DOWNLOAD THIS PAPER (PDF 180KB)
There is little doubt that there are forgeries among the Dunhuang and Central Asian collections although there is little objective proof yet as to their extent. The problem was first made public by Professor Fujieda Akira following his comparison of the Japanese and British collections. In 1998 IDP organised a conference at the British Library to discuss this issue and the papers from the conference were published in 2002 (click here to order).The introduction to this collection is reproduced here.
'if only it were sexier, it might have rated recognition as the world's oldest profession.
Ever since humankind became literate, civilization has been bedevilled by the forger's determination to deceive by mimicking the writing of others.'
Forgery might not be as sexy as sex — after all, what is? — but it comes close. It is a story with all the elements to grip the popular imagination: greed, large sums of money, deceit, sometime violence and, not least, the ability of the ordinary man to bamboozle the greatest expert or most lofty institution. The story of Central Asian Dunhuang manuscript forgeries contains all these elements and, despite almost nine decades having passed since the forgeries started to be produced, the story is yet to be concluded. Whether this is due more to curators' complacency or the forgers' skill is a matter of debate, but the fact remains that we still cannot say with any certainty whether or not there are large number of forgeries among the Dunhuang manuscripts now in collections world-wide, let alone give a foolproof method of detecting them or explain fully how and by whom they were made.
Some contributors to this collection argue that this is overly cautious, that they can distinguish forgery without any doubt, and that it is certain that most of the manuscripts collected after the early expeditions, namely those in St. Petersburg, Japan and a portion of the London collection are forged. Other contributors argue, just as vociferously, that most of these manuscripts are genuine. The scientists would dismiss both claims as subjective and therefore unverifiable and turn to the need for 'objective' testing before proof can be claimed either way. These differences reveal just how little of the story has yet been told and how far there is to go. The purpose of the conference held in June 1997 and these resulting papers, therefore, is not to make decisions but simply to open the debate. For the conference was the first public discussion of this issue. This brings us back to the issue of why it has taken this particular story so long to be told, especially when there was a clear precedent for Central Asian forgeries, the forger himself being unmasked in the decade that the Dunhuang manuscripts were initially dispersed abroad. But the story starts two millennium before this, when the great Central Asian trade routes, later termed 'the Silk Road' started to flourish.
Chinese Central Asia, the area now known as the Chinese provinces of Xinjiang and Gansu, was a key link in the trade routes between China, India, and the Middle East in the first millennium AD. At the height of its prosperity, merchants travelled between thriving city states and smaller oasis towns, villages and settlements. The inhabitants were diverse: Sogdians, Uighurs, Tibetans, Indians, Kucheans, Arabs, Chinese, and others. All these peoples had developed their own written languages, sometimes using existing scripts and sometimes inventing new ones. The Sogdians, for example, used the Aramaic script which was later adopted by the Uighurs to transcribe their language. The Tanguts, a Mongolian peoples who spoke a language belonging to the Tibeto-Burman group, came into the region from the north-east in the eleventh century and subsequently used Chinese as the basis for a Tangut script.
It was not only the administrative and military structures of the empires surrounding this region and its several kingdoms which required language to function, language was also needed to maintain the complex economic structure and several large and well-organised religious communities, not least the Buddhist church. Buddhism had spread from northern India with the development of the trade routes and was dominant in most of the route's city states, such as Kucha and Khotan, as well as in the Chinese empire itself by the seventh century. It was also making inroads in the young Tibetan empire. Its monks and lay adherent travelled throughout the region, either bringing scriptures from India or travelling to India in search of new texts. Texts were translated into the various Silk Road languages. In China, translators and whole teams of copy-editors, proof-readers and editors were employed by imperial command to make accurate translations from the Sanskrit and Prakrit originals, which texts were then distributed to key monasteries in all the regions of China.
Paper had been invented around the Christian era in China and it became the dominant medium for transcribing texts in the first few centuries of the first millennium, replacing wooden strips and silk. The format of the Chinese paper book initially emulated that made of wooden strips, which had been tied together to form a continuous text reading top to bottom, right to left, and then rolled for storage. Paper panels, their size dictated by the paper moulds, were pasted together to form long scrolls. This was the dominant book form for several centuries in China. In the latter part of the millennium, new forms start to appear which evolved into the codex. In other parts of Asia different formats were common. In northern India, for example, and subsequently in Tibet, the pothi ubiquitous. This was a 'book' composed of wide thin, made from palm leaf, joined together by string loosely threaded though holes in the middle of the text so that each sheet, once read, could be laid down flat, and sandwiched between wooden boards.
In central China, fine paper was made for important documents primarily from hemp and mulberry and in standard sizes. By the seventh century, paper-making was a highly refined art. The paper was sized to take ink and dyed to protect against insect and water damage. The ink used was similar to that known today by the misnomer of 'India ink' (it should really be 'Chinese ink) and, once dry was waterproof. Animal-hair brushes were used for writing. There were other centres of paper production, including one at Khotan on the southern part of the Silk Road which survives today. These produced coarser paper made from locally-available fibres. In southern China, bamboo started to be used for paper production.
The majority of the Dunhuang manuscripts in question are in Chinese, with the second largest group in Tibetan and smaller groups in Uighur and other languages. This reflects the political situation of Dunhuang during the period of production. The Chinese empire established garrisons and appointed Chinese officials controlled the routes for much of the period. Their control was constantly challenged by neighbouring empires, including the Arab empire which met with the Chinese army in battle in 751 at the Talas River. In 755 China was plunged into civil war by the rebellion of one of its generals and the emperor was forced to recall all troops posted in the western regions to fight the rebels. The rebellion took almost a decade to quell and, by the end of that period, the Tibetan army had moved into Central Asia and had taken control of the major towns, including Dunhuang. The Tibetans ruled until 848 when they were driven out by a Chinese loyalist. After this, the region, although professing allegiance to the various dynasties that ruled north China once the Tang dynasty fell in 907, was de facto independent. In the tenth century the Uighurs started to have an influence, already controlling a region to the east of Dunhuang. The following century saw the arrival of the Tanguts, but by this time, it is believed, the Dunhuang library cave had been sealed.
The story of the collection of the manuscripts, their placement in the library cave, and its subsequent sealing is full of lacunae and doubts. We know from the manuscripts and from Chinese records that there were over a dozen monastic establishments in and near the town of Dunhuang. At least one of these was at a site called Mogao in Chinese (probably a transcription of a name in local language), a low cliff stretching for over a mile along a small river south-east of the town. Legend tells that a monk had come upon this place in 366 AD and determined to excavate a cave for meditation, to take advantage of the remoteness and beauty of the spot. He was followed by another monk, and in the decades following the Dunhuang rulers and populace started to take an interest and more caves were built, some of them for meditation but, increasingly, more and more as small chapels. There was no local supply of good stone — the cliff itself was formed of a friable conglomerate — and so statues of Buddha, bodhisattvas and monks were made from clay and straw built around a wooden frame. The walls of the caves and the statues were painted using a variety of mainly mineral pigments.
Buddhist monasteries had libraries containing the sacred texts of their faith. The number of texts proliferated as Indian monks brought new texts from India and Chinese monks, such as Faxian in the fifth century, and Xuanzang in the eighth century, travelled to India to collect missing texts. These were translated, sometimes several times over the centuries and as understanding of Buddhist concepts and terminology developed the translations grew better. Xuanzang himself, on his return from India, devoted the remainder of his life to translating some of the texts he had brought back. Monasteries were always striving to fill gaps in their libraries' holdings. The major monasteries in each part of China received imperial copies of texts, made on the finest yellow-dyed paper. Other monasteries had to make do with second-hand scrolls picked up from wherever they could be found. Monks were sent to search for texts in other localities and scribes were paid to make copies.
During Tibetan rule in Dunhuang the direct trade route to China was severed and it was therefore no longer possible to obtain supplies of the fine yellow paper or writing brushes. Indeed, there were periods even before this where the materials were obviously in short supply as revealed by the colophon to one manuscript from Dunhuang now in the British Library. But during the Tibetan period it appears that the shortage was so severe that Chinese administrative documents and other texts (many of which would have been stored in the library of the local government offices), were recycled by the monasteries for copying sutras on the blank verso side. It also appears that locally-made paper was also used more than previously for Buddhist texts.
The head of the Buddhists in this region was a monk called Hongbian and he died not long after the Chinese retook control of Dunhuang. He had probably been instrumental in building Cave 16, a largish cave about halfway along the cliff on ground level. Cave 17, a small cave leading off the right-hand side of Cave 16 entrance corridor, is surmised to have been built after his death as a memorial chapel. A statute of Hongbian was placed against the main wall and flanked by a mural showing two disciples. A stele, made fromstone which must have been imported some distance, recorded his achievements in making the Buddhist faith flourish in Dunhuang.
This was the 860s. All we know now is that sometime after this, perhaps not until the beginning of the eleventh century, perhaps before, the statue was removed into another cave and Cave 17 was filled from floor to ceiling with paintings on silk, manuscripts and printed documents. Some of these are dated and, since the latest date is 10++, it has been assumed that the cave was closed soon after this time. There are several unanswered questions. Firstly, why the manuscripts were placed in the cave. Second, whether the latest manuscript is indeed eleventh century or later and thus, when the cave was sealed. Thirdly, why it was sealed (which may or may not be related to the first question). The first and third questions do not have direct relevance on the question of forgeries, but the second is of paramount importance. I will return to it later.
The area came for a while under Tangut control until the Mongols defeated them and imposed their command over Chinese and Central Asia, extending to the borders of Europe. The overland Silk Road fell into decline and merchants found it safer and cheaper to use the maritime route into China. With the fall of the Mongols (1368) the new Chinese rulers soon abandoned their claims on central Asia being no longer able to control the region, and in 13++ they withdrew to east of Jiayuguan, a border fort east of Dunhuang. There was little activity in the region over this period. The suceeding Manchu Qing dynasty (1644–1911), eager for territorial expansion, again looked west and re-colonised the area west of Jiayuguan naming it 'Xinjiang' or 'New Territories'. Soon after other imperial powers started to realize the strategic importance of central Asia and the Great Game saw Russian, British and Japanese spies. They were followed by adventurers and archaeologists, and the story reaches another crucial juncture.
By the last decade of the nineteenth century China still nominally controlled the area through its appointed officials, but there was growing insurgency both here and in the centre of China. The Chinese government issued travel documents to a whole series of foreign archaeological expeditions, allowing them to excavate objects in the sand-covered desert ruins and to take them out of the country. An Hungarian expedition had visited Dunhuang in 18++ before Cave 17 was discovered, and had told a fellow Hungarian, Marc Aurel Stein, about it on their return. Stein was working for the government of India having taken British nationality, and he was keen to explore the ruins of the Silk Road for himself. At the end of May 1900 he embarked on his first expedition. But he did not travel as far east as Dunhuang: he was an Indologist and more interested in sites to the west which showed Indian influence and yielded Sanskrit and other documents. He was, however, interested in forgeries.
The desert sands had already yielded many manuscripts and even printed documents, which, in these early years, were mainly excavated by locals and offered for sale. When a man called Islam Akhun therefore approached the British Consul resident in Kashgar, George Macartney, in 1896 with a batch of material for sale, it was nothing unusual. He explained in detail where it had been found and Macartney bought it. More was to follow and similar material was offered, at the same time, to the Russian Consul in Kashgar (he and Macartney were not on friendly terms). The material, consisting of both manuscript and printed documents, was sent respectively to Britain and Russia where scholars were invited to examine it. A.F.R. Hoernle, a respected scholar of Indo-aryan languages, examined the British material. His first report was published in 1897. In 1899 he published a fuller report entitled 'Antiquities from Central Asia', with a detailed review of 45 books of block prints, bound in a European style with metal pegs and employing, he writes not less than nine different scripts
As to the language, or perhaps the number of languages, hid in these scripts...
, he continues, it is impossible to venture to express any opinion, before some advance has been made in their decipherment.
What an opportunity for a scholar, the possibility of deciphering scripts and perhaps even discovering a language or languages previously unknown to western scholarship. In his early work on forgeries, Genuine and False, Hans Tietze identifies two facets of human psychology which aid the forger: namely, lack of focus and an eagerness to make an important discovery. Hoernle exemplifies the latter. He was certainly not unfocussed in his examination of the documents and even pointed out certain 'abnormal' features, such as the lack of a system in the order of the books. But his eagerness overcame not only his own doubts, but those of others. In the 1899 article he devotes a section to 'The Question of Genuineness', writing:
'Considering the abundance of the block-prints and the mystery of their scripts, it is not surprising that the suspicion of forgery should suggest itself. It suggested itself to me at an early stage of my acquaintance with the Khotanese books; and I am informed that it has also suggested itself to some of the British Museum authorities and others. But it was not until the summer of 1898 that the suspicion took a more definite shape in a letter, dated the 29th June, 1989, which I received from Mr Bäcklund, Swedish Missionary in Kashghar, in response by me for information on the subject; for at one time, in the course of my examination of the block prints my suspicion had been much strengthened by the observation of the extreme want of order in certain books. This result was subsequently neutralised by the observation of the striking consistency of order in other books. It became clear that, as I have already shown, both phenomena are quite compatible with a general genuineness of the block-print books; and in fact, all the evidence that gradually accumulated has tended to confirm that conclusion.'
He then quotes Mr Bäcklund's letter in full, a model of clear and reasoned argument which, to the modern reader, is most persuasive. The letter reports oral evidence from the missionary's servant and others of hearing from a third party of how the books were made by Islam Akhun and his colleagues, and also enumerates his own suspicions based on the books' lack of resemblance to other old books, the thin layer of rust, strong paper, no marks of handling, paper similar to modern Khotanese paper etc.
Hoernle proceeds to rebut each of Bäcklund's points in turn. The arguments are too persuasive to refute and he therefore resorts to the explanation that there are both genuine and forged items, hence the discrepancies, the forgeries being 'duplicates of genuine books that have been discovered', the 'preparation of a duplicate ..well within the capabilities of a modern Khotanese forger'. Other facets of human psychology reveal themselves in the course of his rebuttal noticeably his underestimation of human ingenuity and greed; 'but the hypothesis that he is capable of inventing not only one but several scripts, and of intricate, but self-consistent systems of their arrangement in books, and finally of binding them after a method, quite unknown in Khotan at the present day, contains more elements of improbability than the hypothesis of the genuineness of the books.' But this, of course, is what he wanted to believe.
He ends:
'To sum up, the conclusion to which, with the present information, I have come, is that the scripts are genuine; and that most, if not all, of the block-prints in the Collection also are genuine antiquities; and that if any are forgeries, they can only be duplicates of others which are genuine.'
Three years later Stein was in Khotan on the track of Islam Akhun, his own doubts not having been assuaged by Hoernle's arguments but instead reinforced by his first year's experience of exploration: 'The grave suspicions which these local enquries had led me to entertain sbout the genuineness of all these 'old books; in 'unknown characters' had been strengthened almost to certainty by my exploration of the winter. Ample as were the manuscript materials which the latter had yielded, and in spite of the great variety of languages and scripts represented among them ... I had failed to trace even the smallest scrap of writing in 'unknown characters. The actual conditions of the sites explored also differed entirely from the conditions under which these queer 'old books' were alleged to have been discovered.' Islam Akhun was elusive but on April 25, 1901, was brought in front of Stein.
'The examination of this versatile individual proved a protracted affair, and through two old days I felt as if breathing the atmosphere of an Indian judicial court ... in the matter of the 'old books', [Islam Akhun] for a long time protested complete innocence.'
However, Stein evetually managed to trick him into contradicting information he had given several years before and
'he was intelligent enough to realize ... that nothing was to be gained by further protestations of innocence ...[and] made a clean breast of it.'
Islam Akhun then provided Stein with a detailed descrption of how the 'old books' were made. He described how he and his accomplices started by copying cursive Brahmi from genuine fragments but 'Islam Akhun quickly perceived that his 'books' were readily paid for, though none of the Europeans who bought them could read their characters or distinguish them from ancient scripts, [so] it became unnecessary to trouble about imitating the characters of genuine fragments. Thus, apparently each individual factory 'hand' was given free scope for inventing his 'unknown characters'.' When the rate of production could not meet demand, the forgers turned to techonology — block-printing — to increase their profit.
Hoernle's dismissal of the forgery argument on the grounds that:
'How can Islam Akhun and his comparatively illiterate confederates be credited with the no mean ingenuity necessary for excogitating [the scripts]?'
was shown to be flawed. Hoernle had made the assumption that the squiggles on the papers were meaningful. Assuming that they were meaningful, he had sought to find patterns which, inevitably, had then emerged: it is in human nature to impose pattern on randomness. As Mark Jones notes:
'Even the most academically and intuitively gifted of all individuals, even the most rigorously organised of institutions, can and will occasionally be wrong. And this is not, or not simply, because knowledge and experience can never be complete, but because perception itself is determined by the structure of expectations that underpin it.'
Stein concluded, in contrast to Hoernle, that
'Islam Akhun was a man of exceptional intelligence for these parts, and also possessed of a quick wit and humour.'
Cave 17 at Dunhuang was probably discovered in June 1900, the same year that Stein embarked on his first expedition. When the entranceway was opened, a cache of tens of thousands of manuscripts, documents and paintings was discovered, stacked from floor to ceiling. There are no contemporary records of the find and the earliest photographs are those of Stein and Pelliot several years later. This is a great pity, as the order of the cave might yield important information as to the nature and importance of the store. The self-appointed guardian of the caves, an itinerant Daoist monk, Wang Yuanlu, presented a few manuscripts and paintings from the cave to local officials, but the majority of the them probably remained in situ, largely undisturbed.
When Stein set out on his second expedition in 1906 he heard rumours of the find and reached Dunhuang in May 1907. A full account of how he persuaded Wang Yuanlu to part with many cartloads of manuscripts for a small sum is given in his expedition report. Many of the manuscripts were brought to him by Wang Yuanlu and he did not make a systematic examination of cave 17. In 1908 when the great French sinologist, Paul Pelliot, arrived, he used the space vacated by the manuscripts taken by Stein to perch on a small stool while he looked through all the remaining items in the cave by candlelight. He took several thousand more, mainly composed of non-Buddhist texts. The manuscripts in question are now in the British Library and the Bibliothèque nationale respectively.
Pelliot left China via the capital, where he showed scholars some of the manuscripts. The government, already aware of the find, were prompted into taking action and Fu Baoshu, an official in the Ministry of Education, was dispatched to Dunhuang with the order to transport all the manuscripts left in cave 17 back to the Ministry for safekeeping. They arrived in 1910. In the same year Fu Baoshu was arrested. Most scholars have accepted the version of these events given by Luo Zhenyu in 1927 (and reproduced in Lou Kamtong and Cheng A-tsai's papers), in which he claimed that a famous bibliophile, Li Shengduo, and several others arranged to steal manuscripts while they were in transit. However, Rong Xinjiang challenges this and makes excellent use of contemporary historical sources in Peking University library to argue that the manuscripts reached Beijing intact. However, he continues, there is evidence to suggest that Li Shengduo arranged the theft later, when the manuscripts were already in the Ministry of Education (Li Shengduo was a high official in the Ministry) and after they had been seen by visiting Japanese scholars. In his challenge, Rong Xinjiang provides a commendable antidote to another failing of human psychology, the acceptance of authority. As he points out,
'Because Luo [Zhenyu] is the most important scholar of Dunhuang studies in the early twentieth century, ...most have accepted without question what he said about the story of the stolen manuscripts.'
If the story of the forgeries is to be completed, then every assumption made about the manuscripts must be challenged and verified by sound evidence, as Rong Xinjiang does here.
A catalogue of Li's collection made in 1935 and published in full in 1958 listed 432 Dunhuang manuscripts (see Appendix 1 to Rong Xinjiang's paper). The present whereabouts of these manuscripts is another matter of debate. Rong Xinjiang again argues that he has found no evidence to confirm, as has been suggested, that some at least are in Taiwan in the National Central Library , as is argued by Lou Kamtong and Cheng A-tsai. He argues rather that the evidence points to them having been sold to Japan but that he has so far failed to find any in the major Japanese collections.
It is the custom in both China and Japan for collectors to add their personal seals to items they acquire. This complicates the picture further as Li Shengduo's seals — he used over twenty — also appear to have been forged at some point and, as Rong Xinjiang points out based on Fujieda Akira's original work, that since it was generally known in the trade that Li Shengduo possessed a fine collection of genuine manuscripts, possibly both the original seals and the forged ones were used on both original and on forged manuscripts to add authenticity.
The story of Li Shengduo is relevant here because the son of a friend of his later wrote that Chen Yi'an, Li's nephew, made copies of the manuscripts in his uncle's collection to earn money. Fujieda has identified a number of manuscripts that he said were written by Chen. As Fields and Seddon point out in their first paper, forgeries are copies made with the intent to deceive and can not, therefore be identified using scientific tests. They prefer to use the term 'fake' to describe manuscripts which are not genuine but have been presented as such, whether or not they were made with this intent. Of course, it is possible that Chen Yi'an produced copies of the manuscripts without intent to deceive and that the copies were later passed off as genuine, but this is a matter for historical research and may be unanswerable because of lack of evidence.
Copying manuscripts and works of art has a long tradition in China. Cohen enumerates the various methods in use, including copying from sight, tracing, tracing the outline and imitation. The last, she notes,
'was considered an original creation and was as faithful as possible to the style of the master.'
The aim was not to deceive — this was not a process of forgery — but to learn. Of course, the same method was used in the western artistic tradition, as Tietze points out:
'The heritage of the past stimulated later centuries, and the acknowledged superiority of the master his contemporaries. To copy paintings, drawing and prints has for centuries been considered the principal method of artistic training and has produced an enormous reservoir of possible sources of error.' (p.11)
In some cases, however, it is apparent that there was an attempt to deceive, but money is not the only motive. Some artists, both eastern and western, seem to take special pleasure in fooling experts and others. A paradigm is the case of Zhang Daqian, a Chinese artist discussed in Whitfield's paper. He spent time at Dunhuang with his students making copies of the wall paintings, but also, according to Whitfield, seems to have taken delight in making Dunhuang forgeries of silk paintings. Tenri Library in Japan has a collection of his manuscript fragments from Dunhuang, some of which must also be open to suspicion. His influence continues to be felt. A 1999 exhibition at the Metropolitan Museum of Art in New York was devoted to arguing that one of the paintings in their collection, 'Riverbank', is not a Zhang Daqian forgery but a tenth-century original. The work was acquired by the Museum in 1997 from a private collector, who bought in from Zhang Daqian in 1956. To help their case, the Museum has hung a 'genuine' Zhang Daqian forgery next to the painting to provide a contrast. Scholars remain divided on the issue.
Cohen also enumerates Zhang Daqian's talents as a forger, and places him in a long tradition of forgery masters in China. Copies were made from the fourth to fifth centuries but, as a number of other authors have pointed out, it is the development of an art market which acts as a catalyst for forging rather than copying. 'Works of art have been counterfeited as long as they have been collected', notes Tietze, and cites examples from Roman times. In China there was a market for calligraphy from the fifth century, for painting by the eighth, and for artefacts by the eleventh, but the Chinese collector came into his own in the sixteenth century. This was when connoisseurship became a high calling for the Confucian gentlemen. The Chinese word for connoisseurship, as pointed out by Jones, is composed of two elements which mean, respectively, 'to discriminate on the grounds of quality' and ' to distinguish true from false'. Chinese connoisseurs were therefore alert to the possibility of artistic forgeries: indeed, they were often the perpetrators.
Forging of texts is a more complex matter. There may be no monetary benefit for the forger of an historical text who wishes, for whatever reason, to create history, nor may there be any personal credit. 'For to forge and counterfeit Books, and father them upon great names, has been a practice almost as old as letters.' Examples of this type of forgery abound in Chinese tradition, from the forging of ancient literary and philosophical texts probably as early as the latter part of the first millennium BC, to the creation of apocryphal sutra throughout the first millennium AD. One of the best known cases is the text Huahujing which tells how Laozi, the father of Daoism, went to the west and became the founder of Buddhism. This was, of course, a weapon to be used 'to prove' the priority of Daoism in the many fierce debates that took place between adherents of the two religions. Scholars are still debating about the date of its creation, although it was realized very early that it was a forgery.
In the case of the Dunhuang manuscripts however, they occupy an interesting position which is neither wholly akin to works of art nor to texts, but something of each. This ambivalence is apparent in their study. As Lancaster notes, early scholarship was interested only in the text so that microfilms were deemed sufficient. This is still the case for many scholars, who show little regard for the manuscript as an object. Yet, at the same time, many of the scholars who study the text, want to hold the original and are therefore not content with facsimiles. In some cases this is religious devotion, since most of the manuscripts are Buddhist sacred texts. Even when the scholars is not a Buddhist, the sentiment informing their desire to touch the document is often most accurately described as spiritual, akin to the art collector who pays 'not for a realised concept, ... but for a direct link with the hand of the creator.'
It is interesting that this desire — to be in direct contact with the past — is one experienced in both east and west, despite the diverging attitudes about the past over the past century. This divergence is best expressed by a musical example used by Jones. He points out that in nineteenth century Europe, performers would adapt Bach for the piano believing that, in this way, they were 'more truly realising his aims.' Nowadays, great store is set by trying to recreate music 'as it was played' as if musical performance were set in stone with only one 'right' way to play it, rather than a living expression of the present, constantly interpretable. This change expresses a paradigm shift in western attitudes towards the past which is now primarily viewed 'as a different country', to quote the famous opening lines from a twentieth-century novel. Of course, this belief that we can 'recreate' the past as it was, is a fallacy. We have no direct knowledge of the past and, as Jones says, 'authentic revival is either self-delusion or deliberate fakery.' This is an attitude explored further with wit and perception by Umberto Eco in his essays on 'hyperreality'.
The nineteenth century musician's attitude is closer to that still prevailing in much of China and Japan, where the past is seen 'as a continuous corridor leading to the present.' This attitude is reflected in the architecture, where a building, despite being rebuilt several times, is still labelled as old as the original and, more relevantly here, where collectors mount old manuscripts and add their own colophons and seals. In the urge for classical statuary seen in late eighteenth-century Europe, ancient statues were 'improved' by the addition of missing parts. Throughout the history of manuscripts, they have been likewise 'improved', with owners, whether they be monasteries or private individuals, adding patches, extra sheets, colophons and seals, remounting them, and replacing badly damaged sections with newly copied ones. Like Bach's music, the manuscript is seen as a 'living' object, with each generation having the right to make it relevant to themselves.
The attitude of textual scholars contrasts with paper historians and others who are only interested in the manuscript as an object and may not even be able to read the text. When it come to looking at the manuscripts from the point of view of forgeries, however, it is apparent that they are treated very much as objects — as works of art (a situation very different from Indian manuscripts, as Lancaster points out). As far as we know, the texts of forged manuscripts are not unique or variant: they are simply reproductions of existing texts. The forgers were not concerned to create history — at least, not by producing variant textual sources — nor, it would appear, were they interested in showing off their personal skills. The majority of Dunhuang forgeries were probably made for one reason: money.
Lancaster discusses this and makes the salient point that, despite China becoming a print culture at a very early stage (the earliest, dated, printed book is a Buddhist text from Dunhuang, The Diamond Sutra, dated 868), the skills for creating manuscripts did not die out. Calligraphy remained a high art and manuscript copies of texts were needed for production of the woodblock for carving. Nevertheless, as a number of authors stress, the skills necessary for producing a good copy of a Dunhuang manuscript were in fairly limited supply. Fang Guangchang, in his paper, suggests that the forger required
'considerable experience in handling genuine Dunhuang manuscripts and to have researched their form and content; ... a detailed knowledge of classical Chinese history; ...well-grounded brushwork skills; ... a solid understanding of mounting techniques; and, finally, a good knowledge of Buddhism.'
In the case of Dunhuang manuscripts, the text is not a major issue since most of the forgeries appear to be of canonical Buddhist works. Even if the forgers did not have genuine manuscripts to copy from, they had copies of the Buddhist canon dating almost as far back as the manuscripts, and whether, therefore, a good knowledge of Buddhism was strictly necessary is a moot point: a scribe should be able to copy anything, whether or not he understands it. However, the importance of the text should not be entirely disregarded. Vetch's paper is a model of historical research based on a forensic examination of the text, and shows us how far a well-trained scholar can proceed using these traditional research methods. In her paper, Scherrer-Schaub also uses historical analysis of Tibetan Buddhist texts for dating, but reinforces it with philological, palaeographical, and codicological investigation. Systematic philological analysis of Chinese Buddhist texts is still to be done and, as Lancaster says, 'Buddhist scholars have given almost all their attention to the printed versions and the lack of comparable attempts to study the stemma of readings from Buddhist Chinese manuscripts means it will take many years to establish this.'
Unfortunately, the early development of printing has resulted in a comparative lack of development in another discipline concerning both Chinese and Tibetan manuscripts which, in the west, is an important tool in the discovery of forgeries: palaeography. Another factor in this, as both Drege and Lancaster note, is that the study of handwriting was approached through calligraphy, that is, as an artistic rather than a scribal tradition. Ueyama uses differences in the signature of a monk appending to several manuscripts to cast doubt on the authenticity of some of them. Ishizuka and Lancaster's second paper both make use of dated changes in the formation of Chinese characters to show that certain manuscripts cannot be genuine and Ishizuka's work on a database of standard forms and changes in the forms will provide an indispensable research tool once it is generally available. This is hardly a task to be undertaken by a single scholar, however, and it is to be hoped that his initiative will lead to a collaborative project enabling acceleration and widening of the work. Already the production of a CD-ROM of the Korean canon, made from woodblocks which were produced in the style of characters still in use at the time of production of Dunhuang manuscripts, is proving a valuable aid for authenticating manuscripts. The fact that it also enables scholars to identify fragments should also have a knock-on effect in this field. Previously, scholars had to be conversant with the Buddhist canon, a lifetime task. Now that such extensive knowledge and memorisation is no longer vital and there is a large body of manuscripts to work with (and more being found all the time, mainly from excavated tombs), it is to be hoped that scholars can turn their attention to developing philological and palaeographical studies in China.
To return to the matter of if, when, and how Dunhuang forgeries were made, few would dispute that there are forgeries in circulation made between 1920 and 1949 when the value of the manuscripts was realized by a new class of collectors and there were still few enough manuscripts in circulation for the forgeries work to beescape detection. Fang Guangchang argues that the market ended after 1949 because the price of manuscripts was state controlled and, consequently, low. However, this is to underestimate the ingenuity of man. Although required to sell important historical artefacts, such as Dunhuang manuscripts, to the state, this does not mean that everyone did so. There was a ready market among Japanese and US collectors. In addition, there is the possibility that forgeries continued to be made outside China. But the main issue here concerns earlier forgeries and there are a number of questions remaining unanswered. Firstly, whether forgeries were made locally to Dunhuang and, if so, when, in what numbers and by whom?
Cohen argues that it is difficult to imagine forgeries being produced before 1909 because, up to then the manuscripts were not worth a great deal and, in any case, there was a ready supply. However, many Japanese scholars, led by the doubts of Fujieda, have challenged the authenticity of all manuscripts acquired local to Dunhuang after the cave was cleared in 1910. This would include all the Russian collection, all the Japanese Otani manuscripts, and those 600-odd scrolls acquired by Stein on his third expedition and his second visit to Dunhuang in 1913. Ishizuka uses an afterword in his conference paper to deny the suggestion that any manuscripts acquired before then are forged, and overwhelmingly scholars have accepted that the bulk of the Stein collection and the entire Pelliot collection, along with the original Beijing collection (not including manuscripts acquired after 1910), are indubitably genuine and can be used as a baseline. But to be consistently rigorous, even this assumption needs to be re-examined, especially as it is the foundation on which all knowledge of the genuine manuscripts is built.
Rong Xinjiang has examined contemporary historical documents in China and these, along with the written and photographic records of Stein and Pelliot provide convincing evidence for the belief that the cave was discovered full of manuscripts. Of course, there have been several 'archaeological' finds which, on later inspection, have been proven to be frauds, the material planted in order to be found usually to make the archaeologist's reputation or to change some assumption. The main arguments against this in the case of Dunhuang are the problem of logistics, the lack of precedent and the lack of motive. Islam Akun's crude forgeries are nothing in either quantity and quality when compared to the tens of thousands of fine paper and silk scrolls from Dunhuang. To produce them would have required an entire army of Zhang Daqians, as well as supplies of the raw materials that evidence suggests were simply not available at that time in China. In addition, this was the first find of its kind. Previous document finds in the desert had mainly been of fragments unearthed from ruined cities. Moreover, there are texts among the documents which were previously unknown. And, although we should not underestimate human ingenuity, it would need some powerful motive to prompt this amount of work, other than the joy of deceit — there was no pecuniary advantage. It may seem pedantic and unnecessary to state these argument, but in a world of forgeries it is all too easy to build on sand and we must always strike down to the bedrock.
The question after this is whether some forgeries started to be made before 1909. Cohen's arguments — lack of a market and lack of skills — are persuasive but not final. The finds were known about by local officials and scholars: the presentation of paintings from the cave to local officials shows that they were considered to have value. Moreover, locals were well aware of the interest of foreign archaeologists in manuscripts. And the ready supply was only 'ready' to Wang Yuanlu, so that anyone else wishing to trade in these manuscripts would have to produce their own. Although the probabilities are against forgeries being made at this time, it cannot, I would argue, be entirely ruled out.
At the 1997 conference, Fujieda suggested that a couple of manuscripts seemingly from Stein's second expedition may be forgeries. Even one such find would destroy our certainty about the baseline collections and force us to re-evaluate all our assumptions. But, even if Fujieda's judgements were correct, our assumptions might still be saved by the problem being that addressed by Fang Guangchang and Drege in his first paper: misattributed provenance.
Firstly, contemporary historical accounts by archaeologists show that Cave 17 was not the only site at the Mogao caves to yield manuscript finds. Many Uighur documents were discovered in Cave 464, Tibetan documents in another cave, and others inside stupas. Second, from a careful examination of institutional records and early catalogues, it appears that most institutions were not as rigorous as they might have been in recording the exact provenance of the manuscripts. Lionel Giles, for example, in his 1957 catalogue entitled Descriptive Catalogue of Cthinese Manuscripts from Tun-huang in fact includes the 600 scrolls acquired by Stein on his third expedition (which were bought in the town, and thus their link with Cave 17 has to be more doubtful than those acquired at Mogao on the second expedition), as well as various other manuscripts not even found at Mogao, including one from a tomb. Drege points out that Kharakhoto documents have been mixed with Dunhuang ones in the St. Petersburg collection, and ++ notes that old Japanese manuscripts were mixed with the Dunhuang finds in various Japanese collection. These are not isolated cases, and examples are given by several authors of similar carelessness.
Fang Guangchang is therefore correct to insist that the first step in the process of identifying forgeries must be to decide what is meant by a Dunhuang manuscript, that is, whether the definition should include all manuscripts found at the Mogao cave site and not simply those from Cave 17. Whatever is decided, it is then of prior importance for this investigation to identify those manuscripts which are definitely from cave 17 and those for which doubts remain about their exact provenance. Since some manuscripts recorded as found at the Mogao caves include some dated much later than the eleventh century, it must be proven that these did not come from Cave 17 before we can hypothesize that the cave was sealed in the eleventh century (of course, it is tempting for scholars to state that they could not come from Cave 17 because they are of a later date, but this way lies circularity and complete lack of certainly or proof). Also, only then will be is possible to identify the baseline set against which other, possibly spurious manuscripts, can be judged.
The problem is highlighted by the case of a manuscript in Tenri Library in Japan, marked by its curators as a Dunhuang manuscript. It has been mounted and extra panels of papers added at the beginning and end for prefaces and colophons to be added by its various owners, a common practice among collectors in China and Japan. This manuscript contains several colophons of nineteenth century collectors which date its discovery to the early part of the century, long before the opening of cave 17. Fang Guangchang accepts its genuineness and notes that one of the colophons places its discovery in a pagoda at Mogao, not in cave 17. Rong Xinjiang is more sceptical about its authenticity. After all, adding colophons and seals to a manuscript, as shown in the case of Li Shengduo's collection, was a common method of giving something authenticity:
'the surrounding apparatus of colophons and seals was at least as much the focus of attention as was the actual pictorial images. Genuine early but unsigned pictures were improved for the market in this way, a practice common from at least the twelfth century.'
There remains the question of whether forgeries were made locally after 1910 and, if so, in what numbers. Again, the assumption that most of the manuscripts obtained from Dunhuang after this date are forgeries, made by several authors in this collection, has to be challenged. Against it is the logistical argument. If it were the case, then it would have meant the production of a thousand or more manuscript scrolls in two or three years, some of which are over seventy foot long, thus requiring a large supply of paper (and preliminary examination of the paper of third expedition Stein scrolls suggests that much of it is not machine-made, inks, brushes and glue, as well as reasonably literate scribes: these are manuscripts in Chinese, many with passable calligraphy, not meaningless squiggles. Moreover, most, if not all, of the manuscripts by this time had, if the argument is accepted, been sent to Beijing and so were no longer available for research and copying.
The alternative would be that not all the manuscripts had been sent to Beijing, but some secreted away by Wang Yuanlu in order to continue funding his repairs on the cave murals. Another possibility is that the manuscripts supplied to later expeditions were not from cave 17 at all, but from another source at Dunhuang. A third possibility is that some of them were forged, but other came from Wang Yuanlu's secret stash. Ueyama usefully quotes part of the diaries of Shochiro Yoshikawa, a Japanese expedition member, showing that he clearly believed he was being offered forged manuscripts and also revealing that he bought manuscripts from a wide variety of people other than Wang Yuanlu, including a farm labourer and a Chinese merchant.
It is too early in our work on the manuscript to decide this answer. Of course, scholars who accept the assumption that most of the later manuscripts are forged would claim proof: that of their own eye. Like Fang Guangchang who asserts:
'when I have been in front of a genuine Dunhuang manuscript, I feel like the Jade Emperor's Demon Detecting Mirror: there is no way you can miss the originals.'
Bonafle's advice might be recalled here: 'Distrust art experts' he said, 'who have never been taken in by a forgery.' The art of the forger, glosses Tietze, is often to make the forged item look more real than the original by emphasizing the features that present viewers wish to see. This point is illustrated by the case study of a manuscript in the St. Petersburg collection presented by Brovenkho. It appears that the original has been burned but closer examination the manuscript revealed the presence of brown dye, presumably added to make the manuscript look older than it is.
The eye of the expert is, nevertheless, a not inconsiderable tool. But to have any validity, the various experts must at least agree, otherwise the subjective basis of their judgement has triumphed over the objective criteria that they are making use of. It was clear from the 1997 that experts did not agree. In his paper, Menshikov, who holds that the St. Petersburg collections are genuine, uses an ingenuous but ultimately sosphistical argument to try to prove his point. Nevertheless, his experience with the manuscripts can not be ignored. It was this 'eye' that Fujieda tried to codify in his ground-breaking work on Dunhuang manuscripts. It was his own eye that led him to realize, as Lancaster explains, that there were differences between many of the manuscripts in Japanese collections and those in Britain and France, and then to spend the following decades examining as many of the British and French manuscripts as he could, on the assumption that these, if any, must be genuine. His resulting codicology is, of necessity, a simplification. Most importantly, he directs his research to Buddhist manuscripts and his parcelling of the manuscripts into four main 'types' with a couple of transition types, ignores too much complexity for some scholars, such as Fang Guangchang and Drege. Nevertheless, in his paper first published over twenty years ago, Drege shows both the strengths and the limitations of codicology. It is dismaying that Drege's original work has not proved a catalyst for international, collaborative work on a much larger scale. Indeed, because Drege's article was first published in French, it appears that it is still not known by some Chinese and Japanese scholars, and this collection was therefore seen to be an ideal place to publish an English version, allowing the author to make updates and amendments. As in Ishizuka's work, the collecting of data cannot be the task of a single scholar. Drege's work to date will be added to the international manuscript database and it is to be hoped that groups of research teams will build on this. Akao's paper is an interim report of another such study.
Fields and Seddon in their first paper discussing the issue of forgeries, accept the importance of the expert's 'eye' and other subjective evidence, but point out that such evidence 'is never without distortion by personal feeling and prejudices' and that 'for a true scientific analysis ... we must rely most heavily on objective evidence.' They thereby chose to ignore much of the contemporary debate on the subjectivity of science. Outnumbered in the conference, their passionate justification of the primacy of science is perhaps understandable, but it must be remembered that science can be a false friend. Firstly, of course, there is the fact that there is good and bad science, and the difference between them depends on the scientists. It has long been recognised that, in the field of radiocarbon dating, there are only a few laboratories world-wide which will produce trustworthy results. Other laboratories may have the same equipment, but its calibration, the preparation of the material for testing and the analysis of the results are all factors relying on human skill. Even if the insturments of measurement are relatively foolproof — and few are — then the design of the experiments and the interpretation of the results is still dependent on the calibre of the scientists. In Kohno et al's paper, which measures the various elements found in the paper of manuscripts, a poor scientist may interpret the presence of lead as significant, whereas it may simply be the result of contamination of the material in the laboratory.
Secondly, science can be used to deceive as well as elucidate and we must never be blinded to its limitations. In the nineteenth century when photography was developed it was really believed that it could not lie: it was a scientific procedure producing purely objective results. The photographs of fairies at the bottom of the garden soon showed the paucity of this argument. Objective tests in common use today, such as thermoluminensce, are also open to the forger's ingenuity. It has been claimed that recently Chinese pots have started to come on the market which, when tested using thermoluminesnce, are 'proven' to be old, but which are, in fact, made from ground up old pot shards. It is not fanciful to envisage a time when radiocardon dates will be similarly suspect. No-one would deny that these are objective tests, but a positive result may not prove the authenticity of the object. In the case of the manuscripts, given that monasteries and other institutions in China often kept supplies of old paper, it is possible (although unlikely given the age and therefore the value of such paper) that some of the forged manuscripts were written on original paper and that radiocarbon testing would give a plausible date for the manuscript being original. A negative test, of course, would prove it to be a fake.
Rischel's work on paper fibre analysis and Lancaster's paper on glyphs are interesting cases of the fuzziness between objectivity and subjectivity, and the parallel fuzziness between scientific and other data. Paper fibre analysis might ordinarily be considered a scientific test but, as Drege points out, there has been little agreement among scientists about the identity of the fibres they have analysed. Lancaster's work, while strictly belonging to the humanistic tradition, is far less open to interpretation and therefore more scientific. Snow's famous 'two cultures' still, unfortunately exists, but if we are to make progress we must recognize the failings of each other's culture. The fact remains that to elucidate the question of Dunhuang forgeries both cultures need to work together to marshal a body of data from a variety of scientific tests and historical, codicological, palaeographic and philosophical research on a large group of manuscripts (the larger the better, hence the need, as Fields and Seddon point out, for international collaboration), evaluate the data and reach a conclusion based on the whole.
A final point is one raised by Seddon at the end of the conference and mentioned in his first paper, namely, how much do we care about identifying forgeries? In other words, how much time, effort and expense are we willing to throw at this problem? In fact, when scholars address the issue of forgeries, they inevitably first have to address the issue of genuine manuscripts and, as this collection of papers shows, problems emerge which have not yet been adequately addressed and are as important for the study of genuine manuscripts as for the study of forgeries. The confusions of mis-attributed provenance, of false colophons and seals, are all cases of this. There are still great steps to be made in historical, palaeographical, philological and general codicological research, and, in these fields, there is no question that further development will be of enormous advantage to scholarship. The light it throws on forgeries will be an additional benefit, rather than the focus of this work.
Whereas science is concerned, many of the methods being developed belong to the same category. They will be of great use in helping us understand better genuine manuscripts. Some tests, however, will always be more focused on authentication, such as radiocarbon dating. This is both destructive and comparatively expensive — at least for public institutions where most manuscripts are held. Like thermoluminesense, it is easy to see that it becoming a test used by collectors in the marketplace to verify their products once the forgery issue becomes common currency and the conspiracy of silence, so long dominant in this field, is broken, Indeed, at least one dealer has already considered using it. But it is another matter whether public institutions will ever be able to justify widespread radiocarbon dating. It is more probable that it will be used for random testing to corroborate other evidence.
But this is still a long way off. The papers in this collection, as mentioned previously, open the debate and pose questions. Now that this issue is in the public sphere and all those involved — the curators, conservators, scholars and scientists — have recognized the need to corroborate on further research, it is to be hoped that the advance in scholarship will be rapid. A full understanding of the historical circumstances of the discovery and dispersal of the manuscripts and a clarification of the provenance of all those manuscripts previously labelled as from Dunhuang is long overdue, and this should be the first aim of future research.