Manuscripts and Practices : Investigating the Tibetan Chan Compendium ( P . Tib . 116 ) 1

The manuscript that I discuss in this paper is one of the most important in Tibetan religious history. Yet it has primarily been studied as a series of individual texts, rather than as a multiple-text manuscript (MTM). This has limited our understanding of why it was written, and what role it might have had in religious practice. In this paper I will argue that a combination of close codicological investigation of the manuscript on the one hand, and an awareness of the socio-historical background of its creation on the other are both necessary. It is my hope that the process of intertwining codicological and socio-historical methods here will be of interest to those working will manuscripts from other traditions as well. This particular manuscript – Pelliot tibétain 116 – was found in a sealed cave in Dunhuang, in Chinese Central Asia. The discovery of this cave in the early 20th century opened up a vast new resource for the history of Buddhism in China, Central Asia, and Tibet. In particular, the study of the Chinese Buddhist movement known as Chan (Zen in Japan) was revolutionized by the appearance of the Dunhuang manuscripts. In the 20th century Sinologists rewrote the early history of Chan based on lost texts preserved among the Dunhuang manuscripts. Even more surprising was the discovery that the manuscripts also contained texts from an extinct Tibetan tradition of Chan. Tibetan history preserved a story of a debate held at the Tibetan imperial court between a Chinese Chan master and an Indian scholar; however, there was no firm evidence that there had ever been a living transmission of Chan lineages in Tibetan before the discovery of the Dunhuang manuscripts.1 Among the forty-odd Tibetan Chan manuscripts from Dunhuang, one stands out as particularly significant. Pelliot tibétain 116 is not only the largest of the Tibetan Chan manuscripts, it also contains the largest number of texts. Since the


Introduction
The manuscript that I discuss in this paper is one of the most important in Tibetan religious history. Yet it has primarily been studied as a series of individual texts, rather than as a multiple-text manuscript (MTM). This has limited our understanding of why it was written, and what role it might have had in religious practice. In this paper I will argue that a combination of close codicological investigation of the manuscript on the one hand, and an awareness of the socio-historical background of its creation on the other are both necessary. It is my hope that the process of intertwining codicological and socio-historical methods here will be of interest to those working will manuscripts from other traditions as well.
This particular manuscript -Pelliot tibétain 116 -was found in a sealed cave in Dunhuang, in Chinese Central Asia. The discovery of this cave in the early 20 th century opened up a vast new resource for the history of Buddhism in China, Central Asia, and Tibet. In particular, the study of the Chinese Buddhist movement known as Chan (Zen in Japan) was revolutionized by the appearance of the Dunhuang manuscripts. In the 20 th century Sinologists rewrote the early history of Chan based on lost texts preserved among the Dunhuang manuscripts. Even more surprising was the discovery that the manuscripts also contained texts from an extinct Tibetan tradition of Chan. Tibetan history preserved a story of a debate held at the Tibetan imperial court between a Chinese Chan master and an Indian scholar; however, there was no firm evidence that there had ever been a living transmission of Chan lineages in Tibetan before the discovery of the Dunhuang manuscripts. 1 Among the forty-odd Tibetan Chan manuscripts from Dunhuang, one stands out as particularly significant. Pelliot tibétain 116 is not only the largest of the Tibetan Chan manuscripts, it also contains the largest number of texts. Since the || 1 The scholarly literature on Tibetan Chan is too extensive to describe here. The important Japanese studies of the 1960s and '70s are summarized in Ueyama 1983. For general discussions of Tibetan Chan see Broughton 1983 and1999. On syncretic movements in Tibetan Chan, see van Schaik / Dalton 2004 andMeinert 2007. collection as a whole is not named, I will refer to Pelliot tibétain 116 here as the Tibetan Chan Compendium (or for brevity, just the Compendium). The texts in the Compendium have proved very useful for understanding the nature of Chan Buddhism in its Tibetan form, and the individual texts have been identified with fragments preserved in polemical works in the later Tibetan tradition. Some of the texts have also been found in other Dunhuang manuscripts. Yet the reasons why they were gathered together here have not been explained, and indeed the codicological nature of this crucial manuscript has barely been discussed at all.

The making of the Tibetan Chan Compendium
The manuscripts from Dunhuang comprise a variety of book forms, including the scroll, codex, pothi and concertina. Pelliot tibétain 116 is a concertina manuscript composed of 124 panels, each of approximately 7 x 29.5 cm; each panel contains four lines of text. The concertina form has been studied by Jean-Pierre Drège (1984) -the term he uses is l'accordéon. Drège identified 263 concertina manuscripts from the two major collections of Dunhuang manuscripts: the British Library and the Bibliothèque nationale de France. Although the majority of the manuscripts in these collections are in Chinese, Drège found that the vast majority of the concertinas (around 90%) are in the Tibetan language. Those few concertinas that contain dates are all from the second half of the 10 th century, and in general, Drège suggested, the concertinas should be dated to the period between the fall of Tibetan rule in Dunhuang and the closing of the cave, that is, from the mid-ninth to the end of the 10 th century.
Many concertinas are MTMs, of some length. Since papermaking technology only allowed for the production of a limited length of paper, these long concertina manuscripts were made by pasting together several lengths of paper. This is essentially the same method used to construct scrolls, except for the folding of the panels. Another difference from scrolls is that the concertina style requires a greater rigidity than the scroll, and thus most concertinas are made with a double layer of paper. Since the sheets of paper are pasted together, the overlap can be seen at the beginnings and ends of a sheet. In the Compendium, the scribe has written over some of these joins, so it is clear that the manuscript was constructed before the scribe wrote on it (Fig. 1). The folds between many of the panels have been torn and repaired at various points. Some have been stitched together with thread, and others have been repaired by gluing strips of paper across the joins (Fig. 2). In addition, there are modern repairs dating to the 20 th century, where some of the more damaged panels have been joined with strips of conservation paper.  It is obvious that the Compendium was well used and much worn by this use. The different types of repair visible on the manuscript suggest that it was repaired more than once. At some point, somebody undertook a major repair by replacing the first and last panels. These two panels are visually different from the rest of the manuscript, being composed of different paper and containing text written in a different hand from all of the other panels (Fig. 3). The person who repaired the Compendium knew enough of the texts it contained to fill in the missing text. In the case of the first text on the recto, this is not surprising, as it is the Prayer of Good Conduct, one of the most popular prayers in Tibetan Buddhism. The last text, however, is a Chan text, and the repairing scribe would have to have found an exemplar to copy, or have been familiar with the text. In fact, it seems that the person who repaired the Compendium was familiar with Chan literature in general. His handwriting also appears on the last page of another Chan MTM, IOL Tib J 710 (which contains two long texts). This scribe may, therefore, have been collecting, repairing and annotating Chan manuscripts. 2 He also wrote a brief Chan text on a single page, Pelliot tibétain 811; this is a very useful manuscript because the scribe's handwriting begins in the headed style and then changes to the headless style. In general, the Tibetan tradition distinguishes between a headed (dbu can) and headless (dbu med) style of writing. The 'heads' are short horizontal strokes that cap many of the Tibetan letters (a characteristic inherited from the late Gupta script of India, which was the main model || 2 We also see this hand in other manuscripts, in both of which only a few pages are in this hand, with the others in the hands of other scribes; for example, Pelliot tibétain 6 has two panels in this hand, and Pelliot tibétain 23 which has the title. The unusual opening curl (mgo yig in Tibetan) in Pelliot tibétain 116, which is composed of three small curls or globule shapes, is also seen in other manuscripts, including IOL Tib J 349, Pelliot tibétain 6, 499, 528, and 814. for the Tibetan alphabet. 3 Thus we have examples of this scribe's hand in both styles (Fig. 4).
Turning to the last panel of the Compendium, we see that it contains notes written in the headless style. Thanks to Pelliot tibétain 811, we can see that these notes are also in the hand of the repairing scribe. 4 The repairing scribe's notes on this final panel are intriguing, but difficult to read, especially as two lines have been deliberately erased at some point. One of the erased lines contains a date: the tenth month of a mouse year. But since the Tibetan calendar of the time was a twelve-year cycle, this does not help us. One passage of text on this panel can still be read (Fig. 5). Here is a provisional translation: In The Little Lamp, two panels are not present. Anyone who makes a copy of the book should include the two missing panels. This is not an omission: the two complete panels were not present. 5 Here the editor seems to be referring to the fact that the end of the manuscript was damaged when it came into his hands. He seems to have removed the fragments of these panels, and replaced them with a single new final panel. The location of the missing panels is confirmed by an interlinear note on the other side of this final panel, which states: 'The two missing panels should be included in between here'. 6  The handwriting of this repairing scribe is characteristic of the latter half of the 10 th century, which helps us to date the repairs to that time. 7 The rest of the manuscript must of course have been old and damaged by the time it was repaired. The handwriting here fits within the mid-ninth century type that is found in hundreds of copies of Buddhist sutras sponsored by the Tibetan imperium (for this reason, I have classified this style elsewhere as 'sutra style') and mostly before in the mid-ninth century. The same handwriting is also seen on many concertina manuscripts, which, given Drége's dating of these manuscripts to after the mid-9 th century, shows that the style persisted after the fall of the Tibetan empire. 8 A further clue to the date of the Compendium is one of the texts within it, the The Single Method of Non-Objectification, which cites a scripture that was translated into Tibetan in the 830s or 840s. Thus the Compendium could date from this time through to the early 10 th century, when the Dunhuang cave was closed. Given the heavy wear and many repairs to the manuscript, it must have been in circulation for several decades at least before being placed in the cave, and thus should date from between the mid-ninth and mid-tenth century. Taking into account that most of the dated concertina manuscripts are from the 10 th century, our more specific estimate would place the date of the creation of the Compendium in period c. 900-950.

The structure of the Tibetan Chan Compendium
Previous scholars have divided up the textual content of the Tibetan Chan Compendium in different ways, resulting in different calculations of the number of texts contained in the manuscript. This reflects a difficulty with MTMs in general, and the fact that in manuscript cultures textual units can be combined and separated along different lines according to the needs of a particular scribe. 9 My own division of the textual content of the Compendium is given in Tab. 1. || 7 On the use of palaeography to date the Tibetan manuscripts from Dunhuang, see van Schaik 2013. And on the use of forensic handwriting analysis to identify individual scribal hands, see Dalton / Davis / van Schaik 2004. 8 Pelliot tibétain 10 is written in a very similar hand to the main body of Pelliot tibétain 116, and that it comprises another copy of the Diamond Sutra. 9 For a different calculation of the text divisions in Pelliot tibétain 116, see Ueyama 1985. For a systematic study of different kinds of scribal alteration of texts (based on medieval European manuscripts) see Dagenais 1994. Furthermore, some texts are themselves compendia; for example, it is not immediately obvious that Text VI should even be considered a text; it is a series of citations of Chan masters, in which each individual citation may have existed in previous versions, but this could well be the first time they had been combined in this particular order. Yet, as a collection of coherent material, placed after another such collection, and with no clear internal signalling of text breaks, it is also justifiable to identify this as a single textual unit. What, after all, is a textual unit? Perhaps the best way to make a convincing case for identifying a chunk of text as unit is to combine an analysis of the layout of the text on the page with textual analysis, and with examples of the same text (or at least a title) found elsewhere. A case in point is Text V, The Single Method of Non-Objectification. This is a single text containing 38 questions and answers, but previous scholars have divided it into several texts based on thematic readings. 11 Versions of parts (a) and (b) of the Single Method exist separately elsewhere among the Dunhuang manuscripts. Some other manuscripts contain a series of questions and answers matching those in the Compendium version, while others contain only a few of the same questions and answers, in a different order and/or interspersed with questions and answers not found in the Single Method.
|| 11 Daishun Ueyama (1983, 331-32) discerned three texts corresponding to our parts (a), (b) and (c) of the Single Method. This was based on a thematic reading of the text; however, as we know that single texts may well contain several thematic sections, a thematic reading is not a sufficient method in itself for distinguishing textual units. Ryutoku Kimura (1980) correctly considered the whole to be a single textual unit under the title found in the explicit, on v47.2. For evidence against Ueyama's consideration of parts (a) and (b) as separate texts, we have another manuscript, Pelliot tibétain 118, which contains part of the Single Method with no visual clue that the scribe considered there to be any distinction between these two parts of the text. Indeed the visual clue in Pelliot tibétain 116 itself is debatable, merely a gap between two sets of double shad (Tibetan punctuation marks usually appearing as long vertical lines) on the same line, at v23.2, and thus not even a line break, which is the minimum demarcation between texts elsewhere on this manuscript. As for the distinction between parts (b) and (c), there is simply no clue on the page at all. Note that Ueyama's (1983, 334) statement that there is text missing between v40 and v41 is mistaken; there is a clear link between the citation of texts in parts (a) and (b) and the précis of the teaching of Chan masters in part (c) in the line which bridges the transition (v40.3-4): 'Thus the simultaneous method does not contradict the words of many sutras; it is also in harmony with the instructions of learned masters.' Following this there is a reference to the master Nāgārjūna as prophesied in the Laṅkāvatāra sūtra which crosses directly from v40.4 to v41.1.

Manuscript
Corresponding Further reason to believe that the Single Method existed elsewhere as a textual unit is found in an early eighth-century Tibetan library catalogue known as the Phang thang ma (see Rta rdo 2003, 58). The listing of a text with the same title in this catalogue suggests that the Single Method was known in Central Tibet. 12 Furthermore, sources from the Tibetan literary tradition that have survived in the transmitted literature also contain some of the same citations found in the Single Method. The Lamp for the Eyes of Contemplation, written in the early 10 th century, contains three scriptural citations and five quotations from Chan masters that are also found in the Compendium. 13 A slightly earlier canonical text, The Meaning of || 12 The catalogue entry is: (833) Dmigs su med pa'i tshul gcig pa'i gzhung (1/2 bam po). Other Chan collections of sutra quotations are found in the same part of the catalogue: (831) Mdo sde brgyad bcu khungs (1 bam po), which also exists in Pelliot tibétain 818; (836) Theg pa chen po gcig car 'jug pa (1 bam po), which is probably the same text as P.5306. Other titles in the same part of the catalogue also appear to be Chan texts. The section is attributed to Khri srong lde btsan, but according to Bu ston this section was headed 'name of author unknown' (see Faber 1985, 49-51). 13 This is the Bsam gtan mig sgron of Nub Sangyé Yeshé (Gnubs Sangs rgyas ye shes).
Nonconceptual meditation of the Simultaneous Approach, shares nineteen scriptural citations and two quotations from Chan masters. 14 Along with the library catalogue, these occurrences suggest that the Single Method and its composite textual units were known in Tibet (and not just Dunhuang) from the late 9 th century, and this popularity is reflected by its inclusion as the central text of the Compendium. 15

Ritual practice and the Tibetan Chan Compendium
The traditional Tibetan narrative of a decisive debate between a Chan monk and a scholarly Indian opponent has influenced the way scholars have approached the Tibetan Chan manuscripts, and the Compendium in particular. In the story, the Tibetan emperor Tri Song Detsen (r. 756-c.800) convened the debate due to antagonism between Chinese and Indian Buddhists at court. The Chinese side was represented by a Chan teacher known as Moheyan 摩訶衍 (a Chinese transliteration of the Sanskrit mahāyāna). The story of the debate is derived from an old historical narrative known as The Testament of Ba, which represents the debate as a battle between exponents of the simultaneous and gradual approaches to enlightenment. 16 Moheyan is portrayed as representing the extreme position of rejecting all Buddhist practice apart from a recognition of the mind's true nature, which is said to lead to 'simultaneous entry' (cig car 'jug pa However, this narrative is contradicted by a similar account in a Chinese Dunhuang manuscript which ends with the emperor endorsing Chan. 18 A number of works on the Compendium and other Tibetan Chan manuscripts have framed the whole enquiry in terms of this debate story. 19 This has led to the Compendium being approached from a purely doctrinal point of view, as if its only reason for being was its relationship with the debate between the opposed positions of the simultaneous and gradual approaches to enlightenment. It is certainly true that some of the texts in the Compendium do show an awareness of this doctrinal tension -but this is not specific to Tibetan Chan; it is also seen in the Chinese Chan manuscripts. In fact, the tension between immediate access to one's true nature and the need for a graduated path of practice crops up again and again in various Buddhist traditions.
So, it may be better to put the debate narrative to one side and look afresh at the uses that the Compendium might have been put to. It may be more useful to substitute social and ritual contexts for doctrinal ones. 20 My reading of the Compendium suggests that the most relevant social and ritual context here is the ceremony of taking the precepts of a bodhisattva. This is a series of vows found only in Mahāyāna Buddhism, and directed to the aspiration of the bodhisattva: to strive for the enlightenment of all sentient beings. The bodhisattva precepts ceremony originated in India, but became especially popular in China, where mass precepts ceremonies were held on specially constructed platforms.
The popularity of these platform ceremonies coincided with the emergence of self-conscious Chan lineages during the 8 th century, so that, as Wendi Adamek has put it, 'Chan can be said to have been born on the bodhisattva precepts platform.' 21 Taking the precepts of the bodhisattva also entailed entering a Chan lineage. The || 18 The Chinese version of the debate story is in Pelliot chinois 4646, titled Dunwu dasheng zhengli jue 頓悟大乘政理決. It was extensive discussed, and translated into French, in Demiéville 1952. The manuscript consists of a series of questions and answers on Chan doctrines, with a preface by the monk Wangxi 王锡 stating that the background to these questions and answers was the patronage of Chan masters by the Tibetan emperor Tri Song Detsen and one of his queens. According to Wangxi there were a series of discussions between the Indian teachers at the Tibetan court and the Chan teacher Moheyan. In contrast to the Tibetan debate narrative, Wangxi concludes his preface with an edict from the Tibetan emperor supporting Chan. 19 See for example Ueyama 1983 andFaber 1985. 20 The ritual content of Chan / Zen Buddhism has been occluded by anti-ritual rhetoric within the tradition, and in Western appropriations of Zen. Some recent scholarship has attempted to redress the balance. See for example the papers in Faure 2003. 21 Adamek 2011, 33. See also Adamek 2007 for a detailed discussion of the historical development of the precepts ceremony in China, with regard to Chan lineages. importance of the platform ceremony in Chan lineages is also evident among the Dunhuang manuscripts. For example, one of the most popular early Chan texts, the Platform Sutra (which is found in several versions among the Dunhuang manuscripts) is constructed around an ordination sermon by the sixth patriarch Huineng 惠能. Another platform sermon by Huineng's disciple Shenhui 神會 is also found in the Dunhuang manuscripts. 22 How then does the arrangement of the texts in the Compendium suggest the context of a precepts ceremony? The Compendium begins with two popular Mahāyāna Buddhist texts, starting with the Prayer of Good Conduct (Skt. Bhadracaryā-praṇidhana) which sets out the aspirations to bring about the welfare and enlightenment of all beings. 23 This is the aspiration of the bodhisattva, which is formalized in Buddhist practice by the ceremony of taking the bodhisattva precepts. The presence of the Prayer of Good Conduct at the beginning of the Compendium is the first clue that the manuscript may have been made for use in such ceremonies.
The Prayer of Good Conduct is followed by the Diamond Sutra (Skt. Vajracchedikā-prajñāpāramitā-sūtra), one of the most popular expositions of the doctrine of emptiness, which states that all things are interdependent, and thus nothing can have an intrinsic essence. 24 In this scriptural text the Buddha repeatedly makes contradictory statements, celebrating the virtuous path of a bodhisattva and the qualities of a Buddha at the same time as denying that they exist. This approach is a || 22 For a translation of the Platform Sutra, see Yampolsky 1967. On the Shenhui sermon in Pelliot chinois 2045, see Liebenthal 1952. 23 The longer title of this prayer is Āryabhadracaryāpraṇidhānarāja. It was translated into Tibetan in the 8 th h century, and has ever since been hugely popular in Tibet. The first complete Chinese translations of the prayer were made by Amoghavajra (Bukong jingang 不空金剛) and Prajñā (Bore 般若) in the 8 th century. See Dessein 2003 for a survey of the literary history of the prayer in China. 24 The presence of the Diamond Sutra in the Compendium is by no means unique; as we saw in Tab. 2 above, the sutra is also found on the reverse side of two concertina manuscripts containing the Chan compilation entitled The Single Method of Non-Objectification: IOL Tib J 707 and Pelliot tibétain 118. The eminence of the Diamond Sutra in these manuscripts challenges an assumption that has been repeated in a number of studies of the Tibetan Chan texts, namely that main influence seen there is from the so-called Northern School. This position, in the English language sources, goes back to Jeffrey Broughton's (1983) introduction to Tibetan Chan. Broughton identified three schools as central to the history of Chan in Tibet: 'the Reverend Kim or Ching-shung lineage, the Wu-chu or Pao-t'ang lineage, and the post-Shen-hsiu Northern lineage, the last of which we might call the late Northern.' In the Compendium the influence of the Southern lineage is further shown by a section on the sayings of Shenhui, who is considered its founder. Equally, the presence in the Compendium of figures associated with other lineages shows that Shenhui's polemics had not resulted in the rejection of material associated with the Northern School. challenge to dualistic concepts of self and other, existence and non-existence, and the like. This use of deliberate paradox as a teaching method had a strong influence on the development of the Chan tradition. 25 The Diamond Sutra occupies a central place in the Platform Sutra, which begins with the story of how Huineng became the sixth patriarch of Chan. Huineng is said to have left home and gone in search of the fifth patriarch of Chan after hearing the Diamond Sutra being recited in the marketplace. Later in the narrative, the fifth patriarch transmits his authority and wisdom to Huineng by explaining the Diamond Sutra to him. After this biographical sketch, the Platform Sutra turns into a sermon given by Huineng in a ceremony of bestowing precepts. This ceremony begins with taking refuge in the Buddha, his teachings and the community of monks and lay practitioners. Then follow the vow of the bodhisattva (equivalent to the Prayer of Good Conduct in the Compendium), and an exposition of the doctrine of the Perfection of Wisdom, with particular reference, again, to the Diamond Sutra. 26 Thus the first and second texts in the Compendium mirror the concerns of ordination sermons like the Platform Sutra: bestowing the precepts of the bodhisattva and orienting those receiving the precepts to a particular scriptural tradition, that of the Perfection of Wisdom in general and the Diamond Sutra in particular. After these two texts, the Compendium turns to less well-known material that is specific to Chan lineages. Text III is an overview of the differences between the Mahāyāna and the lesser vehicles, a popular theme among the Tibetan Chan texts. 27 This text also has an analogous section in Shenhui's sermon. 28 Text IV explains that the correct 'view' (Tib. lta ba, Skt. darśana) is the sameness of all entities.
After this we have three substantial Chan texts, arguably the centrepiece of the Compendium. Text V is the Single Method, a substantial compilation in question-and-answer form, with quotations from sutras and Chan masters. Text VI gathers together the sayings of 19 masters of meditation, some overlapping with the previous text. Text VII is a translation of a Chinese Chan text that is also found in the Chinese manuscripts from Dunhuang. Finally, short texts round off the collection: Text VIII addresses certain faults in meditation practice and their remedies, and Text IX is a poem or song on the ultimate state of reality.
|| 25 See the discussion in Nagatomo 2000. 26 The centrality of the Diamond Sutra to the Platform Sutra has led Christoph Anderl (2013) to suggest that the term 'platform sutra' first referred to the Diamond Sutra itself, and only later became the name of Huineng's text by extension from this use. 27 See for example IOL Tib J 709/10. 28 See Liebenthal 1952, 141-42. Much in this material continues to mirror the platform sermons of Shenhui, and the Platform Sutra itself. In both, the explanation of the perfection of wisdom is followed by an introduction to the view of the Chan, along with a discussion about meditation and how to avoid going astray. Huineng's text cites sutras, while Shenhui's contains an injunction to read the sutras. Both texts end with a song (in Shenhui's sermon, several songs), and so does the Compendium. Therefore I think it is reasonable to say that the logic behind the organization of the texts in the Compendium is the logic of the platform ceremonies which functioned as a monastic or lay ordination into a Chan lineage. 29 Here we should note Alessandro Bausi's concept of MTMs as 'corpus organizers' -Bausi writes that: Far from being conceived as an autonomous and well-defined witness of texts (as it would appear from a purely philological perspective), each of these 'corpus-organizers' acquires its full significance only in mutual relationship to the others. Each manuscript organizes an implicit, but nevertheless material and concrete, evolving knowledge. 30 I would agree that the Compendium represents the state of an evolving knowledge, and add that the manuscript also exists in the foreground of a social picture in which ritual practice is the means by which this knowledge is transmitted from one generation to the next. The specific form of ritual here is the platform ceremony of bestowing the bodhisattva precepts, and the arrangement of texts reflects the method by which the transmission was effected.
The issue of a manuscript's social background brings us to our final question about the Compendium: who paid for it? Unfortunately the repairs carried out on the manuscript by diligent scribes resulted in the complete replacement of the original first and last panel, so we do not have the original colophon. We can, however, get some idea of what kind of colophon might have been appended to the texts by looking at a complete MTM. Pelliot tibétain 98 is a concertina in 81 panels containing a series of dhāraṇī -texts containing magical formulae, recited for various wordly purposes. The last panel contains a colophon giving the date of copying, the name of the sponsor, and the dedication of the merit of writing the texts. || 29 It should be noted that there are other, shorter, MTMs among the Tibetan Chan manuscripts, and some of these (e.g. IOL Tib J 710) have quite a different structure (and therefore, presumably, had a different use) to the Compendium. There are also many multiple-text Chan manuscripts in Chinese (Sørensen 1989, 134). 30 Bausi 2010, 35. At the beginning of the middle month of autumn in the year of the dragon: by the merit of writing these scriptures, may the noble life-force of the patron Ba Tsesyong be increased. May all of his excellent aspirations be fulfilled. May his good works and high position be greatly praised in this world. May his power increase, and may he never ever be struck by any illness or obstacles. May he purify the two kinds of obscuration and complete the accumulations of merit and wisdom. 31 From elsewhere in the manuscript, we know that Ba Tsesyong was a minister. Colophons like this remind us that large and well-made manuscripts like the Compendium needed the funds of a person, or organization of some standing. And although we do not know who sponsored the creation of the Compendium, we can now see that it was copied for a purpose, to be used in Chan precepts ceremonies. The size and relative expensiveness of the Compendium itself shows the importance of the ordination ceremony to the community in which the manuscript was created and used.

Conclusions
Are there any general conclusions regarding MTMs that may be drawn from this investigation into the Tibetan Chan Compendium? Most important, perhaps, is the reminder that the reasons behind the creation of the manuscript, the specific texts written in it, and the sequence in which they are put, may be explicated by attention to the manuscript's socio-historical background. This requires us to attend to the physical nature of the manuscript, including evidence of usage (like repairs), and simultaneously to the choice of texts and their arrangement. Then, we need to look further afield for sources that may provide clues to the sequencing of the texts in the manuscript.
In the case of the Compendium, we found these in Chan ordination sermons like the Platform Sutra. Historical studies of the popularity in China of ordination ceremonies held on ritual platforms then provided the key to understanding the role of the Compendium. Once we understood how the manuscript could have functioned as a part of such practices, the arrangement of the texts within it and the motivation for the creation of the manuscript itself became explicable. || 31 Pelliot tibétain 98: 'brug gi lo ston sla 'bring po'i ngo la// yon bdag 'ba' tse syong tshe ring dpal 'phel/ bsam ba legs dgur grub/ myi 'phan srid mtho/ 'jig rten 'dir che bar grags shinbg btsan la dar pa dang/ nad bgegs cis kyang myi tshogs shig ni phyung/ bsod nams dang/ ye shes kyi tshogs rdzogs so// There is a further extension of this interplay between our attention to the specific features of the manuscript on the one hand, and to the socio-historical background of its creation on the other: the manuscript itself may now serve to illuminate the nature of the social practices for which it was created. Thus the Compendium, now that we have identified it as relating to Chan ordination ceremonies, becomes a unique source for understanding the way these rituals were conducted at the place and time of its creation.