architectural recon-
structions may tend toward idealization, the three millennia
separating the Western Zhou (eleventh century-770 B.C.E.) site
and the buildings of the Forbidden City cannot but cause one to
wonder if the likeness of the drawing to the focal architecture of
Beijing's capital, the one Chinese building complex that may
deserve iconic status,19 is exaggerated.
In fact, the nonevolutionary aspect of certain elements of
the Chinese building tradition cannot be denied. And it has
served China well. The simple but profound visual similarities
one observes in Chinese architecture-elevation on a plat-
form, timber frame, decorative ceramic tile roof, four-sided
enclosure around courtyards-define its image. From Kash-
gar to Kyoto, one is aware one has entered the Chinese
sphere as soon as ceramic tile roofs projecting above low walls
come into view. Adoption of Chinese symbolic space, even in
instances when the symbolism was ignored, enhanced the
process of empire formation at all of China's borders.20
Unfortunately, the presentation of a single building as an
iconic archetype by China's first modern historian of Chinese
architecture to a civilization prone to viewing itself through
idealized, antique patterns has proved an obstruction to the
subsequent study of Chinese architecture.
Creation of the Icon
The years in which Liang Sicheng wrote his most influential
studies of early Chinese architecture coincided with the global-
ization of what for decades had been perceived in China as a
Sino-Japanese conflict, the fall of the Chinese Republic, and
transformation of China into a Communist society, which inev-
itably drew a man central to architectural history and education
in China into the debate about urban modernization. Liang
Sicheng was one of the very few Western-trained Chinese archi-
tects who continued scholarly writing during this period.
In 1941, nine years after the publication of his article on
Tang architecture mentioned above, Liang wrote "China's
6 Reconstruction drawing of a building site at Fengchu,
Shaanxi, ca. 1200 B.C.E. (from Yang Hongxun, "Xi Zhou....,"
25; published with permission of the author)
7 Model of Three Great Halls and Gate, Forbidden City,
Beijing, made in the architecture studio of Edmund Bacon at
the University of Pennsylvania, 1960s, by students T. Davis,
A. Kinsler, S. Olderman, and John Williams (photo: courtesy
of Edmund Bacon)
M x -leff -M
-TYME
I
.
-1 I -,
-/'
,
, "
,
,
, -
ft
-nq
jlb' -., .
232 ART BULL ETIN JUNE 2004 VOLUME LXXXVI NUMBER 2
8 Mogao Cave 444, front facade, dated 926 (published with
permission of the Foreign Affairs Office, Beijing)
9 Toshodaiji Kondo, Nara, 759 (photo: N. Steinhardt)
Oldest Wooden Structure," followed in 1951 by "Ancient
Chinese Architecture in the Dunhuang Wall Paintings" and
"The Toshodaiji Kondo and Tang Architecture" in 1963.21
The focus of the first, in Asia, a magazine published by the
American Asiatic Society with wide readership in the West (in
which Pearl Buck was publishing short stories at the same
time), was, of course, Foguang Monastery East Hall. In the
second, ten times the length of the first, Liang selected
eleven architectural components-halls, towers, corner tow-
ers, gates, gate-towers, arcades, pagodas, platforms, walls, for-
tified walls, and bridges-found in Dunhuang cave murals
through which he could examine early Chinese construction.
Of the twenty-five illustrations, only one was of an actual
building, Foguang Monastery East Hall. The rest were his
own line drawings of paintings he had seen in publications by
Pelliot and others. Liang justified his use of pictures by the
fact that no earlier wooden building had been found in
China.22 Among Liang's illustrations were drawings of fa-
cades in Mogao Caves 427, 431, 437, and 444 (Fig. 8), dated
between 926 and 980, or within a century of the construction
of Foguangsi East Hall.23
In the article of 1963, whose purpose was to relate aJapa-
nese building to Tang architecture, one might have expected
Liang to have discussed Nanchan Monastery Main Hall, by
then known for almost a decade,24 as well as the halls at
Tiantai Hermitage and Five Dragons Temple, published in
1958 and 1959, respectively (Figs. 2-4),25 and thereby per-
haps have reconsidered perceptions of Tang architecture he
had promulgated a quarter of a century earlier through one
building and pictures. Liang mentioned the hall at Nan-
chansi but did not include an illustration, dismissing it as not
relevant to his study due to its lack of interior pillars, roof
style, bracketing, and lack of ceiling.26 He published instead
theoretical reconstruction drawings made by colleagues of
buildings at Daming Palace, constructed by the second and
third Tang emperors in the seventh century.27 In the same
article, Liang included the plan of the Tang capital
Chang'an, presenting it as the source of the plans ofJapanese
capitals of the Nara (710-94) and Heian (794-1185) periods.
Most important, he suggested strong similarities between the
structures of Foguangsi East Hall and the Kondo (main Bud-
dha hall) of the monastery Toshodaiji in Nara (Fig. 9), the
latter constructed in 759 under the supervision of a Chinese
monk named Jianzhen (Japanese: Ganjin, 688-763), who
traveled to Japan to teach and ordain priests in the Vinaya
(Chinese: Li; Japanese: Ritsu) sect. Toshodaiji Kondo was
illustrated in the article in plan, section, and from all four
sides. Through further comparison of its bracketing and
ceiling, Liang showed it to be almost indistinguishable from
Foguangsi East Hall, noting that the hall in Japan had a date
eighty-eight years earlier than the east hall but deemphasiz-
ing the significance of the early date due to the structural
similarities.28 Liang did not explain how city plans or draw-
ings of palaces dated a century earlier than the Japanese hall
helped one understand a ninth-century Chinese Buddhist
hall. By inference one can assume that Liang believed that
structural features were shared by palatial and religious ar-
chitecture in China throughout the Tang period, and that he
had included images of older city planning to verify the
influence of Chinese architecture on Japanese architecture.
To strengthen his presentation of an equation between the
east hall, Nara-period architecture in Japan, and Tang archi-
tecture more generally, Liang included a line drawing of a
Buddhist hall engraved on a stela at the lintel of a ground-
story entrance to Great Wild Goose Pagoda in Chang'an,
dated to the early Tang period, which he had also published
in his article of 1941 (Fig. 10).
The last article on Tang architecture was the first in which
Liang used drawings that had not been made at the Institute
for Research in Chinese Architecture. It is thus all the more
noteworthy that having decided to present research that was
neither his own nor directly associated with him, he still
eclipsed the importance of any Tang Buddhist building other
than the east hall. For someone who had used pictorial
representations for information about architecture, it is,
moreover, hard to imagine that the visual similarities between
Nanchansi main hall and the painting in Mogao Caves 437
and 444 (Figs. 2, 8) or the facades of Caves 427 and 431-the
latter four all published in his article of 1951-had escaped
him. Even though the article of 1963 was written for a com-
memoration of the life of the Buddhist mbnk Jianzhen, the
selection and omission of buildings in order to emphasize the
new, increased importance of Foguangsi East Hall as a me-
dium through which Liang postulated that one could also
THE POLITICS OF CHINESE ARCHITECTURAL IISTORY 233
understandJapanese Buddhist architecture is hard to ignore.
By 1963, not only had Foguangsi East Hall become the single
showcase of Tang architecture in the eyes of China's premier
architectural historian, but also the focus on comparing it
with Toshodaiji Kondo, we shall see below, similarly pre-
sented a skewed perception of both earlier Chinese Buddhist
architecture as well as of the potential importance of other
early Japanese wooden Buddhist buildings to the study of
early Chinese architecture.
Two factors magnified the eventual influence of Liang's
studies of Tang