The National Central Museum in Nanjing (1935–48) was co-designed by the most distinguished architectural historian in twentieth-century China—Liang Sicheng. It has long been regarded, however, as a representative work of “conservative revivalism” in modern Chinese architectural history. Idealizing a Chinese Style: Rethinking Early Writings on Chinese Architecture and the Design of the National Central Museum in Nanjing attempts to demonstrate the creativity of the design process, Lin Huiyin, and the architects’ ideal for a Chinese-style modern architecture. This ideal, Delin Lai argues, is profoundly rooted in the expectation of Chinese intellectuals for a “Chinese renaissance,” for which the Chinese architectural past was studied, evaluated, and more importantly, redefined through a dialogue with the contemporary architectural discourses and historiography formed in the West. The National Central Museum epitomizes this search for an ideal.

In 1946, Liang Sicheng (1901–72), the most distinguished architectural historian of twentieth-century China, wrote:

The Chinese building is a highly “organic” structure. It is an indigenous growth that was conceived and born in the remote prehistoric past, reached its “adolescence” in the Han dynasty (around the beginning of the Christian era), matured into full glory and vigor in the Tang dynasty (seventh and eighth centuries), mellowed with grace and elegance in the Sung (Song) dynasty (eleventh and twelfth centuries), then started to show signs of old age, feebleness, and rigidity, from the beginning of the Ming dynasty (fifteenth century). Now, with the coming of reinforced concrete and steel framing, Chinese architecture faces a grave situation. Indeed, there is a basic similarity between the ancient Chinese and the ultramodern. But can they be combined? Can the traditional Chinese structural system find a new expression in these new materials? Possibly. But it must not be the blind imitation of “periods.” Something new must come out of it, or Chinese architecture will become extinct.1

Did he have an answer, no matter how tentative, to this problem of style, new materials, and Chinese architecture? If so, it may lie in the National Central Museum (Figure 1) in Nanjing, which was designed a decade before he posed the problem.

Figure 1

Su, Yang & Lei Architects, Liang Sicheng Consultant, National Central Museum, Nanjing, 1935–48 (author’s photo)

Figure 1

Su, Yang & Lei Architects, Liang Sicheng Consultant, National Central Museum, Nanjing, 1935–48 (author’s photo)

The National Central Museum was one of the most important buildings created in the early decades of Republican China.2 The product of one of the few architectural competitions held during the period, it had a self-consciously Chinese design, modeled on that of the Liao and Song dynasties, and reflected the historicist thinking of its design advisor, Liang Sicheng. It was the first building design in which Liang participated, and it remains his most significant work because of the “big roof” (da wuding) controversy later in his life.3

Owing to the building’s importance in the history of China’s Republican period, its architectural style has prompted a number of scholarly speculations. Present-day historians who try to explain the designers’ seemingly thorough grounding in Liao and Song architecture tend to view the building’s style as one with a revivalist underpinning. For example, Zhongguo Jianzhu Shi (History of Chinese architecture), the textbook of Chinese colleges, identifies the style of the museum as “Gongdian shi” (palatial style).4 The author of the entry “The Central Museum in Nanjing” in the encyclopedic book Ershi Shiji Zhongguo Jianzhu (Architecture of twentieth- century China) labeled it “fanggu jianzhu” (architecture that imitates the ancient).5 Although these characterizations effectively summarize a key approach to Chinese style architectural design in the twentieth century, they are also misleading. Not only do they imply that the design of the museum passively “imitated” traditional Chinese architecture, an assumption that leads them to a simplistic interpretation of this building, but, more important, they also overlook the efforts of Liang and his colleagues to pursue an architectural idiom that is simultaneously modern and uniquely national.

This study addresses two key questions. How did Liang cope with issues of national identity and modernity in architecture as a Chinese architectural historian and an architect trained in the Beaux-Arts tradition? And what was his conception of modern Chinese culture as a member of the elite during the early Republican period? By carefully identifying the sources of the museum’s design and analyzing the relationship between the design and early writings on Chinese architecture, this article demonstrates the creativity of the design response, particularly in relation to Liang and the architects’ ideal of a Chinese-style modern architecture. This ideal, I argue, is profoundly rooted in the expectation on the part of Chinese intellectuals in the early decades of the twentieth century of a Chinese cultural renaissance. To achieve that goal, they studied China’s architectural past, evaluated it, and, more significant, redefined and justified it by putting it into a dialogue with contemporary architectural discourses and historiography formulated in the West, as well as with criticism of Chinese architecture by Western scholars and the contemporary Chinese public. Seen in this light, the National Central Museum in Nanjing, the most important building dedicated to a cultural institution in the first three decades of the Republic of China, emerges as the epitome of this ideal.

The Design of the National Central Museum

In July 1934 the Ministry of Education authorized the formation of the Preparatory Committee of the National Central Museum (PCNCM). The members of the committee included two geographers, an artist, a historian, a law professor, an economist, an archaeologist, and a physicist, as well as Liang Sicheng.6 Liang, as the only architect and architectural historian, was appointed the “special member” (zhuanmen weiyuan) of the committee and the “project consultant” (gongcheng guwen), responsible for “reviewing all the affairs associated with the project and drafting all the guidelines and contracts.”7 In June 1935 the committee held a limited competition for the design of the museum. Twelve of the thirteen renowned Chinese architects who were invited submitted a proposal.8

Liang had received his master of architecture degree from the School of Fine Arts at the University of Pennsylvania in 1927. He went on to Harvard University to study Chinese architectural history, planning to pursue a doctorate in art history. Dissatisfied with the contemporary treatment of the subject by Western scholars, however, he returned to China in 1928 and established the Department of Architectural Engineering at the National Northeastern University in Shenyang. In 1931 he moved to Beiping (now Beijing) to join the Society for Research in Chinese Architecture (Zhongguo Yingzao Xueshe, SRCA) with his wife, Lin Huiyin (Phyllis Whei-yin Lin, 1903–54), who in 1927 received a bachelor of fine arts degree from Penn as well.9 The SRCA, established in 1929 by Zhu Qiqian (1872–1964), a retired scholar-official in early Republican China, until the 1950s was the only institution in China that studied architectural history. To reform a methodology that had been grounded in the traditional compilation and interpretation of Chinese classical texts, Zhu asked Liang and Liu Dunzhen (Liu Tun-tseng, 1897–1968), an architect who had graduated from the Tokyo Polytechnic in 1921, to spearhead the work of the society by carrying out a series of field surveys to search for the remains of ancient Chinese structures. By the time of the museum competition, their investigation had covered about fifty counties in Hebei, Shanxi, Henan, Zhejiang, and Shandong provinces. In 1934 Liang published Qingshi yingzao zeli, a book on Qing dynasty construction methods.10 Meanwhile, he had also been studying the Song-dynasty architectural manual Yingzao fashi (YZFS) during these years.11 Another work that he had compiled with a student, titled Jianzhu sheji cankao tuji (Pictorial references for architectural design), a collection of details of historical buildings intended for use in the new Chinese-style architectural design, was to be published in November 1935.12

The competition guidelines written by Liang required that the museum “fully employ [the style of] Chinese architecture” “as long as doing so does not affect the function of a modern museum” and providing that it be “as simple and spacious as possible.”13 Implicit in these conditions was a concern that modern function and Chinese style might conflict. Designers were challenged to avoid that conflict.

The proposals were reviewed on 4 and 5 September by a jury consisting of the client’s representative, the chairman of the PCNCM, a representative of the architectural committee, Liu Dunzhen, and Liang. The results of the competition were announced on 6 September.14 None of the proposals dealt thoroughly with the issues of site, function, and architectural form that the committee had expected them to address. The committee reluctantly chose the design by the University of Michigan graduates Xu Jingzhi (Su Gin-Djih) and Li Huibo (Lei Wei-Paak), of the architectural firm Su, Yang & Lei Architects, who rendered the main hall of the museum as an eleven-bay corridor with a Qing-style hip roof (Figure 2). This design, as the report stated, was “the best one among all the proposals in meeting all the requirements and in showing the potential for revision.” Liang, as the “special member” of the museum’s construction committee, responsible for architectural affairs, was appointed architectural consultant, to guide further modification of the design.15

Figure 2

Su, Yang & Lei Architects, “Winning design of National Competition for the National Central Museum, Nanjing, 1935—Chinese Architecture of Ch’ing (Qing) Style” (plate 138 of Chinese Architecture—Past and Contemporary [1964])

Figure 2

Su, Yang & Lei Architects, “Winning design of National Competition for the National Central Museum, Nanjing, 1935—Chinese Architecture of Ch’ing (Qing) Style” (plate 138 of Chinese Architecture—Past and Contemporary [1964])

Liang’s name appeared on major working drawings as the “consultant” (guwen),16 but he must have done more than offer suggestions. Chen Mingda (1915–97), a research assistant in the SRCA, recalled in 1988 that “the architect of the Nanjing Museum (originally the Central Museum) was Xu Jingzhi. Because the design had to include the big roof [of traditional Chinese architecture], and he was ignorant [of that tradition], he came to us. At that time, we as the consultants were very honest and did a lot of substantial work and quite a large number of drawings, unlike those nowadays who are called consultants but do nothing.”17 Chen’s statement reveals the extent of Liang’s involvement in developing the design, which is affirmed by analysis of its details, especially those of the main hall in the museum compound.

Liang’s efforts were instrumental in changing the architects’ design of the main hall. The new design, with nine bays, is in what Liang described as the “Liao and Song Style (Liao Song xingshi)”18 and the architects called the “Liao or early Song style” (Figure 3),19 referring to the architecture of coexistent dynasties of the tenth to the twelfth century, Liao (907–1125) in the North and Song (960–1279) in the South. The construction of the reinforced-concrete structure started in June 1936. Interrupted by the Sino-Japanese War from 1937 to 1945, it was not completed until 1948.20

Figure 3

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the elevation of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 3

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the elevation of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

The style of the building deserves closer scrutiny, given that it was not adopted detail for detail from any historical structure. By the time Liang and his colleagues at the SRCA participated in designing the museum, they had carefully studied a substantial number of buildings dating from the mid-Song and Qing dynasties, as well as two famous construction manuals, the YZFS of the Song (Figure 4) and the Gongbu gongcheng zuofa zeli of the Qing. The style of Liao and early Song dynasty architecture did not exist as a composite style historically but was assembled and classified by SRCA scholars from the historical structures they discovered and identified. Why did Liang and the architects invent a new style instead of simply using either the Qing or mid-Song style?

Figure 4

Liang Sicheng, “Sung (Song) dynasty rules for structural carpentry” (plate 7 of A Pictorial History of Chinese Architecture [1984])

Figure 4

Liang Sicheng, “Sung (Song) dynasty rules for structural carpentry” (plate 7 of A Pictorial History of Chinese Architecture [1984])

To answer this question I follow Liang’s approach to architectural investigation, analyzing the characteristics of the composition and structural elements of the museum by re-examining the data on working drawings. I then compare this information with data on the ancient structures, especially those of the Liao and Song dynasties that Liang and the architectural historians who were his contemporaries investigated, to identify their approach to these models.

The Sources of the Liao-Song Style of the National Central Museum

The main hall of the museum stands on a high platform. It is a ten-column, or nine-bay, corridor building (two side bays were walled in to create rooms) under a big curved hip roof. The roof is covered with yellow glazed tiles; the columns are painted vermilion, more reminiscent of Ming (1368–1644) and Qing (1644–1919) palatial buildings than of existing Liao and Song structures. Yet on the elevation the subtly turned-up central ridge, the fishtail ends of the top roof ridge (chiwei), the shuttle-shaped columns (suozhu) with their slight entasis, the increased height of the columns toward either end of the building (shengqi), as well as the battered corner columns (cejiao, or ch’e-chiao, in Figure 4), are all characteristics of Liao and Song architecture.

A building nine bays wide corresponds in rank to a palatial hall (diange), the highest-ranked structure in YZFS.21 In 1935 only two buildings of that rank dating before the twelfth century had been discovered. One was Daxiongbao Hall, the main hall of Fengguo Monastery (1020), built in the early Liao dynasty in Yi county, Liaoning province; it had been investigated by the Japanese architectural historian Tadashi Sekino (1868–1935) in 1932.22 The other was the main hall, also named Daxiongbao Hall, of Huayan Monastery (1140) in Datong, Shanxi province, which was rebuilt in the early Jin dynasty (1115–1234) and had been investigated by Liang and Liu in 1933.23 The museum’s nine-by-five-bay plan resembles the plans of the main halls at the two monasteries, suggesting that the museum’s design might reference the two largest existing examples to declare, by association, the museum’s status as an important state building. Although Sekino’s report was available to Liang at the SRCA (which actively exchanged publications with Japanese architectural historians), it is likely that the main hall of Huayan Monastery in particular was Liang’s and the architects’ model. Not only is there no record that Liang had ever visited Fengguo Monastery, but also information on the measurements in Sekino’s report was incomplete,24 and, most significant, the museum has more proportional commonalities with Huayan hall than with the hall at Fengguo. For instance, the plans of Huayan hall and the museum both have a 2-to-1 ratio of length to width, whereas Fengguo’s is 1.9 to 1; both have columns with a similar 10-to-1 ratio of height to diameter, whereas Fengguo’s columns have an 8.9-to-1 ratio (information Sekino’s report does not actually provide). Moreover, despite the similar width of the central bays of the museum and the main hall of Fengguo Monastery, the museum overall has the same proportions as the Huayan hall: the width of the museum is 83 percent of the width of the Huayan hall, and the height of the museum’s two central columns is also 83 percent of the height of those in the Huayan hall. In short, the three basic dimensions of the museum, namely, the width, depth, and height of the front corridor, seem to be derived from the Daxiongbao Hall of Huayan Monastery.

A remarkable feature of the museum plan is the elimination of four columns at the center of the grid (Figure 5). Liang, who in his writings named this arrangement the “column-eliminating method” (jianzhu zhi fa), considered it an important stylistic characteristic of Liao and Song architecture. This method was applied not only to the main halls of the Fengguo and Huayan monasteries, but also to other contemporary structures, such as Sandashi Hall, the main hall of Guangji Monastery (1025) in Baodi county, and Daxiongbao Hall (1060) and Sansheng Hall (1128–43) of Shanhua Monastery in Datong.

Figure 5

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the plan of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 5

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the plan of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Outside the main hall of the museum, a wide platform (yuetai) ends at a stairway that descends to a lower platform to the south (see Figures 1 and 16). Liang noticed in his writings that almost all the Liao and Song monuments mentioned above have a yuetai in front, so that its use in the museum made it another stylistic icon. (It also functioned as a roof over the offices beneath it.) Because no balustrades survived for the existing platforms of those ancient monastery halls, the designers looked elsewhere for a model, copying part of a stone balustrade from the Sarira Pagoda of the Five Dynasties period (907–960), which was prior to the Northern Song dynasty, at Qixia Monastery in Nanjing. Excavated by Liu in 1930, this stone relic constitutes the oldest material remains of the balustrade panel with L-shaped patterns,25 whose wooden counterpart was found in Dule Monastery (984), the earliest wooden structure Liang investigated before he discovered the main hall of Foguang Monastery (857) of the Tang dynasty (618–907) in 1937.

Connecting the columns of the museum are lan’e lintels (see 23 in Figure 9) and a long horizontal pupaifang tie beam resting on lintels. Lintel and tie beam combine to form a T-shaped cross section (Figure 6), a structural method visible in many monuments of the Liao, Song, and Jin dynasties (1115–1234), the last of which succeeded the Liao. Liang regarded the straight-cut end of the lintel that protrudes from the top of the side columns, however, as a typical Liao method. But a salient characteristic of the lintel, in addition to its stylistic details, is the proportion of its section (1:2), very close to that of Sandashi Hall of Guangji Monastery (1:2) in Baodi county,26 the Front Gate of Dule Monastery (2:5),27 Bojiajiaozang (Bhaghavat) Sutra Repository, commonly known as the Sutra Library (2:5), as well as Haihui Hall (1:2) of Lower Huanyan Monastery in Datong, all of which Liang investigated.

Figure 6

Corner bracket, lan’e lintel and pupaifang tie beam of the National Central Museum, Nanjing (author’s photo)

Figure 6

Corner bracket, lan’e lintel and pupaifang tie beam of the National Central Museum, Nanjing (author’s photo)

On the pupaifang tie beam are brackets, which are called puzuo in YZFS and dougong in the Qing dynasty and in contemporary China (Figure 7). As in most Liao and Song buildings, three types of bracketing support the museum’s eave: the pillar-top bracketing (zhutou puzuo), the intercolumnar bracketing (bujian puzuo), and the corner-pillar bracketing (zhuanjiao puzuo). As the key element in Liang’s study of Chinese architectural history, dougong was an indicator of the rank of a building, with the dimension of a bracket arm functioning as the basic module for carrying out design and construction. Most important, Liang agreed with Osvald Sirén (1879-1966), a Swedish art historian whom I introduce later in this article, that the change of dougong’s scale from large to small and its function from structural to decorative reveals the stylistic transformation of Chinese architecture following the Liao and Song period.

Figure 7

Su, Yang & Lei Architects, Liang Sicheng Consultant, Working drawing of the pillar-top bracket of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 7

Su, Yang & Lei Architects, Liang Sicheng Consultant, Working drawing of the pillar-top bracket of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

According to YZFS, all the arms of a puzuo set share a section with a standard measurement that is the module cai (ts’ai in Figure 4), or timber unit, in a building’s design and construction. This early manual prescribed eight sizes, or grades, of cai for various ranks of building, from palatial hall to small kiosk. Although the museum’s nine-bay-wide front elevation is similar to that of the main halls of Fengguo and Huanyan monasteries, in absolute size or, better, the size of the element in comparison with that of the same element in other buildings, the museum’s cai is smaller than that of either monastery. The monasteries’ cai measure 30 by 20 (Fengguo) and 29 by 20 centimeters (Huanyan), whereas the museum’s are 26 by 16.5 centimeters (or 10.25 by 6.5 inches, American standards having been adopted in Chinese building construction before 1949). If the cai used by two ancient models ranks first among eight in accordance with YZFS, the museum falls between the second and the third rank, which in the Song dynasty would be appropriate only for palatial-rank buildings of three to seven bays. In relative size, or the size of this element in comparison with that of other elements in the building, however, the museum’s cai is bigger than that of Huayan Monastery. Although the plan of the museum and the height of its corridor are 83 percent of the corresponding measurements of the monastery’s main hall, the cai used by the museum is 87 percent of the hall’s cai. Three Liao buildings with cai of similar measurements might have been the models for Liang and the architects. They are the seven-bay-wide Daxiongbao Hall of Shanhua Monastery, the five-bay-wide Sansheng Hall of the same monastery, and the three-bay-wide Front Gate of Dule Monastery. Among them the cai of the Sansheng Hall is the same as that of the museum, which is 26 by 16.5 centimeters. But in making the zhi (chih in Figure 4), a subsidiary unit that equals the distance between two perpendicularly adjacent bracket arms, the museum might have followed the example of the Front Gate of Dule Monastery because both the gate and the museum share the same ratio of cai to zhi, 2:1.28

Not only the size but also the number of elements in the bracket mattered in determining hierarchy. In the main hall of Fengguo Monastery, a seven-puzuo bracket with four upward projections (tiao) was used as pillar-top bracketing in correspondence with the significance, or rank, of the building. The museum, however, uses a five-puzuo bracket with only two upward projections, traditionally used for a building of the government-office-hall (tingtang) class. Although the main hall of Huayan Monastery also uses the five-puzuo pillar-top bracketing, the museum lacks the lateral bracket arm that in the case of the Huayan hall crosses through the first projected arm. The museum’s simplified combination is called touxin, literally “stolen heart,” in contrast with the standard jixin, “added heart,” in YZFS. Among the Liao and the early Song buildings that Liang investigated, the same type of pillar-top bracketing used in the museum appears on both the five-bay Dacheng Hall of the Confucius Temple in Zhengding, Hebei province, and the low-rank, three-bay Front Gate of Dule Monastery (Figure 8). The pillar-top brackets of the Front Gate and of the museum appear similar, with the measurement of each arm in the museum and the length of its projection very close to those of the corresponding elements of the Dule gate, judging by the measurement system of fen, that is, one-fifteenth of the height of cai, in YZFS. In contrast, the Dacheng Hall of the Confucius Temple might have less relation to the museum design than the gate because Liang’s reports and writings reveal no detailed data or measured drawings of it.

Figure 8

Liang Sicheng, “Plan and cross section” of the Front Gate of Dule Monastery, Ji county, 984 (plate 26c of A Pictorial History of Chinese Architecture [1984])

Figure 8

Liang Sicheng, “Plan and cross section” of the Front Gate of Dule Monastery, Ji county, 984 (plate 26c of A Pictorial History of Chinese Architecture [1984])

Liang regarded the combination of the pillar-top bracket set with the column as “the Chinese Order” (Figure 9). He was particularly sensitive to the proportion of the height of a bracket set to the height of its column. In the Front Gate of Dule Monastery, this proportion is 1:2.5, and in the main hall of Huayan Monastery, 1:3.5. Although the Huayan hall and the Dule gate apparently served as models for the museum’s plan, corridor, and bracketing designs, the proportion of the height of its pillar-top bracket to the height of its column is equivalent to that of neither of the two models. It is 1:3 and identical to that of the main hall of Shanhua Monastery and the Sutra Library of Huayan Monastery. This proportion was achieved by using a relatively bigger cai than that used in the main hall of Huayan Monastery.

The intercolumnar bracketing (bujian puzuo) and corner-pillar bracketing (zhuanjiao puzuo) of the museum are also similar to those of the Dule gate, but are much simpler. The intercolumnar bracketing of both buildings consists of a very short pillar (shuzhu) under a cap block (ludou or lu-tou, 20 in Figure 9) that holds a tie beam (fang, 4, 5, 6 in Figure 9) with the transverse bracket arm (nidaogong or ni-tao-kung, 15 in Figure 9) in relief. Without using the projecting arms that were parts of the intercolumnar bracketing of the gate, however, the bracketing in the same position in the museum (Figure 10; see Figure 16) is in fact even closer to that of a lower-ranked building, Haihui Hall of Huayan Monastery. In addition, although the corner-pillar bracketing of the museum is no doubt derived from that of the Dule gate, it is simpler because it eliminates the diagonal arm cutting across the corner horizontally at a 45-degree angle (xiegong) (compare Figures 6 and 19).

Figure 9

Liang Sicheng, “The Chinese ‘Order’ ” (plate 2 of A Pictorial History of Chinese Architecture [1984])

Figure 9

Liang Sicheng, “The Chinese ‘Order’ ” (plate 2 of A Pictorial History of Chinese Architecture [1984])

Figure 10

Corridor and bracketing of the National Central Museum, Nanjing (author’s photo)

Figure 10

Corridor and bracketing of the National Central Museum, Nanjing (author’s photo)

In the corridor the tail of the upper bracket arm (huagong or hua-kung, 19 in Figure 9) turns into a two-rafter-long beam (rufu, 22 in Figure 9), exposed under the ceiling and above the lan’e lintel (Figure 11, upper structure of the corridor). Connecting the beam and the lintel is a single-arm bracket that rests on a camel’s-hump strut (tuofeng). The interior bracket sets of the museum followed the same method. It is usual for the end of a bracket arm to change to a beam or lintel; that transition can be seen in such Liao buildings as Guanyin Pavilion of Dule Monastery, the Sutra Library of Huayan Monastery, and Sandashi Hall of Guangji Monastery. But it was unusual to combine a camel’s-hump strut with the bracketing, especially the intercolumnar bracketing on the beam across the central bay on the section (Figure 11, upper structure of the central part of the hall), which has two blocks (dou) under the strut. This combination is clearly derived from the Sandashi Hall of Guangji Monastery. Exposing the ceiling frame, a method which was called cheshang luming zao in YZFS, the Sandashi Hall was unique among the Liao and Song structures Liang discovered. (Compare the intercolumnar bracket sets in Figures 10 and 17.)

Figure 11

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the section of the National Central Museum; the upper structure of the corridor and the central part of the hall show the influence of Sandashi Hall of Guangji Monastery (see Figure 17) (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 11

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the section of the National Central Museum; the upper structure of the corridor and the central part of the hall show the influence of Sandashi Hall of Guangji Monastery (see Figure 17) (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Originally the design of the central part of the museum’s hall contained an octagonal corncob-shaped sunken ceiling form. It was termed zaojing, literally “decorative well,” in the Song manual. Yet the museum was modeled, not on the type prescribed there, which is called douba zaojing, a design with exuberant bracket sets, but after the much simpler Liao examples in Guanyin Pavilion and the Sutra Library of Huayan Monastery.29

A simple hip roof (si’e ding) rests on the bracket sets, beams, and lintels, supported by steel trusses concealed by the ceiling. While the fishtail ends of the roof ridge were modeled after those of the Front Gate of Dule Monastery, the two side slopes of the museum’s roof demonstrate the method in YZFS known as outward-pushing ridge (tuishan) (Figure 12). That method changes the curvature of the upper part of each side slope (which should match that of the front and rear slopes), by elongating the roof ridge and turning the side ridges from straight lines into hyperbolas. None of the hip-roof buildings of the Liao and Song period that Liang investigated, such as Sandashi Hall of Guangji Monastery and the main halls of Huayan Monastery and Shanhua Monastery, demonstrates this subtle measure. Nor could the Liao dynasty main hall of Kaishan Monastery (1033) in Xincheng in Hebei province be Liang’s model, for although Liu visited this temple in October 1936,30 he did not record its tuishan feature, which was discovered by another scholar in the 1950s.31 Therefore, Liang must have learned this method from the Song manual or even from the Qing examples.32

Figure 12

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the roof that shows the tuishan method for creating the curve of the four corner ridges of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 12

Su, Yang & Lei Architects, Liang Sicheng Consultant, working drawing of the roof that shows the tuishan method for creating the curve of the four corner ridges of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

As in an equilateral triangle, the height of the roof, or the distance between the ridge purlin (jituan) and the “base,” or the line connecting the front and rear purlins, was called ju (chü) (raise) in YZFS. In the museum, ju measures 1:3.87 of its base. This ratio is close to one-to-four, the average in Liao and early Song buildings, as Liang discovered.33 For instance, in the case of the Sutra Library of Huayan Monastery, this ratio is 1:4.3; in the Front Gate of Dule Monastery, it is 1:3.99. The ratio of the museum is almost the same as the 1:3.81 of Sandashi Hall of Guangji Monastery, whose roofline, as Liang believed, followed, almost to the letter, the prescription of YZFS and was closer than in other buildings to the instruction of the Song manual.34

Unlike any of the aforementioned Liao and early Song models in which rooflines are but slightly concave, the museum’s appears to have a pronounced curvature. Even in comparison with Sandashi Hall, the museum’s roofline is extreme in its concavity. According to YZFS, for instance, the height of each purlin between the roof ridge and the eave should equal that of purlin’s imagined position on a straight line connecting the purlin above it and the eave purlin, minus the rate of depression (zhe or cheh), which is 1/10, 1/20, 1/40, 1/80, and so forth, of ju from high to low (see “Raise & Depress” in Figure 4). In the museum, however, these rates are 1/7.7, 1/14, 1/21, and 1/53 (Figure 13).

Figure 13

Su, Yang & Lei Architects, Liang Sicheng Consultant, Working drawing on the curvature of the roof of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Figure 13

Su, Yang & Lei Architects, Liang Sicheng Consultant, Working drawing on the curvature of the roof of the National Central Museum (Nanjing Museum and the Nanjing Archives of the Municipal Constructions, Nanjing)

Liang and the architects who worked with him had modeled the museum they envisaged (aside from its colors, which were those of Ming and Qing architecture) on the structural and compositional elements of famous buildings dating from the Five Dynasties of the tenth century to the early Jin dynasty of the twelfth—those of Qixia Monastery, Dule Monastery, Guangji Monastery, Huayan Monastery, and Shanhua Monastery—and on the methods of the Song architectural manual YZFS, published in 1103. They elaborated some traditional models but simplified others. In fact the effect they sought to achieve depended more on size, proportion, and color than on the bracketed elements that contravened hierarchical associations (see the  Appendix).

The Design of the Museum and Early Writings on Chinese Architecture

That the design of the National Central Museum resulted from a selection of elements and a meticulous synthesis of ancient building practices indicates Liang’s dominant role in finalizing the design. But questions remain. Why did he adopt such a complicated approach rather than simply follow one of the two familiar sources—the Song or the Qing manual? How significant was the style of Liao and early Song architecture to him? None of the design documents that have survived can explain how and why Liang and the architects made their stylistic decisions, nor do these documents reveal the opinions of other judges of the competition or other members of the PCNCM. Possible answers to the questions might be found in the narrative of Chinese architecture constructed by Liang and his colleagues at the SRCA, especially his wife, Lin, who not only participated in some of Liang’s important field investigations but also contributed to or co-authored some of his works.

Right from the start, Liang and Lin’s writings on Chinese architecture were far more than records of their investigations; rather they demonstrated a great effort to authenticate Chinese architecture in the global and modern context. In other words, they were intended as responses to the dominant Western view and to popular Chinese criticism of Chinese architecture.

Chinese architecture entered the purview of the West in the seventeenth century. Western patrons and architects during the Enlightenment admired aspects of Chinese architecture, as reflected in the chinoiserie designs of the mid-eighteenth century. In the following century, however, especially after Britain defeated China in the Opium War of 1839–42, Chinese architecture fell into disrepute in the Western narrative of architectural history. Two of the most influential works on architectural history, James Fergusson’s History of Indian and Eastern Architecture (1876) and Banister Fletcher’s History of Architecture on the Comparative Method (1896), contributed to this view in important ways. Fergusson, who described Chinese architecture as an enigma because it lacked the monumentality of Western architecture, noted, “There really are no buildings in the country worthy of the people or their civilization.” Believing that color was inferior to pure form in architecture, he wrote,

The Chinese are the only people who now employ polychromy as an essential part of their architecture; so much so, that color is with them far more essential than form; and certainly the result is so singularly pleasing and satisfactory, that for the lower grades of art it is hardly doubtful that it should always be so. It is almost as certain that, for the higher grades of art, color, though most valuable as an accessory, is incapable of the same lofty power of expression which form conveys to the human mind.35

In the fourth edition of his widely circulated textbook (published in 1901), Fletcher placed Chinese architecture in the category of “non-historical styles.” His term echoed Hegel’s “unhistorical history,” a concept the philosopher applied to China and India because he discerned in them neither historical development nor enhancement of rationality over time.36

Liang and Lin’s view of Chinese architectural history countered such Eurocentric views. Lin, in her first academic paper, “On Characteristics of Chinese Architecture,” criticized both Fletcher and Fergusson, without mentioning either historian’s name:

Chinese architecture, though complicated in structure and delicate in art, exhibits a kind of simplicity and frugality in appearance. This is why commoners often misunderstand it as primitive, undeveloped, and inferior, as well as naïve in comparison with the architecture of other systems. This erroneous view originated in the slapdash observations of Westerners and their shallow conclusions. It has influenced the Chinese, making them extremely skeptical and even scornful toward the art of their own country. … The beauty of [Chinese] architecture by no means lies in superficial polychromy and ornaments, or special features. Rather it resides in the basic structural principles that generate this beauty and the incontestable Chinese understanding of the principles for controlling decoration.37

As the passage quoted at the beginning of this article shows, Liang supported this view of the structural significance of Chinese architecture, praising the organic centrality of the building system the Chinese had elaborated. He viewed the structural system, visibly indexed in the external treatment of buildings, as a key characteristic of Chinese architecture: “The study of the Chinese building is primarily a study of its anatomy. For this reason the section drawings are much more important than the elevations. This is an aspect quite different from the study of European architecture, except perhaps the Gothic in which the construction governs more of the exterior appearance than in any other style.”38

In laying out the principles of Chinese architecture, Liang and Lin first argued that Chinese architecture was a xi (system) and constituted a distinctive achievement among the systems of architecture in the world. Second, they viewed the frame structure as an essential characteristic of Chinese architecture, one that it shared with Gothic architecture as well as modern architecture of steel and concrete frame. Third, they saw the bracket-arm set (dougong) as a basic module in Chinese architectural design, its role similar to that of the orders in Greco-Roman architecture. Fourth, they understood that some of Chinese architecture’s unique features, such as the bracket-arm set and the concave roof, had been created for functional purposes. Fifth, they claimed that an architectural style could be measured by the accuracy with which it exhibited its structural system and that the styles exhibiting such systems could help identify the process by which Chinese architecture proceeded from an early period to a late one, and from maturation to decline.39

Liang and Lin probably adopted the concept of xi from both the modern Chinese philosopher Liang Shuming (1893–1988) and the Japanese architectural historian Chuta Itō (1867–1954). In his influential book Dong-Xi fang Wenhua jiqi Zhexue (Eastern and Western cultures and their philosophies, 1921), Liang Shuming disagreed with contemporary Chinese scholars who regarded the West as the model for Chinese modernization. He contended that Chinese culture had its own merit as a unique “system” that complemented the systems of the West and India. He believed that Western culture encouraged individual desire and material acquisition, which caused conflict between human beings and nature and even among human beings. In contrast, Indian culture suppressed individual desire, thus leading to man’s negation of nature and the retrogression of Indian society. He understood Chinese culture, however, as a culture of moderation that emphasized human harmony with nature, and he proclaimed that it would therefore prevail in the world.

This cultural relativist view is also evident in the book Shina kenchiku shi (A history of Chinese architecture, 1925) by Chuta Itō. The most renowned Japanese architectural historian in the Meiji era, Itō attempted to validate Chinese architecture in the global context and legitimated its Japanese derivation. He claimed that Eastern architecture was a system parallel to Western architecture. In this system, Chinese architecture, the only unique living tradition, had influenced a broad area of East Asia inhabited by one-third of the world’s population. For this reason alone it deserved admiration.

Whereas Itō based his argument on the spread of Chinese architecture, the Liangs based theirs on the maturity of structure and aesthetics Chinese architecture had achieved. They were influenced by Osvald Sirén’s book A History of Early Chinese Art, published in 1930, although about a decade later Liang would criticize Sirén for not knowing “the grammar of Chinese architecture” and using the YZFS carelessly.40 Sirén noted that dougong had lost its functional role and become more ornamental since the Ming dynasty (1368–1644): “The particular character and significance of the old Chinese architecture depended primarily on its clear and ingenious wood construction. It was pure carpenters’ [sic] art determined by the special requirements of the material. Each part had a definite function which was not concealed by any superimposed decoration. … But once these were encroached upon by purely decorative tendencies, its vital nerve was severed, and its further possibilities of growth were destroyed.”41

As both Han Baode and Xia Zhujiu have noticed, Liang and Lin’s emphasis on architectural structure reflected the influence of structural rationalism.42 For architectural historians such as A. W. N. Pugin and John Ruskin of England as well as Eugène Viollet-le-Duc and Auguste Choisy of France, from the mid-nineteenth century to the early twentieth century structural rationalism played a dominant role in evaluating modern architecture. Although each of these historians had different expectations of architecture, they were unanimous in emphasizing the primary role of structural expression, the unity between any form in architecture and the corresponding structural system. A more direct influence on Liang’s structural rationalism might have been that of Paul Philippe Cret, Liang’s principal design teacher when he studied at the University of Pennsylvania. Cret, according to Gwendolyn Wright, “stressed the need for students to analyze architectural elements used in past monuments, relying in particular on Choisy’s analytiques, without absorbing a liturgy of styles or monuments.”43

Liang’s stylistic study of Chinese architecture clearly reflected his training at Penn and Harvard in Wölfflinian modes of formal analysis. Heinrich Wölfflin, in revealing visual representation as the result of artistic volition, or Kunstwollen, rather than evolution, challenged Johann J. Winckelmann’s writing of art history and archaeology, with its evolutionary schema of decline and fall. Although Wölfflin’s approach could be taken to justify Chinese architecture as the unique “system” the Liangs believed it was, the couple followed Winckelmann’s schema in describing the development of Chinese architecture. In his History of Ancient Art of 1764, Winckelmann presented a linear history of Greek art, with a four-fold pattern of stylistic development— the Older Style, the Grand Style, the Beautiful Style, and the Style of Imitators—from its beginnings to its greatest achievement, followed by stasis, decline, and end. The Grand Style, represented by the work of Phidias, was characterized by its grandeur, and the Beautiful Style, represented by the work of Praxiteles, by its gracefulness. These two styles stood at the summit of Greek art; the Style of Imitators, as the name suggests, indicated its decline.44

That Sirén was influenced by Winckelmann’s linear history can be seen not only in his History of Early Chinese Art of 1930, but also in his earlier book, Chinese Sculpture from the Fifth to the Fourteenth Century, of 1925, which traced the stylistic evolution from “the Archaic Period” to “the Transition Period,” “the Period of Maturity,” and finally “the Period of Decadence and Reflorescence.”45 These patterns resonated in Lin’s preface to Liang’s book Qingshi yingzao zeli. There she delineated the three periods of the development of Chinese architecture: germination, maturation, and decline. Liang’s Pictorial History of Chinese Architecture, written in the 1940s, evokes Winckelmann’s periodization. After discussing materials found in Chinese archaeological excavations and stone carvings that were the historical remains of an “older” period, Liang cataloged the subsequent development of Chinese wooden-structure architecture as the period of haojin (vigor) of the Tang and Liao dynasties, the period of chunhe (elegance) of the late Northern Song and Yuan dynasties, and the period of jizhi (rigidity) of the Ming and Qing dynasties (Figure 14).46 By using the line of development accepted for Western architecture and thus drawing similarities between Chinese, Greco-Roman, and Gothic architecture, the Liangs established that Chinese architecture was a system as significant as its Western counterparts. Emphasizing the structural rationality of Chinese architecture, they refuted Fergusson’s notion that polychromy was more important to it than form. By delineating a progression parallel to that of Western classical architecture, they also challenged Fletcher’s notion that Chinese architecture is non-historical. Moreover, Liang drew a parallel between Chinese architecture and Greek art in general, as well as Tang and Song architecture and the Grand and Beautiful Styles of Greek art in particular.

Figure 14

Liang Sicheng, “Evolution of the general appearance of timber-frame halls” (plate 20 of A Pictorial History of Chinese Architecture [1984])

Figure 14

Liang Sicheng, “Evolution of the general appearance of timber-frame halls” (plate 20 of A Pictorial History of Chinese Architecture [1984])

Liang’s suggestion of a parallel between Chinese and Greek architecture and his effort to justify Chinese architecture in a global context are most evident in an illustration in his Pictorial History of Chinese Architecture, written in English. His language in this book, including his translation, from Chinese to English, of the legends in all his illustrations, anticipates a Western readership. In the bilingual illustration “The Chinese ‘Order’ ” (see Figure 9), Liang juxtaposed the English title to the Chinese one, but even in the Chinese gave the word “order” in English. This juxtaposition reveals his attempt to equate the bracket-column combination in Chinese architecture with the orders in the Western classical counterpart. What is even more intriguing is that in translating the terms in the legend, Liang used two versions, one that provides the meaning of the original terms, the other, the mere phonetic rendering of the terms. For the key structural parts, such as flying rafters, eave-rafters, beams, and columns, he indicated the meaning, whereas for auxiliary elements—whose names sound strange even to the Chinese—such as ling gong (ling-kung) (literally “command gong”), shuatou (shua-tou, literally “playful head”), and guazi gong (kua-tzu-kung) (literally “melon-seed gong”), he gave the phonetic term. In these ways, he legitimated Chinese architecture by highlighting its structural rationality and obscuring for Western readers the random or irrational aspect of the terms used by traditional craftsmen.

In addition to countering architectural historians of the West, Liang and Lin also resisted disparagement of both Chinese architects of the early twentieth century who recognized the advantage of Western architecture, and the Chinese public who accepted the “erroneous view” of Westerners and considered Chinese architecture obsolete for its functional deficiencies, primitive material, and shabby construction.47 How to justify Chinese architecture in the modern era thus became another challenge for Liang and Lin. Although they did not oppose criticism of Chinese architecture that characterized it as outdated in function and technique, they insisted on its artistic achievement and its value as a unique architectural system: “It is undeniably true that an old architecture will keep its pure artistic value even though its usefulness has become questionable with the progress of people’s lives. Like the pyramids in Egypt and the Parthenon in Greece, the altars, temples, palaces, and halls in Beijing deserve to be honored forever, even though their original practical functions have already been lost.”48

Liang, relying heavily on structural rationalism, attempted to demonstrate the similarity between Chinese architecture and the most contemporary modern architecture. In this regard, Henry-Russell Hitchcock and Philip Johnson’s modernist classic International Style: Architecture Since 1922, published in 1932, provided him with timely support. Hitchcock and Johnson attributed the emergence of the International Style to the development of modern structure, especially the popularization of the steel and concrete frame. Though Liang might not have met Hitchcock when he spent a semester studying art history at Harvard, in 1927–28, while Hitchcock was a graduate student in the same program, he could use Hitchcock’s arguments for the International Style in China. In 1933 Carl Lindbohm, a Swedish architect who claimed to have worked with Le Corbusier, Walter Gropius, and Frank Lloyd Wright, went to Shanghai. His preaching of the International Style of architecture in some influential local newspapers attracted the attention of Chinese architects and the public and generated a debate about modernism and Chinese-style architecture.49 Knowing that the new style was also based on the principle of structural rationalism, Liang noted:

The so-called International Style of architecture, though the name is ambiguous, is in fact honest in its essential principle. … Its most remarkable characteristic is the rational feature resulting from scientific structure. … Those who have acquired a real understanding of the new architecture should know that the ultramodern structural method, though using different material, shares a fundamental principle of Chinese indigenous architecture, namely, to set up the skeleton first and then add the walls. Because of their similarity in [structural] principle, the architecture of the International Style contains many aspects similar to Chinese (or Eastern) architecture.50

Realizing that Chinese architecture and Western architecture had commonality in their structural systems, Liang confidently concluded, “Now is the time for Chinese architecture to be revitalized with new science, material, and structure.”51

Armed with structural rationalism, Liang was able to defend Chinese architecture in the discourse of both international and modern architecture. He set up a hierarchical system based on that criterion, in which the haojin architecture of the Tang, Liao, and Song dynasties ranked aesthetically above that of the Ming and Qing. Embodying both structural rationality and formal beauty, Tang and Liao/Song architecture, Liang believed, deserved to be the paradigm of Chinese classical architecture as well as the wellspring of ideas for “revitalizing” Chinese architecture.

Literally, haojin means strong, masculine physicality, and bold, unconstrained spirituality. Although the word is employed here to describe an aesthetic style, it mirrored the attributes of the Chinese elite’s ideal of modern nation building. Thus Liang’s choice of a style hinged on more than his personal preferences as an architectural historian. As early as 1902, Liang Qichao (1873–1929), Liang Sicheng’s father, whom Joseph R. Levenson has called the “mind of modern China,”52 published a book titled Xinmin shuo (Discourse on the new citizen). He used the arguments of social Darwinism to explain that China owed its repeated defeats to its loss of military spirit. To gain a place on the international stage, China would have to carry out the “blood and iron policy” Otto von Bismarck had implemented in Germany to cultivate the mind (xinli), courage (danli), and physical body (tili) of the people.53 In 1904 he wrote another book, titled Zhongguo zhi Wushidao (China’s Bushido), to promote the military spirit that he believed had prevailed during China’s Warring States period (475–221 BCE) but had withered afterward.54 Lu Xun (1881–1936), another important thinker of twentieth-century China, looked forward to a modern China of spiritual as well as physical vigor. He admired China’s Han and Tang dynasties, noting that they were audacious (hongfang), because both had exhibited an unparalleled open-mindedness, and both were confidently able to absorb foreign art and culture.55

The determination to revitalize the nation’s military spirit made physical education an important constituent of the national agenda during the Republican era.56 It also resonated in the domain of art. A robust style of calligraphy that had begun in the mid-Qing dynasty as a tool of the literati for expression and communication flourished in the late imperial period when China confronted unprecedented foreign invasions. Kang Youwei (1858–1927), Liang Qichao’s mentor and the leader of the Hundred Days’ Reforms in 1898, was the most influential figure in this reform of calligraphy. Believing that a style of calligraphy reflected the spirit of a time, Kang denounced the mellow but meticulous Hall and Pavilion Style (guange ti) that was favored by the court in reviewing the writings of the candidates for imperial examination. Instead, he advocated the angular but forceful Northern Wei (386–556) Stele Style (Wei bei), which derived from inscriptions carved on stone and praised by Kang as being unadorned, solid, vigorous, and powerful (zhuo, hou, xiong, qiang).57 The new trend in calligraphy led to the emergence of the Shanghai school of Chinese painting in the late nineteenth and early twentieth centuries, whereby artists accepted the new taste and adopted new methods of using a brush pen. Lu Xun, who played a significant role in introducing to China in the 1930s the Western woodcut print as a medium that could represent the taste of the working class, also favored the powerful style exemplified by works of Käthe Kollwitz, Frans Masereel, and Carl Meffert. In short, the association of artistic style with power was part of the art discourse in China during the late nineteenth and early twentieth centuries.

Whereas Liang Qichao regarded the Warring States period as the Chinese golden age of Bushido, Lu considered the Han and Tang dynasties as the Chinese golden age of spiritual openness, and Kang identified the Northern Wei as the Chinese golden age of calligraphy, Liang Sicheng found the period from Tang to early Song the most vigorous era of Chinese architecture. As Alan Colquhoun points out in his discussion of the revivalist architecture of the eighteenth century, “Eclecticism took two forms which at first might seem incompatible. On the one hand, different styles could exist side by side. … On the other hand, one style could come to stand for a dominant moral idea and be connected with an idea of social reform.”58 Similarly, in early twentieth-century China, the past was mobilized to advance a vision of the future because the praise of early styles of Chinese art embodied the elite’s expectation of a modern China. So the museum is not simply an imitation of a historical style; it is part of a search for the ideal modern Chinese architecture.

To Liang, Tang architecture and its continuation in the Liao and Early Song dynasties represented the ideal form. Liang’s career as a Chinese architectural historian began with the quest to comprehend the architecture of the Tang dynasty, the zenith of premodern Chinese culture. The first paper he published, “Buddhist Monasteries and Palaces of the Tang Period We Know at This Time,”59 demonstrated his effort to “reconstruct” Tang architecture according to evidence drawn from murals in Buddhist caves at Dunhuang.

In his first published report on an architectural investigation, “Research on Guanyin Pavilion and the Front Gate of Dule Monastery in Ji County” of 1932,60 Liang observed that the pavilion and the gate were stylistically identical to Tang architecture as represented in the Dunhuang murals, which he had studied. Because geographically the Liao area was part of the Tang territory and was separated after Five Dynasties from the Chinese central region, where a new cultural tradition represented by the Song dynasty was emerging, Liang believed that Liao architecture might have preserved more characteristics of Tang architecture than YZFS did of the late Northern Song. This manual, he later said, “had eliminated previous grandeur (shuoda) in favor of a slim and extravagant (xianmi) style, exhibiting a trend toward decoration (xiushi) rather than masculinity (kuiwei).”61 Yet unlike the architecture of the Ming and Qing dynasties, Song architecture was similar to that of the Tang because the dougong bracketing, a key element of Chinese architecture in Liang’s study, played a functional role in both systems and was an “organic” and “rational” part of the structure. By contrast (here Liang agreed with Sirén), in Ming and Qing architecture the dougong had “degenerated” into merely an applied ornament.62

Liang’s quest for an extant wooden example of Tang dynasty architecture drove him in 1937 to search the Wutai Mountains, where, just before the outbreak of the Sino-Japanese War, he made a great discovery, the east hall of Foguang Monastery. He dated the building to the eleventh year of the Dazhong reign of the Tang dynasty, that is, 857 CE.63 Nancy Steinhardt calls this discovery the “crowning moment in the modern search for China’s ancient architecture” and says that the building Liang discovered became his “architectural icon.”64 In 1935, however, before he had a model of Tang architecture, Liang had to resort to Liao and early Song for references in the design of the museum. Even though in Chinese history Liao was a dynasty ruled, not by the Chinese but by pastoral nomads who called themselves Qidan (Khitan/Kitan), Liang claimed it as a representative of Classical Chinese culture because in his view its architecture continued and epitomized that of the Tang dynasty.65

Liang’s admiration of a vigorous style akin to that of Tang, Liao, and early Song architecture is demonstrated in the perspective rendering made by the architects’ office and published by the architect Xu in 1964 under the title “Perspective and Detail of Amended and Final Design of the National Central Museum in Liao Style” (Figure 15). Avoiding the static front elevation of the competition entry, the rendering shows a two-point perspective from the plaza whose sharp contrast between light and shadow dramatizes the “vigor” of the museum’s spreading eave. It is calculated to create the impression of power, masculinity, and monumentality. Significantly, the wide platform in front of the museum hall would never in actuality give an on-site observer the complete view of the corridor shown in the image; the drawing presents what Liang and his colleagues considered the ideal image of the building.

Figure 15

Su, Yang & Lei Architects, Liang Sicheng Consultant, “Perspective and Detail of Amended and Final Design of the National Central Museum in Liao Style” (plate 139 of Chinese Architecture—Past and Contemporary [1964])

Figure 15

Su, Yang & Lei Architects, Liang Sicheng Consultant, “Perspective and Detail of Amended and Final Design of the National Central Museum in Liao Style” (plate 139 of Chinese Architecture—Past and Contemporary [1964])

Guiding Principles behind the Design of the National Central Museum

The synthetic approach to the sources Liang and the architects explored for the museum would have been familiar to Liang as a student of the Beaux-Arts tradition. As Julien Guadet, the chair of the Theory of Architecture at the École des Beaux-Arts, wrote in his monumental work Eléments et théories de l’architecture, the two fundamental issues of architectural design are elements and composition. The elements include the walls, cornices, doors and windows, porticos, and especially the orders. Composition refers to the functional and spatial logic of the plans of different architectural types. Although many of Liang’s contemporaries had attempted to create a Chinese-style modern architecture by combining the plan layout and compositional devices of Western architecture with such elements of Chinese architecture as the curved roof, bracket set, and sculptured plinth, Liang endeavored to define a “pure” language of Chinese architecture. His study of the two ancient architectural manuals and his compilation of ancient architectural details, which he later referred to as the “grammar” and “vocabularies” of Chinese architecture,66 are evidence of this endeavor and of the Beaux-Arts notion of architecture on which it was premised.67

Even though it is clear that in general Liang resorted to what he considered an authentic language of Chinese architecture for the building’s composition, namely, its plan, section, elevation, and structural elements, some crucial details have yet to be explained. What was the rationale for choosing elements from some buildings but not others? Why was a detail as important as the roof curve revised when other elements were not? And most significant, what were the principles that guided the selections and the modifications?

Undoubtedly, the authenticity or “Chineseness” of the design was the first concern of Liang and the architects. Yet further examination of the relationship between the elements they chose and Liang and Lin’s historical writings reveals the link between the Liangs’ writings on Chinese buildings, the design of the museum, and Western architectural preoccupations with an ideal of “classical” beauty and with structural and functional rationalism. In other words, the seemingly pure Chinese style of the museum was in fact shaped by Western aesthetics. To the Liangs, this was not a contradiction, for they understood that the style of the museum was not only Chinese but also universal and modern.

“Classical beauty” is visible in many subtle modifications and in the reference to the classical orders of Western architecture in the design of the corridor. The subtle modifications include the upturned roof ridge, the gradual rise in pillar heights toward the ends of the corridor (shengqi), the outward-pushing ridge (tuishan), as well as the entasis, or slight convexity, of the pillars, and the shape of the lintel (which is flattened on its top and bottom surfaces but has convex sides). These measures would certainly have increased the difficulty of construction. They were key markers, however, of the vigorous and elegant Liao and Song styles.

Liang associated the beauty of the curved lines in Chinese architecture with Greek architecture. For example, in discussing the lintels and beams of Dule Monastery, he said:

As to its delightfulness, a noteworthy point is the entasis that makes the central part of a lintel a convex surface. Although entasis does not change the static property of the lintel, the rigid line becomes rounded. This is the result of assiduous craftsmanship and should not be overlooked. The Parthenon in Athens, Greece, has similar subtle measures. To counterbalance vigor with elegance was a widely accepted principle in the past. So is the aforementioned batter [of the inward-leaning side columns of the Dule gate], which is also visible in Greek architecture. Is it because our ancestors learned from the Greeks?68

As I mention earlier here, Liang, in his writings, regarded the combination of bracket and column in Chinese architecture as analogous to the role of an order in Greco-Roman architecture.69 The increase in the museum’s ratio of bracket to column height to 1:3 was hardly coincidental. According to Fletcher’s History of Architecture, that is the ratio of entablature to a column of the Doric Order, the most masculine of the five orders of Western architecture. By drawing a mathematical connection with the most appropriate paradigmatic Greek order, Liang and the architects achieved a classical Chinese form that was analogous to its Western Greek counterpart (Figure 16).

Figure 16

“Order” and balustrade of the National Central Museum (author’s photo)

Figure 16

“Order” and balustrade of the National Central Museum (author’s photo)

Structural rationalism, in addition to its importance as a criterion for judging Chinese architecture of different periods, also served as a standard for Liang and the architects in their choice of the ideal forms of structural elements from existing Liao and Song buildings. For example, Liang and the architects referred to Sandashi Hall of Guangji Monastery in designing the interior beams and dougong brackets of the museum. Liang once enthusiastically praised the hall for employing the open-frame method (cheshang luming zao), which, with no ceiling to conceal the structural features, “gives the architect a grand opportunity to display his ingenuity in handling major carpentry as an artistic creation.”70 The use of brackets in the hall to replace small posts, he said, “though visible in later buildings, exhibits an unforeseen delicacy of workmanship” (Figure 17).71

Figure 17

Liang Sicheng, “Plan and cross section” of Sandashi Hall of Guangji Monastery, Baodi county, 1025 (plate 28b of A Pictorial History of Chinese Architecture [1984])

Figure 17

Liang Sicheng, “Plan and cross section” of Sandashi Hall of Guangji Monastery, Baodi county, 1025 (plate 28b of A Pictorial History of Chinese Architecture [1984])

Structural rationalism seems also to have shaped their preference for a particular proportion for the section of the lan’e lintel of the corridor. Its height is 20.5 inches, equal to two cai. This measurement follows the prescriptions of YZFS, which specified that “the height of a lintel is twice that of a cai.” Yet its width is not two-thirds of the height, as prescribed, but 10.25 inches, or half the height. This proportion recalls Liang’s comment on the beams and lintels of Guanyin Pavilion and the Front Gate of Dule Monastery: “According to the modern science of beam making, the section should be about two to one in proportion. YZFS’s prescription is three to two, more reasonable than the one based on the Qing method, which is almost a square, with a proportion of ten to eight or twelve to ten. Guanyin Pavilion’s and the Front Gate’s proportions are two to one, the same as the modern way. Is our scientific knowledge becoming backward?” (Figure 18).72

Figure 18

Liang Sicheng, “Evolution of the lan-e and pu-pai fang” (plate 38 of A Pictorial History of Chinese Architecture [1984])

Figure 18

Liang Sicheng, “Evolution of the lan-e and pu-pai fang” (plate 38 of A Pictorial History of Chinese Architecture [1984])

The use of reinforced concrete raised the question whether it was still necessary and “reasonable” to design a structural element in the size and the proportions of wooden models.73 But in this case aesthetic considerations, such as the stylistic authenticity and structural rationality exhibited in the design, led Liang and the architects to sacrifice material sincerity.

As a unique part of Chinese Order, dougong bracketing is clearly the key element in Liang’s study of Chinese architectural history as well as the most characteristic feature of Chinese architecture. Structural rationalism, however, rendered dougong unnecessary in the modern era when reinforced concrete offered a much stronger support for cantilevering than the wooden arms of bracketing. Liang and the architects thus needed to balance contradictory pursuits, that is, to display unique Chinese architectural characteristics on the one hand, and to fulfill the imperative of modern structural rationalism on the other. To do so, they chose their models of pillar-top bracketing and corner-pillar bracketing from the Front Gate of Dule Monastery (Figure 19), a building that, as Liang put it, “is a masterpiece in the use of dougong brackets that meet the highest artistic structural standards,”74 as well as the use of intercolumnar bracketing from Haihui Hall of Huayan Monastery, the simplest of all the Liao and Song examples. The Front Gate of Dule Monastery is low-grade and only three bays wide, and Haihui Hall is also a low-grade building owing to the form of its overhanging gable roof (xuanshan). In adopting the bracketing from these buildings to the high-ranking, nine-bay-wide museum, Liang and the architects abandoned the hierarchical connotation of this traditional element while retaining its unique formal beauty. Furthermore, in a modern republican society, the repudiation of hierarchy could be seen as a modern and universalist choice. In using the gate’s dougong as a model, Liang and the architects highlighted the bracket’s structural-rational purpose over its decorative purpose, for the gate’s dougong is a much simplified type of five-puzuo and touxin. It is made of horizontal load-bearing arms with neither an extra lateral bracket arm crossing through the first projected arm, as in the “added heart” (jixin) method, nor the ang, or lever (17 in Figure 9), an element unnecessary in a concrete structure. For the same reason, it seems Liang and the architects eliminated both the diagonal arms, “a favorite of the Liao architects,”75 and also the diagonal arm cutting across the corner horizontally at a 45-degree angle in the corner-pillar bracketing of the Dule gate.

Figure 19

Corner bracket from the Front Gate of Dule Monastery, Ji county, 984 (photo courtesy of Yang Jing)

Figure 19

Corner bracket from the Front Gate of Dule Monastery, Ji county, 984 (photo courtesy of Yang Jing)

The irregularity of the pillar grid, another noteworthy feature visible in Liao buildings, gets reinterpreted in the museum as well. Although that irregularity may have been a result of using varied structural systems, as Chen later believed,76 Liang named it the “column-eliminating method,” meaning that the irregularity in column arrangement was not a consequence of using special structural systems but rather a measure taken to meet functional requirements or, in Vitruvian terms, “commodity”—appropriate spatial accommodation (shiyong). As Liang wrote in a report of an investigation in 1933: “The most admirable parts of the plans of the aforementioned six buildings [the Daxiongbao Hall, Sutra Library, and Haihui Hall of Huayan Monastery and the Gate, Sansheng Hall, and Daxiongbao Hall of Shanhua Monastery] are the arrangements of columns. They varied according to practical needs, and maintained the architectural principle very well. … [For example,] because the pedestals of statues are placed within the inner circle of columns in the central area of a hall, and worshipers are outside the circle, a big space is needed. Eliminating the columns at the center would satisfy the commodity.”77

Liang had criticized the spatial arrangement of Ming and Qing architecture for maintaining a rigid column grid that lacked the flexibility needed for functional adjustments: “Judging by this single fact, we can see that the architecture of our country had been in decline since the Ming dynasty” (Figure 20).78 Accordingly, Liang and the architects working with him eliminated four central columns of the museum hall, so that the design would demonstrate the architects’ functional orientation (Figure 21).

Figure 20

Liang Sicheng, “Comparison of plan and columniation of timber-frame halls” (plate 21 of A Pictorial History of Chinese Architecture [1984])

Figure 20

Liang Sicheng, “Comparison of plan and columniation of timber-frame halls” (plate 21 of A Pictorial History of Chinese Architecture [1984])

Figure 21

Interior of the National Central Museum (author’s photo)

Figure 21

Interior of the National Central Museum (author’s photo)

The great curvature of the roof demonstrates dramatically the effort to conform the design of the museum to the nature of a modern material, as part of the concern with function as well as rational structure. The upturned roof of Chinese architecture has prompted much speculation in the West. The German scholar Ernst Boerschmann wrote, “The impulse which drove the Chinese to use these curving forms came from their desire to express the movement of life. … By the curving of the roof, buildings are made to approach as nearly as possible the forms of nature—the varied outlines of rocks, trees, etc.”79 It is no surprise that Lin, who believed Chinese architecture was not only structurally rational but also functional, disagreed. She argued:

The roof is originally the most necessary part of a building. … From the beginning it is not merely the building’s cover. Because of the real problems of rain and sunlight, the eave expands. It is not difficult to make a protruding eave, but the deeper the eave, the lower it is, preventing sunlight [from coming in]. Moreover, when rainwater runs down rapidly from the roof, it splashes [the wall] under the eave. To solve these problems, we create the “flying eave,” namely the upturned roof edge with double layers of rafter, which forms a slight curve. For aesthetic reasons, the curvature of the corners of the roof is increased. This “flying eave,” curved on the edge and raised at four corners, is a natural and reasonable structural method. … By and large, the curved roofline, which has long been regarded as a mystery, has no rationale beyond structural principles, unnatural or artificial. [But] it is also successful in both aesthetics and commodity. … Because the roof style developed by the Chinese is unique, distinct from various European traditions, foreigners noticed it from very early on and proposed different interpretations, all of which were quite imaginative. Some said that the Chinese roof was derived from the tents of nomadic times; some said that it symbolized the overhanging branches of the pine tree; some saw the Chinese flying eave as bizarre, some called it childish, some focused on its animal ornaments, uselessly studying their meanings. It is a waste of time to criticize their analyses of the curved roof or the reviews of its decorative principles. The curved roof argues for its own aesthetic and functional value. We can simply acknowledge the undeniable artistic success of our structure.80

To Lin, the “raise-depress” method of constructing a curved Chinese roofline also satisfied the functional point of view. She cited a sentence from Kaogong ji (Records of artificers)—a Chinese classic believed to have been compiled in the early Western Han dynasty (206 BCE–24 CE), based on the remaining text of the Spring and Autumn period (eighth to fifth century BCE)—that reads, “The craftsman who makes the canopy [of a cart] … wants the top zun and the edge bei. Top zun and edge bei can help throw water quickly and far away.” Although no evidence indicates that the curved roofline of Chinese architecture appeared as early as the Western Han dynasty, the precise meanings of the words zun and bei are “high” and “low,” or “up” and “down,” which in Kaogong ji could represent only a pitched roof. Lin, however, interpreted these terms as meaning “steep” and “flat.” Hence, she said, this sentence, “clearly explained the actual function of this kind of roof [built according to the ‘raise-depress’ method].” Moreover, making the top zun can “correct the perspective that downplays the roof ridge, making the roof look lofty and enhancing its beauty in appearance and silhouette.”81

Lin’s view can be seen as the guiding principle in the roof design of the museum. Based on the Song method of “raise and depress,” but exaggerated, this design was not modeled after any existing Liao and Song structure that Liang had investigated, nor did it follow precisely the prescriptions of the Song manual YZFS. Appropriately, the museum’s reinforced-concrete roof appears both thinner and more graceful than its Liao and Song predecessors, which were made of timber and clay. Covered with the kind of yellow glazed tiles used to roof the Ming and Qing imperial palaces, this modern civic edifice had a light and lively visual effect never seen in the heavy and somber Liao and Song religious temples.

Conclusion

The National Central Museum opened to the public on 29 May 1948. Before the building was completed, Liang Sicheng included commentary on it in the last chapter of his book A History of Chinese Architecture, the draft of which he finished in 1944. Without mentioning his own contribution, he wrote: “The Central Museum, designed by Xu Jingzhi and Li Huibo, has very simply and rationally reincarnated Liao and Song architectural form in a modern structure. It is also a significant example of modern Chinese architecture.”82 Xu, moreover, praised the museum in his own book Chinese Architecture—Past and Contemporary, published in 1964: “It was designed along the lines of Liao or early Sung [Song] style. The details are simple and bold. The bracket system is more functional than decorative. It was built in reinforced cement concrete [sic] with the consideration of [the] cantilever principle. The curve of the eave was plotted by increasing the length of the outer columns, and [that of] the roof by using the method of ‘lifting and depressing’ the positions of the purlins. The result was a light and buoyant-looking structure.”83

The museum was a monument of the New Cultural Movement that began in 1916 and led to a fundamental change in Chinese society and culture. One of the leading figures of this movement, Hu Shi (Hu Shih, 1891–1962), called it a “Chinese renaissance” that aimed to re-create Chinese civilization by studying current problems in China, importing new methodologies and new ideas from the West, and “reorganizing the national heritage” (zhengli guogu) by systematically reevaluating and reforming Chinese indigenous civilization. In his Plan for Reordering the National Heritage, proposed in 1923, Hu elaborated a methodology that would make the enigmatic Chinese classics easier to read and apprehend by means of the strategies scholars commonly apply to produce authoritative new editions of important early Western writings: collating, annotating, and adding analytical introductions as well as organizing the text into paragraphs and inserting punctuation.84

The goal of research on traditional Chinese architecture and also of modern Chinese-style architecture designed and built by Chinese architects since the 1920s has resembled that of the New Cultural Movement. As proposed by Zhao Shen, chairman of the Society of Chinese Architects, it was to “synthesize the advantages of both Eastern and Western architecture, to develop the intrinsic beauty of our country.”85 The common approach in efforts to achieve this goal has been to use Chinese architectural motifs or forms with modern material and structures. Two efforts are of great significance. The first, proposed by Lü Yanzhi (Yen-chih Lu) and continued by Yang Tingbao (Ting-pao Yang), and the second, by Liang and Lin, aimed to provide rational principles to codify the Chinese-style designs. Lü and Yang employed Western compositional principles, trying to incorporate Chinese architecture into a universal plan type associated with the architectural doctrines of the Beaux-Arts school.86

Liang, Lin, and their colleagues at the SRCA achieved notable success in identifying the unique compositional principles of Chinese architecture by an in-depth study of the ancient manuals and a wide-ranging investigation of extant structures.87 Liang’s study of the Song manual YZFS and the Qing Gongbu gongcheng zuofa zeli enabled him to pursue zhengli guogu (reorganizing the national heritage) by providing the syntactic order, annotation, and analysis that Chinese traditional writing lacked. Liang thus was able to codify the classical language of Chinese architecture—the “grammars” of the two ancient manuals and the “vocabularies” of elements and motifs.

Like all great buildings, the National Central Museum has an inventive design that neither imitates the past nor represents a “conservative revivalism.” Its idealized approach to a Chinese-style modern architecture could hardly have succeeded had the architects not had such a thorough knowledge of Chinese architecture, nor could this approach be applied to designs intended to serve more complicated functions.88 This example demonstrates, however, the effort that Liang and the architects under his guidance made to “reorganize” the Chinese architectural heritage, to reevaluate it by the standards of modern science and aesthetics, and to create an ideal Chinese-style modern architecture. Made of modern materials, satisfying modern functional requirements, and sharing a similar sense of the nuances of Western classical architecture in composition and the rationalist imperative of modern architecture, the museum embodied the Chinese nationalist elite’s expectations of a “renaissance” of Chinese culture that would originate in a vigorous Chinese tradition but be modified according to Western classical and modern standards.

Appendix: A Chronological List of Ancient Structures and Two Architectural Manuals Mentioned in the Essay

DynastyName of Structure/BookLocationDateInvestigation/Study
Tang (618–907) Main Hall of Foguang Monastery Mt. Wutai, Shanxi province 857 by Liang in 1937 
Five Dynasties period (907–960) The Sarira Pagoda of Qixia Monastery Nanjing, Jiangsu province undated by Liu in 1930 
Liao (907–1125) Front Gate and Guanyin Pavilion of Dule Monastery Ji county, Hebei province 984 by Liang in 1932 
Liao Daxiongbao Hall of Fengguo Monastery Yi county, Liaoning province 1020 by Tadashi Sekino in 1932 
Liao Sandashi Hall of Guangji Monastery Baodi county, Hebei province 1025 (demolished in the 1950s) by Liang in 1932 
Liao Main hall of Kaishan Monastery Xincheng, Hebei province 1033 by Liu in 1936 
Five Dynasties Period to Northern Song (907–1127) Dacheng H all of the Confucius Temple Zhengding, Hebei province undated by Liang in 1933 
Northern Song (960–1127) Yingzao Fashi  1103 by Liang 
Liao Bojiajiaozang (Bhaghavat) Sutra Repository, or Sutra Library, and Haihui Hall of Huayan Monastery Datong, Shanxi province 1140 (Haihui Hall was demolished before 1949) by Liang and Liu in 1933 
Liao-Jin (1115–1234) Daxiongbao Hall of Huayan Monastery Datong, Shanxi province built in 1140, rebuilt in the early Jin dynasty by Liang and Liu in 1933 
Liao-Jin Daxiongbao Hall, Sansheng Hall, and Puxian Pavilion of Shanhua Monastery Datong, Shanxi province 1060–1143 by Liang and Liu in 1933 
Qing (1644–1911) Gongbu gongcheng zuofa zeli  1734 by Liang, with a preface by Lin 
DynastyName of Structure/BookLocationDateInvestigation/Study
Tang (618–907) Main Hall of Foguang Monastery Mt. Wutai, Shanxi province 857 by Liang in 1937 
Five Dynasties period (907–960) The Sarira Pagoda of Qixia Monastery Nanjing, Jiangsu province undated by Liu in 1930 
Liao (907–1125) Front Gate and Guanyin Pavilion of Dule Monastery Ji county, Hebei province 984 by Liang in 1932 
Liao Daxiongbao Hall of Fengguo Monastery Yi county, Liaoning province 1020 by Tadashi Sekino in 1932 
Liao Sandashi Hall of Guangji Monastery Baodi county, Hebei province 1025 (demolished in the 1950s) by Liang in 1932 
Liao Main hall of Kaishan Monastery Xincheng, Hebei province 1033 by Liu in 1936 
Five Dynasties Period to Northern Song (907–1127) Dacheng H all of the Confucius Temple Zhengding, Hebei province undated by Liang in 1933 
Northern Song (960–1127) Yingzao Fashi  1103 by Liang 
Liao Bojiajiaozang (Bhaghavat) Sutra Repository, or Sutra Library, and Haihui Hall of Huayan Monastery Datong, Shanxi province 1140 (Haihui Hall was demolished before 1949) by Liang and Liu in 1933 
Liao-Jin (1115–1234) Daxiongbao Hall of Huayan Monastery Datong, Shanxi province built in 1140, rebuilt in the early Jin dynasty by Liang and Liu in 1933 
Liao-Jin Daxiongbao Hall, Sansheng Hall, and Puxian Pavilion of Shanhua Monastery Datong, Shanxi province 1060–1143 by Liang and Liu in 1933 
Qing (1644–1911) Gongbu gongcheng zuofa zeli  1734 by Liang, with a preface by Lin 

Notes

1.

This article was inspired by a 2003 conference, The Beaux-Arts, Paul P. Cret, and 20th Century Architecture in China, organized by Jeffrey W. Cody, Nancy S. Steinhardt, Tony Atkin, and Xing Ruan and held at the University of Pennsylvania. Professors Wu Hung and Katherine Fischer Taylor responded to the early draft of the paper with constructive criticism. Since then many friends and colleagues have helped me with revisions, especially those undertaken to prepare this English version. Among them are Mrs. Lin Zhu, Drs. Elizabeth Grossman, Swati Chattopadhyay, Stephanie Fay, Wei-cheng Lin, Seng Kuan, Wenming Fei, Yudong Wang, Tonia Eckfeld, Cary Liu, Sun-Chang Lo, Tracy Miller, Chia-Ling Yang, Doris Wai, Mary Byers, Mr. Mark Giorgini, Ms. Mira Rai, Miss Tianjian Lai, and an anonymous reviewer. My study of Liang Sicheng and the design of the National Central Museum in Nanjing has also benefited tremendously from the support of the Asian Cultural Council in New York and Mr. Zhou Jianmin and Ms. Cheng Guihong at the Archives of Nanjing Municipal Construction in Nanjing. I take this opportunity to thank all of those who have helped make this paper’s publication possible.

Liang Sicheng (Liang Ssu-Cheng), preface to Liang Ssu-Cheng, APictorial History of Chinese Architecture: A Study on the Development of Its Structure System and Evolution of Its Types, ed. Wilma Fairbank (Cambridge, Mass.: MIT Press, 1984), vol. 8 of Liang Sicheng quanji (Complete works of Liang Sicheng) (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2001), 207.

2.

The National Central Museum is a large building compound. But in most architectural-historical writings today, the term refers only to its main hall because this hall was the key building in the design that embodied all the requirements of architectural style in the competition program. It was also the part that the architects and Liang concentrated on in modifying the design. Because the focus of this paper is the form of the museum, my use of this term will also be limited to this hall.

3.

Liang was repeatedly criticized for his revivalist ideas in China’s political movements, from the 1950s until his death, and most Chinese still identify his work, and Liang himself, with the “big roof,” even though they may not know of his involvement in the design of the National Central Museum in Nanjing in the 1930s.

4.

Pan Guxi, ed., Zhongguo Jianzhu Shi (History of Chinese architecture) (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2001), 382–83.

5.

Yang Yongsheng and Gu Mengchao, eds., Ershi Shiji Zhongguo Jianzhu (Architecture of twentieth-century China) (Tianjin: Tianjin Kexue Jishu Chubanshe, 1999), 207–8.

6.

Liu Dingming, comp., “Guoli Zhongyang Bowuyuan Choubeichu 1933nian 4yue-1941nian 8yue choubei jingguo baogao” (The report of the Preparatory Committee of the National Central Museum on Its Work from April 1933 to August 1941), Minguo Dang’an (Republican archives), no. 2 (2008), 27–33.

7.

“Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Zhengxuan Jianzhu Tu’an Zhangcheng” (The guidelines of the Architectural Committee of the National Central Museum for the design competition), in Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Jianzhu Tu’an Shencha Weiyuanhui Baogao (The report on architectural design by the jury under the Preparatory Committee of the National Central Museum) (Nanjing: Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui, 1935), 1.

8.

Ibid.

9.

Wilma Fairbank, Liang and Lin: Partners in Exploring China’s Architectural Past (Philadelphia: University of Pennsylvania Press, 1994), 23–30

10.

Liang Sicheng, Qingshi yingzao zeli (Qing structural regulations) (Beiping: SRCA, 1934), in Liang Sicheng quanji, vol. 6 (see note 1).

11.

Yingzao fashi (YZFS; State building standards) is the oldest existing Chinese technical manual on building construction. It was compiled by Li Jie (Li Chieh) in 1100, published by the Northern Song government in 1103, and enforced in 1104. For a general introduction to this book, see Else Glahn, “Unfolding the Chinese Building Standards: Research on the Yingzao fashi,” in Chinese Traditional Architecture, ed. Nancy Shatzman Steinhardt (New York: China Institute in America, 1984), 48–57. For biographical information on Li Jie, see Else Glahn, “Li Chieh,” in Sung Biographies, ed. Herbert Franke (Wiesbaden: Steiner, 1976), 2:523–29. According to a study by the Chinese scholar Cao Xun, “Jie” should be “Cheng,” which corresponds to Li’s courtesy name. See Cao, “Yingzao fashi Chongning ben—Wei jinian Li Cheng Yingzao fashi kanxing 900 zhounian er zuo” (The Chongning edition of Yingzao fashi—In commemoration of the nine-hundredth anniversary of Yingzao fashi by Li Cheng), Jianzhushi 108 (April 2004), 90–100. On Zhu Qiqian, his discovery and republication of Yingzao fashi, his establishment of the Society for Research in Chinese Architecture, and the study of Yingzao fashi by Liang Sicheng and his colleagues in the society, see Lin Zhu, Koukai Lu Ban de damen: Zhongguo Yingzao Xueshe shilue (Opening the door of Lu Ban: A brief history of the Society for Research in Chinese Architecture) (Beijing, 1995); Shiqiao Li, “Reconstituting Chinese Building Tradition: The Yingzao fashi in the Early Twentieth Century,” JSAH 62 (December 2003), 470–89.

12.

Liang Sicheng, preface to Jianzhu sheji cankao tuji (Pictorial references for architectural design), vol. 1 (Beping: SRCA, 1935), in Liang Sicheng quanji, 6:233–36.

13.

“Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Zhengxuan Jianzhu Tu’an Zhangcheng,” in Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Jianzhu Tu’an Shencha Weiyuanhui Baogao, 8.

14.

“Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Zhengxuan Jianzhu Tu’an Shencha Weiyuanhui Baogao,” in Guoli Zhongyang Bowuyuan Jianzhu Weiyuanhui Jianzhu Tu’an Shencha Weiyuanhui Baogao, 1–7.

15.

Ibid., 3. On Liang as “special member,” see the text at n. 7.

16.

These drawings are preserved in the archives of the museum (now Nanjing Museum) and Nanjing Chengshi Jianshe Dang’anguan (Archives of Nanjing Municipal Construction). For help in translating the Liao and Song architectural terminology from Chinese to English, I thank Nancy Shatzman Steinhardt and her pioneering work in her book Liao Architecture (Honolulu: University of Hawai’i Press, 1997).

17.

Chen Mingda, Cheng Mingda jianzhu yu diaoke shilun (Essays by Chen Mingda on ancient architecture and sculpture) (Beijing: Wenwu Chubanshe, 1998), 215–16.

18.

Liang Sicheng, Zhongguo jianzhu shi (A history of Chinese architecture), in Liang Sicheng quanji, 4:216.

19.

Gin-djih Su (Xu Jingzhi), Chinese Architecture—Past and Contemporary (Hong Kong: Sin Poh Amalgamated, 1964), 136.

20.

Lu Haiming and Yang Xinhua, eds., Nanjing Minguo jiangzhu (Architecture of the Republican period in Nanjing) (Nanjing: Nanjing University Press, 2001), 127–32.

21.

The YZFS stipulates that the largest size of cai was for the construction of “the palatial hall (dian) of nine to eleven bays”; the second for “the palatial hall of five to seven bays”; and the third for “the palatial hall of three to five bays and the government-office-hall (tingtang) of seven bays.” Accordingly, we know that in the Song dynasty, buildings of nine and eleven bays were those of the highest rank. Here dian and tingtang also refer to two different structural types. The entry titled “Grading System in Chinese Historical Architecture” (by Fu Xinian) in the Chinese Encyclopedia (1988) specifies that dian was for palaces and temples and tingtang for governmental offices and the houses of commoners. Today, only three buildings of the higher category dated before the early twelfth century have survived: the Daxiongbao halls of Fengguo Monastery and Huayan Monastery, mentioned in this essay, and the Sanqing Hall of the Daoist temple Xuanmiao Guan in Suzhou, which was built in 1175–77.

22.

Tadashi Sekino, “Manshū Giken Hōkohuji Daiyūhōden” (Daxiongbao Hall of Fengguo Monastery in Yi county of Manchuria), Bijiutsu kankyo, no. 14 (February 1932), 37–49.

23.

Liang Sicheng and Liu Dunzhen, “Datong gujianzhu diaocha baogao” (Report on the investigation of ancient architecture in Datong), Zhongguo Yingzaoxueshe huikan (hereafter cited as SRCA Bulletin) 4, nos. 3 and 4 (December 1933), 1–168; in Liang Sicheng quanji, 2:49–176.

24.

Sekino’s report contains data of the plan only in chi (1chi = 1/3 meter) and provides no measurement of the height. For more information on this building, see Du Xianzhou, “Yixian Fengguosi Daxiongbaodian diaocha baoga” (Report on the investigation of Daxiongbao Hall of Fengguo Monastery in Yi county), Wenwu, no. 2 (February 1961), 9.

25.

Liang, Jianzhu sheji cankao tuji (see n. 12), 263.

26.

For data on this building, see Liang Sicheng, “Baodi Guangjisi Sandashidian” (Sandashi Hall of Guangji Monastery in Baodi county), SRCA Bulletin 3, no. 4 (December 1932), 1–50, in Liang Sicheng quanji, 1:249–86.

27.

For data on this building, see Liang Sicheng, “Jixian Dulesi Guanyinge, Shanmen kao” (Research on Guanyin Pavilion and the Front Gate of Dule Monastery in Ji county), SRCA Bulletin 3, no. 2 (June 1932), 1–92, in Liang Sicheng quanji, 1:161–223.

28.

Although according to Yingzao fashi, “All principles of constructing buildings are based on cai,” because Liang did not discover its relation to the overall composition of a building, he applied cai only to the designs of the bracketing and lintel, not to the other parts of the museum. Chen Mingda, in his study of the wooden pagoda of Fogong Monastery in Ying county, Shanxi province (1056) in the 1960s, made the first efforts to discover the influence of cai as a basic module in the dimensions of a building design. (See Chen, Yingxian Muta [Beijing: Wenwu Chubanshe, 1966].) The application of a module system to the design of ancient Chinese architecture continues to be a key issue in the study of Chinese architectural history.

29.

The most likely reason that the decorative well specified in the design was not implemented during construction was postwar financial constraints. But during the renovation of the museum carried out from 2009 to 2013, the ceiling was refurnished with elaborate details and colors as well as three decorative wells.

30.

Liu Dunzhen, “Hebei, Henan, Shandong gujianzhu diaocha riji” (Diary of field investigation in Hebei, Henan, and Shandong), Liu Dunzhen wenji (Collected works of Liu Dunzhen) (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 1987), 3:89.

31.

Qi Yingtao, “Hebei Xincheng Kaishansi dadian” (The main hall of Kaishan Monastery in Xincheng county of Hebei), Wenwu cankao ziliao, no. 10 (1957), 23–29.

32.

Liang had noticed the aesthetic effect of the tuishan method while he was studying Qing architecture in the early 1930s. See Liang, Qingshi yingzao zeli (see n. 10), 40.

33.

Liang, Zhongguo jianzhu shi, in Liang Sicheng quanji, 4:146.

34.

Liang, “Baodi Guangjisi Sandashidian” (see n. 26).

35.

James Fergusson, History of Indian and Eastern Architecture (New York: Dodd, Mead, 1891), 687–88.

36.

Georg W. F. Hegel, The Philosophy of History, trans. J. Sibree (New York: Dover, 1956).

37.

Lin Huiyin, “Zhongguo jianzhu de jige tezheng” (Some characteristics of Chinese architecture), SRCA Bulletin 3, no. 1 (1932), 163–79. Translation mine.

38.

Liang, preface to A Pictorial History of Chinese Architecture (see n. 1), 207.

39.

Lin, preface to Liang, Qingshi yingzao zeli (see n. 10); Liang Sicheng, “Dougong jianshuo” (A brief introduction to dougong), in Liang, Jianzhu sheji cankao tuji (see n. 12), 291–95. Translations mine.

40.

Fairbank, Liang and Lin, 29.

41.

Osvald Sirén, A History of Early Chinese Art (London: Ernest Benn, 1930), 4:72. For Sirén’s influence on Liang and Lin, see Li Jun, “Gudian zhuyi, jiegou lixing zhuyi yu shixing de luoji—Lin Huiyin, Liang Sicheng zaoqi jianzhu sheji yu sixiang de zaijiantao” (Classicism, structural rationalism, and poetic logic: Rethinking the early architectural design and thought of Lin Huiyin and Liang Sicheng), Jianzhushilun huikan, no. 5 (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2012), 383–427.

42.

Han Baode, Ming/Qing jianzhu erlun (Two essays on Ming and Qing architecture) (Taipei: Jing Yu Xiang Chubanshe, 1988); Xia Zhujiu, “Yingzao Xushe-Liang Sicheng jianzhushi lunshu gouzao zhi lilun fenxi” (A historiographical study of the history of Chinese architecture written by the SRCA and Liang Sicheng), Quarterly of Taiwan Social Studies 3, no. 1 (Spring 1990), 6–48.

43.

Gwendolyn Wright and Janet Parks, eds., The History of History in American Schools of Architecture, 1865–1975 (New York: Temple Hoyne Buell Center for the Study of American Architecture, 1990), 24–25.

44.

Johann Joachim Winckelmann, History of Ancient Art, trans. Alexander Gode (New York: Frederick Ungar, 1968), 3:115–43.

45.

Osvald Sirén, Chinese Sculpturefrom the Fifth to the Fourteenth Century (London: Ernest Benn, 1925).

46.

Liang, Pictorial History of Chinese Architecture (see n. 1), 192–93, 207–08.

47.

Lai Delin, “Kexuexing yu Minzuxing—Jindai Zhongguo de Jianzhu Jiazhuguan” (Scientificity and nationality: Double attitudes toward a modern Chinese architecture), Jianzhushi 62 and 63 (February and April 1995), 48–59, 59–76.

48.

Lin, “Zhongguo jianzhu de jige tezheng” (see n. 37).

49.

Shen Tong, “Linpeng Jianzhushi yu ‘Guojishi’ jianzhu xinfa” (Architect Lindbohm and the new architectural method of the “International Style”), Shishi xinbao, 15 February 1933.

50.

Liang, preface to Jianzhu sheji cankao tuji, 1935 (see n. 12), 235–36. Translation mine.

51.

Ibid.

52.

Joseph R. Levenson, Liang Chi-chao [Liang Qichao] and the Mind of Modern China (Cambridge, Mass.: Harvard University Press, 1953).

53.

Liang Qichao, Xin min shuo, in Liang Qichao quanji (Complete works of Liang Qichao) (Beijing: Beijing Chubanshe, 1999), 3:709–14.

54.

Liang Qichao, Zhongguo zhi Wushidao, in Liang Qichao quanji, 5:1376–420. For further discussion of the intellectual relationship between the father and the son, see Shiqiao Li, “Writing a Modern Chinese Architectural History: Liang Sicheng and Liang Qichao,” Journal of Architectural Education 56 (September 2002), 35–45.

55.

Lu Xun, “Kan jing you gan” (Reflections in a mirror), Yusi Weekly, 2 March 1925.

56.

For example, Dai Jitao, an influential member of the Nationalist Party and the chairman of the Organizing Committee of the First National Games, held in 1930, remarked: “We must make physical exercise routine for every family to build up sound bodies and to cultivate a strong spirit. To be free one needs first to be free of disease; to gain equality requires first that one be strong and dauntless. A strong father has no weak son; eugenics produces good species. With a healthy race, the foundation of the country can be solid.” Wu Liande, ed., Zhongguo daguan (China as she is) (Shanghai: Liangyou Tushu Yinshua Youxian Gongsi, 1930), n.p. Translation mine.

57.

Liao Xintian, Qingdai beixue shufa yanjiu (A study on the stele style calligraphy of the Qing dynasty) (Taipei: Taipei Municipal Art Museum, 1993), 197.

58.

Alan Colquhoun, “Three Kinds of Historicism,” in Modernity and the Classical Tradition: Architectural Essays, 1980–1987 (Cambridge, Mass.: MIT Press, 1989), 6.

59.

Liang Sicheng, “Women suo zhidao de Tangdai fosi yu gongdian” (Buddhist Monasteries and Palaces of the Tang Period We Know at This Time), SRCA Bulletin 3, no. 1 (March 1932), 75–114.

60.

Liang, “Jixian Dulesi Guanyinge, Shanmen kao” (see n. 27).

61.

Liang, Zhongguo jianzhu shi (see n. 18), 92.

62.

Liang, “Jixian Dulesi Guanyinge, Shanmen kao” (see n. 27), 168–69; Zhongguo jianzhu shi (see n. 18), 92. Since Liang, many scholars have contributed to the further study of Tang architecture and its development in the Liao and Song dynasties. Their studies reveal that a more nuanced development of Chinese architecture than that recounted by Liang and Lin. Here I can mention but a few of the other scholars’ studies: Nancy Shatzman Steinhardt, Liao Architecture (Honolulu: University of Hawai’i Press, 1997); Fu Xinian, ed., Zhongguo Gudai Jianzhu Shi, vol. 2 (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2003); Guo Daiheng, ed., Zhongguo Gudai Jianzhu Shi, vol. 3 (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2003); Jiren Feng, Architecture and Metaphor: Song Culture in the Yingzao Fashi Building Manual (Honolulu: University of Hawai’i Press, 2011); and Tracy Miller, “The Eleventh-century Daxiongbaodian of Kaihauasi and Architectural Style in Southern Sanxi’s Shangdang Region,“ Archives of Asian Art 58 (2008), 1–42; “Northern Song Architecture in Southern Shanxi Province,” Journal of Song-Yuan Studies 29 (1999), 91–109; “Something Old, Something New, Something Borrowed: Regional Style in Tenth Century Timber-frame Architecture,” in The Five Dynasties and Ten Kingdoms, ed. Peter A. Lorge (Hong Kong: Chinese University Press, 2011), 165–219.

63.

Liang Sicheng, “China’s Oldest Wooden Structure,” Asia 41 (July 1941), 374–77, in Liang Sicheng quanji, 3:361–64.

64.

Nancy Shatzman Steinhardt, “The Tang Architectural Icon and the Politics of Chinese Architectural History,” Art Bulletin 86, no. 2 (June 2004), 228–54.

65.

I came up with this idea during a conversation with Professor Douglas Fix on 21 March 2006, at Reed College, Portland, Oregon. Although Liang discovered the main hall of Foguang Monastery—an exemplary building of Tang architecture constructed in 857—two years after the Central Museum competition, he was unable to make use of it to revise the museum design because by then its reinforced-concrete structure was almost completed. See Lu and Yang, Nanjing Minguo jiangzhu (see n. 20), 130. The working drawings preserved in the museum’s archives also demonstrate that the architects made no changes when construction resumed after the war. So it is safe to say that the discovery of the monastery had little effect on the design of the museum.

66.

Liang Sicheng, “Zhongguo jianzhu de tezheng” (Characteristics of Chinese architecture), Jianzhu xuebao, no. 1 (January 1954), 36–39, in Liang Sicheng quanji, 5:179–84.

67.

As a matter of fact, this Beaux-Arts notion is also visible in Sir Banister Fletcher’s History of Architecture on the Comparative Method, where the “comparative” part of each chapter discusses plans, or the “composition,” and walls, openings, roofs, columns, moldings, and ornament,or the “elements.”

68.

Liang, “Jixian Dulesi Guanyinge, Shanmen kao” (see n. 27), 188. Translation mine.

69.

For more on the issue of classical orders and the analogy in Chinese architecture, see Zhao Chen, “‘Minzu zhuyi’ yu ‘gudian zhuyi’—Liang Sicheng jianzhu lilun tixi de maodunxing yu beijuxing zhi fenxi” (Nationalism and classicism—An analysis of the contradiction and tragedy of Liang Sicheng’s architectural theory), in Zhongguo jindai jianzhu yanjiu yu baohu, ed. Zhang Fuhe (Beijing: Tsinghua University Press, 2001), 2: 77–86; Li, “Writing a Modern Chinese Architectural History,” (see n. 54); Lai Delin, “Liang Sicheng, Lin Huiyin Zhongguo jianzhushi xiezuo biaowei” (A historiographical study of Liang Sicheng and Lin Huiyin’s writings on Chinese architectural history), 21st Century, no. 64 (April 2001), 90–99.

70.

Liang, A Pictorial History of Chinese Architecture (see n. 1), 225.

71.

Liang, “Baodi Guangjisi Sandashidian” (see n. 26), 272. Translation mine.

72.

Liang, “Jixian Dulesi Guanyinge, Shanmen kao” (see n. 27), 169. Translation mine.

73.

For instance, Liang’s Penn colleague Tong Jun (Tung Chuin), who received his master of architecture degree in 1928, held the view that material ingenuity is a positive value in architecture. He criticized “Chinese revival” architecture in 1945: “The only similarity between the Chinese wooden construction system and the reinforced-concrete method is their structural principle, which is frame-like rather than case-like. In terms of the cost-effectiveness of materials, however, the optimal span of a wooden frame is only half that of the steel one. Hence the proportions of all the elements will be totally different. Judging from the size of the elements alone, to imitate the palatial structure with reinforced concrete is a waste, and [the palatial structure thus] should not be taken as a model. Moreover, it will coat what is already a sufficient number of concrete beams and columns with paint, and cover what is already a sufficient number of roof terraces with tiles. Indeed, it can hardly be called rational.” See Tong, “Woguo gonggong jianzhu waiguan de jiantao” (Rethinking the appearance of the public buildings in our country), Neizheng zhuankan 1 (October 1945), in Tong, Tong Jun wenji (Collected works of Tong Jun) (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2000), 1:118–21. Translation mine.

74.

Liang, Zhongguo jianzhu shi (see n. 18), 107. Translation mine.

75.

Liang, A Pictorial History of Chinese Architecture (see n. 1), 228.

76.

Chen, Cheng Mingda jianzhu yu diaoke shilun (see n. 17), 215–16.

77.

Liang and Liu, “Datong gujianzhu diaocha baogao” (see n. 23), 155–57. Translation mine.

78.

Ibid., 157.

79.

Ernst Boerschmann, Chinesische Architektur (Berlin: Verlag Ernst Wasmuth, 1925), 2:49, quoted in W. Perceval Yetts, “Writings on Chinese Architecture,” SRCA Bulletin 1, no. 1 (July 1930), 1–8.

80.

Lin, “Zhongguo jianzhu de jige tezheng” (see n. 37), 163–79. Translation mine. Lin must have acquired this functional view of the roof in Chinese architecture from James Fergusson. See James Fergusson, Illustrated Handbook of Architecture: Being a Concise and Popular Account of the Different Styles of Architecture Prevailing in All Ages and Countries (London: John Murray, 1855), 140.

81.

Lin, preface to Liang, Qingshi yingzao zeli (see n. 10), 23. Translation mine.

82.

Liang, Zhongguo jianzhu shi (see n. 18), 216. Translation mine.

83.

Su, Chinese Architecture (see n. 19), 136.

84.

Hu Shih, Chinese Renaissance (Chicago: University of Chicago Press, 1934); Hu Shih, Hu Shih koushu zizhuan (An oral autobiography of Hu Shih), ed. and trans. Tang Degang (Taipei: Zhuanji Wenxue Chubanshe, 1981), 165–79; Geng Yunzhi, Hu Shi xinlun (A new look at Hu Shi) (Changsha: Hunan Chubanshe,1996), 66.

85.

Zhao Shen, “Fakan ci” (Forward the journal), Zhongguo jianzhu (Chinese architecture), preview issue (November 1932), 1.

86.

Lai Delin, “Zhezhong beihou de linian—Yang Tingbao jianzhu de bili wenti yanjiu” (An ideal underlying eclectic design: A study on the proportions of Yang Tingbao’s architecture), Yishushi yanjiu 4 (2002), 445–64; Delin Lai, “Searching for a Modern Chinese Monument: The Design of the Sun Yat-sen Mausoleum in Nanjing,” JSAH 64, no. 1 (March 2005), 22–55.

87.

Chen Mingda once said that the goal of the society was “to find the [doctrines of] ancient Chinese architecture.” Chen, Cheng Mingda jianzhu yu diaoke shilun (see n. 17), 215–16.

88.

As a matter of fact, with the exception of Jianzhen Memorial (Yangzhou, 1963), which commemorated the Tang dynasty monk Jianzhen and was designed in the Tang style Liang had learned from his study of the main hall of Foguang Monastery, Liang’s other designs did not emphasize stylistic purity. In 1954 he proposed the idea of “architectural translatability” (jianzhu keyilun) in the design of a Chinese-style architecture. He suggested that by substituting Chinese architectural elements for the corresponding Western elements of a Western-style building, the Western style could become Chinese. This idea, which demonstrated the influence of the Beaux-Arts dichotomy between “composition” and “elements,” was exhibited in numerous designs after the 1870s. See Lai Delin, “Liang Sicheng ‘jianzhu keyilun’ zhiqian de Zhongguo shijian” (Architectural translation and the notion of architectural translatability in modern China), Jianzhushi 137 (February 2009), 22–30; “Goutu yu yaosu—xueyuanpai laiyuan yu Liang Sicheng ‘wenfa-cihui’ biaoshu ji Zhongguo xiandai jianzhu” (Composition and Elements: The Beaux-Arts origin and Liang Sicheng’s “grammar and vocabularies” study of Chinese architecture and its influence on modern Chinese architecture), Jianzhusi 142 (December 2009), 55–64; Li Hua, “Cong Buza de zhishi jiegou kan ‘xin’ er ‘Zhong’ de jianzhu shijian” (“A ‘new’ and ‘Chinese’ architecture in relation to the Beaux-Arts tradition as a system of knowledge”), in Zhongguo jianzhu 60 nian (1949–2009): Lishi lilun yanjiu (Sixty years of Chinese architecture [1949–2009]: History, theory, and criticism), ed. Jianfei Zhu (Beijing: Zhongguo Jianzhu Gongye Chubanshe, 2009), 33–45.