The Metropolitan Museum of Art held an extraordinary all-day symposium December 11, 1999 on the attribution of a painting, shown above, it has attributed to Dong Yuan (active 930 to 960s), a painting that has been described by its former owner, C. C. Wang, a noted dealer, collector and artist, as the "Mona Lisa" of Chinese paintings, a description quoted by The New York Times when it ran a long article on its acquisition as a promised gift by Oscar L. Tang to the museum on its front page in 1997.

The painting's attribution to Dong Yuan, however, has been challenged by some leading Chinese art experts, such as James Cahill, the author of many books on Chinese painting, and Sherman Lee, the retired director of the Cleveland Museum of Art and the author of a leading textbook on Far East Asian Art. Cahill has argued that the work is a 20th Century forgery by Zhang Daqian (1899-1983), a noted collector and painter from whom C. C. Wang acquired the work. Both Cahill and Lee participated in the symposium, and C. C. Wang sat in the first row of the museum's auditorium and had his daughter read a statement reaffirming his belief in the painting's authenticity.

Mr. Tang is the brother of the wife of Wen Fong, professor emeritus at Princeton University who has been consultative chairman of the museum's Chinese Painting, Department of Asian Art, since the early 1970's when he orchestrated the acquisition of 25 allegedly early Chinese paintings from C. C. Wang. Many of the paintings in that acquisition have also had their attributions by Wen Fong challenged (see The City Review article), but this symposium did not address the attributions of them and focused only on the "Riverbank" work.

Wen Fong has orchestrated many of the museum's acquisitions of Asian Art for a generation with the support of Thomas P. F. Hoving, the museum's former director and an alumnus of Princeton University, and Douglas Dillon, a former chairman of the museum. While the amount of funds spent by the museum on Wen Fong's acquisitions has not been publicly disclosed, one of the major purchases was of the Harry Packard Collection of Japanese Art, and the museum said at the time that that acquisition used up its acquisition funds for the entire museum for five years. Most of the Packard Collection has not yet been shown at the museum, although some attractive works are now on display in the Japanese Art galleries.

The City Review was the first news organization to publish a report suggesting that there were doubts about the "Riverbank" painting (see The City Review article). Subsequently, Carl Nagin wrote an article for The New Yorker magazine in which he cited comments by James Cahill and Sherman Lee that were critical of the museum's attributions. In his September 3, 1999 review, entitled "A World of Passion, Stroke by Quivering Stroke," in The New York Times of this exhibition, Holland Cotter noted that the attribution of "Along The Riverbank," had created something of a "furor" in the art world and that "Phrases like 'Chinagate' began to pop up on the Internet," an apparent reference to the series of articles in The City Review on the attributions of Chinese art at the Metropolitan Museum. Cotter ran another article about a week prior to the symposium that suggested that it, the symposium, would prove inconclusive.

In addition to participating in a panel discussion, each of the symposium's participants read remarks that were included in a new museum publication, "Issues of Authenticity," edited by Judith G. Smith and Wen C. Fong ($19.95) that was made available that day and they also answered questions from the standing-room only audience. While the participants apparently had read the symposium's papers beforehand, the vast majority of the audience had not and several of the presentations that purported to defend the dating of the work in question were very technical with many slides.

Although the general tenor of the symposium was quite civilized, there was little doubt for most of the audience that jammed the museum's large auditorium to capacity that the occasion was momentous and there was probably general agreement that the museum did itself proud in affording two of its most important critics in this controversy, Mr. Cahill and Mr. Lee, ample opportunity to public present their case. The museum, on its behalf, put forward several experts of its own to defend its case for the work in question.

Despite the fact that the symposium took up most of the day and each speaker was able to present very good slides on large screens to embellish their presentations, it was not easy going for the audience, which was well-behaved and patient. Many of the speakers simply referred the audience to their papers for more detail and some of the questions from the audience to the whole panel went largely unanswered. Most speakers stuck to their "guns" and did not indicate that the day's presentations had substantially altered their opinion of the work. Robert Harrist Jr., who moderated the afternoon sesssion's panel discussion, indicated he was leaning toward not accepting Riverbank as a tenth-century work. Jerome Silbergeld, the professor of at history at the University of Washingtonon who moderated the session's morning panel discussion, on the other hand, said that the painting might be "early," added, however, that it was "very problematic" that it was by Dong Yuan.

Nevertheless, there was no befuddlement in the air. Clearly, the challenges to the work's attribution were disturbing and not casual. Wen Fong issued a long statement in addition to his prepared remarks in the book issued that day that was a direct, further "reply" to Mr. Cahill's long "indictment" of the work.

While Wen Fong and his disciples sought to establish that the work was of sufficient age and style to possibly be of the Song Period, definitive attribution to Dong Yuan clearly was difficult since the few other works attributed to him are quite different in style.

Thus, one must conclude that the museum has backed off somewhat from the very grandiose claims that have been made for it, claims that fly in the face of casual observers who would be hard put to consider the work masterful in comparison with many other uncontested early Chinese paintings. Moreover, its condition is definitely less than pristine. The "scientific" examination of the work is really inconclusive as Zhang Daquian had access to old seals and old silks and was clearly very skilled.

In any event, the most that can be said on its behalf is that it might be an interesting example of Dong Yuan's early style, but that apart from its size it is neither spectacular nor great especially given its poor condition.

Cahill and Lee are such distinguished experts that one cannot cavalierly dismiss their qualms and Wen Fong and C. C. Wang bring a lot of questionable baggage to the table. Wen Fong claimed that the museum's 1973 acquisition of 25 Chinese paintings from C. C. Wang represented the best of his collection. There are issues of "conflicts of interest" in Wen Fong's and C. C. Wang's dealings that are disturbing and of the type that most institutions assiduously avoid especially when there are substantial monies involved, as there have been in the C. C. Wang acquisitions. While an immensely colorful and very talented person, Zhang Daqian, shown below in a computer drawing based on a photograph, was notorious and contemptuous in his manipulation of collectors.

There are important reputations at stake in this controversy and the Metropolitan Museum of Art has not wavered in its steadfast support of Wen Fong. The decisions to hold the symposium and issue the new publication are laudable, extremely so. It has not, of course, renounced Wen Fong's interpretations of Chinese Art History, but the symposium and the publication of its companion book certainly acknowledge that there are problems and major dissents.

There are several possible interpretations of the Riverbank controversy: (1) Wen Fong, C. C. Wang and Zhang Daqian conspired to defraud the museum and the public with the possible subsequent collusion, witting or unwitting, of Douglas Dillon and Thomas P. F. Hoving; (2) Wen Fong and C. C. Wang are wrong and have been duped by a Zhang Daqian "forgery"; (3) Wen Fong and C. C. Wang are wrong in their aggressive attribution to Dong Yuan, but the work is an interesting, early work from an historic rather than an aesthetic perspective; (4) Wen Fong and C. C. Wang are right and the museum has an interesting and large, but not great, work by a rare early master.

As presented in the symposium book and his additional "paper" replay to Cahill's criticisms, Wen Fong's "defense" was rather weak in that it did little to address, let along destroy, the quite devastating arguments presented by Cahill and others. Wen Fong's primary defense was that an intellectual intepretation of the work as a heroic and political statement about reclusion at a time of some social and political turmoil, a thesis that is interesting and possible, but has little bearing on the question of its authenticity. Such an argument plays on the almost mythic stature of Dong Yuan since so few works survive that are attributed to him and since those are in a variety of styles, so much so that it is hard to even begin to speak of a Dong Yuan "style" per se.

It is possible that a scholar can overturn conventional wisdom in his field with his brilliant "discoveries" and new "interpretations." Wen Fong's aggressive attributions have been part of his "interpretation," but while it may sound/read cogent, its logic was not very compelling, especially in light of the many problems presented by Cahill and others. Many of those problems, it is true, stemmed more from the work's painting's "modern" history than the work itself, but they nonetheless were disturbing in trying to assess the merits of the specific work.

As much as some observers might feel that Wen Fong has over-reached, the museum's commitment to establishing significant new galleries to non-Western cultures can only be applauded and Wen Fong deserves considerable credit for that even if his critics are right in their disagreement over attributions.

As one of the world's, foremost cultural institutions, the Metropolitan Museum cannot afford to have its dedication to the highest standards of scholarship in serious question. The issue is far more important than the recent controversy at the Brooklyn Museum of Art over the "Sensations" exhibition that involved questions of taste.

Are the C. C. Wang acquisitions by the Metropolitan tainted?

Some observers may argue that even if some of them may have been ambitiously misattributed, the museum still has acquired some gems and, moreover, done its role in helping to awaken greater Western interest in non-Western cultures and traditions. Had Wen Fong and the museum exercised some curatorial conservatism, they might have avoided the substantial embarrassment that has ensued of which the symposium is ample evidence of the severity of the problem.

Several speakers at the symposium maintained that such issues cannot be put to a simple vote. Who is eligible to vote? Who determines who are the experts?

Art experts are not certified or licensed. Some are heads of departments at universities, or curators at museums. Others may work at auction houses, or be dealers, or collectors.

Much of Wen Fong's support comes from his former students. Presumably Cahill has similar supporters.

While much of the debate might seem esoteric to many, there is no reason to assume that such controversies cannot be reasonably resolved given sufficient airing. The symposium was a good, indeed historic, start and the papers lay out important groundwork for study.

The museum should convene another symposium in a year or so. This symposium devoted about three-quarters of its time, however, to reading of papers that despite the fine multi-media presentations, were not easy to follow. Too little time was left for debate and for questions from the audience. The next symposium should simply have Cahill and Wen Fong and a moderator, presumably Jerome Silbergeld, whose lively manner was most welcome. An Internet hook-up should be arranged for the heads of Chinese Art departments at major universities and museums as well as established dealers in early Chinese paintings, and perhaps even some well-known art critics, from around the world to listen in and finally vote on whether they accept the attribution to Dong Yuan, the dating of the work, and whether they believe it is the handiwork of Zhang Daqian. The votes should be "yes," "no," or "undecided," but comments should also be additionally welcomed, and the complete results should be published. Far from being an exercise in an underappreciated field in the West, such a continued investigation of "Riverbank," and, perhaps, other early Chinese works at the Metropolitan, or elsewhere, would serve an important educational and cultural purpose in demonstrating the complexity and importance of attributions in the art world. The Metropolitan and Wen Fong have now crosssed over the threshold of embarrassment and will only be well served by continuing this worthwhile endeavor that will bolster confidence in institution's commitment to the highest standards of scholarship.

Connoisseurs of Western art can begin to appreciate Eastern art, but it is not something quickly learned and it really requires fluency in the language since the appreciation of Chinese painting involves not only the painting, but also the calligraphy and the seals. The latter two considerations, however, cannot be as important as the painting itself, but lend important poetic and historical understandings that sometimes can be very interesting and important.

This article will summarize, with extensive quotations, below the major commentaries in the book by many of the speakers.

Mr. Cahill, professor emeritus, history of art, the University of California at Berkeley, maintained in his essay, "The Case Against Riverbank," that Along the Riverbank is "a modern fabrication produced by the painter and collector-dealer Zhang Daqian." Cahill notes that until recently he had maintained only that the work "must be a recent fabrication, probably by Zhang; now, having spent much more time with the painting and with the literature and arguments that have accumulated around it, I am even more certain about Zhang's authorship."

"Those who recognize Riverbank as a forgery by Zhang Daqian, and who point to serious representational flaws in the work, acknowledge that it is the masterwork among his forgeries and stands very high in his oeuvre as a whole. Although virtually every element in Riverbank can be found in his other paintings, whether done under his own name or as forgeries, nowhere elese does Zhang combine these particular elements into such an impressive and imposing whole. While in no other fabrications he employed one or another established brush manner (cunfa), it was his brilliant move in Riverbank to avoid distingushable brushstrokes, altogether in the earth areas, instead brushing on the ink smoothly for a dramatic effect of light and shadow. Those trained the Chinese mode of brushwork connoisseurship will immediately, and somewhat unconsciously, associate this stylistic feature with an early date, prior to the development of texture-stroke systems. This aspect of Riverbank, along with the dynamic energies generated in its composition (by means that are anachronistic but no less effective), its avoidance of the unnatural patterning (such as 'folded' hillsides) seen in most of Zhang's other forgeries of early paintings, and the sheer quanity of entertaining detail (trees, figures, architecture) constitutes the complex of qualities that persuades many viewers...to say that Zhang Daqian could not possibly have painted it. ...As highly versatile as Zhang Daqian was, he could not altogether avoid incorporating traits of his own twentieth-century style intohis forgeries....Zhang...made mistakes in Riverbank that can be detected and that together rule out an early date for the painting," Cahill asserts in the beginning of his essay.

Cahill presents 14 "counts of an indictment against Riverbank, the first eight of which he considers to be very important. His last "count," however, an "alternative" history, is very interesting and intriguing and perhaps the most important.

Cahill argues that "the upper part of Riverbank is completely anachronistic, aside from being sloppily executed, as if it were unfinished or finished off quickly." "In tenth-century landscape painting, mist or fog, when present at all, is restricted to small areas. Nowhere can one see fog spreading so extensively that the tops of mountains disappear altogether or (as here, in the upper right) hover ambiguously in the far distance with no indication of how they continue below....The whole area is essentially unreadable, and not because of damage or repainting. Perhaps Zhang meant it to be read as the murkiness of a rainstorm, but it is nonetheless anachronistic. So many features of the painting point to later periods that one can properly term Riverbank a pastiche....Riverbank is not consistently in any period style."

Cahill has suggested that a river and the road in the upper part of the painting create "visual confusion," adding that "A good early artist would not have permitted such visual confusion. But in Zhang Daqian's landscapes, as this and other instances will show, rivers and roads winding out of the deep distance often exhibit this kind of ambiguity."

Cahill's maintains that "Riverbank has serious, indeed, fatal, structural flaws and is filled with representational inconsistencies." "A good early artist would not have permitted such visual confusion. But in Zhang Daqian's landscapes, as this and other instances will show, rivers and roads winding out o the deep distance often exhibit this kind of ambiguity," Cahill maintained.

Cahill notes similarities between Riverbank and several of Zhang's works such as The Wei River, collection unknown, Mount Qingcheng, and Immortals' Dwellings at Hyayang, (Chang Hsu Wen-go Collection, Taipei). Cahill quotes Fu Shen's 1991 catalogue for an exhibition of Zhang Daqian's works that Zhang believed that "in this work he became the equal of Dong Yuan."

"In spite of Zhang's claim of a Dong Yuan model," Cahill continued, "however, this compositional type has no true precedent among extant early paintings. The closest to it, noted also by Fu Shen, is a landscape in the National Palace Museum, Taipei, ascribed to Dong Yuan, and titled Summer Mountains Before Rain….Fu dates the painting to the late Ming dynasty (1368-1644); in my 1980 Index of Early Chinese Painters and Paintings, I describe it simply as an "interesting Ming work in Dong Yuan's tradition.' It may well have inspired Zhang Daqian, but it far too late to serve as an early and supportive parallel for the composition of Riverbank."

He cites "Zhang's forgery of Temples among Streams and Hills, a painting attributed to the late-tenth-century artist Juran, which was once owned by the Hong Kong collector J. D. Chen (Chen Rentao), who published it in 1955." "It is apparently based on a Ming work in the Juran manner (although attributed to Dong Yuan)," he continued, "in the collection of the Freer Gallery of Art…but since that picture was unpublished in Zhang's lifetime, it is unclear how he could have known it. Zhang's copy is furnished with seals of Emperor Huizong [r. 1101-25], the siyin half-seal (an inventory seal of the early-Ming imperial collection, used in the period 1374-84), and seals o the early Qing-dynastry (1644-1911) collector Wang Shimin (1592-1680). It differs from its model in two notable ways: the blurry brushwork, a feature of Zhang's style discussed below, and the top of the hill, which, unlike in the model, has no distinct crest but simply disappears into the dark silk. Why Zhang allowed this to happen is a mystery; it is, as we have seen, also a feature of Riverbank, and one found nowhere in early Chinese painting. A clue to understanding this curious feature of Zhang's forgeries may be recognized in another of the works done under his own name from the late 1940s, Mountain Temple and Drifting Clouds in the Manner of Dong Yuan…, painting in 1947. Here, he employs a compositional device favored, and probably invented, by the late-Ming painter and critic Dong Qichang (1555-1636), creating dynamic energies in the construction of an ascending mountainside, with the heavily modeled forms pushing this way and that, thrust answered by counterthrust, and then containing them (barely) with a simple, flattening contour line at the top. Deleting even this inconclusive ending (an easy move) would leave the mountaintop unbounded, as in Zhang's forgery of the Juran Temples among Steams and Hills, discussed above, and in Riverbank. In its dynamic build-up of the mountainside, Mountain Temple is strikingly similar to Riverbank main features the same slanting and serrated flat plateaus and the same rows of trees diminishing upward in size and clarity. These forms are conspicuous also in other of Zhang's early forgeries, such as the composition ascribed to both Liu Daoshi (active late 10th century…) and Guan Tong (active ca. 907-23…). One may even feel that here Zhang has given himself away by using so blatantly the materials of his 'early' forgeries in a picture done under his own name - materials that, to emphasize the point once more, are not to be seen in truly early landscape paintings. In his long inscription on Mountain Temple, Zhang writes that he painted it by 'consulting' three works by Dong Yuan in his own collection. Two years later, in the inscription on Immortals' Dwellings…, he would write that he had at last mastered Dong Yuan's style. What he had really mastered in the course of producing this impressive series of landscapes was not so much Dong Yuan's style as a system of forms and compositional devices (making up an artist's 'brush method,' in the traditional and inadequate Chinese term) that would serve as a repertory for his forgeries of tenth-century masters. A spurious Juran handscroll titled Myriad Ravines…, which was sold at auction in 1987, is another of Zhang's pastiches, filled with antique-looking inventions that again fail to come together into a coherent picture. Its imagery includes the now-familiar river winding out of an ambiguous distance, widening and narrowing arbitrarily, crossed and paralleled by roads; the zigzagging, flat-topped bluffs, here elongated into bizarre forms; a profusion and diversity of trees; soft, blurry dotting; and, near the end, the steep slope with rows of trees that swoops dramatically upward - and disappears. The scroll could have been a kind of warm-up for Riverbank, Zhang's masterwork in the genre, in which these motifs are handled with considerably more finesse. A hanging scroll in the same style, Dense Groves and Layered Peaks…in the Liu Haisu Art Gallery, Shanghai, also inscribed to Juran and purporting to be from the Xuanhe collection of Emperor Huizong (with the appropriate seals and title), must have been done by Zhang around the same time….It bears an authenticating inscription by the collector Wu Hufan (1894-1968), who saw the painting in 1951. The scenery it presents, like that of the handscroll Myriad Ravines, is highly mannered and unnatural; but since there was no agreement on how a Juran painting should look, the forger had considerable latitude. Wind on the River, a handscroll once attributed to the Northern Sung master Yan Wengui (active ca. 970-1030) that sold at auction in March 1999, is another work that has many features in common with Riverbank - the profusion of windblown trees and diverse types scattered throughout the composition, the fishnet pattern on the water, the zigzag mesa, and the mysterious lighting - not to speak of the familiar array of impressive 'old' seals.

"Some supporters of Riverbank content that its brushwork confirms its authenticity as a tenth-century work. This contention is odd in view of the fact that there is in a sense no brushwork in most areas of the painting representing earth surfaces - that is, no traces of the brush having been put down and moved as to leave distinct brushstrokes. Instead, the ink is rubbed onto the silk smoothly, without separate and visible strokes. In this, of course, it departs fundamentally from the systems of brushwork commonly seen in early Chinese landscape painting, where the brushstrokes, even when they seem strange and sloppy (as in Wintry Groves and Layered Banks, ascribed in Dong Yuan in the Kurokawa Institute of Ancient Cultures…or overlap and interweave (as in The Xiao and Xiang Rivers, also attributed to Dong Yuan, in the Palace Museum, Beijing…, still read as distinct strokes. Needless to say, these two paintings along with others attributed to the artist, are themselves problematic both in their dating and in their relationship to Dong Yuan, so that for Riverbank to look like them is certainly not a count against it. The argument has been made that the brush technique in Riverbank represents a stage in the history of Chinese landscape painting before texture-stroke systems were developed, and that in this respect Riverbank agrees with two generally accepted tenth-century paintings, A Chess Party in the Mountains…, by an unidentified artist, from a Liao dynasty (916-1125) tomb at Yemaotai, and The Lofty Scholar Liang Boluan…, by Wei Xian (active ca. 960-75). But a study of details of these works betrays important differences. In the Liao tomb painting…the earth masses are powerfully sculpted by brushstrokes that have the double function of shading, as they are applied densely or thinly to render light and shadow, and texturing, imparting tactility and earthiness. Moreover, while they are not uniformly distinct, the brushstrokes tend to be linear, as if made by a dragged brush, and direct the movement of the viewer's eye over the curving surfaces, thus enhancing the three-dimensionality of the forms. The result is a landscape in which spaces are strongly hollowed out, enclosed by convincingly rendered concavities in the earth masses….We see a similar effect achieved in a well-known landscape painting on a lute plectrum guard, in the Shoso-in Treasury, Nara. The same is true in Wei Xian's Lofty Scholar…, where the ink is brushed onto the earth masses so as to render light and shadow and tactility, as well as direction - in this case, mostly upward for an effect of height. Again, the masses are sculpted in a readable way….The application of ink on the foreground rocks is somewhat looser than in the upper area, but is still done in dabs and flecks, or in some places as directional streaks. There is no real ambiguity. Let us compare Riverbank…, where the almost strokeless rubbing of ink onto the silk produces an undifferentiated texturing and a light-and-shadow modeling so inconsistent as to make both the forms themselves and their interrelationships unreadable in some places. This is not a deliberate and expressive manipulation of geological forms and their lighting, as in Guo Xi's (ca. 1000-ca. 1090) Early Spring…; it is the outcome of a lack of full control, and results in an effect of arbitrariness. Often we cannot be sure whether one form is in front of another or behind it. For example, if we try to read the middle-ground conglomerate of earth masses and what are presumably rocks, we are continually frustrated. We find a comparable blurriness not in any early painting but in the British Museum's Dense Forests and Layered Peaks…, now recognized as a Zhang Daqian forgery of a work by Juran. It is in a different style from Riverbank, but the application of ink is similar, with large and small brushstrokes merged with rubbed-on ink into an atmospheric obscurity. The same kind of brushwork, more visible when on paper, can be seen in Retreat in the Yayi Mountains after Wang Meng, a signed 1954 work by Zhang Daqian. The silk ground in the British Museum picture and Riverbank increases the blurriness. A representational problem caused by this 'strokeless' application of ink is that forms, unless clearly bounded, tend to merge confusingly with surrounding areas, as can be seen in the lower part of the British Museum painting where the trees merge with the hillside behind. This confused merging of forms is what prevents us from making out the top of the thrusting bluff that terminates so indecisively the middle-ground landmass of Riverbank…: it is not set off visually from the equally confused area meant to appear behind it. It is because of such merging, in fact, that there are so many places in this part of the painting where we are not even sure what we are supposed to be looking at. Richard Barnhart has likened the rendering of earth forms without contours in Riverbank to that in Zhao Gan's (active ca. 960-75) handscroll First Snow along the River…in the National Palace Museum, Taipei. This comparison strikes me as unfortunate, since Riverbank comes off so badly in this respect against the Zhao Gan work, in which the forms are consistently distinct, never blurring or fusing ambiguously as they do in Riverbank."

"While one can find landforms that function dynamically in early Chinese landscape compositions," Cahill argues, "there are none quite like those in Riverbank, which lunge diagonally and are countered by masses lunging in the opposite direction. The artist, I submit, was very familiar with this compositional method as it had been developed by Dong Qichang and his followers, and used it, perhaps unconsciously, in this inappropriate context. Juxtaposing Riverbank with Dong Qichang's landscape painting A River in Chu..., we observe that in certain respects the two works have more in common with one another than either has with any genuine tenth-century landscape. Besides the domination of both compositions by pointed, volumetrically rendered earth masses that engage tin vigorous diagonal thrusts and counterthrusts, in both paintings small, blockly masses are placed at the feet of slopes, and trees of strikingly varied types are lined up in the foreground. (The foreground trees in Riverbank are meant to recall those in Dong Yuan's Wintry Groves and Layered Banks, but the passages are actually quite dissimilar.)…The overly strenuous, muscular forms in Riverbank must be distinguished from the diagonally pushing rock and earth masses seen in some early landscapes, such as those in the rocky background of Shaka, the Historical Buddha, Preaching on Vulture Peak…, perhaps ninth century in date, in the Museum of Fine Arts, Boston. Here, the dynamic forms are limited to a particular area of the composition, and are not expanded to become an organizing principle for the whole. Zhang Daqian's fondness for over-animated compositions based on diagonally disposed masses is seen in others of this forgeries, such as Clear Morning over Lakes and Mountains…, attributed to Liu Daoshi, and (the same composition) Drinking and Singing at the Foot of a Precipitous Mountain…, ascribed to Guan Tong, as well as in another would-be Dong Yuan, a handscroll titled Waiting for the Ferry in Summer, in the National Palace Museum, Taipei, one of Zhang's clumsier and least convincing productions. The heavy shadowing in the ravines and crevices that separate the earth masses in Riverbank, enhancing the overall dramatic effect of the composition, is used to the same effect in a landscape by early-Qing artist Wang Hui (1632-1717…), falsely attributed to Xu Daoning (ca. 970-ca. 1051/52). Nothing of the kind can be found in early landscapes; the lighting in Guo Xi's Early Spring, for instance, is very different. Zhang Daqian could not resist dramatizing, making his 'early' pictures more visually exciting than they properly should be. And Riverbank is indeed exciting in this way. It is as if a Cézanne landscape were dropped in among a group o Claude Lorrains; to a twentieth-century viewer, it would stand out as far more stimulating than its companions. Returning to the quieter world of genuinely early Chinese landscape, consider the Song-period work in the Juran manner, Xiao Yi Seizing the Lanting Manuscript…. We can argue about the date - I have put it as late as the Southern Song (1127-1279) - but the landscape forms, while they belong distinctively to the Juran manner, are stable and earthy; the trees diminish convincingly and do not attempt to exhaust the entire repertory of tree types; the whole composition is undramatic, clearly readable, and - exactly because it does not lay claim to being the Mona Lisa of Chinese paintings - deeply satisfying."

Cahill argues that Riverbank's lighting is visually stimulating to some observers but "here, again, the artist reveals his own time, unintentionally." "The drama is achieved not only through pronounced contrasts of light and dark on the forms (often with transitions that are too sudden), but also by the creation of areas of unexplained luminosity in the picture. The lighting is not naturalistic, but rather gives the effect of an unnatural glow, as though the forms themselves were radiating light. This is what so strikingly distinguished Zhang Daqian's forger of Juran's Dense Forests and Layered Peaks…in the British Museum from the Shanghai picture that is its model…, an old work in traditional style, however one may date it. One sees the same lighting in another recognized Zhang forgery, Drinking and Singing at the Foot of a Precipitous Mountain…, a gainting formerly attributed to Guan Tong in the Museum of Fine Arts, Boston, in which the cliffs flanking a dark ravine are suddenly and strangely sunlit, and in the previously mentioned handscroll Wind on the River, attributed to Yan Wengui. This treatment of light is no less striking in Zhang's forgeries of figure painting, as seen, for example, in his Laozi Passing the Barrier…, ascribed to an unidentified artist of the Tang period (618-907). The painting, as I pointed out in a 1991 symposium on Zhang Daqian, is based on a wood-block print in a Japanese picture book of 1818. In all these paintings, Zhang was playing to a widespread belief, which is not without bias, that early Chinese artists sometimes used illusionistic lighting effects of a kind that all but disappear after the early Song…."

"The painting that Zhang Daqian took to be a great work by Dong Yuan (or, more likely, that Zhang himself created), Along the Riverbank at Dust, in the National Palace Museum,Taipei, is overinhabited in just this way, a feature that convinces me even more of Zhang's likely authorship. In any case, the work is obviously too late to serve as an early precedent for Riverbank. And finally, the forgery attributed to Juran, Dense Groves and Layered Peaks…, is similarly overpopulated, the point where its status as an image of reclusion is severely compromised. Zhang Daqian apparently lacked the restraint that prevented early landscapists from enlivening their pictures with a profusion of active, attention-drawing figures….One of the differences between the British Museum's Dense Forests and Layered Peaks, Zhang's forgery of a Juran work, and its older model in the Shanghai Museum…, is that whereas the man in the Shanghai version is self-contained, exhibiting a proper Chinese decorum, his counterpart in the British Museum picture leans moodily toward the water, gazing over his shoulder, as if warding off the unease generated by his vibrant, indistinct surrounding. The scholar in Riverbank…is closer to the British Museum image. As [Maxwell] Hearn writes, he 'serenely faces the storm - a vivid metaphor for the political chaos sweeping across the land at the time.' The problem, at least to my eye, is that he does not simply express his serenity, he also dramatizes and projects it. Those familiar with Zhang Daqian's acknowledged works know how the central figure in his landscapes, besides frequently looking like the artist himself, express more self-awareness than do those in early paintings."

Cahill states that he is not an expert on calligraphy and reports the views of other experts about the writing on the painting. He quotes Daniel Bryant, a specialist in Chinese language and history at the University of Victoria as quoting Suzuki Kei, a leading Japanese authority now retired from teaching at Tokyo University, as being 'troubled by the discrepancy between the low status of the position [allegedly held by Dong Yuan] in the palace garden administration' and the high status assumed for Dong Yuan by the eleventh-century writer Guo Ruoxu, in whose entry on the artist this title first appears. [Hironobu] Kohara goes further, pointing out that although there is no evidence for the meaning that the title had in the Southern Tang state under which Dong Yuan served, the section on official titles in Songshi (History of the Song Dynasty) identifies the position as that of a minor functionary, with 'the same duties as that of an eunuch,' definitely not an honorable status. It is impossible, Kohara believes, that Dong Yuan would have used this title with his signature, and he speculates, 'I think Zhang [Daqian] got this knowledge from Guo Ruoxu's book, but could not check the role of the title beyond the officer's name.' Kohara also points out that the presence of the word chen (your subject), written in smaller characters and set above and to the right of the signature, follows court practice in the Qing period. It is also true, however, that chen sometimes accompanies signatures on Song paintings….Thus it is not impossible, only very unusual, that the term would appear on a tenth-century painting. In his 1989 article on the handscroll The Three Worthies of Wu…in the Freer Gallery of Art, Fu Shen discusses the calligraphic style of the signatures and inscriptions on Zhang Daqian's forgeries, finding them (as others have) to be apparently from a single hand, even when the works on which they appear span centuries. The style, Fu writes, is intended to look like pre-Song writing (Wen Fong considers the Riverbank signature to be in the style of the eight-century calligrapher Yan Zhenqing (709-785), but it is in fact 'a single personal style,' which in some aspects resembles Zhang's own. The writing on Zhang's forgeries is sometimes said to have been done by Zhang's third wife, Yang Wanjun, but Fu Shen believes it is by Zhang himself. Kohara observes that the calligraphy of the Riverbank signature is close to the writing on the British Museum painting Dense Forests and Layered Peaks."

Cahill generally dismisses the significance of seals for purposes of authentication. He does, however, quote a letter in the November 1998 issue of Orientations magazine by Steven P. Gaskin that took issue with a letter in the July-August 1998 issue of the same magazine by Kathleen Yang that was "In support of The Riverbank."

"'The seals she…shows us are actually a rather damning bit of evidence, since there are sufficient differences between the ones on the painting and the 'authentic' examples shown that we must seriously suspect they are later additions,'" Cahill quotes Gaskin. Cahill quotes a subsequent letter to him from Gaskin that went into further, specific detail and Cahill wrotes that "The distinctions he makes between these seal impressions appear to my eye to be accurate," adding that "More than individual matches or mismatches, however, what arouses suspicion is the impressive panoply of old seals, some of which (including the siyin half-seal) appear regularly on Zhang's other forgeries and follow the same pattern of distribution over the centuries, especially in the Song-Yuan period…."

"Both Richard Barnart and Wen Fong attempt to identify Riverbank with a painting mentioned by the thirteenth-century scholar Zhou Mi (1232-1298) in his Yunyan guoyan lu (Record of What Was Seen as Clouds and Mists) as a work he saw in the collection of Prince Zhao Yuqin (late 13th century)," Cahill notes. "But," he continued, "as Ankeney Weitz points out in her study of Zhou Mi's book, in a section of modern 'misuses' of that book by collectors eager to match paintings they own with those recorded in it, the identification is 'somewhat suspect' (an understatement!) since 'according to Zhou Mi's text this painting should have been a 'short handscroll,' not a larger hanging scroll.' As she also notes, Barnhart himself had rejected the identification earlier because of this discrepancy, but in his 1983 book, Along the Border of Heaven: Sung and Yuan Paintings from the C. C. Wang Family Collection, changed his mind. Like the owner of Riverbank, C. C. Wang, he accepted the identification on the basis of seals of Jia Sidao (1213-1275) and Zhao Yuqin. Weitz comments, 'In my opinion, these collectors' seals could be reproductions. Further, the presence of the seals of Zhang Daqian, a master forger, should raise at least some doubts about the scroll's purported provenance.' If Riverbank had come down to us lacking seals indicating that it was known to important writers and collectors - as, for instance, seems to be the case with the Shanghai Museum's Bamboo, Old Trees, and Rock, attributed to the tenth-century painter Xu Xi (active 943-75) - its absence from old catalogues and record books would be easier to understand. However, the seals on Riverbank indicate that during the period from the late Song to the early Ming, the work passed through the hands of prominent owners. It is hard to believe, then, that it would not have been noticed by major collectors of later times and recorded and commented on as one of the notable relics (mingji) of Chinese painting. In the seventeenth century, when surviving early masterworks were increasingly being identified and written about, and in the eighteenth century, when many early landscape paintings had passed into the imperial collection and examples were rarely to be seen outside that collection, Riverbank should have enjoyed fame as a major signed work by a much sought-after master. But it somehow escaped the notice of Dong Qichang and all the collectors and connoisseurs who followed him, until it was published in 1956 by Zhang Daqian and in 1957 by Xie Zhiliu (1910-1997).

Cahill notes that Wen Fong maintains that Dwelling in Seclusion in Summer Mountains by Wang Meng (ca. 1308-1385) had as its model Riverbank. "To my eye, there is not much resemblance between the two paintings," Cahill asserts, adding that "A relationship between Riverbank and the 'half-landscape' ascribed to Dong Yuan in the Ogawa Collection in Kyoto, Travelers amid Streams and Mountains….is suggested by both Wen Fong and C. C. Wang, who describes the two paintings as 'very close.' But again, I fail to see the 'close' resemblance. Moreover, the Ogawa picture has its own problems of dating and origin and can therefore scarcely serve as firm support for another problematic work."

In this section of his "indictment" of Riverbank, Cahill discusses at length a letter from the painter Xu Beihong (1895-1953) that has been used by some supporters of Riverbank to support its authenticity, or, as Cahill puts it, "at least lessen or eliminate the likelihood that it was painted by Zhang Daqian, by proving that the painting existed in 1938, before Zhang began his career as a forger of pre-Song paintings."

Admitting to be addicted to "thrillers," Cahill argues that a serious forger can "construct" an impressive "history" for a specific work and he cites three Zhang Daqian forgeries as examples: the "Guan Tong" in the Museum of Fine Arts in Boston that was "mounted and boxed so as to make us believe that it came frm the storehouse (kara) of some old Japanese family where it has been kept for centuries; the 'Liang Kai' Sleeping Gibbon in the Honolulu Academy of Arts, bearing a title purportedly writeen by a famous thirteenth-century scholar and an inscription by a modern collector claiming that it was bought by his grandfather in 1892; the 'Han Gan' Horses and Groom in the Musée Cernuschi, Paris, accompanied by an inscription written by Pu Ru (Pu Xinyu; 1896-1963), a member of the Manchu imperial family, stating that it was found in a cave at Dun-huang in 1900 and subsequently remounted in Japan. I would suggest that the so-called history and provenance of Riverbank should be seen as another notable product of this kind of project - a construction to be admired in itself, but not to be believed."

Cahill focuses on the work's "recent" and "official" history, according to which the painter Xu Beihong "acquired the work in 1938, in Yangshuo, Guangxi Province." "Later that year, Zhang Daqian saw the painting when he visited Xu in Guilin and was 'so enamoured of it' that Xu allowed him to take it to Sichuan 'to research and authenticate it.' Xu Beihong described these events briefly in a letter written shortly afterward to a Mr. Sun…Some years later - in 1944, according to an inscription by Xu Beihong…- Xu agreed to give up Riverbank in exchange for a painting by Jin Nong (1687-1763) owned by Zhang. Riverbank remained in Zhang's possession until it was acquired in about 1968 by C. C. Wang in exchange for twelve paintings by later artists."

Cahill suggests a different version - a "complicated swindle" that "requires the complicity of only a few people: Zhang Daqian himself, Xu Beihong, the scholar-painter Xie Zhiliu, and possibly Ding Xiyuan (in addition to others on Zhang's team who assisted in making and aging the painting, forging the seals, and so forth)."

"As to why Xu would consent, we could speculate that some material consideration may have been involved, but this need not have been the case, since Xu, who had known Zhang Daqian for years (Zhang had been hired by Xu to teach at Nanjing University from 1934 to 1936), could have done it purely out of old friendship. But we can also counter the question by asking: why did other prominent collectors and connoisseurs agree to help Zhang Daqian by providing provenances and recent histories for his modern fabrications? These include Wu Hufan, for the 'Wu Wei' handscroll in The Iron Flute in the Shanghai Museum and, along with Ye Gongchuo (1881-1968), the 'Liang Kai' Sleeping Gibbon in the Honolulu Academy of Arts; Pu Ru, for another version of the Sleeping Gibbon, the 'Han Gan' Horses and Groom in the Musée Cernuschi, Paris, The Three Worthies of Wu in the Freer Gallery, and quite a few others. Wu Hufan's inscription on the Sleeping Gibbon claims that the work was purchased in Hangzhou in 1892 by his famous grandfather Wu Dacheng (1835-1902), whose (forged) seal appears on the painting - another obliging attempt, like that of Xu Beihong for Riverbank, to establish a recent history for the work that would eliminate the possibility of Zhang's authorship. Pu Ru reportedly wrote colophons for Zhang's forgeries without ever seeing the works; Zhang would send him a request, with the wording he wanted, and Pu would write out and sign and inscription, which was later mounted with the painting. Along the Riverbank at Dusk, a painting once attributed to Dong Yuan that was owned by Zhang and is probably another of Zhang's forgeries, is accompanied by inscriptions by Wu Hufan, Xie Zhiliu, Pu Ru, and the famous older Shanghai collector Pang Yuanji (1864-1949). Surely these people were too sharp-eyed to believe in the paintings they were 'authenticating.' C. C. Wang once told me (in another context) that members of this group commonly did such things to 'help each other out,' lending their names to spurious works to make them more believable. The device was effective, since foreign collectors and museum curators in the 1950s still accepted too easily the cultural authority of famous Chinese connoisseurs….It is clear from the well-documented record that if Xu Beihong did indeed assist Zhang Daqian in constructing a recent history for Riverbank, as I believe he did, he was doing only what appears to have been a common practice at the time, one not even considered especially reprehensible."

"The story in the letter about Zhang's taking Riverbank away at this time to 'study it' was necessary to explain why the painting dropped out of sight for some years - why, in other words, Xu did not show it proudly to others, write about it, and make it known, as he did with his beloved Eighty-seven Immortals…., a long handscroll depicting a procession of figures that in Xu's opinion was the work of a Tang artist, for which he composed a jubliant colophon soon after acquiring it," Cahill writes, adding that Xu had acquired Eighty-seven Immortals a few months after Riverbank according to the "official" history although "it is clear from other evidence" that he had actually acquired the previous year with money and his own paintings from "an unidentified German woman who is said to have been the daughter of a long-time resident of China whose collection of paintings she inherited." "It is somewhat mysterious that whereas Xu provides a detailed story of this acquisition, in his colophon and elsewhere, neither he nor anyone else gives any clue abut the source or circumstances of his acquisition of Riverbank, a painting he himself called 'perhaps the preeminent Dong Yuan under heaven.'….The question of what happened between 1938…and Riverbank's 'reappearance' constitutes the first of the gaping holes in the official account. Because nothing at all happens, for a dozen years or more. There is no record of Xu's mentioning the painting to anyone or writing about it again until 1950 - and even that date is uncertain, since it appears in Xu's own inscription, on a Jin Nong landscape….Nor does Zhang Daqian, so far as we know, show it to anyone or write about it during these years. Riverbank needed to disappear for some years, at least until the late 1940s or early 1950s when, I believe, it was in fact being made….C. C. Wang himself offers a variant account of Riverbank's history, which, if reliable, would clear up everyone's doubts on the matter. To the question put to him in a series of unpublished interviews with Joan Stanley-Baker in the 1970s, 'Could the signature be a work of Zhang Daqian?' he responded: No, I saw the work before Zhang Daqian bought it. I saw it when Xu Beihong bought it from a German. It was an exchange. Xu Beihong exchanged works with the German (a merchant or official) in Shanghai, and I remember he got two works. One is this Dong Yuan, the other a handscroll on silk, [it] was called Bashiqi shenxian [Eighty-seven Immortals]. I think it was an adaptation [by] a Southern Song artist of the Wu Zongyuan scroll in my collection. Later Zhang Daqian for some reason bought it from Xu Beihhong. Later I bought it from Zhang Daqian. But C. C. Wang, good connoisseur though he is, misremembers on just about every point. Xu Beihong, according to Xie Zhiliu's account, acquired Riverbank in Yangshuo, not in Shanghai. Xu cannot have brought it from the German (a woman, not a man) from whom, according to the official account, he acquired not Riverbank but the Eighty-seven Immortals, a transaction that took place some time after he had met with Zhang Daqian in Guilin and let Zhang take away Riverbank….Finally, C. C. Wang cannot have seen Riverbank inXu's possession in Shanghai before Zhang 'bought it,' because, even in the official version, Xu never had the painting in Shanghai.

Cahill continues his "alternative" history of Riverbank by suggesting that the Xu Beihong subsequently informs some people of the letter, "said to be now be in an unidentified private collection in Japan, and is accessible only in the 1988 Taipei Fine Arts Museum catalogue reproduction, and in a photograph provided by (who else?) Ding Xiyuan." "Xu's confidants must have included Xie Zhiliu, well known to be Zhang Daqian's disciple and close friend. It would appear that Xie entrusted Ding Xiyuan with the information and charged him with the responsibility of transmitting it. Ding, who reports a conversation with Xie Zingliu in June 1996, shortly before Xie's death in 1997, does indeed publish the letter again…at the suitable moment, along with a lengthy account of the whole fabricated history. By that time, Xu Beihong, Zhang Daqian and Xie Zingliu had all departed this world, leaving only Ding to carry out the plan and to say, with Ishmael, that he alone survives to tell the tale."

"Another very shaky episode in the fabricated history of Riverbank is Xu Beihong's fascination with a Jin Nong landscape painting, Returning by Boat in a Rainstorm…, and his willingness to accept it in exchange for Riverbank. This, too, we learn from Ding Xiyuan, as personally related to him by Xie Zingliu, and also, in a shorter version, from Xu's widow Liao Jingwen, who mentions the exchange of the two works in the English version of her preface to the 1991 volume of reproductions of old paintings in Xu's collection, in which the Jin Nong is reproduced. Xu apparently could not rest, however, with having the story and its crucial 1938 beginning recorded only in his letter and in the memories of his friends. He had to get it all written down once more, and did so in the inscription he wrote in 1950 (or dated to that year) on the Jin Jong painting….Roughly translated, it reads: This is one of the rarest works in Chinese painting. In my lifetime I've seen Fan Zhongli's [Kuan]'s Traveling among Streams and Mountains, the anonymous Song-period Snow Scene [?], Zhou Dongcun [Chen]'s North Sea [now in the Nelson-Atkins Museum of Art, Kansas City], and this painting - they can be called the four pillars of extant Chinese landscape painting in all the world. The four [Five Dynasties] masters Jing [Hao], Guan [Tong], Dong [Yuan], and Ju [ran] are much praised in the world. Works by Jing and Dong can still be found today; those by Juran are hard to find. They all would be difficult for me to choose. [?] In July 1938 [Zhang] Daqian took away with him from Guilin my large Dong Yuan painting. In the spring of 1944, when I was living in Chonking, Daqian, knowing that I especially loved this unusual Jin Nong painting in his collection, asked [Zhang] Muhan to bring it to me [in exchange]. I was very happy, because it is the quality of the painting that I value, not the name [of the artist]…."

"Instead of supporting or corroborating the evidence of his letter, this inscription, by being so excessive and unnatural, calls it seriously into question. Xu Beihong, like other collectors, often wrote inscriptions on the mountings of paintings he owned, expressing his enthusiasm for them and recording information about them. Here, he begins and ends with extravagant praise for Jin Nong's landscape, ranking it (absurdly) with the Fan Kuan and two other paintings as the four great originals among extant Chinese landscapes. He makes the claim that he values the quality of a painting more than the name of the artist; an extraordinary Jin Nong, he implies, is worth as much to him as a large, genuine Dong Yuan, even one that he had earlier, in his letter, called 'perhaps the preeminent Dong Yuan under heaven.'…What seems obvious is that Xu was planting his story once more to 'corroborate' the rest and to try to make the exchange seem plausible - as it decidedly is not, on the face of it. Kohara believes that the Jin Nong is itself a Zhang Daqian forgery and it may be. (I saw it years ago, and did not doubt it.) The question of authorship is not crucial to my argument; in any case, the picture is certainly a modest work, especially in relation to Riverbank. Liao Jingwen writes of the exchange, 'Because of the fact that paintings by Dong Yuan…were hard to find and also because of the big size of that landscape, it cot [i.e., was worth] over ten thousand times more than the piece of Jin Nong in terms of money.' We are supposed to believe, as Liao Jingwen did and as Ding Xiyuan argues, that the story portrays Xu rising admirably above material concerns. But, as Kohara remarked to me, 'Xu Beihong was not stupid,' and stupid he would have to have been to agree to such a trade. We should note that Xu's account, which states that Zhang Daqian asked his younger half-brother Zhang Muhan to bring the Jin Nong to him in Chongking in the spring of 1944, does not agree with Xie Zingliu's account as told to Ding Xiyuan, in which Xie himself served as the go-between. Since it is unlikely that either Xu or Xie would have a lapse of memory about such a momentous event, it would appear that they simply failed to coordinate their stories. Zhang Daqian, then, 'acquired' Riverbank in 1944. Even after this, however, he did not, as one might expect, present his new acquisition proudly to the world. If a genuinely old painting of this size and complexity, bearing the signature of a major master, had fallen into his hands at this time, he would have trumpeted its importance, proclaimed it the find of the century, and enlisted his friends to write colophons for it. None of these happened. Zhang is mysteriously - and as regards the official account, fatally, silent about the painting in the years that follow; it is as if the work did not exit. And for some of those years, it truly did not, because he had not yet created it. But his silence about it continues even after the painting must have come into existence. In 1947, he writes in a colophon, 'I now own three Dong Yuans,' but does not include Riverbank among them. In 1949 he names two others in his collection as works that permitted him to 'understand [Dong Yuan's] brush method,' but again, Riverbank is not one of them. Fu Shen acknowledges that Zhang seems never to have made a copy of the painting or even discussed it. And when, in 1955, Zhang published the first volume of his Dafengtang mingji (Taifudo meiseki), with reproductions of the most important old paintings in his collection, Riverbank is again absent. Instead, Dong Yuan is represented there by the far less interesting and, in fact, quite pedestrian Along the Riverbank at Dusk…, at best a Ming painting and at worst another of Zhang's fakes. Riverbank is included at last, without commentary, in the fourth volume of Dafengtang, published in 1956. The explanation for all this silence that supporters of the painting offer is that Zhang 'did not recognize the importance of Riverbank.' This explanation is clearly inconsistent with Xu Beihong's statement in his letter that in 1938 Zhang was 'so enamored of it that he insisted on taking it with him'; it is also inconsistent with estimates of the painting made by Xu Beihong, and later by Xie Zingliu….How could Zhang be so determined to acquire Riverbank and then set so little store by it? The truth of the matter, I submit, is that he knew perfectly well that the sudden appearance of another previously unseen 'ancient masterpiece' in his hands would be greeted with deep suspicion, as indeed it should be. He had already introduced another would-be Dong Yuan, Along the Riverbank at Dust, which is probably one of his own fabrications, with great fanfare and a story about how he had seen it in 1938, dreamed of it for years, and finally acquired it for a huge price in 1945. Newspaper accounts hailed the event and the unprecedented price paid; five major connoisseurs wrote colophons for the painting to accompany two by Zhang himself. Zhang could not pull the same trick again without arousing suspicion; he needed another stratagem."

Cahill notes that Xie Zhiliu reproduces Riverbank in 1957 in Tang Wudai Song Yuan mingji (Famous Paintings Surviving from the Tang, Five Dynasties, Song, and Yuan), and in 1979 "would write of Riverbank as presenting Dong Yuan's 'original appearance,' that is, his early style, while the better-known The Xiao and Xiang Rivers…, and two other handscrolls ascribed to Dong Yuan, Summer Mountains in the Shanghai Museum and Awaiting the Ferry at the Foot of Summer Mountains… in the Liaoning Museum, are said to belong Dong Yuan's transformed style from a later period. Still later, in 1989, along with relating yet again the story of Zhang Daqian's acquisition of the painting in the exchange with Xu Beihong, Xie would add a 'middle period' for Dong, represented by Residents of the Capital City…in the National Palace Museum, Taipei, a picture now usually dated to the Yuan period."

"The record of fabrication for both the painting and its history is too long and too intricate to be easily laid to rest. But laid to rest it must be, in the end, if we are to remove this interloper from our histories of early Chinese landscape painting," Cahill concludes.

In his essay on "Notes on the Recent History of Riverbank," Hironobu Kohara notes that the delivery of the Jin Jong painting to Xu Beihong is not recorded in The Life of Xu Beihong: A Memoir…., written by Liao Jingwen, the second wife of Xu Beihong. Furthermore, "the most comprehensive documentation of Xu's life yet published, the Xu Beihong nianpu (Chronology of Xu Beihong), edited by Xu Boyang and Jin Shan, also does not record this event."

Kohara discusses inscriptions on Eighty-seven Immortals, a Song dynasty painting, now in the Xu Beihong Memorial Museum in Beijing, that was Xu's favorite. Xu Beihong had the painting stolen from him in 1940 but recovered it two years later and requested Zhang Daqian and Zie Zhiliu to add their own colophons to the scroll. "…mysteriously, neither mentions Zhang's 1939 acquisition of Riverbank. It is hard to believe that, in writing about early paintings they owned and admired, Zhang and Xie would fail to mention Riverbank, which was supposedly a Northern Song painting and which was ostensibly already in Zhang's collection. Following a traditional format of inscriptions for calligraphy and painting, Zhang and Xie are straightforward in their colophons and do not hesitate to cite other works. In this case, had they known of it, the work they would have most likely mentioned would have been Riverbank. There was no reason for Zhang to keep Riverbank a secret. The cause of this improbable omission is that neither Zhang Daqian or Xie Zhiliu could have written about Riverbank, as the painting did not exist at that time."

Kohara also discusses a letter purportedly writted by Xu to Zhang's elder brother, Zhang Shanzi (1882-1940), that mentions "Dong Yuan's large painting" and states that it has a signature by Dong Yuan, a seal of the Song palace, and collectors' seals of Jia Sidao and Ke Jiusi, and whose silk is "badly damaged."

"Compared with an example of Xu's writing from 1934…, which shows his graceful and fine calligraphic style, the writing in this letter appears sluggish and loose. That it is not from the same hand is clear from the two characters 'Beihong,' which are noticeably inferior in execution to those in Xu's own writings. The letter is a poor and senseless forgery by Zhang Daqian and displays no attempt to imitate clearly the style of Xu Beihong's hand. While Zhang Daqian expended great energy and attention to detail in producing paintings, he had a tendency to make slovenly and careless mistakes out of his contempt for viewers and his disregard for their level of judgment….The letter also mentions with affection Li Zuhan and his sister Qiujun, two close friends of Zhang Daqian. When in Shanghai, Zhang always stayed with the Li family. He attended Qiujun's birthday celebration and purchased a tomb with her, pledging to be 'buried together when death comes upon us.' Zhang also promised to write a tomb inscription for her. So close was their friendship that when Qiujun died, in 1971, Zhang mourned deeply, considering it 'the loss of the best friend of [his] whole life.' It is inconceivable that Xu Beihong, who did not know them, would have referred to Li Zuhan and his sister in such an intimate manner. For all the reasons given above, there can be no doubt that the letter is an outright forgery made by Zhang. Consequently, the story about his receiving the Dong Yuan from Xu Beihong is fundamentally unsubstantiated. In short, all the episodes surrounding Riverbank are fictions, written and acted out by Zhang Daqian himself."

Bohara observes that Xu Beihong was "respected both inside and outside China as a patriotic painter who led a virtuous life" and therefore "at first glance, it may seem inconceivable that he would take part in a conspiracy involving Riverbank."

However, Bohara suggests that legal proceedings in 1944 against Xu Beihong involving his pending divorce from his first wife, Jiang Biwei, may have provided motivation as it required that he give 100 of his paintings and one million yuan to his former wife and he "was having a hard time getting together the money." He concludes his essay by stating that "the account written by Liao Jingwen is perhaps her desperate effort to keep secret the scheme that had made her marriage possible."

Bohara also says that the whereabouts of Riverbank prior to 1938 "remain completely unknown." He notes that according to Selections from the Xu Beihong Painting Collection, Xu did purchase two paintings in Giulin: a landscape handscroll by Tang Yin (1470-1523) and an anonymous Ming portrait of an imperial concubine, both of which bear Xu's inscriptions, dated 1930 and 1951, respectively." "Only Riverbank lacks documentation regarding its acquisition," he added.

In his essay entitled "Riverbank: A Recent Effort in a Long Tradition," Sherman Lee notes that the word for a Chinese landscape is shanshui, or mountain-water picture," adding that "The water content is paramount." He compares Riverbank with Zhao Gan's (active ca. 960-975) First Snow Along the River and finds the water treatment in Riverbank to be "tiresome to look at and must certainly have been tiring to paint." "Nowhere does it dance and flatten in response to variations in the surface tension. It is not the shui observed in early works; only a modern could fail to see the varying tension when observing water in nature. In Riverbank we are confronted with a terra incognito rather than a terra firma. The closer we look at the details, the vaguer and more insubstantial the forms and shapes become. Whether a receding plane of earth or a vertical rock sheet, we are confronted with nothing. We do not perceive lines and boundaries, but an obscure, rough shape implying something vague; we strain to read the structure of the cliff, a rock, a mist-shrouded point, but are confronted with nothing. Surely there is nothing comparable to this treatment of substance in any of the handful of works surviving from the Tang, Five Dynasties, or early Northern Song periods. Wonders succeed wonders. A distant river becomes a path. A thatch roof appears, its structure sadly misunderstood. Scraggly old deciduous trees appear at random, the old holes dimpling their surfaces in a simulation of random selection. But some of the trees bear leaves, others are leafless with old twigs that jut forth, bare to our puzzled gaze. And the population of the area seems, when legs are visible, to be struck with a rare disease of thick ink - charitably due to retouching but perhaps not, since they seem to be of one family. These are discrepancies of style and representation. Similar discrepancies of appearance in the matter of the seals and the signature in the painting as well as the history of the painting are a piece of the whole, which matches the style and the methods of representation in the work. The result is a morass of starts, false starts, and half starts that point inexorably to a modern pastiche all too familiar to many of us and unworthy of serious consideration by our serious colleagues."

Mr. Silbergeld, a former student of Wen Fong, served as moderator of the morning session's panel discussion and quickly stripped off his tie and jacket to reveal a football referee's shirt and used sporting analogies to enliven his talk. His comments and conduct were received with great amusement by the panel and the audience.

Early in his remarks, he said that "Whatever doubts I might harbor about Riverbank, I commend the Metropolitan Museum for having acquired it and brought it into the public domain." "The purposes of a museum are not the same as those of an academic institution, but they greatly overlap and mutually support each other, and in the quest for deeper understanding - for paradigms and rules to guide us - I believe this acquisition will serve its purpose." He noted that for museums "a deferred judgment is often no option at all."

"If we look at early texts that mention Dong Yuan…, we find sufficient material to support the authenticity of Riverbank, but we also find sufficient material to impeach it. One thing that all the evidence demonstrates is that Dong Yuan was not regarded as a foremost artist by the judges of his own time, and not because he was obscure or because his paintings were unknown. While good enough to be called upon to paint for his emperor, it remained for historical standards to change, radically, before he enjoyed a great reputation in later times….while the Song imperial painting catalogue…speaks of the 'heroic vigor' of his brushwork, his 'perilously towering peaks,' his 'serried ranges and precipitous cliffs,' Dong's first great champion, Mi Fu (1052-1107), wrote that he 'was not a skillful painter of towering peaks.' Most scholars today acknowledge that we have no original, representative works of Dong Yuan: We know his 'typical' style only through copies, while Riverbank, at best, is presented as early and atypical, absent the brush technique - the 'hemp-fiber' texture stroke - that was his stylistic signature for later centuries of admirers….My question to all is, Riverbank aside, what should be expect in our search for the original Dong Yuan? Maybe he was not all that good a painter….Maybe, like our ancestors in the Olduvai Gorge, he was simply good genetic material out of which later material could eventually be fashioned. And with Riverbank in mind, perhaps lowered expectations and a tempered rhetoric are more appropriate to its assessment."

Silbergeld notes that Richard Barnhart has "written of Riverbank that 'its importance to the history of Chinese landscape painting can scarcely be overstated' and that it is 'one of the treasures of world art,'" but adds that "Strangely, then, in Barnhart's recent chapter on Five Dynasties and Song painting in Three Thousand Years of Chinese Painting, this painting goes entirely unmentioned by him. More strangely, in his book Along the Border of Heaven, Barnhart makes Dong the 'leading landscape master' in the Painting Academy of the Southern Tang and turns the academy painter Zhao Gan (active ca. 960-75) into his 'faithful student and follower.' In the book, Along the Riverbank, which accompanies the present exhibition, Fong follows this formulation while Hearn does not. Since Dong Yuan's distinguishing social characteristic was his scholar-amateur status, which made him an eligible model for later socially exclusive amateur artists, his radical transformation in to a leading court artist by Barnhart and Fong highlights the historical revisionism that has remade Dong Yuan's reputation again and again."

Silbergeld notes that while C. C. Wang maintains he has no doubts that Dong himself painted and signed the work, "others have been more doubtful and readier to separate the judgement of the painting from its signature." "Barnhart, for example, has written that 'there is no way to confirm or deny the authenticity of the signature.' Hearn concludes that 'no reliable signatures of Dong Yuan survive to corroborate the authenticity of this inscription,' but he dates the inscription to 'no later than the early Song period' based on a stylistic analysis that places it within the eight-century manner of Yan Zhenqing (709-785) as practiced by such later artists as Yan's eleventh-century champion, Ouyang Xiu (1007-1072). To other doubts about the signature, I would add this reservation, that the use of Yan Zhenqing's style might be about a century premature here, Yan's star not yet having risen high, and certainly not at this southern court. Riverbank is preseumably and peculiarly dedicated in Yan-stuyle script to the Southern Tang emperior Li Jing (r. 943-61), whose son and successor on the throne, Li Yu (r. 961-75), found Yan's calligraphy crude and objectionable - in Li Yu's own words, "like an uncouth farmer facing forward with arms folded and legs spread apart.'"

As regards the painting's seals, Silbergeld observed that "a careful examination of the early seal impressions for patterns of silk penetration and surface wear bolsters Hearn's conclusion that the Zhao Yuqin and Ming inventory seals '[support] the painting's existence in the thirteenth century,' but adds that "Yet, the fact that all the other purportedly early seals have no demonstrated match elsewhere and are possibly more recent in date leaves lingering the question of how and when such patterns of wear were established."

Silbergeld dismisses attempts to show that the painting's title of Riverbank is very old: "So the title is specious, and there is no reason to conclude that any historical references to a Dong Yuan painting named Riverbank, such as Zhao Mengfu's poem, coincide with the Metropolitan painting."

Silbergeld notes that some of the proponents of the work as an authentic Dong Yuan argue that the work is "an expressive statement of the recluse-scholar theme" and "a vivid metaphor for the political chaos sweeping across the land," but asks whether it is "not possible to view this painting less metaphorically, as less of an allegory of political crisis and instead something closer to face value?" "This more tepid interpretation would leave open several questions raised by the 'Dong Yuan' signature, as to how the retirement theme is intended to relate to the emperor addressed in the signature as recipient of the painting. Surely, the lowly servitor Dong Yuan did not paint this about his own retirement, so the painting can hardly be self-expressive; the signature, if valid, would not allow for Fong's conclusion that Riverbank was 'an embodied image of the artist's self.' And while early records tell us that Dong Yuan's royal boss….did indeed plan for his own future rural retirement, can we imagine Dong Yuan blowing him there on the winds of apolitical storm? So, while still clutching my whistler and posed to use it, my third question is": can we really be comfortable with the notion of this allegory of 'polit8cal choas,' in Hearn's words, being depicted for and presented to a sitting emperor? Would that not be an inflammatory political critique, implying that the emperor and his court officials were all failing I think, not up to the task of maintaining harmony in the State?…The notion that Dong Yuan, either as vice-director of the royal park or as a court painter, would have produced something so politically sensitive rouses my suspicion and causes me to raise the whistle to my lips…..In short, the 'political crisis' thesis and the dedicated inscription seem incompatible. The more credence one gives to the former, the less one is able to accept the latter, and it is the signature, primarily, that offends. Unintentionally I think, the defenders of this thesis have impeached the purported authorship of the painting with their interpretation of its theme."

Cahill's objection that the work has too many figures is dismissed by Silbergeld who notes that it has fewer figures per area of painting than several other early works. Silbergeld also downplays Cahill's concerns about the work's "spatial ambiguity," citing some other early examples.

Silbergeld argues that the painting's inscription "suggests the painting is not early," adding that "So, too, does the interpretation of the theme: if this is an early Dong Yuan - and not too early - then the Southern Tang was also young, "a prosperous and relatively stable haven from unrest" in Wen Fong's words."

"Where, then," Silbergeld asks, "is its imminent collapse? Conversely, the stylistic dating, if we were to accept it, could show that the interpretation of the theme is wrong. Clearly, there is confusion on the court. So, whistle. And buzzer."

Silbergeld argues that some critics may be too rigid in their attitude towards anomalous works and concludes that in his personal opinion, "this might well be early painting, even tenth century." However, he maintained that an attribution to Dong Yuan was "Very problematic." "An important painting for us today and a very important acquisition for the Metropolitan Museum? Yes, I think so. Yet that is just one more opinion. You can have a lot of fouls called against you but still win the game; and this game is not yet over."

In his essay, "Riverbank: From Connoisseurship to Art History," Wen Fong notes that "as younger scholars have attempted to find new approaches to Chinese art, they have turned for inspiration to a variety of techniques ranging from literary criticism, structuralist anthropology, and ethnography to an examination and material and visual culture." "Compelling though such new strategies are (and indeed they have both enriched and energized our ways of looking at Chinese art), there can be no real understanding of an individual work, either as art or as material object, without our first ascertaining, as accurately as possible, the date and circumstances of its production. Precisely because the history of Chinese art is fraught with imitations and forgeries, connoisseurship remains the only means we have with which to determine the authenticity and significance of an individual work as evidence of historical inquiry. More than simply a tool for the art market, connoisseurship - the examination of the visual evidence of a work of art - alone enables us to decipher what art can tell us about history," Fong wrote.

Of course, art has other purposes than to "tell us about history": it can tell us about an individual's personal expression and vision; it can tell us about different ways of viewing objects, subjects, and themes; it can be didactic, as in political art; it can be stimulating as in stylistic explorations; and it can just be lovely, something that inspires reflections about beauty and the value of beauty.

"Because the history of Chinese painting is still a relatively young discipline," Fong continues, not bothering to observe that as far as histories of painting go it is just about the longest, "it is possible for different critics to arrive at contradictory assessments of the same work or sequence of works. Such is the case with Riverbank…, which, if our attribution of it to the tenth century is correct, becomes critically important to our understanding of early Chinese landscape painting….To come to terms with a work of art about which there are conflicting judgments, as there are with Riverbank, we must first examine the critical assumptions and methods on which these judgments are based. James Cahill, in a 1981 article 'Some Aspects of Tenth Century Painting as Seen in Three Recently Published Works,' written at a time when 'the real nature of tenth-century painting [was] a central enigma in Chinese painting studies,' analyzes three tenth-century works: A Chess Party in the Mountains…, by an unidentified artist, in the Liaoning Provincial Museum; The Lofty Scholar Liang Boluan…, attributed to Wei Xuan (active ca. 960-75), in the Palace Museum, Beijing; and A Flour Mill, also attributed to Wei Xuan, in the Shanghai Museum. He points out that '[the] new way of organizing the composition [in these paintings] represents an important advance over the older, additive space-cell mode in which movement from one discrete unit of space to the net was achieved only with difficulty; it opens the way to the spatial unification of Song [landscape] painting.' He adds that 'after the Yuan, this mode of representation seems to have been virtually forgotten, and the Ming-Qing understanding of Five Dynasties style is limited, largely, to the familiar motifs and mannerisms associated with Dong Yuan and Li Cheng.' Cahill thus follows the broad conceptualization of the history of Chinese art sketched out earlier by Max Loehr, who saw the fifteen hundred years between the Han dynasty (206 BC-AD 220) and the end of the Southern Song (1127-1279) as the 'phase of representational art in China, revealing a progress from images of isolated objects toward images of the visual phenomenon of pure space.' For Cahill, Riverbank fails to fit into his conception of this progression of representation art. 'The whole composition,' he writes, 'is full of spongy, ambiguous forms and spatial contradictions.' In Riverbank, what must be a distant river turns imperceptibly into a road with people walking on it. Furthermore, the whole upper part is obscure, neither the top of the main mountain nor the distant ones to the right being really readable - not because of damage or concealing mist (which would be anachronistic in so early a work) but because they were never clearly painting in.' For Cahill, Riverbank fails, to his eye, to be representationally convincing as a tenth-century work. As a critical concept, however, mimetic representation of nature, or naturalism, is of itself ahistorical. Because the rendering of nature changes in the course of history and in different cultures, the way a critic sees it today may not be the way the painter saw it. Therefore, our task is to place what we see in a historical context. There are scholars today who refuse to accept what they term a 'purely formal approach to style.'"

Wen Fong quotes Martin Powers citation of Wu Hung, "who also considers stylistic analysis ill-suited to dealing with problems of early Chinese art. Rooted as his work in the traditional Chinese antiquarian study of engraved inscriptions on ancient monuments, Wu challenges any analysis that interprets '[Chinese] stylistic evolution in terms of the accumulation of means to conquer the third dimension' as one that has 'taken the evolution of Western painting as a model.' He strongly objects to what he perceives as a Western 'Orientalist' approach which, 'either consciously or unconsciously, equates [Chinese stylistic development] with post-Renaissance painting that employs linear perspective as the most powerful means to create pictorial illusions.' The death of authenticated and dated monuments led early Western students of Chinese art mistakenly to follow Western analogies in their characterization of Chinese stylistic developments."

"In attempting to date a Chinese painting," Fong continues, "we must focus on defining period styles in Chinese painting history. At the same time, to avoid a teleological argument we must not assume that stylistic changes are predetermined. In truth, our approach to Chinese painting history through period styles, formulated on the basis on individual works that derive from stylistic permutations, mixed influences, and artistic creativity, can be conceived only with hindsight. I have long discussed with James Cahill the need for defining period styles in Chinese painting history. In describing style, modern art historians deal not only with techniques, form elements or motifs, and compositional patterns, which are earmarks of individual creations, but also with form relationships and visual structures (the way forms are visualized in painting), which are group or 'period' characteristics. In Chinese painting, traditional critics use form elements or motifs - for example, Dong Yuan's (active 930s-960s) round rocks and 'hemp-fiber' texture strokes - to define the basic idiom of an individual style. When a copyist or a forger imitates or appropriates an ancient style, he easily captures its basic form elements or motifs and compositional patterns, but in combining such elements or motifs to create an new effect or to find the solution to a new problem, he inevitably creates form relationships and visual structures more characteristic of his own time. Through an analysis of the form structure - the way forms are conceived and organized on the picture plane - we can demonstrate how pictorial conventions were fined and altered within a certain period and over time, either to create a closer approximation of perceived reality as transmitted or to express meaning through new techniques, emphasis, or distortion."

Fong's paper, and subsequent addendum, see below, devote considerable space and graphics to explaining and investigating planar features of spatial representation in some Chinese paintings. It is certainly an interesting exercise, but even if his examples are correct, if not brilliant, it is not without some problems when applied to issues of attribution. A key premise of Fong's argument is his belief that a copyist or forger "inevitably creates form relationships and visual structures more characteristic of his own time" than of earlier times he may be trying to represent. This may be true in some instances, but it is not necessarily so in all instances. Some writers try to give the impression that forgers are not all that competent and that many of their "successes" are due more to a particular culture's willingess to accept new "discoveries" than to take a careful look at the works in question which they, the writers, maintain are obviously "fakes" stylistically. Van Meegeren's Vermeers are often cited and while it is true that most of them appear quite crude and unpoetic despite their subject matter and Van Meegeren's incorporation of traditional Vermeer elements, there are many other works by other copiers/forgers of works by famous artists that occasionally not only pass muster in other generations but also are exceedingly well done. It is an overstatement to assume that all fakes are obvious to connoisseurs just as it is a serious mistake to assume that all artists are always consistent and at their best. A proliferation of doctoral candidates in art history has led to the occasional over-enthusiasms of some fledging "experts" to make a name for themselves by "exposing" fakes, that is, by often arguing that some works long accepted in a particular artist's "canon" should not longer be considered "authentic." These "experts," of course, may sometimes be right but they sometimes are not, especially as some of their arguments assume that there one or more brilliant forgers for every artist in the textbooks and even for artists not included in most textbooks and that there was a sufficient market for such forgeries when they were allegedly executed. This comment is not meant to imply that all works are suspect or that none should be, but simply to suggest that art history is not a science and that not all experts are equally talented and not all connoisseurs experts, etc. The art market, of course, wants as much "certainty" as possible and the careful examination of attributions is a very worthy pursuit. Labelitis, however, is a difficult disease to cure. It is hard to overcome cultural preferences for big names over small names even when sometimes the lesser-known artists are capable of producing very fine, if not better, works than the legends, on occasion. Consistency of style, materials, subjects, etc., surely helps reinforce one's certainty about specific works, but true connoisseurs will actually find aberrations, or inconsistencies, often more interesting, and more human. Most artists are actually human, not gods.

Most Westerners "new" to Eastern art definitely have a lot of agonizing in front of them to focus their "eyes" on its different culture. Some might suggest that these difficulties are insurmountable for many, especially those stepped in Western traditions, and presumably the reverse might be true, some Easterners might have trouble with Western art. Part of the debate over Riverbank and over Wen Fong's attributions touch on such sensibilities, an unspoken suggestion that perhaps only the Chinese can understand their own art. Such an interpretation may have some value as a generality, but Cahill and Sherman Lee are about as steeped in Chinese art as any Westerner could ever hope to be as it has been the primary intellectual focus of their lives for many decades.

Fong continues in his essay to note that "When Cahill compares Riverbank with works by the twentieth-century painter Zhang Daqian…, he points only to superficial form elements and motifs and compositional patterns." "He writes that 'Zhang was fond…of long, continuous, winding movements connecting near to far along zig-zag streaks of white that can be either roads or rivers - and it is often unclear which.' This describes Zhang's imitation of the early-Qing 'dragon-vein' compositional formula, which, in Zhang's hand, operates as a two-dimensional surface movement on the picture plane. In the case of Riverbank, however, as befits a far earlier landscape, the painter was concerned with achieving recession in space. Because the Chinese never developed an anatomical approach to figural representation or an approach to space based on linear perspective, the mastery of illusion in Chinese pictorial representation must be understood on its own terms, not on those of post-Renaissance Western painting. Over the centuries, Chinese painters mastered the modeling of forms and the representation of the illusion of depth and movement in space by developing a number of pictorial conventions that suggested three-dimensional space on the surface of the two-dimensional picture plane….Before the adoption in the eighteenth century of a Western-style linear perspective, with a single vanishing point, Chinese painters developed a technique of isometric perspective based on direct observation, using diagonal lines that are either parallel to each other or fan out, without converging toward a vanishing point. During the Six Dynasties period, from the third to the sixth century, painters experimented with both architectural and mountain motifs…to create space cells for figures to inhabit. By elevating the view to a bird's eye perspective, the painter was able to show the interior of a building as well as the broad panorama of a wide vista. Another device that evolved in early landscape painting was the use of overlapping triangles to suggest recession in space. Hawks and Ducks…, a painting on a lute plectrum guard dating to the eighth century in the Shoso-in Treasury, presents series of overlapping sloping mountains, with each sequence going in a single direction limited to three or four steps, after which the chain breaks, jumps to a higher level in the picture plane, and starts again. In this compartmentalized treatment of space, the foreground, the middle distance, and the far distance occupy three separate levels on the picture plane, each tilted at a different angle. This treatment is exemplified in one panel of a recently discovered wall painting of a six-panel landscape screen…from a tomb in Fuping County, Shaanxi Province, dating to about the eighth century. The steeply rising mountain view is depicted in three stages. In the first a mountain path is framed by overlapping slopes that lead rightward from the bottom of the picture. In the second the path turns in the middle distance, with sloping mountain forms stepping first to the left, then to the right. And in the third, toward the top of the painting, a river is framed by overlapping forms that twist again to the left. The landscape elements in Riverbank are organized in precisely the same way, with sloping curves from two sides of the mountain path forming a series of interlocking V-shaped pockets that twist and turn into distant space….Cahill's description of 'what must be a distant river [that] turns imperceptibly into a road with people walking on it' is a misreading of the scene. The passage in question repeats the compositional device seen in the Tang screen painting….The zigzag mountain path in the valley leads the eye upward into the distance; the path ends by the banks of the river, and the river extends the eye into the far distance as it winds its way through space. In both instances, the additive, compartmentalized, interlocking V-shaped space pockets reflect the conceptual approach seen in Tang and early Song landscape painting."

Fong's essay reproduces this wall panel and description of the river's winding twists and turns is accurate, although he neglects to mention that the top of the painting has clearly defined mountain tops beneath a substantial portion of sky in which is a very distinctive and interesting cloud formation, elements not present in Riverbank, an example of having perhaps too much cake or at least of demonstrating that selective comparisons are not always fully convincing, or, more important, relevant.

Fong continues that "In stating that 'concealing mist…would be anachronistic in so early a work,' Cahill ignores the well-known textual evidence that credits Dong Yuan with the invention of evening light and reflection in early Chinese landscape painting." "Describing the work of Dong Yuan and his follow Juran (active ca. 960-975), the Northern Song write Shen Gua (1030-1093) worte that 'paintings by Dong Yuan and Juran should best be seen at a distance, because their brushwork can be sketchy….If one examines [a work] close up, it does not seem remarkable, but if one looks at it from a distance one can vaguely see a village in the obscure depths of the picture's evening light, while the peaks of the distant cliffs seem to reflect the color [of the setting sun].' In the upper part of Riverbank…, while the main mountain peaks on the left disappearing into the mist 'seem to reflect the color [of the setting sun],' the peaks over the distant valley at the right, rendering in sketchy, blurry brushwork, dissolve artfully into the surrounding mist at their base."

Given Riverbank's less than pristine condition and general lack of color, it is difficult to find reflections now of the colors of a setting sun and the Shen Gua quote in Fong's essay does not credit Dong Yuan with inventing "evening light and reflection," but merely comments on its perceived presence.

"Some years ago," Fong continued, "I proposed that, between the eighth and fourteenth centuries, the mastery of spatial representation in Chinese landscape painting evolved in three stages. In the Tang and early Northern Song, from the eight to the early eleventh century, space in compartmentalized…" individual elements, each a discrete image viewed frontally, are seen one by one, and organized on an additive basis. By the end of the Northern Song, in the late eleventh and through the twelfth century, unconnected silhouetted forms are ranged continously through space, unified by mist or the atmosphere that surrounds them, as seen in Li shi's (Master Li; active ca. 1170) Dream Journey through the Xiao and Xiang Rivers….Finally, during the Yuan, in the late thirteenth and fourteenth centuries, the spatial integration of landscape elements is achieved through foreshortening and by the addition of a receding ground plane, as used by Zhao Menfu (1254-1322) in Autumn Colors on the Qiao and Hua Mountains….The forms are perceived as connected masses, the technique of fused brushwork suggesting forms seen through atmosphere. In Riverbank, the distant peaks shrouded in mist remain additive and compartmentalized. The mountain forms, built up with softly rubbed but tactilely exacting modeling strokes and a graded ink wash, diminish in size and clarity as they dissolve in surrounding mist. By the late eleventh century, as seen in Guo Xi's (ca. 1000-ca. 1090) Early Spring…, spatial continuity is suggested through a modeling technique that explores the impression of a diffused atmosphere. In Guo Xi's painting…, the thickening and thinning outlines of forms and modeling strokes, in values ranging from transparent blue gray to charcoal black, are applied one on top of the other so that the ink tones run and fuse together to create a wet, blurry surface. In drawing trees, Guo alternates between dark and light ink tones, effectively suggesting an enveloping veil of mist. Compared with the zigzag shoreline it he deep distance of Riverbank…, which are flat silhouettes piled up vertically on the picture plane, the distant shorelines o the left side of Early Spring…recede in a foreshortened perspective. Like Dong Yuan, Guo Xi was famous for depicting light in his paintings, with landscape forms emerging from and receding behind dense, wafting mists. Unlike European landscape painting since the seventeeth century, where light emanates form a fixed outside source and is an external organizing principle, light in both Riverbank and Early Spring emanates from within the forms, which are constructed as a harmony of alternating patterns of dark and light, texture strokes and ink wash, concavity and convexity. The absence of light as an external organizing principle in Chinese painting and the construction of landscape forms as a harmony of alternating patterns of brush sand ink and dark and light may be seen to reflect the Chinese view of the universe as characterized by spontaneous creation and alternations of phenomena without an external creator. When we reach the Ming and Qing periods, we find landscape paintings less concerned with the representation of ancient styles. In this exploration, as they turned increasingly to calligraphic brushwork, they focused on problems of surface organization and decorative pattern. In Landscape after Wang Meng…, by Wen Zhengming (1470-1559), dated 1555, in the National Palace Museum, Taipei, we see that the very complexity of details demands new organization through pattern and stylization. The piling up of elements on the vertical picture plane harks back to the archaic use of a bird's-eye perspective….It was a work that we date to the early Qing period, Myriad Ravines with Wind in the Pines…, with a typical early-Qing-dragon-vein composition, that the modern painter Zhang Daqian used as his compositional model in making a forgery of Juran's Dense Forests and Layered Peaks…, now in the British Museum. According to Fu Shen, Zhang used a photograph of Myriad Ravines, which he obtained in early 1951, as the model for the composition, while for the details he relied on his familiarity with two other well-known Juran attributions, Seeking Dao in the Autumn Mountains, in the National Palace Museum, Taipei, and Living in the Mountains, in the Saito Collection, Japan, both of which can be dated stylistically to the Ming period. Once the forgery has been exposed, Zhang's hand becomes embarrassingly evident. For all his dash and presence as a skilled painter, Zhang as a forger imitiating an early style can only interpret and transform ancient conventions and techniques through the visual structures of his own time. He uses such well-worn Juran brush idioms as round calligraphic hemp-fiber texture strokes and large round moss-dots, but he builds his forms as though from a copybook. The entire picture is flat and uninspired, a retreading of the early-Qing-dragon-veing composition. Compared with details in Myriad Ravines…, Zhang's brushwork in Dense Forests…is fluent but flat, vibrant but slick. He uses ink wash to create Western-style chiaroscuro, with light shining externally from one direction to enhance the overall effect, a technique that dates his painting to the twentieth century. Significantly, Zhang's modern forgery of Juran, who was a follower of Dong Yuan, has nothing to do with either the form or the content of Riverbank, a work we attribute to Dong Yuan."

It should be noted that the British Museum painting referred to above is included in the museum's current exhibition on the C. C. Wang paintings close to Riverbank. The Qing dynasty painting after which it was copied, however, is not included in the exhibition but is reproduced in the symposium book and is startling similar to Zhang's British Museum "Juran" painting. Zhang's "Juran" painting in the British Museum is a more satisfying composition in large part because of the treatment of the distant mountains near the top and the stronger emphasis on the double waterfall in the middle of the painting. It should be noted, furthermore, that the painting that Wen Fong and the Metropolitan attribute to Dong Yuan has a cascading double-waterfall that also is prominently highlighted in the left center of the composition, although Fong does not comment on the double waterfalls although his essay does reproduce details of them in the British and Shanghai museum paintings.

Fong proceeds to discuss Chinese art history methodology and writes that "Modern stylistic analysis stakes claim to universal validity by treating the formal characteristics of an image as solutions to generic problems of delineation, modeling, and composition." "The methods and techniques art historians employ to analyze art forms are independent of the culture or the period that produced those forms. It is no accident that the introduction of modern stylistic analysis coincided with the expansion of the history of visual arts to embrace non-traditional subjects including crafts, decorative art, and non-Western arts. Wu Hung, of course, correctly urges that we not misinterpret Chinese art by applying evolutionary theories based on the Western post-Renaissance model. Thus, it is all the more important that the stylistic development of Chinese painting be understood on its own terms, and that stylistic analysis as a scholarly discipline be expanded and applied to images from all non-Western cultures."

[Wu Hung is the author of Monumentality in Early Chinese Art and Architecture, published in 1995 by the Stanford University Press.]

Postmodern dialectics and rhetoric often tend to overcomplicate issues while presumably leading to a more sophisticated, historical understanding of cultural objects and their creation. Many such analyses are quite interesting and revealing, but they often tend to be rather obscurantist, convoluted and difficult exercises in stretched logic that sometimes results in placing greater emphasis on the interpreter than the artist, or, more importantly, the work of art in question. A work of art is primarily important as art, not as an historical curiosity. Most people care about art for its aesthetic values, not for its non-aesthetic baggage, if any. Academics, of course, have a variety of agendas and certainly the more the study and learn about a specific work of art the better we can appreciate both it and our reaction to it. There is no question that art-responders are burdened with considerable cultural baggage that indeed does reflect their times and not necessarily those of the artist, but when we are dealing with great art - the Mona Lisas of the world - we have upted to ante to preclude merely interesting art and focus on great art, which conceivably means art that transcends mere temporal value. Part of the reason why some people can quite a bit about the attributions of Chinese paintings at the Metropolitan is that rather grandiose claims have been made about them by Wen Fong and others. The museum, Wen Fong and his cohorts deserve great credit for bringing greater attention to Chinese paintings in New York but if there are problems with what they are highlighting those problems do not help the cause of multiculturalism, art history, and the proper role of cultural institutions in society and whether they should be held accountable for their public actions.

Fong writes that "one might note that Hegel's teleological conception of history survived among some early-twentieth-century Western writers on Chinese art, such as Ludwig Bachhofer, particularly the view that stylistic change manifested a change in zeitgeist, or collective spirit." "What can Chinese art tell us about Chinese history? If Riverbank is indeed a product of tenth-century China, it manifests the environment that shaped it, and as such it carries a significance beyond what historical records can convey. By imbedding the image in its original social, cultural, and material context, we may attempt to reconstruct the historical conditions that produced it. Between the fall of the Tang dynasty in 907 and the founding of the Northern Song in 960, during the period known as the Five Dynasties and the Ten Kingdoms, North China was ruled by a succession of ephemeral dynasties, none lasting more than sixteen years, while central and South China were partitioned among regional military commanders into more than a dozen coexisting kingdoms. The Southern Tang kingdom (937-75), which ruled much of the Lower Yangzi River delta region from its capital in Nanjing, was a prosperous and relatively stable haven from unrest. There, in the mid-tenth century, under the patronage of two rulings, Li Jing (r. 943-61) and Li Yu (r. 961-75), a renaissance of art and culture took place. The end of the Tang dynasty, in the early tenth century, is considered a major watershed in Chinese history. While the Tang had ruled its extensive empire by delegating power to regional military commissioners, the succeeding Song dynasty established a centralized bureaucracy staffed entirely by scholar-officials. The founder of the Song, Taizu (r. 960-75)…., was a military officer and had little interest in Confusian learning. Proclaimed emperor by soldiers under his command in 960, Taizu conquered the central Yangzi valley in 963, Sichuan in 965, and Guangzhou in 971, annexing all but two remaining independent states before he died in 976. The kingdom of the Southern Tang was founded by Li Sheng (r. 937-43), an adopted son of Xu Wen, a powerful general of the kingdom of Wu (902-37), whose throne Li Sheng eventually usurped. Claiming descent from the Tang emperor Xianzong (r. 806-20), Li Sheng considered his own royal house heirs to the glory of the Tang. Sheng's grandson Li Yu became the ruler of the Southern Tang in 961, was captured by the Song army in 975, and died in captivity three years later in the Song capital, Bianjing. In contrast to the martial Song emperor Taizu, Li Yu, meek and scholarly in demeanor, was known for excessive, luxurious living, and as a composer of lyric songs. The momentous shift in the tenth century from a society ruled by hereditary privilege, one that automatically conferred official status on sons of officials, to a society governed by a central bureaucracy of commoners was accomplished by the rise of a new scholar-official class selected through a system of civil examinations. As ruler of the Southern Tang, Li Yu, who was both a learned Confucian scholar and a devout Buddhist, embarked on a campaign to build schools and promote the civil examination system. After his ignominious surrender to the Song, Li Yu was severely criticized by Song scholar-officials for his indolent and licentious living, even as his talent as a lyric poet was grudgingly acknowledged….Blessed with material prosperity and having no real political or military ambitions, the ruling elite of the Southern Tang indulged in an indolent and irresponsible lifestyle, building extravagant palaces and filling them with vast numbers of women attendants, musicians, and entertainers."

Riverbank, Fong continues, "by contrast, is a painting about reclusive living, a subject that echoed ancient Confucian ideals. Reclusion, the deliberate non-participation in government affairs by scholar-officials, was considered a political act. After the fall of the Tang many scholar-officials, rejecting political involvement, retired to the mountains. The theme of reclusion was popular in both North and South China. It is exemplified in the anonymous A Chess Party in the Mountains…, dated to about 950-85, and The Lofty Scholar Liang Boluan…, by Wei Xian, a colleague of Dong Yuan in Nanjing. Lofty Scholar was the left-most panel of a six-panel screen that once decorated a Southern Tang palace hall. There is reason to believe that Riverbank, too, was the left panel of a horizontal screen composition, similarly executed for the Southern Tang court. The monochome landscape of Riverbank represents a new genre and a radical departure from Tang figural painting as well as the fully colored style of the Tang palace tradition continued in Zhou Wenju's In the Palace and Gu Hongzhong's The Night Entertainment of Han Xizai. Also, unlike monumental landscape compositions by such leading eleventh-century masters as Fan Kuan (active ca. 990-1030) and Guo Xi…, which are characterized by a heroic, epic style more representative of the imperial taste of the Northern Song emperors, Riverbank, a political landscape, is suffused with memory and imagination."

Fan Kuan's painting of a monumental moutainscape is perhaps the most famous picture in Chinese painting and manifests a drama and cohesion and artistic quality far beyond the merits of Riverbank.

"Indeed, the art and literature of the Southern Tang," Fong continued, "is characterized by a deep, melancholic longing for the past, and it is this melancholy that so dramatically pervades the majestic landscape of Riverbank. The passages of lush, moist southern Chinese landscape in the painting evoke, in their descriptive lyricism, the song-poem of Li Yu. A mountain path runs behind the pavilion into the hills, disappearing into a misty river valley in the deep distance….Wild geese fly into the darkening sky. And the journey continues beyond the distant rivers, leading far away to the sojourner's homeland in China's lost northern plains. Visual representation, far from being a transparent medium for capturing objective reality, involves choices with distinct ideological and political implications. Although Riverbank depicts the rich diversity of mountain and tree forms in nature, the rendering of nature in the painting does not replicate real landscape. Many of Dong Yuan's contemporaries painting and wrote about paintings like Riverbank. One of them was the landscape painter and theorist Jing Hao (ca. 870-930) who, writing about 'divining the emblems [xiang], or archetypes, of objects and capturing the 'truth' [in landscape painting],' believed that the painter could discover in nature the oral order that had been lost in the human world. The complex mountain forms in Riverbank echo Jing Hao's detailed description of 'mountain peaks with pointed tops…flat tops…round tops, and ranged with connected peaks,' and Jing's belief that 'the [forms] of the mountains and rivers are mutually generative, their breath forces causing one another to grow,' a belief reflecting the Neo-Confucian philosophy that defined learning and self-cultivation as 'the investigation of things leading to the perfection of knowledge.' There is in Riverbank a wide range of different species of trees - pines, cypresses, deciduous paulownias - all beautifully depicted as magnificent botanical specimens, 'each different from the other in form and character.' Pine trees, Jing Hao wrote, 'possess the virtuous air of a superior gentleman [junzi].'….Indeed, for the Neo-Confucian philosopher, just as the great pine was the lord of the mountains, the recluse scholar-artist was now the sovereign of the moral universe. This connection with morality served as the artist's personal confirmation of selfhood during a time of trouble. Unlike in Western tradition…, in which painting has been pursued as a science, Chinese paintings attempted to grasp the expressive nature of the image. A Chinese painter links his pictorial representation through calligraphic brushwork to his own physical self. In Riverbank the artist depicts a gathering storm gusting across the landscape with violently stretched trees and turbulent waves energized by an expressive calligraphic brushwork, brilliantly transforming his representation of reclusive life, metaphor for a safe haven in a threatening world, into an embodied image of the artist's self."

"If the foregoing shows the rewards of a historical approach to Riverbank," Fong continues, "the visual evidence of Riverbank offers an even greater contribution to our understanding of early Chinese art history. It is well known that figure painting as a genre peaked during the Tang dynasty, to be succeeded by the development of monumental landscape painting during the early Northern Song. Riverbank affords us a picture of the period between the two defining chapters of Chinese art history. The late Tang and the short-lived Southern Tang of ninth- and tenth-century China are generally perceived as constituting a period of political and social turmoil, of moral decadence and artistic decline. In Riverbank, however, we see an unprecedented engagement with complex natural forms that was never again matched in later landscape painting. The faces of the mountains, gouged and hollowed out by natural forces, are conceived as wholes, divisible from within, constructed in a harmony of alternating yin and yang patterns o solid and void, convex and concave; boulders of different shapes and descriptions, all spectacular but never unnatural or contrived in appearance, seem to grow out of each other, giving form to streams and pools of gushing water. The development of naturalistic representation and spatial realism in Riverbank signals the beginning of a period that witnessed the most brilliant scientific and technological achievements in Chinese history. Beginning in the late tenth century, the Neo-Confucian 'investigation of things' stimulated the objective transformation of nature even though the phrase does not imply a modern scientific method) and spawned, in addition to monumental landscape painting, many new, realistic genres such as architectural and flow-and-bird paintings. By 1300, with the emergence of literati landscape painting in the early Yuan period, there was a shift toward calligraphic self-expression, and the objective representation of nature never returned as a dominant theme in Chinese painting. In interpreting the visual evidence of a work of art, there can be no purely formal approach to style for we must deal with both the form and the expressive intent of the work. Although Riverbank inherits the compositional schemata, and therefore the visual structure, of a Tang landscape screen…, its profound conception of the landscape as a diagram of cosmic 'truth' has transformed the Tang composition into a new art form. Its patient engagement with a rich diversity of natural forms, painted with a sincerity and rigor unsurpassed even by Fan Kuan or Guo Xi, illuminates for us today the meaning and significance of the textual evidence of the early landscape treatises of Jing Hao and Guo Xi. By contrast, Zhang Daqian, a popular artist of early-twentieth-century China, painted only what he knew best. Like the renowed seventeenth-century painter Wang Hui (1632-1717) before him, Zhang, in imitating an ancient style, would first master the brush idiom associated with the style and then animate it with his own bravura brushwork, turning it into a dramatic extravaganza. If we compare his forgery of Juran's Dense Forests and Layered Peaks with Riverbank, for all Zhang's natural affinity for flamboyance and virtuoso brushwork, what he could not imitate was the ancient master's tireless and obsessive search for representational 'truth.'" Thus ends Wen Fong's essay in the symposium book.

"I have set forth all that I know about Riverbank in the book Along the Riverbank and the volume published for this symposium, Issues of authenticity in Chinese Painting.' Here, I would like to reply to Jim Cahill's queries about-or, more dramatically, 'Indictment' of the painting.

"In my published work, I have discussed how different new approaches to Chinese art have all but abandoned stylistic history and connoisseurship . Last month, at the James Haley Memorial Lecture in Princeton, Jim Cahill gave a compelling presentation of his "Thoughts on the History and Post-history of Chinese Painting." Referring to younger scholars who are on the whole disinclined to take part in the project of mapping out the sequence of stylistic development in Chinese art history, Cahill observed wittily that to discredit the task of assembling such a history before it has been accomplished would be like 'abandon[ing] the practice of architecture before we [have] built our city.' Our common task, he stated unequivocally, remains that of building and writing Chinese art history. That evening at Princeton, Jim Cahill graciously dedicated his lecture to me. I would like to return that honor by dedicating my remarks this afternoon to Professor Cahill. I do so in the name of the common enterprise he and I have shared all our lives that of building and writing Chinese art history. From the time we first met more than forty years ago, Jim and I have argued endlessly about one issue or another, always in the belief that it is through such continuous dialogue and the constant testing and verification of hypotheses that progress in our field of study is made.

"I will begin my reply to Cahill's queries about the authenticity of Riverbank with his teacher Max Loehr's conceptualization of Chinese art history, formulated first in 1964 and developed in 1970. Loehr proposed a periodization of post-Han Chinese painting, divided into three phases: (1) Representational art (Han through Southern Song, 206 B.C. to about A.D. 1300), when style served as an instrument to recover or capture reality; (2) Supra-representational art (Yuan, 1279-1368), when "the objective representation of Song was discarded for subjective expression"; and (3) Historically oriented art (Ming and Qing, 1368-1911), when "styles of the past began to function as subject matter.

"Over the years, Jim Cahill, drawing on Loehr's insights that Yuan painting had discarded objective representation for subjective expression and that in the arthistorically oriented art of Ming and Qing, "styles of the past began to function as subject matter," has described brilliantly, in book after book, how, in Ming and Qing paintings, the uses of style became the principal stratagem for bringing about radical new stylistic directions. He developed this theme in his essay entitled 'Chinese Painting: Innovation after 'Progress' Ends,' in the 1998 catalogue of the Guggenheim Museum exhibition China: 5000 Years. And Loehr's conceptualization remains central in Cahill's Princeton lecture, 'Thoughts on the History and Post-history of Chinese painting.'

"In my own book Images of the Mind (1984), I pointed out that Loehr's concept of Chinese art history, in which each phase is defined in terms of its "changing content," gives no historical stylistic description. Might it not be possible for us to reintegrate form and content by relating style to its motivating content, and thereby arrive at a historical account of the content of the form in Chinese painting? While form without content does not explain why style changes in history, Loehr's 'changing content' without stylistic description fails to account for how such changes are expressed in individual paintings.

"In Images of the Mind, I tried to give Loehr's concept of a representational art from the Han through the end of the Song a stylistic and art-historical description. Because of the traditional Chinese critical attitude against form-likeness (xingsi) in painting, some scholars are reluctant to use mimetic representation of nature as a criterion for judging Chinese paintings, Wu Hung, for example, has strongly objected to what he perceives as a Western 'Orientalist' approach to Chinese painting, which, "either consciously or unconsciously, equates Chinese [stylistic development] with post-Renaissance painting that employs linear perspective as the most powerful means to create pictorial illusions. Because Chinese painters never developed a scientific approach to space based on linear perspective with a single vanishing point (they adopted the Western-style linear perspective system in the eighteenth century), the mastery of illusion in Chinese pictorial representation, especially that during Max Loehr's representational phase of art, must be understood in terms of its own mechanics of visualization, that is, the way forms are conceived and organized on the picture plane.

"Over the centuries, Chinese painters mastered the modeling of forms and the representation of the illusion of depth and movement in space by developing a number of pictorial conventions that suggested three-dimensional space on the surface of the two-dimensional picture plane. I have proposed that between the eighth and the fourteenth centuries, the mastery of spatial representation in Chinese landscape painting evolved in three phases:

PHASE I

"In the Tang and early Northern Song, from the eighth to the early eleventh century, as seen in the eighth-century painting Hawks and Ducks…, the composition proceeds from front to back in three separate stakes. Individual elements, each a discrete image presented frontally without physical depth, are organized additively, motif by motif, in the vertical picture plane. Even juxtaposed elements do not physically connect or relate to one another.

"A simple device that evolved in early Chinese landscape painting…was the use of overlapping triangular mountain forms, neatly staggered in echelon, to suggest recession in space. I emphasize the word 'suggest,' because without foreshortening, there is no actual description of physical depth or a receding ground plane. In Phase 1, each sequence of overlapping triangular mountain forms is limited to three or four steps, after which the chain breaks, jumps to a higher level in the picture plane, and starts again. In this compartmentalized treatment of space, the foreground, the middle distance, and the far distance occupy three separate levels on the picture plane, each tilted at a different angle. Recession is thus suggested by a broken ground plane, but the lines or directions of recession never converge into a single vanishing point; they either diverge or parallel one another.

PHASE 11

"By the end of the Northern Song, in the late eleventh century, and through the twelfth century, as seen in Li shi's (active ca. 1170) Dream Journey through the Xiao and Xiang Rivers…, both the modeling and composition techniques change to accommodate and match more closely the natural vision that collects landscape forms in a unified spatial continuum. The modeling of the forms, relying heavily on the use of ink wash, softens to a generalized surface to suggest atmosphere; and overlapping triangular mountain forms are ranged from front to back in a continuous sequence, fanning out into the distance like a deck of playing cards…. The lines of recession parallel one another in what is sometimes called parallel perspective. Disconnected silhouetted mountain forms dissolve artfully into surrounding mist at their bases, which gives the impression of a unified vision without actual physical depth.

PHASE III

"Finally, during the Yuan, in the late thirteenth and fourteenth centuries, as seen in Zhao Mengfu's (1254-1322) Autumn Colors on the Qiao and Hua Mountains…, dated 1296, the landscape is depicted as a spatially integrated, physical environment. Foreshortening is discovered: the elements are no longer superimposed along the vertical picture plane; those far away are seen behind and through those in front. The ground plane recedes continuously, physically linking the landscape elements…. The forms are perceived as connected masses with actual physical depth, the technique of fused and blurred brushwork and ink wash, in mixed values and tones, suggesting forms seen through illusionistic atmosphere.

"In a recently discovered wall painting of a six-panel landscape screen in a tomb in Shaanxi Province, dating to the eighth century…, which shows the visual structure of Phase I, the steeply rising mountain view is depicted in three stages. In the first, a mountain path is framed by overlapping slopes that lead rightward from the bottom of the picture. In the second, the path turns in the middle distance, with sloping mountain forms stepping first to the left and then to the right. And in the third, toward the top of the painting, a river is framed by overlapping triangular mountain motifs that twist again to the left. Each of these triangular motifs is viewed frontally without physical depth. Recession is suggested here by a broken ground plane,. with each sequence of stepped mountain motifs tilted at a different angle; because they do not describe physical depth, the larger triangular motifs remain in the frontal picture plane.

"The landscape elements in Riverbank…are organized in precisely the same way, with sloping curves from two sides of the mountain path fronting a series of interlocking V-shaped space pockets, each viewed frontally without physical depth, that twist and turn into suggested distant space. Proceeding in three stages, the foreground elements lead first toward the right; in the middle distance, the mountain path turns to the left; and, at the top, a river winds its way into the far distance. But lacking the description of physical depth, the large sloping curves remain in the frontal picture plane.

"The relationship between the archaeologically discovered eighth-century landscape screen…and the tenth-century Riverbank may be characterized by Gombrich's formulation of artistic process as 'schema and correction.' The perception of the painter, which is reflected in the painting he creates, is the result of modifications, or corrections, of inherited schemata or conventions, and is thus both culturally and historically determined. In Riverbank, the spatial ambiguities that so confound the modem viewer are inevitable limitations of the Phase I visual structure, which uses additive, compartmentalized, interlocking V-shaped space pockets to suggest recession without actual physical depth. Without the use of foreshortening and a receding ground plane, the compartmentalized, V-shaped space pockets, by definition, remain in the frontal, two-dimensional picture plane. Cahill's description of 'how the river winding out of the distance in the upper part of the painting turns all but imperceptibly into a road with people walking on it'…is a misunderstanding of how the additive visual structure of a tenth-century painting should be read ('above' is to be read as 'behind'). The same visual ambiguity may be observed in Guo Xi's (ca. 1000-ca. 1090) Early Spring…dated 1072, which may be placed in Phase II of the development of spatial representation in Chinese landscape….Viewed objectively, the structural formations of the mountain masses simply do not work, but, as a landscape of the imagination, the spellbinding vision of the central peak surrounded by mist and subsidiary peaks soars upward on the picture plane like the writhing body and limbs of a great dragon.

"The rendering of nature in early Chinese landscape painting must be understood on its own terms, and not on those of post-Renaissance Western painting. This is not a matter of subjectivity in seeing, but a question of understanding the painting's visual structure, and the painter's own mechanics of visualization. I have long argued with Jim Cahill, a specialist in later Chinese painting, about the need for defining period styles in early Chinese painting history. In my review of his prize-winning book on Ming and Qing paintings, The Compelling Image (1986), I wrote: 'Cahill's new modes of analysis mask an unwillingness (or inability) to be specific about stylistic descriptions. He speaks of 'structures of meaning' and 'aesthetic'structures,' but never attempts a rigorous analysis of the form relationships and visual structures (the way forms are visualized and organized) in painting that could lead to a better understanding of the successive changing visual structures in history.'

"To avoid a teleological argument in my stylistic description of Riverbank, I propose to borrow George Kubler's concept of reconstructing stylistic sequences as "linked solutions," which in art history involve both form and content. There is reason to believe, as I have demonstrated in the book Along the Riverbank, that Riverbank was the left panel of a screen composition executed for the Southern Tang court. The tenth-century artist of Riverbank employs the simple schemata of the Tang landscape composition…to tell the story of reclusive living, using every narrative device available to him - scurrying figures, windblown trees, choppy waves, darkening sky, and wild geese flying - to dramatize his depiction. At the same time, he stretches the limits of the inherited schemata or convention to create a closer visual approximation of perceived reality. This emphasis on story telling, which Cahill finds out of place in an early landscape…, thus reflects the changing content of landscape painting, between the late Tang and the early Song, from the narrative to the expressive. By the eleventh century, the monumental landscapes of Fan Kuan (active ca. 990-1030) and Guo Xi expressed a symbolic cosmic vision of the natural and human world.

"The visual structure of Riverbank…, however, remains the same as that of the Tang wall painting. Individual bulging mountain boulders…, like those of other reliably dated tenth-century paintings…, are conceived frontally as overlapping triangles ranged in echelon. The exterior outline of a large boulder is conceived as a flat triangle, with its base sitting horizontally across the picture plane; while the interior structure of the boulder, which shows a series of overlapping triangles, suggests recession in space, the top of the boulder is inevitably folded back into the frontal picture plane. This bothers Cahill…, who writes: 'Often we cannot be sure whether one form is in front of another or behind it. For example, if we try to read the middle ground conglomerate of earth and what are presumably rocks, we are continually frustrated.'

"The boulders are shaded naturalistically, with a softly rubbed brushwork that blends with and disappears into a graded ink wash, without a clearly defined texture pattern. In his "Indictment," count 5, Cahill writes: '[There are] no traces of the brush having been put down and moved so as to leave distinct brushstrokes. Instead, the ink is rubbed onto the silk smoothly, without separate and visible strokes.' This, in fact, is a perfect description of the tenth-century modeling technique, which employs brushwork as a transparent medium for modeling forms before it developed into cun or texture patterns with formulaic brushstrokes. The mountain forms of Riverbank…capture the geological peculiarities of Jiangnan, or southern China, where rocky boulders are covered with rich earth. The complex shapes of the mountain, gouged and hollowed out by natural forces, are built in three distinct steps: a light, sensitive outline is used to delineate the rugged form; soft modeling "without separate and visible strokes" defines the creviced surface; and shaded ink wash is applied to fill in and round out the form. In a painting excavated from a tenth-century Liao tomb…, which shows the early cun texture pattern, individual brushstrokes remain relatively formless and flat. By the late eleventh century, Guo Xi's thickening-and-thinning texture strokes…and modeling ink wash are applied simultaneously and fused together to create wet, blurry atmospheric surfaces. By comparison, the forms of Riverbank and the Liao tomb painting are tactile rather than visually illusionistic. Once it occurred, this movement from the tactile to the visual was never reversed. No later landscape painter, however detailed and intricate his technique may have been, ever showed the same fascination for, and engagement with, the natural complexity of landscape forms as the artist of Riverbank.

"Finally, a word about the much remarked 'ambiguity' or 'unreadability' of the upper part of Riverbank…, which I have addressed in my essay in the symposium volume. Suffice it here to say that, representationally, the crest of the mountain peak on the left disappears into the surrounding mist in exactly the same way as the bottom of the distant peaks on the right dissolves into mist; the dissolving mountain form is described simply by a graded ink wash that fades into the surface of the silk. In the painting The Xiao and Xiang Rivers…, attributed to Dong Yuan (active 930s-60s), parts of the mountaintops are also hidden behind mist. In both cases, the treatment of mist is conceptually simple and uncomplicated when compared to that of Guo Xi's Early Spring, where tree and mountain forms are seen through patches of mist wafting across them.

"While Song painting took the representation of the objective world as its subject, Yuan painting marked the end of representation. The real subject of Yuan landscape painting is the artist's inner response to the world. In depictions of the theme of reclusive living from the fourteenth century, narrative elements all but disappear. For example, Ni Zan (1306-1374), in Empty Pavilion in Pine Grove..., dated 1354, painted only an empty pavilion. In Wang Meng's (ca. 1308-1385) Dwelling in Seclusion in Summer Mountains, dated also 1354..., which I take as clearly based on Riverbank (…my answer to Cahill's "Indictment,'..' count 12, that 'there is no Yuan painting based on Riverbank),' the Yuan master paints his own vision of mountain living. But in their visual structure, both Ni Zan's and Wang Meng's mountain forms, typical of landscape paintings of the fourteenth century (Phase 111), are illusionistically integrated three-dimensional masses that sit firmly along a continuous receding ground plane. In his famous Hermitage in the Qingbian Mountain of 1366…, Wang Meng expresses his dark vision of impending doom, when the struggle between the forces of the future Ming emperor Zhu Yuanzhang and his chief rival, Zhang Shicheng, was nearing its climatic end. Although we are not invited to wander through Wang Meng's landscape, its illusionistic visual structure shows integrated mountain masses moving from front to back along a receding ground plane.

"The visual structures of Ming, Qing, and modern Chinese paintings, even after Cahill's 'history' or 'progress' ends, can be precisely analyzed and described. They are not the same as those of the earlier periods. When a copyist or a forger appropriates an ancient style, he easily captures its form elements or motifs and large compositional pattern, but not its visual structure. In his forgery of Juran's (active ca. 960-975) Dense Forests and Layered Peak…, for example, Zhang Daqian (1899-1983), the twentieth-century artist, sees tenth-century landscape forms through the illusionistic visual structures of Huang Gongwang (1268-1354) and Wang Meng of the fourteenth century and the "dragon-vein' compositional formulas of Dong Qichang (1555-1636) and Wang Hui (1632-1717) of the seventeenth-century, and then models his forms using the Western-style chiaroscuro technique of the twentieth century. In using tenth-century elements or motifs to solve new pictorial problems (George Kubler's "linked solutions'), the forger inevitably creates form relationships and visual structures characteristics of his own time.

"Without a firm view of the changing visual structures in history, however, Cahill has real difficulty in distinguishing what he perceives as representational ambiguity, unreadability, and blurry brushwork in Riverbank ("Indictment," counts 1, 3, 4, and 5) from similar qualities in works by Zhang Daqian.I would like to explain why I believe these qualities are different.

"Compared to Riverbank…, Zhang's brushwork is no longer representationally descriptive; it is calligraphically abstract and formulaic. For Zhang, whose sole preoccupation was the styles of the past, painting has little narrative or expressive content, and his landscapes can be representationally faulty. The crucial difference between Zhang's work and Riverbank, however, lies in the illusionistic integration of Zhang's landscape forms through the use of foreshortening and a consistent, diagonally receding ground plane. In Zhang's work, broad brushstrokes, applied with mixed ink tones, are fused and blurred to create variegated illusionistic surfaces in the foliage and mountains. In his own paintings, such as Mountain Temple and Drifting Clouds, dated 1947..., the mountain forms are likewise seen as illusionistically integrated masses with foreshortened shoulders and sides, and their broad three-dimensional bases sit squarely on a continuous receding ground plane.

"One last note about Cahill's 'Indictment': In count 6, he compares the "animated landforms" in Riverbank…with an early seventeenth-century work, A River in Chu…by Dong Qichang. Dong's work is indeed the antecedent of Zhang's painting with Wang Hui's 'dragon-vein' compositional formula acting as the intermediate link between them. What separates Dong Qichang's work…from Riverbank…, however, is the illusionistic visual structure that has replaced the schemata of additive overlapping triangles of the earlier painting, the archaic frontality of which is so puzzling to the modern eye that it can never be recaptured by a modern forger.

"The stylistic features of the later Chinese paintings that belong to Cahill's post-historical or post-post-historical phases cannot be confused with those of the earlier periods. Between Riverbank…and Zhang Daqian…lies one thousand years of art history. To confuse the two may be a simple slip in connoisseurship, but it makes an enormous difference for art history."

Wen Fong's addendum, which has footnotes and illustrations not included here, ends here. In one of his footnotes, Wen Fong says that he has "no comment on Cahill's indictment count 13, 'the too good to be true' phenomenon,' and count 14, 'an alternative recent history.'"

While Wen Fong's essay and addendum provide interesting historical background, they do not really address many of the concerns of Cahill, Lee and Kohara, nor remove their very serious doubts about Riverbank.

In his concluding essay in the symposium book, "Connoisseurship: Seeing and Believing," Robert E. Harrist Jr., associate professor of the Department of Art and Archaeology at Columbia University and the moderator of the symposium's afternoon panel discussion, wrote that "connoisseurs standing in front of the same painting frequently describe very different perceptions, which are shaped by their own training and experience and, more elusively, by their temperament."

"Thus," he continued, "after careful study, Sherman Lee concludes that Riverbank…, attributed to Dong Yuan (active 930s-960s), is 'a modern pastiche all too familiar to many of us and unworthy of serious consideration by our serious colleagues.' James Cahill is convinced the painting is a forgery by the twentieth-century painter Zhang Daqian…. Hironobu Kohara also argues that the painting and a letter ostensibly by the modern painter Xu Belhong…that would support its authenticity are clever deceptions by that same master forger. Qi Gong states that 'the idea that Riverbank is a forgery made by Zhang Daqian is, of course, ludicrous, and not worthy of further discussion.' Wen Fong and Shih Shou-chien [professor, Graduate Institute of At History, National Taiwan University, one of the participants in the symposium] accept the painting as an authentic work of the tenth century, and Maxwell Hearn offers a carefully argued technical report to support an early date for the painting. Adequate explanation of the differences in these scholars' interpretation of the same painting would require extensive intellectual or psychological biographies of each, containing revelations that perhaps we can look forward to when they write their memoirs. For now, we cannot say why they see as they do; we can only try to make sure we understand what they are telling us. Obviously, not all of them are right…, conclusions about the attribution and dating of paintings must be either correct or incorrect, close or way off. Yet, barring the announcement of some sensational new archaeological discovery-let us say, a tenth-century tomb containing a close copy of Riverbank that also bore an inscription stating that it was based on a painting by Dong Yuan - or the deployment of some new technology establishing irrefutably the painting's date, the only information we have for settling the question of when Riverbank was made comes from what our eyes tell us and from what connoisseurs help us to see by translating their perceptions into words."

In his discussion of "Problems of Knowledge, Language, and Representation," Harrist writes that "The notion that there is no such thing as an innocent eye is a commonplace of art-historical discourse; but this loss of innocence results not only from what we have seen but also from what we have heard and read - what Stephen Little calls the 'catechism' produced by the mechanisms of academe, the museum, and the art market. Like a change in the fighting of a gallery, or a shift of vantage point produced as a spectator moves, the words of one viewer can change radically what another viewer perceives. Any who doubt the power of a master rhetorician to alter, at least temporarily, how one sees a familiar work of art should sample the confoundingly brilliant essay by Gombrich 'proving,' nonsensically, that one of the prophets on the Sistine Chapel ceiling was painted by Raphael rather than by Michelangelo. Just as a descriptive title or label shapes the way a viewer responds to a painting, statements by authorities as eminent and as eloquent as Sherman Lee or James Cahill or Wen Fong become part of the experience of seeing a painting. A viewer who believes himself independent in his judgments naturally resolves to 'make up his own mind,' testing the statements made by others against the evidence of his own eyes. But even for an assertive viewer trained to look critically at works of art, or at any other visual phenomena, sorting out what one actually sees from what one has been prompted to believe is an extraordinarily difficult mental feat. This is why Rudolf Arnheim has argued that 'once a work is suspected of being a fake, it becomes a different perceptual object.' Equally important, and sobering, is Arnheim's reminder that '[u]nder such conditions, even a bona fide original may exhibit suspicious features.' The task of the connoisseur is made more difficult by the notorious instability of the language in which works of art are discussed. As Michael Baxandall has pointed out, we do not see linearly, and 'the linear form of our discourse is curiously at odds with the form of its object, whether this is considered to be the work of art itself or our experience of it.' A cluster of especially slippery terms recurring in many contexts in this collection of papers refers to the representational goals of Chinese painters. The authors write of images that look 'natural,' 'unnatural,' or 'naturalistic'; they discuss 'representational flaws,' 'accuracy,' and 'verisimilitude.' As familiar as these terms may seem, their value depends entirely on the precision with which they are used. Fong zeroes in on this problem, stating that 'mimetic representation of nature, or naturalism, is of itself ahistorical. Because the rendering of nature changes in the course of history and in different cultures, the way a critic sees it today may not be the way the painter saw it. Therefore, our task is to place what we see in a historical context."

"The enterprise of sorting out originals from copies addressed by Stephen Little [Pritzker Curator of Asian Art, The Art Institute of Chicago, and another symposium participant] and Maxwell Hearn." Harrist continued, "is a matter of precise fine tuning in comparison with the problem of dating Riverbank. Simply put, Fong, Shih, and Hearn all compare Riverbank with works of the tenth century to argue that it dates from that period; Cahill uses some of the same comparisons to argue that it does not, while Lee and Kohara echo his conclusions. Do these radically different opinions about Riverbank arise solely from the idiosyncrasies of these connoisseurs, or is there something about the painting itself that breeds confusion? Does the real mystery of Riverbank derive not from its suspect provenance but from the approach to representation displayed in the painting? Although Fong and Shih argue that the brushwork and modeling of forms are consistent with works from the tenth century, most attempts to date Riverbank have focused on motifs in the painting and on its compositional structure. Thatched buildings overlooking the rippled surface of a river or take, foreground trees drawn with sharp contrasts of light and dark, misty views of distant scenery, and lively narrative detail in Riverbank do evoke memories of various paintings believed to date from the tenth century, including Zhao Gan's (active ca. 960-75) First Snow along the River…, widely accepted as an authentic work from around 960-70, and Wintry Groves and Layered Banks…, long believed to be a good example of Dong Yuan's style, though possibly a copy made after his lifetime. Beyond these easily spotted - and easily replicated - motifs, the pictorial structure and representation of landscape forms in Riverbank are far more difficult to place in the history of Chinese painting. Recognized works of tenth-century landscape vary so greatly that a skillful art historian could use them to argue for many different and conflicting visions of what painting in this period was really like, but one trait that artists of tenth-century paintings seemed to share was that of composing landscapes in ways that would guide the viewer's eye through pictorial space. This intention is apparent in A Chess Party in the Mountains…, a hanging scroll discovered in a tenth-century tomb in Liaoning Province. The ground plane at the bottom of the painting, framed by tree-clad embankments, is clearly defined by the placement of pine trees and three walking figures. The figures approach a passageway behind a middle-ground bluff that will lead them up to a pavilion framed by the central peak and by another mountain on the left. The spires of remote mountains loom in the distance. Although the two paintings have otherwise little in common, Wintry Groves and Layered Banks, like A Chess Party in the Mountains, displays a highly legible landscape composition. In this large hanging scroll, the eye moves easily from the foreground sandbar, to the earthen embankment and bamboo-sheltered house in the middle ground, then across the water to the hills and marshland above. Paths, bridges, buildings, and trees function not only as parts of the scenery but also as pictorial markers strategically placed by the artist to indicate the relative positions of landscape forms. The result is a fictive space governed by a carefully controlled structural logic. In Riverbank, as in other tenth-century landscape compositions, the artist arranges in an orderly sequence the foreground embankment, diagonally thrusting rocks on the other side of the river, and the buildings beyond. But the landscape itself erupts into a rolling field of visual ambiguities. Some of these pictorial effects are addressed by Sherman Lee and by James Cahill in several counts of his 'Indictment.' Consider in particular the area on the left of the composition around the small waterfall cascading into a sequence of ponds. The U-shaped rims of the ponds rise on the right to join an eroded bluff. The connection between these formations, as well as that between the fissures of the lower part of the bluff and the light-toned land on top of the bluff, challenges the viewer's eye. Are the lighter vertical passages to be read as receding, foreshortened shapes or as jagged contours parallel to the picture plane? Another strikingly ambiguous passage occurs in the upper right of the painting: is the towering peak a continuation of the land form below, bordered by the diagonal line of the road, or is it part of another formation rising from the point where the road meets the zigzagging river? One of the most confusing passages appears in the middle ground. The large rock rising upward diagonally toward the right, which partially blocks the view of the hermit's retreat, dissolves into a passage of pale ink in which the rock and the architecture seem to coexist in the same indeterminate space. Just to the left, setting off the lighter tone of a tree trunk, is a large area of dark ink: is this amorphous shape a rock, a dark cloud, or something else? Representational ambiguities certainly appear in tenth-century landscape paintings. A geographer or mapmaker attempting to diagram the topography depicted in the Liaoning hanging scroll would find this an impossibly frustrating task: like an architectural drawing that does not 'square up,' the placement of the landscape forms does not accord with the principles of geology. It is impossible, for example, to say exactly where the base of the central peak is located in relation to the middle-ground bluff or whether the small, flat-topped boulder in the lower right is parallel to the picture plane or tilted at an angle. But these ambiguities are counteracted by the overall structure of the painting, which is anchored visually by large, legible divisions of the landscape. Similar effects appear in the great monuments of Song landscape painting, such as Guo Xi's…Early Spring. The frequently noted disjunctions between areas of landscape in this work reveal the artist following his own advice (recorded in the Linquan gaozbl [Lofty Ambitions among Groves and Streams]),namely, masking the geological continuities between forms in order to increase the visual effects of spaciousness and height; but these ambiguities are framed and held in check by the bold silhouettes of the larger configurations of the landscape. It is the absence of unifying compositional devices in Riverbank that make the painting so different from works dated securely to the tenth and eleventh centuries. Passages that might have helped to structure the landscape, such as the paths in the upper right, actually increase its instability. In addition to the much discussed 'road that turns into a river' the central section of the path on which a figure in a straw raincoat walks is sealed by rock formations painted in such a way that it is impossible to tell which form overlaps which. Riverbank differs from the various tenth-century works with which it has been compared in that visual ambiguity seems to be a fundamental pictorial principle rather than an incidental element in the painting. Many of these effects are produced by the artist's use of light and dark in ways that obscure rather than clarify the shapes of landscape formations. In the faceted cliffs rising upward diagonally in the center of the painting, just above the hermit's retreat, patterns of light and dark cause the forms of the landscape to jump back and forth in space. Even more perplexing is the artist's handling of the relationship between the peak on the right and the mountains in the far distance. Because the tonalities of ink used for the distant mountain and foliage are exactly the same as those used for the peak, which should lie somewhere in the middle distance, the two planes seem to merge, collapsing pictorial space in this passage in a manner more typical of Ming than of Song painting. Contrast this passage in Riverbank with the painting of peaks in the upper left of Wei Xian's (active ca. 960-75) The Lofty Scbolar Liang Boluan…, dated to the tenth century, in which different tonalities of ink indicate the relative positions of the mountains. Even in a work from two centuries earlier, a painting on a lute plectrum guard in the Shoson-in…, the pictorial challenge of distinguishing near, middle, and far distance is addressed with greater clarity. This reading of the representation of landscape in Riverbank leads to several possible conclusions: I am misinterpreting what 1 see and the painting fits in the context of tenth-century landscape; I am correctly interpreting what I see and our understanding of tenth-century painting must be amended to incorporate the representational practices found in Riverbank; I am correctly interpreting what I see and Riverbank is not a tenth-century painting. The third conclusion, which I am inclined to accept, obviously raises a hot of new questions. If not tenth century, when? If not Dong Yuan, who? While the debate framed by these symposium papers would seem to force upon us a choice between dating the painting to the tenth century or the twentieth century, between judging it an early masterpiece or a modern forgery, future studies will surely explore the very great likelihood that the real origins of the painting lie somewhere else within the spacious continuum of one thousand years that separates the lifetime of Dong Yuan from that of Zhang Daqian. Awaiting such research, which could have enriched the symposium discussion, I believe that respectful agnosticism, more easily tolerated by the academic art historian than by the collector or curator, may be the most appropriate response to the mystery of Riverbank. If proof of the authenticity of Riverbank should come to light, how would this revelation change our understanding of the history of Chinese painting? As Shih Shou-chien notes, when compared with paintings that have been attributed to Dong Yuan for many centuries, Riverbank 'stands virtually alone.' Admitting the painting to the canon of tenth-century painting would require that we find some way to come to terms with the style of this new, anomalous work. But once that was settled, once the painting was assimilated into a new history of pictorial styles, what should we do next? After the authenticity of the painting had been proved, there would remain wide-open debate about its significance in the original cultural and art-historical context in which it was produced. This conclusion is apparent from a close reading of the papers by Fong and Shih, who accept the tenth century date. Both scholars address the meaning of the painting, and both associate the image of the man sitting in the pavilion with the theme of reclusion. Fong detects in Riverbank 'a deep, melancholic longing for the past' and interprets the waterside retreat as a safe haven in a threatening world, signified by a 'gathering storm gusting across the landscape with violently stretched trees and turbulent waves.' Shih's reading of the painting stresses its cozy domesticity and defines its thematic focus as the 'family life of a recluse.' The fantastic imagery of the landscape symbolizes, in Shih's view, not a melancholy or threatening domain but 'an idealized world within the hermit's mind.' Not only do their readings of the landscape differ in subtle but significant ways, but Fong and Shih also approach the issue of artistic creativity from different perspectives. Ultimately, in Fong's view, the painting, though produced for the Southern Tang court, is an 'embodied image of the artist's self.' Shih says nothing about the status of Riverbank as a work of artistic self-expression; he links the significance of Riverbank not to the artist's self but to the tastes of its patrons, Southern Tang rulers who were attracted to the idea of living as hermits. These different readings of Riverbank offered by Fong and Shih may not be as hard to reconcile as the assessments of the work's origins voiced by Fong and Cahill, but they do illustrate that if the question of authenticity could be resolved, interpretations of the painting would remain in flux….The papers in this volume have advanced the study of Chinese painting by examining a host of intriguing issues, but we are many Sundays away from resolving all the problems of connoisseurship that still confront us. In the face of these challenges, it is important to keep in mind that, although it would not make a banner headline or a juicy 'Talk of the Town' item, the real news in our notoriously difficult field may be that through the labors of countless scholars a consensus of opinion does exist concerning the attribution and dating of a huge corpus of paintings. It is the contested works that keep the field vital by forcing us all - believers, doubters, and agnostics - to look and argue more carefully and to ponder with fruitful self-reflection the foundations of the discipline to which we have devoted our lives."