Manuscripts under the microscope


What I like about working with manuscripts is that there are so many different ways to approach them. You can read the texts written on them (and sometimes that’s as far as you get) but you can also look at their shape and size, how they were put together, how the writing is laid out on the page (codicology) and the style of the writing itself (palaeography). You can get into their materiality, feel the rough and smooth sides of a page, their coarse and fine fibres, the subtle patterns of laid and chain lines. If you’re lucky, you can find out who wrote them, who owned them and how they were used, who repaired and re-used them, and so on.

I like to think this isn’t just the idle curiosity of somebody who’s spent too much time around books. While most studies of the early Tibetan manuscripts from Dunhuang and other Central Asian sites are focussed on the texts, there’s a lot more we can find out from the physicality of the thing itself. Sure, we can discover what a text is about by reading it and comparing it with other texts. But there are a lot of things we won’t know, like who made the manuscript, who used it, and what it was used for. If we can get some kind of answers to those questions about the manuscript, our understanding of the text will be enriched. Or to put it another way, if we want to know the meaning of a text, we should look at how it was used.

* * *

A few years ago I started working with Agnieszka Helman-Wazny, a scientist specializing in the study of Tibetan and Central Asian paper. Agnieszka’s speciality is the microscopic analysis of paper fibres. She also looks at the patterns left on the paper by the process of making paper (such as the fine pattern of ‘laid lines’) and other aspects of the technology of papermaking. Gradually we developed a plan to bring the results of her analysis of the paper in the Tibetan manuscripts from Central Asia with the work I had done in the palaeography of the manuscripts, and of course the contents of the texts as well. We selected a group of fifty manuscripts, put everything we could find out about them into a table, and looked at the patterns that emerged.

One of the most interesting results was this: those manuscripts that had been brought to Dunhuang from Tibet itself, were made in quite different ways from those that were made locally at Dunhuang. Though our sample was limited, this opens up the possibility of ‘fingerprinting’ a manuscript to find out where it was made.

It looks like the manuscripts made in Dunhuang and other Central Asian areas inhabited by the Tibetans during the 8th and 9th centuries were made with rag paper, which is probably mainly recycled textiles. The technical apparatus of papermaking was a mould made from a sieve made from bamboo or reeds arranged on a wooden frame, which leaves the tell-tale pattern of laid lines on the finished paper. The advantage of this kind of mould is that you can lift out the piece of paper and leave it to dry while you begin to make another one. On the other hand, in places like Tibet, Bhutan and Nepal, the method to this day is to use a wooden frame with a cloth backing stretched across it. With this kind of mould the paper cannot be removed until dry, so the paper dries on the frame. This is obviously a slower method, and the paper produced this way does not have the laid lines characteristic of the sieve method.

DaphneTwo manuscripts, letters that we already thought may have been originated from Tibet, did turn out to have been made on a woven mould. Also, they were not made of rag paper, like the locally produced Central Asian manuscripts, but paper made from the Daphne or Edgeworthia plants, which grow along the Himalayas. As well as these letters, a sutra manuscript written in the archaic ‘square style’ also turned out to be composed of Daphne fibres.

Then there are the big Perfection of Wisdom manuscripts that were brought to Dunhuang to be used as models by the local scribes who had been ordered by the Tibetan emperor to produce copies. The Perfection of Wisdom manuscripts made in Dunhuang are composed of rag paper made on a sieve mould, like other locally made manuscripts. But those that were brought in are composed of Paper Mulberry (Broussonetia) fibres and were made on a woven mould. Paper Mulberry is not native to Central Tibet, but it is found in Eastern Tibet, so perhaps these Perfection of Wisdom manuscripts were produced in the Eastern regions of the Tibetan Empire, before being brought to Dunhuang. This would give us a triangle of geographic locations to which we can assign the manuscripts: Central Asia, Central Tibet and Eastern Tibet.

*  *  *

Though I can’t put the article in which we published our results on this site, I am going to make it briefly available for download here. Of course, the 50 manuscripts that we studied are a tiny proportion of the Central Asian manuscripts in Tibetan, so more work needs to be done to confirm what we have suggested. As well as using these results to pin down the geographical origin of early Tibetan manuscripts we can also say a bit more about what ‘Tibetan paper’ means in this early period. If we can begin to speak of a type of paper with specifically Tibetan characteristics, it is a paper composed of Daphne or Edgeworthia (from Central Tibet) or Paper Mulberry (from Eastern Tibet), made on a woven mould — a technology that continues to the present day.

*  *  *


Helman-Wazny, Agnieszka and Sam van Schaik. 2013. “Witnesses for Tibetan Craftsmanship: Bringing Together Paper Analysis, Palaeography and Codicology in the Examination of the Earliest Tibetan Manuscripts.” Archaeometry 55.4: 707–741. DOI: 10.1111/j.1475-4754.2012.00687.x

Iwao, Kazushi (forthcoming). “On the Tibetan Śatasāhasrikā-prajñāpāramitā sūtra from Dunhuang.” In Scribes, texts, and rituals in early Tibet and Dunhuang (eds. B. Dotson, K. Iwao and T. Takeuchi). Wiesbaden: Reichert Verlag.


1. Paper mulberry (Broussonetia sp.) fibres stained with Herzberg stain, found in IOL Tib J 1560.

2. A large-size ‘floating’ mould, constructed with a wooden frame and attached woven textile, placed in water (a stream) in Gyantse, c. 1910–1920. Photo 1112/2 (139), © The British Library

3. The flower of the Daphne plant.

4. Sheets of paper left to dry on individual moulds on the mountain slope near Tawang, Arunchal Pradesh, 1914. MSS Eur/F157 (324), © The British Library.

*  *  *

Final Notes

Paper made from the Daphne and Edgeworthia species  is shog shing or dung lo ma in Tibetan. There is also a Tibetan paper made from the roots of both the Stellera chamaejasme species (re lcag pa in Tibetan) and, more seldom, Euphorbia fisheriana (re lcag gi rtsa ba in Tibetan). It has been suggested that re lcag pa is the “original” Tibetan paper. Though we did not find any of this paper in our study, finds from Tibet itself may help to confirm whether it was produced during the Tibetan imperial period or later. Also, it is hard not to oversimplify this complex research, and I had better clarify here (as we did in the article) that the rag paper in the Dunhuang manuscripts was also often made with the addition of Paper Mulberry and/or Hemp. Agnieszka Helman-Wazny’s continuing work on the Chinese manuscripts from Central Asia will no doubt add much more to our knowledge of Central Asian papermaking.

*  *  *


Monks and Mahāyoga

As you probably know, after the collapse of Tibetan imperial power towards the end of the 9th century, the lineage of monastic vows (the vinaya) died out in Central Tibet. During the ensuing dark period, if the traditional histories are to be believed, the lineage of the vows survived only in the far northeast of the Tibetan cultural area. Now if that is true, we might hope to see some corroborating evidence among the Dunhuang manuscripts — and I think we do. Several manuscripts that (judging by their handwriting) seem to be from the post-imperial period contain classic texts on the monks’ vows, such as the Vinaya-vāstu and the Prātimokṣa-sūtra (see for example IOL Tib J 1).

It seems likely that these manuscripts, like most manuscripts, were made to be used.  They are the kinds of extracts and summaries that would have been part of the ceremonies of taking and renewing of the monastic vows. Which is to say, they were probably written and used by Buddhist monks. These monks who were maintaining a monastic lineage which may well have died out in Central Tibet, but was very much alive here in the northeast, in places like the mountain retreat of Dantig, or the walled city of Tsongka.

If we accept that the Dunhuang manuscripts containing vinaya texts were used by Buddhist monks, then an interesting issue arises: were these monks also writing and making use of the many tantric manuscripts also to be found in the Dunhuang collections, including those Mahāyoga texts containing violent and sexual imagery? If they were, then the problems involved in monks practising tantric rituals must have come up here, before they were explicitly discussed by Atiśa, who famously addressed the issue a century later.

*  *  *

I was thinking about this after looking at one of the biggest Mahāyoga manuscripts in the Dunhuang collections, a manuscript so big that it begins in the Pelliot collection in Paris (Pelliot tibétain 42), continues in the Stein collection in London (IOL Tib J 419) and ends back in Paris again (Pelliot tibétain 36). Clearly it had already broken into three parts before Stein and Pelliot arrived at the cave in Dunhuang. Put the three back together, and you get a major ritual, involving torma offerings, teachings, the visualization of mandalas, and a violent ritual of liberation (sgrol ba). The liberation ritual has recently been discussed in detail in Jacob Dalton, who describes it as “clearly the most violent text to emerge from the library cave at Dunhuang.” Surely not the kind of thing for monks?

Looking at the manuscript again recently, I noticed some text that had been added to the end of the ritual, either by a different scribe, or by the same scribe writing less carefully. This text turns out to be a summary of the vinaya, beginning like this:

The vinaya of the hearers is divided into eighteen different sects. Of these, the one that exists in Tibet is the system of the Mūlasarvāstivādins.

Fair enough — this agrees with what the Tibetan historians say, and indeed the fact that the massive Mūlasarvāstivādin vinaya is the version of the monastic vows that was preserved in the Tibetan canon. On the other hand, I think this is the first time I have seen the fact mentioned in a Dunhuang manuscript. The text goes on to enumerate the different classes of vows in the vinaya of the Mūlasarvastivādins. Maybe it was a kind of primer for new monks.

So why is this text written on the last pages of a major Mahāyoga ritual? Perhaps so that the monks performing the ritual should do it in the context of their Buddhist vows (and thus certainly not taking the violent and sexual imagery of the texts literally). Or as a rebuke to the text by a shocked monk: this is what Buddhism is about, not that! I don’t know, but I suspect the former is more likely than the latter. Everything we know about tantric Buddhism in India and Tibet suggests that it was thoroughly accepted in the monastic context. What remained uncertain and shifting was the exact nature of the relationship between the monastic vows and tantric practices, and issue that received much discussion later in Tibet in the “three vows” literature (these being the monastic vows, the bodhisattva vows and the tantric samaya vows). The juxtaposition of texts here suggests that similar negotations were already taking place in the northeast of Tibet in the tenth century.

*  *  *


Jacob Dalton. 2011. The Taming of the Demons: Violence and Liberation in Tibetan Buddhism. New Haven: Yale University Press.

Ronald Davidson. 2005. Tibetan Renaissance: Tantric Buddhism in the Rebirth of Tibetan Culture. New York: Columbia University Press.

Carmen Meinert. 2006. “Between the Profane and the Sacred? On the Context of the Rite of ‘Liberation’ (sgrol ba).” In Michael Zimmermann (ed.), Buddhism and Violence. Lumbini: Lumbini International Research Institute. 99-130.

Sam van Schaik and Imre Galambos. 2012. Manuscripts and Travellers: The Sino-Tibetan Documents of a Tenth-Century Buddhist Pilgrim. Berlin: de Gruyter.

Defining Mahāyoga


Some years ago I was chatting to someone at a conference about the work I had been doing on Mahāyoga texts in the Dunhuang collections. “But what,” he asked, “is Mahāyoga anyway?” Though the later Nyingma tradition has perfectly good answers to this question, I couldn’t give him a satisfactory answer about what it meant in early (pre-11th century) Tibet. It would be good to know, because references to Mahāyoga often crop up in the Dunhuang manuscripts. Moreover, if the Testament of Ba is to be believed, when the Tibetan emperor Tri Song Detsen set up a massive Buddhist translation project in Tibet in the 9th century, he specifically banned Mahāyoga texts from being translated.

In fact it turns out that the Tibetans, during their first exposure to tantric Buddhism, had also asked themselves what Mahāyoga stands for. I found the answer in a tatty manuscript containing a text called A Summary of the View of Mahāyoga According to Scripture. So I translated that text and wrote an article all about Mahāyoga in early Tibet, which was published in 2008. I’ve finally scanned the article, and it’s now on the Author page of this site, or you can just click here.

New Publication: The Origin of the Headless Script (dbu med) in Tibet


A new collection of articles, Tibeto-Burman Languages IV, edited by Nathan Hill, is just out. Here you will find studies of fascinating languages such as Mon, Nam, Lepcha, Pyu, Tangut, and of course Tibetan. The collection also contains my study of the development of the “headless” script in Tibet. Traditionally it has been said that the headed (dbu med) and headless (dbu can) scripts were invented during the seventh century based on two different Indian scripts. In this article I argue, based on an extensive survery of the earliest examples of Tibetan writing, on rock, wood and paper, that the headless script was a later development from the headed script. This can be shown to have happened through principles of “cursivization” seen in scripts all over the world. Click here to download the article.

Ancient Tibetan Seals

Recently I spent some time at the British Museum looking at objects from the Tibetan forts of Miran and Mazar Tagh, where many of the Tibetan manuscripts that I work with at the British Library come from. The reason for the somewhat arbitrary division of textual and non-textual material from the same site across two different institutions goes back to the British Library Act of 1972. While the decision to separate the texts from the things made sense in that the British Museum simply didn’t have the space to house an ever-increasing collection of books, it also resulted – in the case of the finds from Miran and Mazar Tagh – in letters being separated from the pens that were used to write them, and the seals that were used to stamp them.

As for those seals (and if you are wondering, that’s what is pictured above), I saw these for the first time recently when Catrin Kost, who has been cataloguing the British Museum’s artifacts from Central Asia, asked me to have a look at them. When the seals (three in total) were first brought to the Museum by Aurel Stein, he had impressions taken of them, and these were read by L.D Barnett. The latter’s readings were not bad, but he let his imagination run away with him in the case of this seal, which he read as a Tibetan version of the Western name Anthony (‘Ang to ne)! In fact, these Tibetan syllables seem to be a Chinese name (Wang to ne). As you can see from the image, the name ends with a swastika, and sits on a lotus blossom. The three seals are all made from animal horns, with a hole bored through the midpoint, probably so that they could be carried on a string or leather thong. I’m not sure what animal such horns might belong to – though this might not be clear from the photo, they are very small, only about 3 inches long.
We find stamps made with this kind of round seal in lots of documents from Miran, Mazar Tagh and Dunhuang. The round seals are always used by private individuals, in contrast to the larger, more impressive square seals used by officials (see this post for some examples of those). These personal seals would function like a signature on legal documents, like contracts and receipts. Pictured below is a receipt for the repayment on a loan of wheat (IOL Tib J 844), with the seals of three people involved in the transaction.
Since they were found in a Tibetan imperial fort, these seals should date to the first half of the ninth century — so they are by far the earliest examples of Tibetan seals. As far as I know, later examples are all made from metal, as in this picture from a Chinese exhibition (but note also the presence of holes in the middle of many of the seals, and the leather thong tied through one of them):
So, let me finish with an attempt to read the names on the three seals.
1. MAS.606 (pictured at top): this seems to be Wang To ne, certainly a Chinese surname, and probably a Chinese personal name as well.
2. MAS.607 (pictured above): Written in the four spaces marked out by a swastika, we have the Tibetan clan name Gnyos followed by Lha la brtsan. I’m quote sure about the brtsan but not so much about the lha la, if anybody else wants to try their hand, or eye. Barnett read it as khal.
3. MAS.608 (pictured above): This seal is now quite damaged, but if you click on the link to the IDP site, you can see a the imprint taken by Stein. The name I see is ‘O nam Gnyan lha, a clan name (‘O nal) that we also see in the Miran documents, and a common imperial period personal name (Gnyan lha). Similarly to the first seal, the name is flanked by swastikas and sits above a lotus.
*  *  *
Further Reading
British Museum: Search the Collection – if you type in the seals’ numbers here (e.g. MAS.606) you can see the full description by Catrin Kost, which includes my own readings.
Aurel Stein’s description of the Miran site and its artefacts, as well as the seal impressions are in: Stein, Mark Aurel, Serindia, Oxford: Clarendon, 1921 (vol.1, p.480).
The beautiful new photographs of the seals shown here are all by IDP photographer Rachel Roberts, and are (c) The British Museum.

Blood writing

There is something compelling about the idea of a text written in blood. The 20th century Chinese writer Lu Xun once said “Lies written in ink can never disguise facts written in blood.” Here the phrase “written in blood” is metaphorical — Lu Xun was talking about the killing of student protesters — but resonates with Chinese history, as China does actually have a tradition of writing in blood. The tradition was especially present in Buddhism and the earliest surviving examples we have are from the Dunhuang manuscript collections. For example, there is a booklet containing the Diamond Sutra (S.5451) with the following colophon (as translated by Lionel Giles):

Copied by an old man of 83, who pricked his own hand to draw blood [to write with], on the 2nd day of the 2nd moon of bingyan, the 3rd year of Tianyu [27 Feb. 906].

Using one’s own blood to write Buddhist sutras is an ascetic practice, that can be included in along with other, more drastic ascetic practices that were practiced in China over the centuries, including slicing off parts of one’s flesh, burning oneself with incense, burning off a finger, or even complete self-immolation (on which, see the book by James Benn in the references). Much later, in the 17th century, a Chinese writer defended the practice of blood writing against its detractors:

Those disciples of “crazy wisdom” (kuanghui 狂慧) belittle it [blood-writing] as [involving] “corporeality” (youxiang 有相). But among the root causes of beginningless birth and death, none is deeper than the very perception of the body (shenjian 身見)… This [practice of blood-writing] is called paying reverence to the Correct Dharma; it is also called using the Dharma to make offering to Buddha. The Lotus and Śuramgama [sutras] have profound praise for incinerating one’s limbs and fingers, as well as the merits from burning incense [into one’s body]. The practices of severing the limb of afflictions and burning the body of ignorance are situated precisely in this very flesh and blood.

So, what about Tibet? It is my impression that this kind of extreme ascetic practice in general, and blood writing in particular, is historically less common among Tibetan than Chinese Buddhists. The manuscript pictured above (IOL Tib J 308) therefore strikes me as an exception. It certainly looks like it is written in blood: the colour is reddish-brown, and appears to congeal in some places. In fact, it looks much more like blood than the writing in the book by the 83-year-old man, which looks like ordinary ink. In that case, perhaps, the old man just added a few drops of blood to the inkwell.


Recently, I had the chance to have the ink in this Tibetan manuscript examined by Renate Noller, a specialist in pigment identification at the Bundesanstalt für Materialforschung und -prüfung. Her results are yet to be published, but this particular ink turned out to have a very high iron content. Now, there are inks made with iron (in the West, iron gall ink was particularly popular, and was used, for example, by Leonardo da Vinci in all his manuscripts), but that tends to darken with time to a browny-black, and lacks the clotted quality of this manuscript. If you look closely, you can see that the scribe was dipping his pen very frequently, that the ink went down very thickly and then ran out after a couple of letters.

The text that is (perhaps) written in blood in IOL Tib J 308 is the Sutra of Aparimitayus, a very popular text in Tibet, on the visualisation and the mantra of a deity representing long life and rebirth in a pure land. In the 840s thousands of scrolls of this sutra were written at Dunhuang at the behest of the Tibetan emperor, to ensure his long life through the religious merit generated by copying the sutra. This manuscript is not one of those, and to judge from its archaic orthography and “square” style, may be even older than them. Still, the motivation for copying the sutra is probably the same. If it was written in blood, this act would have given a greater value to the act of copying of the sutra, and thus to the merit generated by doing so.

*  *  *


James Baskind. 2007. “Mortification Practices in the Obaku School“, in Essays on East Asian Religion and Culture, edited by Christian Wittern and Shi Lishan, Kyoto.

James Benn. 2007. Burning for the Buddha. Honolulu: University of Hawai’i Press.

John Kieschnick. 2000. “Blood Writing in Chinese Buddhism.” JIABS 23.2: 177–194.

John Kieschnick. 2002. The impact of Buddhism on Chinese material culture. Princeton: Princeton University Press.

Jimmy Y. Yu. 2007. “Bodies of Sanctity: Ascetic Practices in Late Imperial China“. Dissertation prospectus, Princeton University. (Source of the 17th century passage above.)

Jimmy Y. Yu. 2012. Sanctity and Self-Inflicted Violence in Chinese Religions. New York: Oxford University Press.

New Publication: Old Tibetan Texts in the Stein Collection Or.8210


Just out — a new catalogue of 88 Tibetan manuscripts from the Dunhuang cave, held here at the British Library but previously neglected because they were catalogued in the sequence dedicated to Chinese Dunhuang manuscripts (Or.8210). This catalogue was begun by Tsuguhito Takeuchi of Kobe University, who passed it on to Kazushi Iwao, who worked with me directly on the manuscripts at the British Library. The catalogue includes some manuscripts that I have previously written about here, such as this divination text, this text on bon-po funeral rituals, and this fragment of the Testament of Ba. The catalogue is the first in a new series published by the Toyo Bunko: Studies in Old Tibetan Texts from Central Asia. You can download PDF files of the whole book at the Toyo Bunko’s website, here.