In 1900, the first year of this century, a Daoist monk named Wang Yuanlu who lived near the Mogao Caves at Dunhuang, discovered the secret "library cave," which was sealed and covered by fresh murals about 1000 years ago. At first, Wang's discovery did not attract much attention. Only a few Chinese scholars heard the news and collected some manuscripts and silk paintings from the secret cave. For example, Ye Changzhi, the chief official of Education Department of Gansu Province, collected five volumes of Buddhist sutras and two scroll-paintings. However, this exciting news soon attracted the attention of foreign explorers. The newly discovered old Chinese cultural relics were just the things they were looking for. When Sir Aurel Stein came into the small Dunhuang town and heard the exciting rumor, the discovery of many antiquities in one of the Mogao caves, he decided to visit the secret cave right away. Besides the relics discovered in the library cave, he found hundreds of caves decorated with paintings and sculptures. The Dunhuang Caves were "re-discovered" by the Western world.

Part One
Western SCHOLARSHIP

1. Reports from the Explorers
In 1907, Aurel Stein visited the Dunhuang caves and took some 26 boxes of antiquities from the secret "library cave," including manuscripts, printed books, paintings on silk and paper, and some ritual objects. In the next year, Paul Pelliot came to Dunhuang and searched the antiquities left by Stein and collected the most important Chinese texts that were ignored by Stein due to his lack of knowledge of Chinese language and history. They both, however, photographed some of the mural paintings and painted sculptures remaining in the caves. Stein published his photographs in the Thousand Buddhas: Ancient Buddhist Paintings from the Cave-temples of Tun-huang on the Western Frontier of China (London, 1921) and Pelliot edited his personal records on the caves and published a part of them with his photographs in Les Grottes de Touen-houang (Paris, 1914-24).

The photographs published by Stein and Pelliot have long served as primary visual sources for the study of the Dunhuang caves. Although these photographs are in black and white, they are the only visual materials available for those art historians who could not visit the site in person.

Stein and Pelliot respectively numbered the Mogao Grottoes and provided very brief introductions. As a professional explorer, Stein's interest and knowledge were mainly in northern India and nomadic Central Asia. Facing the complicated Dunhuang caves filled with Chinese inscriptions, he deeply regretted for his ignorance of Chinese language and history. He could only use the very limited literature translated by his Chinese interpreter and his own experience in the "styles" of Khotan paintings, which sometimes could be related to the Dunhuang murals, to figure out the chronology of the caves. Naturally, he analyzed the "styles" of Dunhuang cave art by comparing them with the art works remaining at sites located to the west of Dunhuang, and he traced the "origin" of Dunhuang art to ancient Greece and Rome. This conclusion was probably not only derived from a contemporary theory of the "origin of the world civilizations" popular in the western world, but also due to his limited knowledge of Chinese art history.

An excellent sinologist, Pelliot contributed to establishing the basis for the study of the Dunhuang caves in the West. He personally recorded many Chinese inscriptions that remained on the murals and examined some of the inscriptions against the reliable historical texts. His knowledge of Chinese literature and other languages enabled him to do research beyond brief description of the works. His personal diary on the inscriptions, written in 1908, is still quoted by modern scholars as "first hand records." However, Pelliot's approaches mainly concentrate on literary values of the inscriptions in the study of Chinese history. Like many Chinese scholars and western sinologists, Pelliot's interest was not on the artistic quality of the caves, but their historical significance in general.

Based on the primary reports and photographs published by Stein and Pelliot, western scholars started to study the art works remaining in the Dunhuang caves. The major trends of western art historians in the studies of Dunhuang art are "formal analysis," "iconographical interpretation," and "material borrowing" (i.e., to borrow specific materials from the Dunhuang caves to solve the difficult problems in Chinese art history).

Formal Analysis
Although Stein and Pelliot published their photographs of the Dunhuang caves and some paintings on silk discovered from the "library cave" (Cave 17) at Dunhuang, their literary records are not sufficiently detailed for art historians to do a comprehensive study. However, some scholars started to approach the Dunhuang materials, using the method of "formal analysis," a method developed in the field of European art and was very popular in the first half of this century. When Wolfflin's disciple Ludwig Bachhofer formulated a developmental scheme of Chinese painting, he unified Wolfflin's polar categories into a pair of key concepts expressed as "two-dimensional representation versus three-dimensional representation," which he then employed as basic, unquestionable criteria to measure the development of Chinese art: "The effort of many centuries had paved the way for 15th century European painting to render space as an all comprehensive and unconfined unity......In the Far East too, realism proceeded step by step and centuries passed, before a preliminary goal was reached."

Armed with his assumption and concepts on the "evolutionist process" of Chinese art, Bachhofer published a series of articles partially dealing with the mural paintings of Dunhuang ("Die Raumdarstellung in der Chinesischen Malerei des ersten Jahrtausends n.Chr." [Space conceptions in Chinese painting during the first millennium after Christ]. Munchner Jahrbuch der bildenden kunst, Band VIII, 1931. Translated into English by Harold Joachim, manuscript in Rubel Library; "On the Origin and Development of Chinese Art." Burlington Magazine 67, 1935 p.251-64; "Chinese Landscape Painting in the Eighth Century," Burlington Magazine 67, 1935, p.189-91.). He also published a book on Chinese art history (A Short History of Chinese Art. Washington: Pantheon Books, Inc., 1946. 139 pp. 136 illus.). Because human figures are the main motifs in early Dunhuang caves, Bachhofer seems tentative in constructing his "chronological order" for these caves at first. Finally, he uses the small decorative mountains in the narrative scenes as his key to date the whole cave. For example, he finds that "there are small sections or 'cells' of space, glued together like objects" appearing in the narrative illustration painted on the eastern wall of Cave 135 (Pelliot number, Dunhuang Cave 428). Based on his analysis of these space "cells," he dates this cave to 530-540. Unfortunately, this cave has been firmly dated to the Northern Zhou period (556-580) through comparing the inscriptions with the manuscripts discovered in the "library cave."

Bachhofer's method of formal analysis might seem very attractive for some art historians who did not have any background in Chinese language and history, because this method does not require an intensive investigation of the context of the works of art but simply watching them with "sharp eyes." However, this method has been proven to be inadequate in the field of Chinese art, particularly in Dunhuang art. Unlike the study of Indian art, in which the mistakes made by formal analysts are not easily recognizable by contextual research because of the lack of historical records and dated inscriptions, we have numerous ancient texts and inscriptions related to Chinese art. In the case of Dunhuang art, we have various manuscripts discovered from the "library cave" and many inscriptions remaining on the wall. These literary sources are invaluable to our study of the cave art. If we simply continue the method of "formal analysis" in Dunhuang studies, our conclusions would be less accurate.

Some sinologists recognized the vital limitation of the method soon after Bachhofer published his book of Chinese art in 1946, and they started to criticize this method. John A. Pope severely criticized the art historians who simply employed this method in their studies and claimed:

Lacking the essential tool of language and relying on stylistic formulae which had been tried and proved in the European field, the art historian ventured into the vast and complex field of Chinese art. The results of his efforts provide ample evidence that his equipment has been unsuited to the task.

It seems that Pope's criticism did not bother some of the art historians in this field. We have found that this method has been carried on with slight changes.

Basil Gray had long worked on the paintings on silk and paper brought to London from Dunhuang by Aurel Stein in 1908 and studied the photographs of the wall paintings taken soon after by Paul Pelliot. The photographs taken by the Vincents in 1948 also enriched his knowledge on the Dunhuang caves. Decisively, Gray visited the Dunhuang caves for seven days in May 1957, thereafter, in Arthur Waley's words, he "has established himself as the leading Western authority on the Tun-huang paintings." His book, Buddhist Cave Paintings at Tun-huang (Chicago: the University Press, 1959. 86 pp. 70 pls.) raised enthusiastic attentions in the West. This book has been reviewed by at least four famous scholars in the field of Chinese art: Bachhofer, L. (Ars Oriental, V. 1963, p.311-314), Demieville, P. (TP, XLVIII, 4-5, p.463-474), Tucci, G. (East & West, XI, 2-3, p.191) and Drake, F.S. (JOS, V, 1-2. 1965, p.236-240).

Gray's approaches to the Dunhuang caves are more comprehensive and intensive than Bachhofer's. His effort to reconstruct the history of Dunhuang provides art history students with a historical background of the complicated caves, although the "local history" is too simple (only three pages) to support his study on the complicated caves. Gray's contribution is his publication of some color photographs of the caves, which at least can help students understand the richness of colors of the Dunhuang murals.

Gray's study includes two problems, vague description in "stylistic development" and mis-identification based on fragmentary visual materials.

(1) "Stylistic development." Like Bachhofer, he assumes that stylistic changes from the early period to the later ones constitute a "development," from two-dimensional representation to three-dimensional representation, but he bases his formal analysis on the "influences" from Central Asia on the Dunhuang caves, not the caves themselves. Awkwardly, he uses the "chiaroscuro" technique, which emphasizes concave-convex effects by light-dark colors, as a standard to distinguish the paintings. Gray's comments on the caves are very brief. His comparison between the mural paintings of Central Asia and the Dunhuang caves is fragmentary. Bachhofer objected to Gray's ignorance of many important works of Central Asia in his review on this book.

(2) Mis-identification. Gray made some mistakes in his iconographical approaches toward the paintings. Sometimes his mistakes are so striking that even the formalist Bachhofer, whose major concern is only with the "style" could recognize his mis-identification of the scenes of the "Jataka of Deer Ruru." (see Bachhofer's review in ARS ORIENTALIS, 1963.) Gray also inherited some mistakes made by Arthur Waley. For example, the "fighting scene" in Cave 217 was identified by Waley as "military exercises" and Gray continued this identification. This "fighting scene" is a part of a large painting representing the Amitayurdhyana Sutra. If one puts this scene back to its original location, it would be not difficult to rightly identify it as a "true fighting scene"--the fighting between Prince Ajatasatru's troops and his father's bodyguards.

Shih Hsio-yen also tries to analyze the early caves at Dunhuang with this method of formal analysis ("New Problems in Tun-huang Studies." In Proceeding of the International Conference on Sinology, Section of History of Art, Taipei, 1981, pp.221-232. Also in Cheng Te-k'un ed. Studies in Chinese Art, The Chinese University Press, Hong Kong, 1983). Shih's approach, however, slightly differs from the previous western scholars. She takes the "brushwork" into consideration besides the "space representation." For example, when she analyzes the western wall of Cave 285, she considers that this painting was made by "a person with training in the practice of brush and ink, and the calligraphic line." Shih's study of the Chinese painting techniques added some Chinese characteristics into this method of analysis that was developed on the basis of western art.

Marylin Rhie's studies in Chinese art mainly concentrate on Buddhist sculpture. She defines her method of research way as follow:

The methodology for the study is one of comparative stylistic analysis in conjunction with historical documentation. In a step-by-step, complementary process, the art of one region helps to clarify the developments in the art of another region. Where parallel developments match, there is established, like rungs in a ladder, a relation which, along with other such matching relations, can result in clarifying or solving more satisfactorily some chronological gaps and in understanding more accurately the developments and movements of regional styles in the art of the time.

With this method, she has dramatically connected Dunhuang sculptures with many similar sculptures in various regions, including central China, Japan, Korea, Central Asia and India, and shed light on some high quality sculptures which have long been considered chronologically hopeless objects. Her success lies in her deep awareness of the importance of documents and inscriptions in Chinese art studies. Because the chronology of most caves at Dunhuang has been firmly established by the continued efforts of the scholars in various fields of several generations, she used some Dunhuang sculptures as the standards by which to date problematic sculptures and refreshed our ideas on these works. However, her study still seems fragmentary and the historical approach is too general. Sometimes, she picks up a detail from a monk's robe and compares it to a Buddha's garment which could be located some thousand kilometers away. This kind of "free" selection may surprisingly broaden our view but also mislead our thinking at the same time.

Iconographical Interpretation

Arthur Waley's Catalogue of Paintings Recovered from Tunhuang by Sir Aurel Stein (London, 1931) is the first comprehensive study of the paintings on silk and paper discovered in the "library cave" at Dunhuang. His interpretations are mainly based upon an intensive textual investigation of Buddhist sutras. In the early period, most scholars of the West and the East interpreted the Dunhuang paintings by connecting them only with Buddhist sutras. Because the subject matters of Dunhuang art are essentially Buddhism, this kind of iconographic studies is necessary at the beginning. However, if scholars limit themselves only at the Buddhist texts and ignore other texts and archaeological evidence, some motifs would be never interpreted.

British Museum started to catalogue all the paintings recovered from Dunhuang by Aurel Stein before these paintings were divided into two parts, under an agreement between the government of India and the British Museum, in 1918. Therefore, Waley's catalogue contains every painting of the Stein collection. He carefully recorded the conditions of these paintings and translated all the inscriptions on the painting. This catalogue has long been considered the standard text of the study of Dunhuang art among the Western scholars.

In "Literary Evidence for Early Buddhist art in China," (Artibus Asiae, XVI, 1-2, 1953) Soper translated many Chinese records on the history of Buddhism into English and used these texts to interpret Buddhist iconography. Soper's translation has been used as background material for the study of early Buddhist art in China since the 1960s when it was published. Because some iconography in Chinese Buddhist art is not based on Buddhist sutras but historical records, or even some oral legends, scholars need knowledge other than Buddhist sutras. Soper's intensive introduction of non-sutra materials enables many western scholars to interpret Chinese Buddhist iconography in various ways.

Soper himself interprets some "famous images" painted on a piece of silk, which was cut into two parts and kept in London and New Delhi respectively, by connecting these images with Chinese historical records, biographies of Buddhist monks, and travelers' dairies ("Representations of Famous Images at Tun-huang," ARTIBUS ASIAE, vol. XXVII,4, 1966). Soper's rich knowledge of Chinese history and excellent skill in Chinese language enable him to use various literary sources other than Buddhist sutras. He also links his interpretations to the local history and emphasizes the historical possibilities of his interpretations. Therefore, his conclusion seems much more convincing than that of the other scholars who only identify their images through Buddhist scriptures.

Roderick Whitfield re-investigated the iconographical identifications of the paintings on silk and paper collected in the British Museum, which were made by Arthur Waley fifty years ago, and published his new catalogue of the paintings with the assistance of Japanese scholars. Whitfield compiled most new achievements acquired by both western and eastern scholars and corrected some mistakes made by Waley. His focuses are chronology and iconography. To build up the chronology of these paintings, he uses the paintings with dated captions as standards of comparison with the others, focusing on such details as costumes, decorative patterns, and brushwork. To identify the contents of the paintings, he mainly compiles the previous scholarship and makes them as clear as possible. On the whole, Whitfield's approach seems conventional, but the systematic introduction on these very important paintings with high quality plates enables other scholars to do further research.

"Material Borrowing"

Almost every scholar wants to "borrow" something from the Dunhuang treasures to use in their own concentrations when they first encounter these objects. Because of the richness of Dunhuang caves, art historians working on different specialties can always find something interesting to bring to them. In particular, scholars working on Chinese landscape painting find they can really enjoy a new chapter for their narration on the history of Chinese landscape if they take Dunhuang materials into account. However, different scholars have different attitudes towards the cave art at Dunhuang.

Sullivan's books on Chinese landscape paintings, (The Birth of Landscape Painting in China. University of California Press, Berkeley & Los Angeles, 1962; Symbols of Eternity: The Art of Landscape Painting in China. Stanford University Press, Stanford, California, 1979; and Chinese Landscape Painting of the Sui and TUang. Berkeley: University of California Press, 1980.), have used landscape scenes in Dunhuang murals to discuss the history of Chinese landscape painting in general. Because the landscape paintings of pre-Tang periods (before 618) all disappeared except some Dunhuang murals, Sullivan has to rely on these only "reliable" works. But he basically considers Dunhuang paintings "Provincial and unrepresentative," therefore, he fails to connect these works with the related historical records. His study is mainly "descriptive," i.e., to describe the appearance of the so-called "landscape elements," and all the other issues are ignored. He follows Bachhofer in analyzing the progress of "space cells" in early landscape paintings of the Dunhuang caves and simply uses the works as examples to fill up the "time-gap" from post-Han to the end of Tang (220-907).

On the contrary, Anil de Silva argues for the importance and representativity of the Dunhuang landscape paintings:

Again it has been asserted that nearly all the art of Tun-huang, especially that of the earlier periods, was a provincial art. Study of the caves does not bear this out. Many caves may have been painted by local men who were not great artists, but this could be true anywhere, even in the capital. If the art of Tun-huang is to be dismissed simply as provincial art, without discrimination, then Chinese art in the capital must have reached a level never attained elsewhere in the world. The truth is that the importance of Tun-huang as a center of religion and learning--one has only to think of the size of the walled-up library--makes it probable that in many aspects of painting, and sculpture it was a center which gave lead to the rest of China.

Silva's idea derives from her personal study of the caves. She worked at Dunhuang for several weeks with her photographer Dominique Barbois. Her Chinese Landscape Paintings in the Caves of Tun-huang takes the Dunhuang murals as the basis to studying the landscape paintings of China in general. This fully illustrated book narrates the history of Chinese landscape paintings with Chinese art theories on landscape. She tries to locate the Dunhuang materials into the framework of Chinese art history and interprets some art phenomena in a Chinese contest. She attacks the art theory established by formalists and argues that Chinese art "does not represent a striving towards realism." In her opinion:

The essential form of this art, as of Gothic art in Europe, is that of juxtaposition without the limitation of linear perspective; the picture expands, revealing itself like a panorama. One of its essential qualities is that the beholder participates in the painting and identifies himself with what is represented. He must not remain detached for in doing so he would be able to grasp only part of the whole; he who sees "only a single aspect of the Tao will not be able to comprehend its totality," and the meaning will escape him for it is uncontrolled, suggestive, mysterious, stimulating, and "sharpens perception."

Silva's approach clearly differs from conventional studies on the Dunhuang caves and achieved a deeper understanding on the values of the cave art at Dunhuang. She not only uses Dunhuang materials to study the history of Chinese landscape paintings, but also to study the Dunhuang materials in a comprehensive context. However, her efforts to connect the Dunhuang visual materials with literary texts remain at the "primary level," because the literary sources she selected often can not be directly linked to the art works and her connection seems too "farfetched." Many important records could have been included, but she fails to use them except the records kept in the famous works of Chinese literature.

The Approaches of the New Generation (1980-> )

Judy Ho luckily visited the Dunhuang caves several times in the early 1980s and stayed in Cave 249, which she chose as her dissertation topic, for some days. This unique chance enabled her to put the partially published pictures into their original visual context and to consider the cave as a whole complex. Ho's dissertation, Tunhuang Cave 249: A Representation of the Vimalakirtinirdesa (Yale University, 1985), begins with an examination of the textual sources and contextual factors such as interpretive traditions, devotionalism, artistic models and the tastes and interests of the Chinese audience. The central part of her study is a reconstruction of the narrative sequence in Cave 249. In conclusion, she argues that the variable phenomenon of Sinicization is the determining factor in the unity and diversity of the illustrations of the sutra in Cave 249 and in other pre-Tang examples.

Ho has surpassed the conventional way viewing the Dunhuang paintings "piece by piece" and researched the works in their original visual context. This is a crucial step toward understanding the Dunhuang art in a comprehensive and complete manner. Probably too confident at her new interpretive way, she simply took a single Buddhist sutra, the Vimalakirti-nirdesa, as the textual basis of the cave. To put everything in this cave together, she had to "cut the feet to fit the shoes." She identified the two small figures located on the west slope as Vimalakirti and Manjusry (they could be, but there is no strong evidence), and then based her interpretations of all the other motifs on this uncertain identification. Because the whole dissertation is based upon an uncertain identification of two very small figures on the ceiling of the cave, if this identification is wrong (which is possible, but most scholars do not agree with her identification), the whole interpretation would be nonsense. Therefore, when Mr. Duan Wenjie judged that this dissertation should have failed, I agreed with him in some aspects.

Stanley Abe's "Art and Practice in a Fifth-Century Chinese Buddhist Cave Temple" (ARS ORIENTALIS, vol. 20, 1990) is based on parts of the author's Ph.D. dissertation, "Mogao Cave 254: A Case Study in Early Chinese Buddhist Art," University of California, Berkeley, 1989. In Abe's mind, "studies of Buddhist art often focused on the task of aligning imagery with Buddhist texts." Therefore, he approached Chinese Buddhist art in another way--a functional approach--by focusing on Cave 254 at Dunhuang. Abe discussed the relationship between Buddhist art and practice in Cave 254 and regarded "the art and architecture of a complete religious sanctuary in terms of those who designed, produced, and used the space and its decoration." Basing his discussion upon the context of the local history of Buddhism in the Liangzhou area, including Dunhuang, Abe considers the art project of Cave 254 as "an eclectic combination of imagery and practice." After an intensive discussion about the activities--visualization, the recitation of the names of the Buddhas, circumambulation, and some type of oral recitation--that probably took place in this cave, he concludes: "Cave 254 is evidence of how elements of architecture, painting, and sculpture were incorporated to meet the functional and pragmatic needs of both clergy and lay Buddhists at the Mogao site."

Abe recognized the limitation of simple iconographical study through Buddhist texts and started to search for the role of "producer" and "user" in the formation of the caves at Dunhuang. This kind of "functional approach," could probably lead to some new interpretation of the artistic form of the works. For example, he explains the "ambiguities" in the visual presentation of the Mahasattva Jataka in terms of its function. He thinks "that this painting was not expected to be the primary means by which the narrative was to be conveyed. These could be left to an oral presenter. As a visual adjunct to the narrative, the painting was thus free to emphasize the dramatic and emotional aspects of the tale without consideration for strict narrative logic." As a new interpretation of the artistic appearance of the painting, either wrong or right, it provides us with not only a fresh idea on the painting but also a new angle to approach other narrative paintings remaining in the caves. However, Abe's study remains at the "primary level." He limited himself to studying a single cave and barely made the very necessary comparative studies between Cave 254 and other contemporaneous caves at the same site and at other sites. Since Cave 254 was discussed as an "isolated case," the "credibility" of his conclusion needs to be reconsidered. Even for Cave 254 itself, some problems still remain unsolved. For example, we still do not know who built the cave and why. Abe's study has not built up the link between the cave and its patronage. Also, the inscriptions written beside each of the Buddha figures suggest that a viewer can easily understand the motifs and watch the paintings by himself, why is an "oral presenter" needed? The Mahasattva Jataka is only a small section of the whole project and its so-called "ambiguities" could be the artist's invention in narrative representation. The process of "image making" is as important as "image viewing" even in terms of religious requirement. Why should we only concern the "user" not the "creator"?

Victor Mair is a scholar working mainly on Chinese literature rather than art history, but his discussion of the relationship between some Dunhuang paintings and popular literary works seems very interesting (Painting and Performance, University of Hawaii Press, Honolulu, 1988). Based on intensive investigations on the "transformation texts" (bianwen, a genre of Chinese popular literature), he connects these texts with some paintings on silk and paper and some wall-paintings comprised of similar contents. He then concludes: "Essentially, it was the task of the artist to represent these manifestations [of the Buddhist figures] on paper or silk, or in wall-paintings, in which case they may be called pienhsiang ('transformation scenes or tableaux'). The pien-storyteller would then use the pienshiang as an illustrative device during his performance." Mair's reconstruction of the story-recitation tradition in China and its history in the whole world explains the function of some Dunhuang paintings in a different way. Mair tries to answer how some narrative paintings were used and their relationship with literature. This functional approach seems continued by Stanley Abe who is concerned not only how but why these paintings were used.

Wu Hung takes his new approach on the same subject--the relationship between Dunhuang art and Dunhuang literature--by focusing on a series of paintings consisting of the same theme: the "Subjugation of Demons." For a long time, researchers have been confused by the complicated arrangement of the episodes of the tale, because the episodes depicted on the large-sized painting seem illogical; to read these scenes according to the literary order, it is necessary to shift our gaze from one corner to another, to cross the whole width of the painting (which could be forty feet across), or to "scan" the complex composition to search for a minute detail. Our eyes and mind would spin until we got totally dizzy and finally gave up. However, all "Subjugation" murals created during the late Tang and Song have such an irregular composition. To resolve this and other problems, Wu Hung bases his new interpretation on two propositions: (1) devotional art is essentially an art of image-making rather than image-viewing, and (2) the process of image-making has its own logic that differs from those found in writing and oral recitation. He then figured out: "what the painter first presents to us through laying down an overall symmetrical composition, is the topic of the story: 'Raudraksa's contest with the sage' or the 'Subjugation of demons.' This topic is explained, enforced, and enriched by numerous secondary images painted in different spatial blocks. While within each block a limited narrative sequence may be established based on the original text, the links between these blocks must be invented according to their new spatial relationships. Although such a mural could not have been used in bianwen performances, it does tell the story, but in a different way. It is not a literary narrative created by writers or storytellers, but a pictorial narrative created by painters."

Wu Hung's success in interpreting the "Subjugation" paintings provides us with a model for the studies of the numerous illustrations of Buddhist sutras that exist in the Dunhuang caves. In particular, his two important propositions about the relationship between "image-making" and "image-viewing" in devotional art enlighten us to reconsider our methods in interpreting Chinese Buddhist art. This case study sharpened our method in studying Dunhuang art and literature in general.

PART TWO

EASTERN SCHOLARSHIP

Chinese scholars began their studies on Dunhuang art by writing notes on some of the paintings on paper and silk from the "library cave." Pelliot brought some manuscripts and paintings of Dunhuang to Beiping and showed them to Chinese scholars in 1909 on his way home. Many famous Chinese scholars, including Wang Guowei, Lo Zhengyu and Jiang Fu, immediately recognized the importance of these antiquities and started to publish their copies and notes. Although their interests were basically in the literary manuscripts, some of the scholars recorded and discussed a few works of art. In 1909, Jiang Fu wrote a note on the Vaisravana image, which was printed in the Five Dynasties period, and published it in his Shazhou wenlu. (Records on the manuscripts from Shazhou [Dunhuang]). This is the first record by a Chinese scholar on Dunhuang painting. It carefully documents the inscriptions accompanying the images. Jiang connects the iconography of the painting with a folk story recorded by a Buddhist monk Nian Chang in Fozhu Tongji and concludes that the donor Yuanzhong's motive in engraving the image is a continuation of the Tang tradition.

Wang Guowei (1877-1927) discussed two paintings from Dunhuang in Guantang jilin (Collected works of Guantang [i.e. Wang Guowei]), chap. 20. His interest in the paintings clearly stems from a historianUs perspective. He displays a stunning mastery of a vast array of sources with which he bears upon the iconography of two patrons' images in a painting from Dunhuang. His only clues are the inscriptions. The minimalist inscriptions do not deter him from summoning up a wide range of related historical sources out of which particular historical circumstances were reconstructed with accuracy. Implicit in the finding is of course the light it throws on the painting of the Guanyin image. Only Wang Guowei leaves such social history of art for us to pursue. For him, wresting a wealth of information out of a minimum of clues is all what a seasoned historian is after.

Chen Wanli, a professor at Beiping University, accompanied Warner on the expedition trip to Dunhuang in 1925 and stayed in the caves for three days. Much of this three daysU sojourn at was published when he returned to Beiping [Chen Wanli: RDunhuang qianfo dong sanrijian suode zhi yinxiangS (Impressions after a three daysU visit at the Caves of Thousand Buddhas of Dunhuang) in Xixing riji (Dairies of a journey to the West)]. On the basis of the "Li Family Stele" inscription [on the occasion of renovation of Tang], Chen determined the beginning date of the building of the cave temples to be Jianyuan the 2nd year (366 A.D.). This dating is still followed by the later scholars up to this day. He also documented inscriptions from twenty strong caves. Moreover, Chen charts a hypothetical trajectory for future inquiries into the Dunhuang caves. His proposals include the determination of the cavesU dates, the iconographical studies of the paintings, the chemical analysis of the pigments, the stylistic lineage, reconstruction of the historical circumstances surrounding the Dunhuang art, and the comparison between the clay sculpture and stone sculpture, etc.

In 1931 He Changqun published a long article, RDunhuang fojiao yishu zhi xitongS (The system of the Dunhuang Buddhist art) in Dongfang zazhi (Journal of the Orient), 28, no. 17. He Changqun had never been to Dunhuang. His study was based on the materials published by Pelliot and Stein, etc. But his article was the first comprehensive work on the art of Dunhuang written by a Chinese scholar. He traces the foreign influences from the Western regions to China, and gives an account of the development of Chinese Buddhist art. That social context spurred rise of the Chinese Buddhist art is clearly spelled out. Social misery is the major reason for the flourishing of Buddhist art. He ChangqunUs marks a turning point in the study of the Buddhist art of Dunhuang. The interest in the inscriptions accompanying the paintings as clues to historical inquiries shifted to the concern with the pictorial art itself. In fact, many of He ChangqunUs motifs have been reiterated in later studies: the problems of influence, the social origin of the Buddhist art, etc. He ChangqunUs article seemed too ahead of the scholarly community to warrant sufficient attention. It did not cause a stir.

The methods of Japanese scholars in the study of Dunhuang materials, both texts and visual works, are basically the same as Chinese scholars'. Therefore, I group Japanese scholarship and Chinese scholarship together in this review. The major trends and focuses in the study of Dunhuang art in the eastern intellectual tradition can be summarized into five aspects: identification of the images, rephrasing the caves, analysis of the styles, social-historical interpretation, and searching for the "origin."

Identification of the Images

Japanese scholar Eiichi Matsumoto published his monumental work, Tonkoga no Kenkyu (Studies on Tun-huang Paintings) in 1937 (2 vols., Toho-Bunka-Gakuin, Tokyo). This book compiles about 400 plates including the mural paintings of the caves, the paintings on silk and paper collected in the British Museum, the Musee Guimet and some private collections, and some mural fragments removed from some caves in the Eastern Turkistan area by Le Coq. Matsumoto carefully identified most of the paintings by connecting them with Buddhist sutras and some art history texts. His rich knowledge of Buddhist texts enables him to do this incredibly hard work. Although his method is merely to interpret the paintings, episode by episode, through related textual evidence, his identifications of the complicated contents of the paintings have long served as a "primary text book" for art history students studying the art of Dunhuang. When I started my research on the cave art, the first book my teacher Director Duan required me to read is this Tonkoga no Kenkyu. Indeed, this book formed the solid basis for the studies of Dunhuang art for all eastern scholars.

Zhou Yiliang continued Matsumoto's approach by discussing the deep relationship between Dunhuang murals and Buddhist sutras. ("Dunhuang bihua yu fojin" [Dunhuang murals and Buddhist sutras]. Wenwu cankao ziliao, 2, no. 4,1951). Since He Changqun, any study of Dunhuang art in China has referred to the relationship between the paintings and the Buddhist sutras. ZhouUs is the first attempt to address the issues fully. Zhou defines bianxiang (sutra painting) as denoting both iconic paintings of Buddhist deities and narrative paintings drawing on subject matters from Buddhist scriptures. He also notes the relationship between bianxiang and the popular sermons of the time. He does not discriminate between the sujiang (popular sermons), jiangjinwen (sermon notes), and bianwen (sutra stories) which had a complex relationship among themselves.

In 1960 Sun Zuoyun distinguished some images in the Dunhuang murals from Buddhism and considered them to be of Chinese traditional motifs (RDunhuang bihua zhong de shenguaihuaS [The paintings of spirits and monsters in Dunhuang paintings]. Kaogu, 6, 1960). Sun's iconography study re-interprets some images of monstrous creatures seen in the Dunhuang paintings, especially those in the Cave 249 and Cave 285, and identifies them with some Chinese mythological texts and archaeological evidence. To Sun, they are mostly from a Chinese mythological repertoire and had to do with popular beliefs concerning the afterlife and immortality--a legacy from the Han culture. Even such a figure as the flying apsaras, normally thought to be purely an import from India, is seen by Sun as a descendant from the Chinese feathered-man motif in the pictorial genre of Ascension to Heaven. In short, Sun is arguing for an indigenous Chinese tradition that informed the Dunhuang art.

Shi Weixiang, a senior research fellow of the Dunhuang Research Academy, has spent most of his life in the caves. His iconographical approach differs from other scholars in the detailed investigation of historical records rather than Buddhist sutras. This new angle enables him to identify some complicated paintings which are extremely difficult to identify. For example, the southern wall of Cave 321 was tentatively identified by previous scholars as the Buddha's Preaching in Mountain Grdhrakuta, but many scenes within the painting had remained unclear until Shi Weixiang found the truth ("Dunhuang mogaoku de baoyujinbian" [The illustration of the sutra of the treasure rain in the Mogao Caves of Dunhuang]. 1983 nian shueshuhuiyi lunwenji, shikuyishu bian. Lanzhou: Gansu Renmin Chubanshe). This article is ShiUs most celebrated tour de force in his iconographic study of a wall painting. The painting in question had hitherto been mis-interpreted as representing the Sermons on the Mount Ghridhrakuta. Shi points out that the picture actually is related to the Treasure Rain sutra which was corrupted and put into political use in the early Tang as political propaganda for the Empress Wu Zetian in her campaign for the throne. The images of the sun and the moon actually allude to the empressUs name. Shi then proves how this could have been relevant in the local region by looking at the wide range of WuUs campaign and the local governorUs involvement in the conspiracy of omen-making. In this way, Shi convincingly shows that a wall painting with a sutra subject matter is actually pictures of a political landscape. He broke through the limitations of the conventional method of iconography, which only focuses on finding the Buddhist sutras that correspond to the paintings, and found his answer in historical analysis.

After a long term effort by researchers, the contents of most caves have been identified, and the results were published in the monumental book, Dunhuang mogaoku nierong zhonglu (Catalogue of the contents of the Mogao caves at Dunhuang), ed. by the Dunhuang Research Academy.

Rephrasing the Caves: Methods and Results

The 492 caves remaining on the cliff are a big mess. One cave was probably built in the fifth century, the cave next to it may belong to the Yuan Dynasty (1271-1368). Therefore, the Mogao caves are called "a mis-bound book of history." It took a long time for Stein and Pelliot to figure out a very brief chronology of the caves. In fact, chronological study is the basis of all the studies in the caves. The most comprehensive study started in the 1940s, when the Dunhuang Art Research Institute (later Dunhuang Research Academy) was established by the Chinese government at the site.

In 1944, Li Yu wrote a report of his one-year investigation of all the caves, RMogaoku gedong neirong zhi diaochaS (A survey on the contents of the Mogao caves) (unpublished, Archive of Institute of Dunhuang Cultural Relics). The survey is a comprehensive and definitive iconographic and chronological study of the Mogao cave temples. Although this report has never been published, it serves as a basis for later scholars to rephrase the caves and to interpret the murals. Shi Weixiang and his colleagues re-investigated each cave in the 1960s and corrected many mis-identifications of the contents and mis-locations of the dynasties of some caves. The report finally came out as the complete catalogue of the 492 caves (Dunhuang Mogaoku neirong zhonglu [the complete catalogue of the Mogao caves, Beijing: Wenwu Chubanshe, 1982). Their method is very comprehensive and flexible. It takes every possible element into consideration: historical condition, the dated inscriptions, the shape of cave structure, the color of mural, the costume of donor portrait, the artistic style, etc. In fact, it is a kind of intuition mixed with their knowledge that makes the final decision on dating.

Probably one will doubt the credibility of the intuitive dating, however, the so-called scientific typological study confirms the accuracy of the dating. Trained in the Archaeology Department of Beijing University, Fan Jingshi and Ma Shichang started to use the archaeological method of typology to re-date the early caves at Dunhuang. Unexpectedly, their results exactly match the old chronology made by previous scholars with their intuitive method of dating ("Dunhuang Mogaoku beichao doku de fengqi" [the chronology of the caves of the Northern Dynasties in Mogao Caves at Dunhuang], in Dunhuang yanjiu wenji, Lanzhou: Gansu Renmin Chubanshe, 1983) This phenomenon may be understood in two ways. firstly, because the typologists based their new approach on and limited by the old chronological, their result only confirms the old one. Secondly, their method of typology seems very "technical" rather than "scientific." This technical analysis of the selected isolated details, such as the folds of garment, the number of petals of decorative flowers, or the shape of the nose of human figures, etc., can only reveal some "changes or differences" among the caves without any suggestion of chronological order unless the differences are connected with the caves that are already dated. However, this method can reveal a detailed changing process within a chronological framework.

Both the intuitional method and the typological study rely on the "standard caves," which have been surely dated. The number of standard caves are increasing with the continuity of Dunhuang studies. Some inscriptions containing donors' names can be related to the dated manuscripts which have the same names discovered from the sealed library cave. Japanese scholar Terukazu Akiyama compiled all the dating information concerning the Dunhuang paintings from various sources: cave inscriptions, historical records, dated documents, captions remaining on paintings on silk and paper. His important article, "Tonko kaiga no hennen shiryo (Further chronological materials concerning the Dunhuang paintings)" (Tokyo daigaku bungakubu bunka koryu kenkyu shisetsu kenkyu kiyo, 1975), provides us with some 65 pieces of dating information, from 354 to 1035. Akiyama's research is basically textual and historical.

Analysis of the Styles

In the early 1940s, some Chinese artists visited Dunhuang and were shocked by the gorgeous murals of the caves. Zhang Daqian, a very famous contemporary painter, spent two years in the remote site and made some 200 copies of the murals. In 1943, his copies of the Dunhuang murals were exhibited in Chongqing, the temporary capital of China during the Sino-Japan war. The catalogue edited by Luo Xinzhi and Liu Junli (Zhang Daqian linmo Dunhuang bihua zhanlan muci [Exhibition catalogue of Zhang DaqianUs copies of the Dunhuang murals], Chongqing, 1943) contains Zhang DaqianUs commentary on Dunhuang paintings. It registers a practicing artistUs sensibility towards the stylistic properties of the paintings. Writing in the vein of the traditional Chinese poetics, Zhang offers a descriptive taxonomy of the changing style of Dunhuang art. The terms are vague and intuitive but compact and terse; they carry a considerable weight of traditional Chinese humanistic discipline which makes sense to the those well-versed in a similar discourse.

Since 1957, Jin Weinuo, a professor of the Central Academy of Fine Arts, has published a series of articles dealing with Dunhuang art (RDunhuang benshengtu de neirong yu xingshiS [The iconography and form in the Jataka pictures in Dunhuang], Meishu yanjiu, vol. 3, 1957; RDunhuang bihua zhong de zhongguo fojiao gushiS [The Chinese Buddhist stories in the Dunhuang mural]. Meishu Yanjiu, vol.1, 1958; etc.).

Jin Weinuo has made some progress in content identification and stylistic analysis. He identifies the murals in Cave 323 as depicting scenes or episodes from Chinese Buddhist history. Most of them could be found in the historical texts. He also sees in these pictures some cognitive values in connection with medieval social life. He also discusses the compositional and stylistic properties of these paintings.

Jin Weinuo's study seems more specific and accurate than that of Zhang Daqian and other artists. In particular, his study on the narrative painting of the Jetavana Garden (RDunhuang bihua zhiyuan jitu kaoS [On the story of Jetavana Garden represented in the Dunhuang murals], Wenwu, 1958.10: 8-13) emphasizes the Chinese way of representation.

This is a study of the pictorial representation of the Jetavana story at Dunhuang. The author notes that the Chinese way of representing the story is from the very beginning different from that of India. It initially takes the form of the continuous narrative. As it develops, it evolves into the compositional form of a symmetrical confrontation between the two opposing parties. Jin suggests that in the late Tang, paintings of Buddhist narratives such as Jetavana story is based on the bianwen, a kind of dramatized popular narrative based on Buddhist sutras rather than on the sutras themselves. RThe breeze of life had dispelled the religious atmosphere.S

Duan Wenjie's approach on the style of Dunhuang art focuses on the issue of "foreign influence and Chinese tradition." (Dunhuang Zhaoqi bihua de minzu chuantong yu wailai yingxiang [the national tradition and foreign influence in the early murals of Dunhuang], Wenwu, vol. 12, 1978). Duan distinguishes two styles, the "western style" and the "style of central China," and discusses the social backgrounds of the emergence of the two styles. To Duan, the changing process of styles is parallel to local history. His social-historical interpretation of the styles differs from simple formal analysis.

Socio-historical Interpretation

Chinese scholars showed their interests in historical interpretation at the beginning. Jiang Fu and Wang Guowei used the visual materials from Dunhuang as a kind of historical record to discuss the issues of Chinese history.

Pelliot was the first person to document the inscriptions in the Dunhuang caves. In 1925, Chen Wanli on a Warner-led trip to Dunhuang recorded inscriptions from 20 caves. Shi Yan went to Dunhuang in the spring of 1943. He spent half a year copying inscriptions. The result was his book Dunhuang shishi huaxiang tishi (Inscriptions from the rock caves at Dunhuang) published in 1947.

What is interesting to us is his study on the images of patrons. He notes that the patron images of Tang times are inflated into life-sized portraits. The inscriptions of names are inlaid with gold. The head-gear and dress style become elaborate and flamboyant. Here Shi exercises a kind of symptomatic reading of images and observes an internal discrepancy between the avowed religious function of the patron portraits and the actual social function of these same images. Paintings were made to demonstrate religious piety, but the excess of flaunting oneUs own magnificence displays oneUs own sense of dignity; it therefore undermines the original import and intention of such images.

Shi Yan would have qualified as what we now call the social historian of art. His inquiry includes an exhaustive study of the prominent family clans of the region as registered in the inscriptions in the caves. Eight prominent families have been identified. He goes further to study the gongde zhu (merit owners) who were the main patrons of the caves and who consequently became the owner of particular caves for their donation.

In 1955, Wang Xun published RDunhuang bihua he zhongjiao yishu fanying shenghuo de wentiS (Dunhuang wall paintings: the problems of reflection of life in religious art, in Meishu, vol. 11, 1955). In the background lurks the ideological pressure from rigid leftist Marxism with its insistence on class and historical progress. Wang XunUs article is therefore an apologetic treatise on the RprogressiveS aspect of the Dunhuang painting. Patrons of Dunhuang paintings may have been the Rruling class,S but this takes nothing away from the positive aspect of the paintings so long as they reflect an uplifting yearning of the historical era. Further, the paintings depict daily life, they are therefore in the mode of realism which was ideologically sanctioned by the leftist Marxism.

Li Zehou is a philosopher and an intellectual historian, author of Critique of KantUs Critical Philosophy and a handful of books on Chinese intellectual history including a Critique of Confucius. Well-versed in both Western philosophical discourse and Chinese intellectual history, an ardent exponent of situating Chinese issues in the Western discourse and also a superb man of letters, Li acted as an enlightenment figure in the Chinese cultural landscape in the late 70Us and early 80Us. His book on the history of Chinese aesthetics--Mei de licheng (Path of beauty)--was widely perused by scholars of all humanities fields, especially among the young generations.

Kant, Hegel, plus Marx--with some early 20th century Western aesthetical theory such as Clive BellUs Rmeaningful formS thrown in--constitutes LiUs basic conceptual framework in his critique of medieval Buddhist art. According to Li's theory, religion is opium--there goes the Marxist echo--but it is also an escape from, and even a protest against suffering. The story of the Chinese reception of Buddhist art is formulated by Li as a process from an Ranti-rational ecstasy of mysticismS to the ultimate triumphant sobering Rrationality and historicism.S The Buddhist images of different ages therefore took on different implications. In his article concentrating on Dunhuang Buddhist art, RShen de shijian fengmao (The secular appearance of the gods)S (Wenwu, vol. 12, 1978), Li divides medieval Buddhist art into three periods: (1) Wei (2) early Tang (3) later Tang, Five Dynasties, and Song. To Li, the restless and vibrant pictorial configuration of the Northern Wei bespeaks a propensity towards an ecstasy which is at once a condensation of suffering and a sublimation. The art is a negative mirror of reality: the uglier the reality, the more beautiful the form. Early Tang art takes on a new phase. Buddhist images have been unveiled from the mystifying mask of ambiguity and meditation. The figural grouping assumes a hierarchy in correspondence with the hierarchy of the world. The cave turns from the Northern Wei vihara to the palatial interior. The Buddhist universe becomes a mirror of the secular political order. From the middle Tang onward, Buddhism was increasingly secularized. The boundary between the divine world and the secular world is increasingly transgressed and blurred. Buddhist deities are modeled after the palace ladies. Thus is spelt the end of Buddhist art.

Li is one of the foremost critical historians of art whose broad strokes cut a wide swath and earned him a large following.

Duan Wenjie's RXinxiang de lishi--Tan Dunhuang bihua de lishi jiazhi" (Notes on the historical value of wall paintings at Dunhuang. In Lanzhou daxue xuebao zhesheban, vol. 2, 1980) is concerned with how much historical reality we can dig out of the wall paintings at Dunhuang. The basic assumption is of course that art RreflectsS reality. In DuanUs view, art reflects reality in several ways. There are those Rdirect reflections,S such that we can read the patronsU images in a straightforward way. There are RindirectS reflections: the painting of five hundred robbers becoming Buddhists had to do with the crackdown down on the peasantsU uprising under the reigns of Western Wei and Northern Zhou. There are also RrefractionsS: the set of nine pin (levels) in the Buddhist universe actually has its correspondence in the secular social-political order.

Shi Weixiang's RShizu yu shikuS (Regional elite clans and the caves) in Dunhuang yanjiu wenji (Essays on the studies of Dunhuang. Gansu renmin chubanshe, 1982) concentrates on the relationship between local elite clans and the caves at Dunhuang. In this article he argues that the perpetuation of Buddhist cave-building and image-making at Dunhuang is motivated by the Rhistory of production relationshipS (p. 151). Dunhuang was ruled by regional elite clan families. Some of these families came from the central plains and settled here. Each family would RownS a cave and generation after generation, the cave as the family shrine would get rebuilt. The Li family owned Cave 331 and 332 built in Tang. The Yin family came to prominence in the Sui-Tang period and owned cave 217. In the pictures of patrons led by monks in their veneration of Buddha images, the monks are mostly members of the same prominent families. Thus Shi aptly quotes Marx: RReligion no longer appears as the basis, but as the manifestation of secular narrownessS (Marx, ROn the Jewish QuestionS).

A close follow-up of the social and political history of the region allows Shi to shed light on some of the Buddhist wall painting with a striking boldness. Zhang YichaoUs taking back of the Dunhuang region from the Tibetans, according to Shi Weixiang, gave some traditional Buddhist sutra painting a new meaning. RThe large-format of the Raudraksa Subjugating Demons,S Shi points out, is Ra manifestation of the felicity of the triumphant people of Shazhou in celebrating their victory over the Tibetan slave-owners.S

With a positivist historianUs muscle armed by the armor of the Marxist critical-analytic apparatus, Shi is peerless in his social history of Dunhuang art. He has published a number of iconographical studies in which he bears his command of vast range of historical canonical texts, Buddhist sutras, Dunhuang manuscripts, and stele inscriptions, upon some wall paintings and convincingly solved some tough iconographical problems. In this more traditional scholarship, he is almost second to none. But it is above all his social history of art that marks him apart from other Chinese scholars. It is not that others do not quote Marx profusely or with less enthusiasm; it is ShiUs solid grip on the social historical materials and his ability to relate these materials intimately to the wall paintings that has made him a compelling presence.

The Genetic Theories

In 1944, Xiang Da delivered a lecture on the "origin" of Dunhuang Buddhist art at Lanzhou University. (RDunhuang fojiao yishu zhi yuanyuan jiqi zai zhongguo yishushi shang zhi di weiS [The origin of the Buddhist art of Dunhuang and its place in the history of Chinese art. Dunhuangxue jikan, 2, 1981. A lecture given in Lanzhou, 1944. Note taken by Shui Tianming). In this lecture, Xiang Da asserts the Indian origin of the Dunhuang painting. In 1951, he published RMogao, Yulin erku zakaoS (Notes on Mogao and Yulin caves. Wenwu cankao ziliao, 2, no. 5, 1951) and addressed this issue again.

Xiang Da is an ardent exponent of the foreign influences informing the Dunhuang art. He bases his argument on the following grounds: (1) The similar ways of preparing the walls for a kind of tempera painting in India, Kucha, Turfan, and Dunhuang; (2) The similar ways shared by the Indian painters and Chinese painters in sketching the base; (3) the similar use of pictorial plasticity in the modeling of the human bodies.

More interesting are the authorUs historical accounts of the four aspects of paintings in the making rather than the central thesis about the influence and origin, because the same theory has been established half a century ago. Xiang offers an in-depth scrutiny of the medium of the Chinese medieval mural painting in terms of its physical compositions, the actual process of preparations, technical procedures, compositional conventions, scales of proportions, and stylistic properties.

Chang Shuhong's RDunhuang yishu de yuanliu yu neirongS (The genesis, genealogy, and iconography of the Dunhuang art. Wenwu cankao ziliao, 2, no. 4, 1951) is a comprehensive survey of the routes of influences involving India, Central Asia, and China. He traces the northern and southern routes through which Gandhara art spread to the East. But he also took care to show that the Tang artistic style with its emphatic use of color and lineament also spread westward. DunhuangUs assimilation of the western-region style was relayed into central China via two routes. The southern route went through Maijishan, Jin Zhou, Guangyuan, Da Zhu, till it reached Leshan; the northern route traversed through Yungang, Longmen, Gongxian, Tianlongshan, and reached Xiangtanshan. The closer it got to the Central Plains, the more traditionally Chinese in style. Chang emphasizes the inertia of the traditional culture, the legacy of the Han culture. In his view, Indian art was a tributary, not an origin of the Chinese Buddhist art. The Chinese reception of an alien art largely hinges on the Chinese native mental equipment and cultural disposition. The change of the Amitabha into Amitayus during the early years of Northern Wei suggests that the Chinese were ready to embrace this religious system only because since the Qin-Han period they had the traditional longing for afterlife and immortality; it was with this mind-set that they appropriated the Buddhist symbolic system which was transformed by Chinese hands into a cult of immortality.

Su Bai, a professor and long term chairman of the Department of Archaeology of Beijing University, initiates the theory of "the eastern origin of Dunhuang art." ("Canguan Dunhuang di 285 hao ku zhaji" [Note on a visit to Cave 285 of Dunhuang], Wenwu cankao ziliao, vol. 2, 1956; "Chun Yue Zhun Fa Liang nianxiang dao de wenti" [the questions associated with Yue Zhun and Fa Liang], Dagongbao zhaigang fukan she zhounian jinian wenji, Hong Kong, 1980) His theory is based on an assumption that any local culture must be lower than the culture of capital and that the local cultures must grow up under the influence of the capital. This assumption caters to the ideology of a communist government whose major concern is to keep strengthening the absolute power of the central government. It is the only "theory" that was not criticized by the "revolutionary scholars" during the Cultural Revolusion (1966-1976). Su Bai argues that because the Yungang caves, located near the Northern Wei capital Pingchen, were built after 460, the earliest group of Dunhuang caves must be later than this year, and the artistic quality of the local art must be lower than that of the capital art. To Su Bai, the power of influence on regional art is decided by its archaeological relationship with the capital: the closer, the stronger. Being a remote site, Dunhuang art must have taken its "root" from the Yungang art, the "capital art." This genetic theory was initiated by a professor living in Beijing, the red capital, who only took his tour to Dunhuang twice and stayed in the caves for several days, but it is very "influential" politically. This phenomenon, on the one hand, reflects the relationship between scholarship and politics in modern China; on the other hand, it conveys the strong national "feeling" of some Chinese scholars with their identity as Chinese.

The scholars living at Dunhuang seem more interested in finding the truth, not in modern political ideology. In 1982, Shi Weixiang published his RDunhuang fojiao yishu cansheng de lishi yijuS (The historical grounds for the rise of Buddhist art at Dunhuang. Dunhuang yanjiu, vol. 1, 1982).

Early in the 50's, Chang Shuhong had argued that the Han-Jin tradition was the root of Buddhist art at Dunhuang. But the argument is more of a polemic without an analytic backing; it therefore ended not with a bang but with a whimper. ShiUs work is much more solidly grounded in the local history. He studies the immigration pattern, the composition of the local population, the hierarchy of the local feudal social economic structure. The particular regional culture bred the imperative that for a religion to be embraced, it had to be transfigured to accommodate the local needs and to speak the local dialect. Shi connects the Dunhuang caves with the newly discovered tombs of the Jin Dynasty (265-420) in the Hexi Corridor, which are very close to the early caves in architectural form and decorative motifs. Duan Wenjie also argues for a local origin of the Dunhuang caves by focusing on the "Daoist motifs" appearing in the Buddhist caves. He considers the combination of two religious traditions to be a special regional phenomenon.

CONCLUSION

In comparing the studies of Western scholars with the Eastern scholarship, we can see some important differences

: (1) Western formalists take "space representation" as the key for analyzing the formal changes in Dunhuang art; their goal is to establish a chronology of the caves. Eastern style analysts consider the "form" of artworks to be a combination of shape, color and "brushwork"; their goal focuses on the relationship between the changes in style and the changes in Chinese history. The Western method is detailed but too fragmentary. The great danger of this methodological approach is that it assumes a priori existence of a "universal evolution pattern" and thus transforms Chinese art history into an analogue of Western art history. But it has never been proven that Dunhuang art developed along a path resembling that of Western art. The Eastern method is comprehensive but too general and vague. This kind of intuitive analysis omits details and assumes a stereotype of "dynastic style". The vivid variations of the representational form of Dunhuang art thus become always a "reflection" of some "dynastic styles". The characteristics of Dunhuang art are covered by a thick curtain of the dynastic styles.

(2) Western scholars, except for a few young scholars who very recently started to consider the issues of sociology, seem have ignored a sociological approach to the study of Dunhuang art. On the contrary, Chinese scholars have connected the Dunhuang caves with social changes. However, their studies are based on a historical framework drawn by Marx, and their conclusions are always the same: Dunhuang art reflects the conflict of social classes, and the driving power of the development of the art is the change in economy. This kind of sociological approach ignores the concrete facts related to Dunhuang art. Some of the conclusions they have reached are not convincing at all.

On the whole, the studies of Dunhuang art in the West seem now to have arrived at a "turning point". It is a time to rethink our limitations and advantages and to experiment with some new interpretative methods.