23

1. The Dīrghāgama Manuscript

The Prāsādika-sūtra (DĀ 15) and the Prasādanīya-sūtra (DĀ 16), which are the subject of this book, are two paired sūtras from the Yuga-nipāta (the section dealing with thematically paired sūtras) contained in the manuscript witness of the Dīrghāgama (Long Discourses) of the Mūlasarvāstivāda¹ tradition. The Dīrghāgama is a compilation of ancient, canonical Buddhist discourses transmitted in Sanskrit and written on birchbark folios in the Gilgit-Bamiyan Type II script, also known as Proto-Śāradā. This collection had been lost for centuries and was likely rediscovered somewhere within the border area of Afghanistan and Pakistan in the 1990s. Among the longest extant manuscripts found in the area, originally consisting of over 450 folios, it contains forty-seven individual texts. One cannot specify with certainty the location of the find spot of the manuscript because it only came to the attention of scholars after it had appeared on the rare book market in London, where it was identified by Kazunobu Matsuda and shortly thereafter subjected to preliminary examination by Lance Cousins, Somadeva Vasudeva, and Klaus Wille.

While it is regrettable that the provenance remains unknown, it is believed to be either another part of the cache of manuscripts found in the 1930s at the Gilgit site in Pakistan, which was historically part of the area we refer to as Greater Gandhāra, or to have been found nearby within the vicinity. Based on paleographical analysis and radiocarbon dating, the manuscript is thought to date from between 676–776 CE,² and the script, suggesting a production after the sixth century, allows us to conclude that it may have been copied sometime around the eighth century of the Common Era.³ While this particular manuscript witness of the Mūlasarvāstivāda Dīrghāgama was likely created in the eighth century, the Dīrghāgama as a work of oral literature is much older and was likely composed centuries before.⁴

Currently, over half of the Dīrghāgama manuscript is split into four private collections, two in Japan and one each in Norway and the United States. The whereabouts of the rest of the manuscript remain a mystery, one that will hopefully be solved in time. The folios in three of the four private collections, while fragmentary in many places, have fortunately been subjected to high-resolution scans,⁵ allowing scholars to study the texts independently of the physical folios, which remain housed with their respective owners.

While we are fortunate that high-quality photos of the manuscript are available, the folios comprising the Prāsādika and Prasādanīya sūtras are damaged throughout, in some places quite heavily, and the text itself is often problematic, displaying instances of negation where there should be none and the opposite along with no small amount of textual ambiguity. As Hartmann notes:

This has compounded the already philologically complicated process of creating the reconstruction, much like piecing together a jigsaw puzzle. However, I have been successful in fitting together damaged sections of the manuscript and have been able to reconstruct the missing passages based on the many Sanskrit and Pali parallels to the text that may be found within extant Buddhist literature; whenever I have a textual parallel, I have been able to reconstruct the text with a reasonable degree of confidence, and the great majority of both have been reconstructed. However, in the rare instances when both the manuscript is damaged and there are no extant parallels, such portions of the text must remain lost until either the missing part of the manuscript is found or a similar parallel is discovered.

The Prāsādika-sūtra and Prasādanīya-sūtra together run to twenty-six folios, with the Prāsādika-sūtra spanning seventeen (fols. 274v5–290r4) and the Prasādanīya-sūtra ten (fols. 290r5–299v3). The folios of the Prasādanīya-sūtra, while damaged throughout, are all extant. The Prāsādika-sūtra on the other hand was missing nearly a quarter of its folios when I began working on it. I have been able to identify and partially restore all of the folios and partially or completely reconstruct the text of all previously missing folios. Some I was able to piece together: some folios from various fragments were missing their folio numbers; others I discovered were obscured by other folios due to the folios being stuck together. The reason so many folios were stuck together was made clear to me when Kazunobu Matsuda kindly sent me pictures of the folio bundles he had 6managed to photograph when he first encountered them in the possession of the Sam Fogg rare bookseller in London where the Dīrghāgama manuscript first became known to scholars. One of those pictures (fig. 1) shows that the manuscript bundle had been separated into at least three smaller bundles at some point in its history. It is not clear whether the manuscript was divided after it was found or some time in the centuries before that, but the bundles must have been together in one bundle for a long period in the past, since layers of adjacent folios are fused together and stuck to the topmost folios in the bundles. This fusing of birchbark folios is not uncommon in recovered manuscripts from Greater Gandhāra, especially for folios at the top or bottom of a manuscript bundle.

The Prāsādika-sūtra was split between the top and bottom of two of the bundles, where folios seem to have been especially in danger of becoming fused together, due perhaps to exposure to the elements. Due to their being so close to the top and bottom of these bundles, they appear to have suffered an unusual amount of damage compared to those of some other sections of the manuscript when the individual folios were eventually separated. Fortunately, my digital restoration of the manuscript via Photoshop has been successful in allowing for the majority of the damaged folios to be restored. However, several folios still remain fused and should be reexamined by a skilled conservator.

Due to the damaged and fragmentary quality of the manuscript folios making up the two sūtras, I have had to spend considerable time and effort identifying fragments from folios that were present on other, unrelated folios and digitally restoring these fragments to their correct place in the manuscript. As noted, I have had great success using Adobe Photoshop as my main tool for digital restoration of folios. Of the twenty-six folios containing the Prāsādika-sūtra and Prasādanīya-sūtra, well over half have undergone digital restoration, ranging from small fragments of a few akṣaras (the graphemes used in Brahmic scripts) to an entire folio (see figs. 2–4). The most common digital restoration consists of fragments from one line up to one fourth of the folio in length being moved from the folio on which they rested to their correct folio in the manuscript. Fragments that have been digitally restored are generally not commented upon in the transliteration, which reads the folios in their most complete state. Indeed, the digital restorations I have created supersede all other images of the folios they represent. While I have identified the correct location and digitally restored some fragments to their proper folio, in the transliteration 7I have also retained them as fragments in the place where they were found in the manuscript, as removing them from the folio on which they rested would have damaged the fidelity of that folio. In such cases, these fragments are noted in the transliteration with their proper location given. Fragments that remain unidentified are naturally also noted in the transliteration. Such fragments are generally quite short, often consisting of no more than a few illegible akṣaras.

FIGURE 1. Folio bundles of the Dīrghāgama manuscript with the uppermost folios fused together. Photo courtesy of Kazunobu Matsuda.

FIGURE 2. Damaged section of folio 298r7.

8

FIGURE 3. Digital restoration of folio 298r7 using fragment that had been fused to 299r7.

FIGURE 4. Partially restored folio 285r restored from five disparate fragments found in different locations of the manuscript and birchbark cells that had been transposed onto other folios. Note that the akṣaras in the upper right corner are upside down due to their having been restored from transposed birchbark layers.

1. To be more precise, it is not certain whether this Dīrghāgama manuscript should be classified as belonging to the Sarvāstivāda or Mūlasarvāstivāda tradition. While it is clear in terms of vinaya that the Sarvāstivāda and Mūlasarvāstivāda traditions had unique transmission, it is not clear whether this was the case with their āgama transmissions nor what, if anything, the distinctions between these traditions may have been if they shared āgamas or had separate āgama transmissions. Hartmann has discussed this issue in the past (see Hartmann 2014, 140n5) and has suggested the hyphenated term (Mūla-)Sarvāstivāda to make this ambiguity clear. Anālayo, however, rejects such terms and suggests that differences in āgama transmissions between Sarvāstivāda and Mūlasarvāstivāda traditions may be explained as resulting from the transmission lineages of these traditions having been formed regionally following ordination lineages and that the oral transmission of these works played a key role in these distinctions (see Anālayo 2020a, 404ff.). It might be noted that the beginning of the Prasādanīya-sūtra edited here is quoted within Śamathadeva’s Abhidharmakośaṭīkopāyikā (see DĀ 16.1–2 below), which suggests to me that this Dīrghāgama manuscript may cautiously be considered as Mūlasarvāstivāda. For reasons such as this, throughout this book I refer to the Dīrghāgama manuscript witness as a Mūlasarvāstivāda work but the inherent ambiguity in such terminology should be implicitly understood by the reader.

2. Personal correspondence via email between the author, Kazunobu Matsuda, Jens-Uwe Hartmann, and Jin kyoung Choi on November 19, 2019. This new radiocarbon dating has resulted from current research being conducted by Fumi Yao. Yao also had the Bhaiṣajyavastu-Uttaragrantha manuscript she has worked on tested, with a remarkably similar result of 671–770 CE. This may speak to a strong relationship in the production of the Dīrghāgama manuscript and the Mūlasarvāstivāda Vinayavastvāgama manuscripts found in Gilgit.

3. Cf. Hartmann and Wille 2014, 137, and Hartmann 2014, 155. Previous radiocarbon dating of the Dīrghāgama manuscript discussed in these two publications give a range of 764–1000 CE. These ranges are based on testing commissioned by the Sam Fogg dealer (first reported in Allon et al. 2006, 279) and are now superseded by the new radiocarbon dating performed in 2019 as part of Yao’s research described in note 2 above.

4. The Long Discourse collections are held to be buddhavacana by all Buddhist traditions and are therefore considered to have been spoken by the historical Buddha Gautama, transmitted orally for generations, and only later compiled in writing as a distinct collection and eventually translated into various languages. This of course makes discussion of composition somewhat complicated. In this process of textual transmission, many manuscript witnesses of these texts were produced via a copying tradition that spanned hundreds of years.

5. Unfortunately, the folios in the Hirayama Collection are only available in low-resolution images.

6. Hartmann 2014, 155.