CLP 188: 2005 “Pictorial Integrity: The Readable Image as Indicator of Authenticity in Chinese Painting”

CLP 188 (2005) “Pictorial Integrity: The Readable Image as Indicator of Authenticity in Chinese Painting”

(Note: this paper was included, along with another, in my article  "Chinese Art and Authenticity," in: Jason C. Kuo, ed., Perspectives on Connoisseurship of Chinese Painting (Washington, DC: New Academia Publishing, 2008) 33-66. There it is accompanied by illustrations, and so is easier to follow.  For the other paper on authenticity, see CLP 45 .)


Issues of authenticity continue to disturb Chinese painting studies, even while most in the field prefer, understandably, to turn their attention elsewhere. It seems worthwhile to set forth here, in general terms and for a wider readership, an argument that some of us have been making for a long time, but that still meets with surprisingly strong resistance. In its simplest form it goes like this: Good Chinese paintings are among other things good pictures; a painting with serious representational mistakes, when ascribed to a respected master or an early date, should be held in suspicion of being a copy or forgery. This is because where the original artist was depicting objects (in whatever style), the copyist or forger was replicating or imitating pre-existing artistic forms, and is likely, somewhere in the picture, to have misunderstood their representational intent and garbled them, rendering them unconvincing or even unreadable. We should be able to distinguish such garbling from expressive distortion, amateurish awkwardness, and other legitimate factors that work against realistic portrayal in genuine paintings. And we should recognize the irrelevance, in this context, of the old cliché about how Chinese artists "do not represent the outer appearances of things," and so forth. Purposeful departures from verisimilitude are not to be equated with pictorial blunders of the kind discussed here.

Most of what is quoted or summarized below has been printed already in some document, but several of these are not easily accessible to general readers. In citing uses of this method by others than myself, I am not attempting to judge whether they are right or not in their conclusions, but only considering the kind of argument they make and the criteria they use.

In a trial conducted in New York in 1956 in which the dealer Walter Hochstadter sued the collector-dealer C. C. Wang, alleging that Wang had knowingly given him bad paintings in an exchange they had carried out, Sherman Lee testified on behalf of Hochstadter, relating in particular why he had rejected a certain work (a handscroll representing "Zhuangzi's Butterfly Dream" allegedly by the Yuan master Liu Guandao) when Wang had offered it to his museum for purchase. These are excerpts from Lee's testimony, taken from the transcript of the trial:

P. 224: "In judging a painting, I first study the painting itself, and if it seems to stand up internally, then I go to the external evidence, and if that will hold water, then in most cases I am reasonably satisfied about the authenticity of the painting. If there may be rather peculiar things in the colophons or seals and the painting is good, that may have a slight effect on my judgment, but I have bought paintings which have definitely been fiddled with, in colophons or seals, but the paintings are in my opinion absolutely right. So I would say that the primary thing is the painting and the other material is secondary."

P. 235 (showing a slide): "in this, for example, this vase here, the thing is flat. There is no weight, no solidity . . . It is just a shape which is placed there flat on the surface ..."

P. 239: "Now, there is one particular point in this painting where I think ... we have what you might call a scientific proof of there being something wrong. " [He points out a place where the drawing of the wooden bench is, he believes, wrongly continued over what he takes to be a false, deliberately-made tear in the silk.] "And the result is he [the forger] made a mistake. Now, somewhere along the line these people usually do make a mistake. And this is just the kind of thing that I try to find because ... that is, I think, incontrovertible proof that this painting is a copy, or I would say, in my opinion, a forgery."

244: "This kind of thing, once my suspicions were aroused . . . then you start looking, because sooner or later you find a place where they just make--sometimes a subtle mistake--or sometimes a very crude mistake, as you have here."

Later in the trial (transcript, p. 596), C. C. Wang responds to Lee, saying that he himself doesn't look at paintings that way, and that seals are more important to him. "He [Lee] used the Western way," he says, "and not the Oriental way . . . It is not the way I learned in China, what was my own experience. It is entirely different . . . that is why he said some other paintings, he thinks . . .it is unbelievable." (I would read these partially and faultily transcribed words to mean: "Some other judgments Lee has made on this basis seem to me unsound, so I question the method.")

The way of reading and judging paintings described by Sherman Lee has been in use, then, for nearly half a century (and surely much longer, in unrecorded practice.) Why hasn't it been more generally adopted, when it is clearly so effective, sometimes even (as Lee claims, and I believe) decisive? An answer may be suggested by what Carlo Ginsberg writes about the "Morellian method" proposed in the 1870s by Giovanni Morelli, a way of analysis aimed not so much at determining authenticity as at identifying individual artists' hands in European, especially Italian, painting. We should examine, Morelli maintained (in Ginsberg's words), "the most trivial details that would have been influenced least by mannerisms of the artist's school: earlobes, fingernails, shapes of fingers and of toes." Although some of Morelli's new identifications of paintings were "sensational"--a picture that had been taken as a copy after Titian came to be recognized as "one of the very few authentic works by Giorgione"--Morelli's method, Ginsberg writes, was "heavily criticized, in part, perhaps, because of the almost arrogant certainty with which he applied it."  This raises a shocking possibility: can it be that Sherman Lee, with his "incontrovertible" and "scientific proof," and I myself with similarly positive claims, have been seen as--but no, perish the thought! Still, something other than academic disagreement must inspire the decades-long, heated opposition to this very reasonable procedure.

A two-day "Post-mortem Symposium" to reconsider some of the paintings that had appeared in the Chinese Art Treasures exhibition from the National Palace Museum in Taiwan, organized by myself and held in the auditorium of Asia House Gallery in New York on October 4th-5th, 1962, was attended by virtually all the major scholars in the field, along with many who were then graduate students--it was the first in a succession of grand gatherings enjoyed by our highly specialized community. Later I sent out to all who had participated a summarized transcript of the discussions (checked and corrected by the speakers), a little-known but crucial document in the history of our field. Landscape paintings occupied most of our attention, but near the end we turned to the hanging scroll "A Literary Gathering" attributed to Emperor Huizong (no. 31 in the Chinese Art Treasures catalog.)  

The first to offer opinions on it were Sherman Lee and Laurence Sickman, both of whom began by calling it an academy work, probably of Huizong's time. Lee compared it (with a slide) to a copy in handscroll form by Qiu Ying of the Ming, and observed that in the version ascribed to Huizong, "the folds of the garments have weight" and that the details show "observation of actual objects, e.g. in the table setting," whereas those in the Qiu Ying picture are "symbols rather than actually real." Sickman commented on "the clean structure of the table in the hanging scroll [Literary Gathering] which is confused in the handscroll {Qiu Ying.]" John Pope, a specialist in Chinese ceramics, then rose to claim that one of figures in the hanging scroll is holding "what can only be an early Ming blue-and-whiite dish," and the discussion turned to the identification of ceramics depicted in the painting--a telling criterion of age, obviously, since the painting cannot be older than the youngest identifiable and datable object in it. Later Alexander Soper, commenting on a different painting (attributed to Zhao Yan, "Eight Riders in Spring," catalog no. 11) and arguing for a post-Song date for it, said this: "The "Riders" balustrade can be fitted into an evolutionary sequence of architectural details as rendered by painters. Its dryness, flatness, and lack of reasonable articulation are typical late transformations of qualities that in Sung were still understood and appreciated."

These observations are directed, to be sure, toward somewhat different issues: a genuine painting vs. a forgery, an original vs. an honest copy, an early work vs. a later one. But the criteria for making the distinctions are more or less the same, depending as they do on whether or not things in the painting are represented with real understanding based on observation, or are secondary forms imitated or copied from earlier pictures.

The method is less applicable to landscape paintings, for obvious reasons: the elements of landscape, less fixed in form than figures and artifacts, are less susceptible to "wrong" representation. Nevertheless, when we find, for instance, tree groups in which the individual trees cannot be disentangled, or spatial anomalies that appear inadvertent, not deliberate, or a water surface pattern that does not adapt to the flow of the water, or a river that turns into a path with people walking on it--all these and more were pointed out by Sherman Lee and myself as pictorial afflictions in the "Riverbank" landscape attributed to the tenth century master Dong Yuan--when the picture has so much wrong with it, we have reason to deny the work an early date or great-master authorship, since good early artists were too skilled and too respectful of nature to commit blunders of that kind.

That the "Admonitions to the Court Ladies" scroll attributed to Gu Kaizhi (c. 345-406) exhibits visual anomalies indicating that it is a copy has been pointed out by a succession of writers during the century since its acquisition by the British Museum in 1903; Charles Mason, writing about the recent history of the scroll, states that those who believe it to be a copy typically cite "several passages in the scroll . . . where visual inconsistencies suggest the hand of someone imitating a design rather than painting a picture."  The most recent in this succession is Chen Pao-chen, whose sharp=eyed observations, delivered at a 2001 symposium on the "Admonitions" scroll, identify a number of "copyist's errors" in it, from the misunderstanding of multilayered garments to the misattachment of supports for the canopy of a palanquin.  And again, voices of angry opposition were heard from those who want to see the scroll as an original.

A similar examination with a similar outcome can be used, and has been. for the "Nymph of the Luo River" scroll also ascribed to Gu Kaizhi, even though the issue here is still another: a version in the Liaoning Museum believed to be an earlier and accurate copy of a lost original vs. two less faithful and later copies, in the Palace Museum, Beijing and the Freer Gallery of Art. (I am told that the late Wai-kam Ho delivered a paper at a symposium at the Liaoning Museum in November, 2004 arguing for a pre-Song dating for their scroll.) Out of many comparisons of details that could illustrate the differences, I offer the boats from the Liaoning and Freer scrolls (Figs. 6, 7). In the Liaoning picture, two flat bands of cloth hang over the canopy that covers the platform on top of the boat, their near ends blown slightly outward by wind, their far ends blown more strongly so as to almost touch the near side. In both the Beijing and Freer pictures, the near ends hang straight down; the far ends do the same, but are drawn as if they were hanging straight down from the near side above—negating, in effect, the depth of the canopy. The feature of the wind-blown bands, that is, readable in the Liaoning version, is misunderstood by the copyists. In the Liaoning and Beijing versions, thick wooden struts are laid at intervals on the outer deck, forming a kind of horizontal ladder against which the boatmen’s feet push (a feature still to be seen on poled boats today). The painter of the Freer version, with no understanding of this construction, draws parallel lines readable only as flat boards set into the deck, useless to the boatmen. In these, too, there can be no question of which most faithfully reproduces the lost original.

(A section of this paper that dealt with the different versions of the paintings ascribed to Shen Zhou and Du Jin is deleted here, since it largely duplicates a section in the American Academy paper. I ended this section by noting that after I had pointed out the “two-legged tripod” anomaly in the National Palace Museum version of the Du Jin composition, I wrote that I would like to say “Q.E.D.—that is, I have proved my case.”)

Another manifestation, admittedly, of Ginsberg's "almost arrogant certainty," with the predictable response: all three of my adversaries found ways (not the same ways, and in part mutually contradictory) to argue that the Palace Museum version is nonetheless, in some sense, a "real" Du Jin: hand of studio assistant, "cut-rate" version made within multiple production, etc. (I will not attempt to list or sum up the counter-arguments; seriously interested readers should consult the complete text.) We are still, that is, far short of agreement.

The buyer of such a painting in the artist's time would surely have complained about such representational flaws; connoisseurship based heavily on brushwork and professing to ignore readable imagery as a criterion of value belongs mostly to a later period and to the world of prestigious name-artist collecting. As all readers of Song-period writings on painting know, getting the image "right" was a major criterion at that time for identifying the best artists and pictures. In Japan, where imagery continued to be a major concern and brushwork in the Chinese sense was only imperfectly understood, kanteika or authenticators would often make sketch-copies from works they were called on to judge, in part to catch representational infelicities. And even within the orthodox tradition of connoisseurship in the later periods in China, no uniformity of practice can be assumed; there must always have been those who wanted, among other things, a good picture, even at the risk of being derided by their fellows for displaying such a philistine taste. I do not even believe, after spending many hours looking at paintings with the late C. C. Wang, paragon of traditional Chinese connoisseurship, that he truly was inattentive to the "scenery" (his dismissive term for representational content) in the paintings he judged.

And all of us, to the extent we are able, apply to the painting our understandings of individual style and period style, to see how the painting fits these, besides taking into account what Lee calls "collateral evidence"--seals, inscriptions, signatures, provenance, etc. The recognition of what I call "pictorial integrity," a quality I take to be characteristic of good paintings of all periods and styles, does not replace these additional factors but augments them, permitting us to make judgments with more assurance, on solider ground.

Attendees at the 1962 "Post-mortem" symposium referred to above were dismayed to hear two leading authorities, Max Loehr and Alexander Soper, diverging by a millenium in their dating of the first painting considered, "Emperor Minghuang's Flight to Shu" (Catalog No. 1), Loehr making it eighth century, Soper eighteenth. At that time most of us were confident that this was only a symptom of the fledgling state of our field, and that we would reach a greater degree of consensus in years to come. But the "Riverbank" controversy of 1999, and the Gu Kaizhi symposium of 2001, have both revealed gaps as wide, or nearly so, in datings proposed by major scholars who took part. Either gap could be largely closed, I think, by a more widespread adoption of the method advocated here. One may hope that specialists of the next generation, looking beyond arrogance, will be willing to do that.

• < Prev
• Next >