FROM BIRCH BARK TO DIGITAL DATA: RECENT ADVANCES IN BUDDHIST MANUSCRIPT RESEARCH

Transcription

1 FROM BIRCH BARK TO DIGITAL DATA: RECENT ADVANCES IN BUDDHIST MANUSCRIPT RESEARCH Papes Pesented at the Confeence Indic Buddhist Manuscipts: The State of the Field Stanfod, June Edited by Paul Haison and Jens-Uwe Hatmann

2 ÖSTERREICHISCHE AKADEMIE DER WIS SEN SCHAF TEN PHILOSOPHISCH-HISTORISCHE KLASSE DENKSCHRIFTEN, 460. BAND Beitäge zu Kultu- und Geistesgeschichte Asiens N. 80 Heausgegeben on Helmut Kasse

3 ÖSTERREICHISCHE AKADEMIE DER WIS SEN SCHAF TEN PHILOSOPHISCH-HISTORISCHE KLASSE DENKSCHRIFTEN, 460. BAND Fom Bich Bak to Digital Data: Recent Adances in Buddhist Manuscipt Reseach Papes Pesented at the Confeence Indic Buddhist Manuscipts: The State of the Field Stanfod, June Edited by Paul Haison and Jens-Uwe Hatmann

4 Vogelegt on w. M. ERNST STEINKELLNER in de Sitzung am 8. Oktobe 203 Diese Publikation wude einem anonymen, intenationalen Pee-Reiew-Vefahen untezogen. This publication has undegone the pocess of anonymous, intenational pee eiew. Die ewendeten Papiesoten sind aus chlofei gebleichtem Zellstoff hegestellt, fei on säuebildenden Bestandteilen und alteungsbeständig. Alle Rechte obehalten. ISBN Copyight 204 by Östeeichische Akademie de Wissenschaften Wien Duck und Bindung: Duckeei Fedinand Bege & Söhne GesmbH, Hon Pinted and bound in the EU

5 TABLE OF CONTENTS PAUL HARRISON AND JENS-UWE HARTMANN, Intoduction... ii RICHARD SALOMON, Gāndhāī Manuscipts in the Bitish Libay, Schøyen and Othe Collections... MARK ALLON, The Senio Khaoṣṭhī Manuscipts... 9 COLLETT COX, Gāndhāī Khaoṣṭhī Manuscipts: Exegetical Texts HARRY FALK AND INGO STRAUCH, The Bajau and Split Collections of Khaoṣṭhī Manuscipts within the Context of Buddhist Gāndhāī Liteatue... 5 OSKAR VON HINÜBER, The Gilgit Manuscipts: An Ancient Buddhist Libay in Moden Reseach JENS-UWE HARTMANN AND KLAUS WILLE, The Manuscipt of the Dīghāgama and the Piate Collection in Viginia JENS BRAARVIG, The Schøyen Collection KAZUNOBU MATSUDA, Japanese Collections of Buddhist Manuscipt Fagments fom the Same Region as the Schøyen Collection LORE SANDER, Dating and Localizing Undated Manuscipts... 7 KLAUS WILLE, Suey of the Sanskit Manuscipts in the Tufan Collection (Belin) JENS-UWE HARTMANN AND KLAUS WILLE, The Cental Asian Sanskit Fagments in the Pelliot Collection (Pais) KLAUS WILLE, Suey of the Identified Sanskit Manuscipts in the Hoenle, Stein, and Skine Collections of the Bitish Libay (London) JENS-UWE HARTMANN AND KLAUS WILLE, Futhe Collections of Sanskit Manuscipts fom Cental Asia SHIN ICHIRŌ HORI, Fom the Kathmandu Valley to the Taim Basin DUAN QING, Indic and Khotanese Manuscipts: Some New Finds and Findings fom Xinjiang PAUL HARRISON, Ealie Inentoies of Sanskit Manuscipts in Tibet: A Synoptic List of Titles SAERJI, Indic Buddhist Manuscipts in the People s Republic of China: The Peking Uniesity Poject HELMUT KRASSER, Indic Buddhist Manuscipts in Vienna: A Sino-Austian Co-opeatie Poject, with Methodological Remaks on Śāstic Utexts LUO HONG, Sanskit Manuscipt Pojects in the China Tibetology Reseach Cente YOSHIYASU YONEZAWA AND JUNDŌ NAGASHIMA, The Sanskit Manuscipt Reseach Poject at Taisho Uniesity MICHAEL HAHN, Vaious Aspects of Dealing with Buddhist codices unici PETER SKILLING, Reflections on the Pali Liteatue of Siam BHIKKHU ÑĀṆATUSITA, Pali Manuscipts of Si Lanka

6 Paticipants in the confeence Indic Buddhist Manuscipts: The State of the Field held at Stanfod Uniesity, June 5 9, 2009 Font ow (fom left to ight): DUAN Qing (Beijing), Richad Salomon (Seattle), Iene LIN (Stanfod), Collett Cox (Seattle), Michael Hahn (Mabug), Helmut Kasse (Vienna), Haunaga Isaacson (Hambug), Oska on Hinübe (Feibug), Kazunobu Matsuda (Kyoto), LUO Hong (Beijing), Saeji (Beijing). Second ow: Jundō Nagashima (Tokyo), Pete Skilling (Bangkok), Shin ichio Hoi (Tokyo), Jens Baaig (Oslo), Loe Sande (Belin), Jens-Uwe Hatmann (Munich), Ingo Stauch (Belin), Klaus Wille (Göttingen), Yoshiyasu Yonezawa (Tokyo). Thid ow: Mak Allon (Sydney), Paul Haison (Stanfod). In absentia: Hay Falk (Belin).

7 The Bajau and Split Collections of Khaoṣṭhī Manuscipts within the Context of Buddhist Gāndhāī Liteatue HARRY FALK (BERLIN) AND INGO STRAUCH (LAUSANNE). Intoduction Duing the last decade the collections of Gāndhāī manuscipts being studied by the Ealy Buddhist Manuscipts Poject in Seattle hae been supplemented by two new manuscipt finds which ae said to oiginate fom the nothwest of Pakistan, o moe pecisely fom the Bajau distict in the Khybe Pakhtunkhwa (fome Nothwest Fontie Poince). The Bajau Collection is named afte the find-spot on the nothen edge of the egion Bajau on the bode with Afghanistan, while the Split Collection has no attested place of oigin. Since some moe pats coming fom the same souce hae sufaced in the antiquities maket, pointing at a segmentation o split-up of the oiginal lot, the tem Split Collection is used to allow futhe acquisitions to find a home unde the same oof. Both new collections contain impotant additions to the hitheto known copus of Gāndhāī liteatue, among them two texts which clealy belong to the Mahāyāna banch of Buddhism. The Bajau Collection houses what is to this date the lagest manuscipt of a Mahāyāna sūta in the Gāndhāī language. Although it esembles in many egads aious genes of ealy Mahāyāna including Pajñāpāamitā texts and Pue Land Buddhist texts like the Akṣobhyayūha it is not identical with any known sūta. The Split Collection contains the ealiest known manuscipt of a Pajñāpāamitā text. The contents of both collections can consequently be used to get some new data fo one of the most contoesially debated questions in the histoy of Buddhism, namely the doctinal and institutional contexts in which ealy Mahāyāna aose. This question is of couse closely connected with the poblem of the specific chaacte of Gandhāan Buddhism and its liteay poduction. It is only possible to ealuate the eidence of a gien text o text collection by appopiately consideing this extended backgound of Gāndhāī Buddhist liteatue and defining the position of the espectie textual mateial theein. Howee, despite the enomous pogess made in Gandhāan studies in the last decade mainly thanks to the wok done by the membes of the Ealy Buddhist Manuscipts Poject unde the guidance of Richad Salomon such a suey of Gāndhāī liteatue which subsumes the main aailable data fom the pespectie of liteay genes is still a desideatum. Ou pape will theefoe intoduce the new collections studied in Belin i.e. the Bajau and Split Collections by contasting thei inentoy of texts with the cuently aailable copus of Gāndhāī liteatue. Thus it will also poide a base fo futhe eseach by oganizing the athe dispaate infomation about this copus accoding to the aailable publications in a concise and systematic fom. Due to the ongoing o sometimes een not yet initiated eseach on many of the texts cited in this suey the infomation must be egaded as peliminay, to be supeseded by a futue compehensie histoy of Gāndhāī liteatue which emains to be witten. Befoe coming to this oeiew of liteay genes it is, howee, necessay to gie a shot intoduction to the main physical featues of the Bajau and Split Collections and the histoy of thei discoey and eseach. 2. Gāndhāī Studies in Belin 2. The Bajau Collection of Khaoṣṭhī manuscipts The Bajau Collection of Khaoṣṭhī Manuscipts was discoeed in 999 in the uins of a Buddhist monastey nea the illage of Mian Kili in the Di distict. The monastey itself is

8 52 HARRY FALK AND INGO STRAUCH situated on the opposite bank of the ie Rud which maks the bounday of the moden disticts Di and Bajau. Hence it was decided to name the collection Bajau Collection. The monastey has not been excaated no is it descibed in seconday liteatue. Its position can be best explained by the existence of a tade oute connecting the Swat alley ia two passes towads the Kuna alley and thence towads Nangaha o Cital. In the yea of thei discoey the manuscipts wee handed oe to M. Nasim Khan, at that time Assistant Pofesso in (now Pofesso and Head of) the Depatment of Achaeology of the Uniesity of Peshawa. In Mach Apil 2004 the authos of this aticle spent a couple of days at the Peshawa Uniesity fo a seies of lectues. Duing this peiod a lage set of bich-bak fagments wee shown to them by M. Nasim Khan who had secued them, unolled them and put them into glass fames accoding to the method which had been descibed by Richad Salomon in his gound-beaking monogaph on the Bitish Libay bich-bak scolls (999). Soon the idea came up to establish a special manuscipt poject unde a moe compehensie umbella named Pak-Geman pojects, which wee supposed to compise also achaeological, ethnological and geogaphical components. As a fist step towads cataloguing the manuscipts they wee gien peliminay signatues in Apil 2004 by Ingo Stauch, with a iew to claifying the numbe of manuscipts and scibes and the inteelationship of the fagments. Afte a pepaatoy phase of 8 months, the Bajau Collection Poject financed by the Deutsche Foschungsgemeinschaft (DFG) stated its wok in Octobe 2005 at the Feie Uniesität Belin with Hay Falk as its supeiso, Ingo Stauch as Scientific Collaboato and Andea Schlosse as Student Collaboato. On the Pakistani side, M. Nasim Khan and Sohail Khan eceied equipment and financial suppot and epesented the poject in Peshawa. Duing the initial phase fom Octobe 2005 to May 2007 high-esolution digital images of the fames wee pepaed by the Pakistani patnes which still see as the basis fo ou econstuction and editing wok. On this basis a fist peliminay catalogue with sample extacts and tanslations was completed and published online on the homepage of ou poject, which went public in August This online aticle was eisited and slightly ewoked in May 2008 accoding to a seies of aluable comments, paticulaly fom the side of ou colleagues fom Seattle, with whom a athe close coopeation deeloped (Stauch 2007/8). The main esults of this peliminay suey wee published in Stauch Due to seeal easons the intended collaboation with the Pakistani side failed. Since 2007 the Bajau Collection Poject has theefoe been acting independently fom the Depatment of Achaeology of the Peshawa Uniesity. The focus of the poject s pesent wok is diected towads the final edition of fou manuscipts fom a Śāakayāna canonical o paacanonical backgound (BajC, 3, 5,3) and towads the fist exploation of the contents and chaacteistics of the collection s most impotant text, the Mahāyāna sūta BajC 2 (Stauch 200). The Bajau Collection compises altogethe fagments fom about 9 diffeent bich-bak scolls, containing aound 22 diffeent texts witten by at least 8 diffeent scibes. They ae now peseed in 35 fames. Thei extent is athe diffeent anging fom a small manuscipt measuing only 6 cm in length (BajC 5) up to a lage scoll oe 2 m long (BajC 2). It cannot be uled out that futhe eseach will eeal that some of the fagments and scibes teated hee sepaately ae in fact identical. The numbes ae theefoe poisional.

9 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 53 Figue : Pat of the Bajau bich-bak manuscipt 2 containing an ealy Gāndhāī Mahāyāna sūta with the junctues indicated, patially econstucted. Coutesy: The Bajau Collection Poject, Feie Uniesität Belin. The fomats epesented in the Bajau Collection coespond to the aieties known so fa. Out of the 9 scolls eleen belong to the so-called long and naow type of scoll meas-

10 54 HARRY FALK AND INGO STRAUCH uing 8 cm in width. Thei length can ay consideably. Up to 45 cm they ae fomed by a single sheet of bich bak. Scolls exceeding this length ae glued togethe foming theeby the so-called composite scolls (see Figue peceding page). The emaining manuscipts belong to the wide and shot fomat type, which is usually wide than 20 cm. Often they hae been folded in the middle afte haing been olled up, and consequently hae been boken etically. This damage has often esulted in the loss of consideable potions of one of the manuscipt s sides. In most cases the scolls ae inscibed on both sides with a single text. Thee ae, howee, a few manuscipts whee the eese was oiginally left uninscibed. Some of these uninscibed eeses wee late on used secondaily fo othe texts. 2.2 The Split Collection In 2005 Hay Falk was shown a few olls of bich bak by a collecto in Pakistan. They had allegedly been inspected ealie by a manuscipt deale fom Euope and ponounced fakes. Little was said about thei poenance, with hints towads Mohmand o Bajau. It soon tuned out that the oiginal find was much lage than the few olls pesented. Cetainly, the witing isible on the olls did not suppot the idea of a fogey. The popieto consented to hae the scolls opened fo a thoough inspection. This eealed that the mateial contained pats of fie diffeent texts, consisting of fou scolls and one single flat sheet. The olls wee opened and the fagments put into doublesided glass fames. All the olls ae no wide than 4 cm. Rolls which wee oiginally wide than this hae lost the excess mateial. Thei length anges fom 2 cm to 90 cm. None of them is joined to a second sheet. The quality of the bich bak is not unifom. The single sheet is of the sot of well-aged and thick bak which is, compaatiely, unusually well-peseed without flaws in its textue. The fou scolls, howee, wee made fom athe young bak, which tends to delaminate. The single sheet and all of the scolls ae inscibed at least patially on both sides. All scolls wee inscibed in diffeent hands. The handwiting is not unifom. Eey hand has its own paticulaities in espect of foms of lettes as well as modes of othogaphy. Small fagments fom two olls wee subjected to 4 C tests. One esult encompassed almost the whole ange of the last two centuies BC, with a peak in ca. 84 BC (Two sigma ange, cal., BC with a pobability of 95%). The second test yielded 74 AD as the cental date in a Two sigma ange fom 47 to 27 AD with a pobability of 8%. While the fist esult does not concu with what we beliee we know about the stages of Khaoṣṭhī and thei chonology, the second is well in accod with pesent-day iews on the deelopment of this scipt. 3. The Bajau and Split Collections within the context of Gāndhāī liteatue 3. The copus of Khaoṣṭhī manuscipts The Bajau and Split Collections continue a seies of discoeies which wee spaked off in 994 by the lage manuscipt find which was late acquied by the Bitish Libay. It is now commonly known as the Bitish Libay Collection. The following yeas witnessed futhe discoeies, which can help us gain a compehensie pictue of the liteatue which was cuent in Gandhāa between the fist centuy BC and the fouth centuy AD. By now the following manuscipt collections and single manuscipts ae aailable 2 : 2 The following aticles contain useful sueys of some of these collections: Glass 2004, Salomon

11 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 55 Name Siglum Manuscipts/ scibes Date (c. AD) Main souce of infomation Manuscipt collections Bitish Libay Collection BL 28/2 Salomon 999 Senio Collection RS 24/ 2 Salomon 2003, Allon 2007b Bamiyan fagments of the Schøyen and othe piate collections MS > 50 / > Allon and Salomon 2000, Allon et al Cental Asian fagments (Pelliot Collection, Oldenbug Collection) PC, OC 5 8/ Salomon 998, Voob ea- Desi atoskai a 2006 Bajau Collection BajC 9/8 2 Stauch 2007/8, Stauch 2008 Split Collection SplitC 5/5-2 Falk 20 Unpublished piate collection +x/+x 2 Allon and Salomon 200: Single Manuscipts Khotan Dhamapada KhDhp / 2 Bough 962 Libay of Congess Scoll LC /2 2 Salomon and Baums 2007 Uniesity of Washington Scoll UW / 2 Glass 2004: 4f. Beside these majo collections and manuscipts thee is small numbe of athe fagmentay emnants of Khaoṣṭhī manuscipts (cf. Glass 2004: 42, Salomon 999: 58 68). Accoding to thei poenance they can be diided into two goups. The fist of them compises manuscipts fom Cental Asia, as e.g. a ey small pape fagment discoeed by Sen Hedin in Loulan which does not, howee, allow fo any meaningful eading (Conady 920: 3, 9, pl. 38, no. 36). The texts on two othe pape fagments fom the Otani Collection (Hasuike 2004: 95f., no. 60) show a stong tendency towads Sanskitization and pobably belong to the late phase of Khaoṣṭhī witing in the 3d 4th c. AD. 3 The second goup is epesented by Khaoṣṭhī fagments fom Afghanistan like the fagment on palm-leaf discoeed in 834 by Chales Masson beside numeous othe fagmentay pieces in one of the sites at Haḍḍa nea Jalalabad (Wilson 84: between pp. 53 and 54, pl. III, no. ). The pesence of Khaoṣṭhī manuscipts in this egion is also confimed by the excaations of J. Bathoux in and Mustamandi in , which bought to light many small Khaoṣṭhī fagments which emained, howee, unpublished (Salomon 999: 63 65). The othe Afghanistan cente of Khaoṣṭhī witing is Bamiyan, whee the manuscipts of the Schøyen Collection ae said to hae thei oigin. Fom the same egion ae fou small Khaoṣṭhī fagments on palm leaf discoeed fom Shah-i Zuhak, twele kilometes away fom Bamiyan (Pauly 967: pl. 4, nos. E H). Anothe hitheto unpublished palm-leaf folio which was pat of the collection of the Kabul Museum can pobably be identified with the manuscipt emains discoeed by Joseph Hackin in a cae thity-fie metes east of the geat Buddha of Bamiyan (Salomon 999: 66). Thanks to a photogaph made by Hebet Hätel in the 970s (cf. Sande, infa) it is possible to detemine the language of 2006b, Allon 2007b and We will not conside hee the few documents witten in the late Fomal Khaoṣṭhī fom the Kuca and Tufan oases in Xinjiang which seems to hae been in use till the 7th c. AD. Fo a peliminay suey of these documents see Sande 999: Two of them hae been ead and tanslated by LIN Meicun (2004).

12 56 HARRY FALK AND INGO STRAUCH this text as Buddhist Sanskit and to fix its date to the late phase of Khaoṣṭhī, i.e. the 3d 4th c. AD. 4 Duing his stay in Peshawa in 2008 Ingo Stauch was shown anothe small palm-leaf fagment. Its exact poenance is not known, but accoding to its appeaance and the infomation aailable fom the owne it pobably belongs to the goup of Khaoṣṭhī manuscipts fom Afghanistan. It is published hee fo the fist time (Figue 2). 2A Figue 2: A new fagment of a Khaoṣṭhī palm-leaf manuscipt, seen in Peshawa in 2008 (Photogaph: Ingo Stauch) It eads: Side A: ///? [śatha] sakṣe pati? /// /// [śa]tha ki ṣ[a pa]? /// Side B: /// (*śa)thu pahana samodha /// ///? śa ma go 4-- [śath]. /// The text cannot be identified on the basis of the peseed potion. Peliminay study eeals that it epeatedly efes to the Buddha, called hee by his epithet Skt. śāstā (G. śatha/śathu ( teache ). 5 Othe identifiable lexemes include sakṣe, Skt. sākṣī ( witness ), pahana, Skt. pahāṇa ( abandoning ), and samodha(*na), Skt. samaadhāna, P. samodhāna ( collocation, combination ). Accoding to the inseted numeal 6, the text consists of at least seen segments. It is cetain that thee ae many moe as yet unpublished fagments of Khaoṣṭhī manuscipts in piate collections. 6 Most of them ae athe too small and fagmentay to allow any futhe conclusions about thei chaacte o contents. Theefoe they will not be consideed in this suey. Despite the gowing amount of texts it has to be sumised that they epesent only the tip of the icebeg of a much lage liteay tadition. It is pesently impossible to detemine whethe they ae epesentatie of Gāndhāī liteatue as a whole. Moeoe, the pesent peliminay state of eseach, with many texts still unedited and unstudied, peents us fom dawing definitie conclusions. Nonetheless, it is aleady possible to use the aailable infomation fo a suey of the pesence and distibution of liteay genes within Gāndhāī Buddhist liteatue. 2B 4 The Buddhist Sanskit chaacte is e.g. indicated by foms like kāeṃti (Skt. kāayanti) (3B, line 2) and catāi ayasatyāni (3A, line 3). 5 Fo śathu as nom. sg. m. cf. Lenz 200: 42. Altenatie Gāndhāī aiants include śastu and śasta. 6 Some of these fagments and the aailable eidence on lost manuscipts ae discussed by Salomon (999: 59 65).

13 The Bajau and Split Collections of Khaoṣṭhī Manuscipts The genes of Gāndhāī liteatue The table below summaizes the eidence which will be discussed in moe detail in the subsequent paagaphs. The figue in the espectie field indicates the numbe of texts belonging to this gene. Ms collection/single ms BL RS MS PC/ OC BajC SplitC UC KDhp LC UW Total Suggested date (c. AD) Numbe of manuscipts > x > 35 Śāakayāna Canonical Texts Vinaya Pātimokṣasūta 2 2 Kamaācanā 2 2 Vinaya elated naatie pose *5 *5 Āgama sūta/ese texts Dīghāgama Madhyamāgama Saṃyuktāgama CM Ekottaikāgama 3 3 Kṣudakāgama 3 M +7 CM 2 M M 3 Unidentified/unspecified M 2 Abhidhama (see below) Paacanonical (Śāakayāna) texts Scholastic texts/commentaies Scholastic teatises Commentaies Unspecified texts Rakṣā sūtas/dhāaṇī Aadāna/pūayoga (collections) Buddha paises/stota 2 4 Miscellaneous/undetemined texts Vese texts 2 3 Pose texts 6 8 Unspecified texts/fagments 3 3 Mahāyāna texts Sūtas 3 6 Scholastic texts *3 *3 Non-Buddhist texts Nīti texts Secula documents Key to ms collections and single mss: BL = Bitish Libay; RS = Senio Collection; MS = Bamiyan fagments of the Schøyen and othe collections; PC/OC = Cental Asian fagments of the Pelliot and Oldenbug Collections; BajC = Bajau Collection; SplitC = Split Collection; UC = Unpublished Collection; KDhp = Khotan Dhamapada; LC = Libay of Congess; UW = Uniesity of Washington Since some manuscipts contain moe than one text sometimes een texts belonging to diffeent genes the numbe of manuscipts is not identical with the numbe of texts. Many of these manuscipts ae as yet unedited and unstudied it is theefoe not always possible to detemine the exact numbe of texts contained in them with a sufficient degee of cetainty. In these cases the figue efes to the numbe of manuscipts whee the

14 58 HARRY FALK AND INGO STRAUCH espectie textual gene is attested. The actual numbe of texts might consequently be somewhat highe. Supescipt C indicates that a cetain text is attested as pat of a commentay. Supescipt M points to the metical chaacte of a text. Uncetain attibutions ae indicated by an asteisk *. With egad to thei inentoies the collections show distinctie diffeences. The Senio Collection contains only sūtas o closely elated naatie texts fom an assumed Vinaya backgound, witten by a single scibe. It is theefoe possible to designate this collection as an intentional compilation of canonical texts, pobably intended fom the outset fo itual buial inside a stūpa. On the othe hand, the Bitish Libay, Bajau and Split Collections and most pobably also the less well-peseed collections fom Bamiyan and Cental Asia ae athe heteogeneous compilations which compise a multitude of Buddhist liteay genes. The manuscipts aying states of peseation as well as thei abitay inentoy seem to indicate othe puposes. Although it is possible that the manuscipts as we got them had been itually buied afte use, 7 thei pimay pupose was cetainly diffeent and seems to point to an actie use, eithe as pat of a monastey libay o in the pesonal possession of an indiidual monk. The data pesented in the table aboe will now be expounded in moe detail by concisely intoducing the hitheto known texts accoding to thei gene affiliation. Special attention is gien hee to the Bajau and Split Collections, which ae being studied by the authos of this aticle. Moe infomation about the othe manuscipts and manuscipt collections can be found in the aticles by Mak Allon, Collett Cox and Richad Salomon in the pesent olume. Each intoductoy paagaph is followed by a synoptic list containing all hitheto known texts o manuscipts of the espectie gene. As indicated aboe, texts which ae peseed only within a commentay ae maked by supescipt C. They actually belong to the categoy of Scholastic texts and commentaies, but, of couse, poide positie eidence fo the ciculation of the texts which they ae commenting on. Sometimes, howee, texts appea as well as pats of a commentay as they do as uncommented independent units. Fo distinguishing between these two aieties, these texts (e.g. the Dhamapada and the Athapada) ae cited twice: once as pats of a commentay and once as independent texts. The titles of the sūtas ae geneally indicated in the fom of thei Pāli (P.) o Sanskit (Skt.) paallels. If no paallels ae aailable, a hypothetical Skt. title is gien (*), followed by its pesumable Gāndhāī equialent, if the editos o eseaches of the text decided to intoduce such a designation. All titles ae listed in the ode of the Latin alphabet. The sign peceding a bibliogaphical efeence means that the wok in question contains a complete edition of the manuscipt o text Canonical Śāakayāna liteatue The tem canonical is used hee without any implication fo the assumed shape of the Gāndhāī Buddhist canon. This categoy indicates that diect paallels of the espectie texts ae pats of canonical collections of othe Buddhist taditions. Consequently, the table aboe as well as this annotated suey ae not meant to suggest that the texts included hee once fomed such a coheent body of texts, which could be identified as a canon o potocanon. Moeoe, it cannot be taken fo ganted that all the texts can be attibuted to a single and clealy identifiable Buddhist school (cf. Stauch 2008: 4f.). Although thee is a cetain amount of eidence which points to the Dhamaguptakas as oiginatos of these texts, we can hadly exclude the paticipation of othe schools. As fa as the epigaphical 7 See the discussion in Salomon 2009, Stauch 2008: 04f., and Stauch, fothcoming, a.

15 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 59 eidence indicates, apat fom the Dhamaguptakas thee was a stong pesence of Saāstiādins in Geate Gandhāa, including the egion of Haḍḍa fom whee at least the Bitish Libay Collection is said to hae come Vinaya texts Until ey ecently it has epeatedly been suggested that the witing down of this textual gene stated in a much late peiod of Buddhist histoy, i.e. fom the 4th c. AD onwads. Neithe the fist known collections of Khaoṣṭhī manuscipts no the Kuṣāṇa peiod manuscipts fom Easten Tukestan contained any texts fom a Vinaya. Although it cannot be excluded that some of the naatie texts fom the Senio Collection ae eflectie of a Vinaya backgound (cf. Allon 2007a: 22), the fist indisputable eidence fo witten texts fom a Vinaya could be identified among the manuscipts of the Bajau Collection, which contains two diffeent esions of a pat of the Pātimokṣasūta and two Kamaācanā fomulae. Both texts ae not tansmitted as pats of a Vibhaṅga o Skandhaka, but as independent texts. They ae not only the oldest Vinaya manuscipts known so fa thei impotance fo the textual histoy of this gene is augmented by the fact that they ae een olde than the ealiest Chinese tanslations of Vinaya liteatue, which go back to the 4th c. AD Pātimokṣasūta The Bajau fagment 3 contains two diffeent esions of the beginning of the Naiḥsagika Pācittiya section of the Pātimokṣasūta. Both esions ae witten by the same scibe on the obese and eese of the scoll. Accoding to a compaatie analysis conducted on the basis of the Pātimokṣasūta texts peseed in Indian languages and in Chinese and Tibetan tanslations the Gāndhāī esions ae not identical with any of these paallels. Nonetheless it is possible to attibute them to two clealy distinct banches in the tansmission of the Pātimokṣasūta. The esion on the obese elates to the Dhamaguptaka/ Kāśyapīya Pātimokṣasūtas, while the text on the eese seems to be moe closely affiliated with the Saāstiādin/Mūlasaāstiādin tadition (Stauch 2008: 6f.). The diffeences between both esions and thei canonical paallels allow us to chaacteize the Gāndhāī tadition as a witness to a fomatie state in the textual genesis of the aious nikāya esions of the Pātimokṣasūta which had not yet eached thei fully deeloped and late canonized shape. The eason fo this side-by-side existence of two diffeent esions is had to explain, but points to the fact that within a monastic community moe than one esion of this basic text could be known. They could influence each othe and thus esult in the emegence of new and contaminated textual foms Kamaācanā The second Vinaya text of the Bajau Collection is epesented by a ey small manuscipt containing the Kamaācanā fomulae fo the appointment of the distibuto of lodgingplaces (śayanāsanagāhaka) on the obese and that fo the ceemony of taking up the eteat of the ainy season (aṣopagamana) on the eese. Again the compaison of both Gāndhāī esions with thei paallels in Kamaācanā collections and Vinayaibhaṅgas of aious schools cannot help to establish a definite school identity. Combining this eidence 8 Fo the Dhamaguptakas see the eidence discussed by Salomon 999: 67 7 and Stauch 2007 with a moe ecent efeence to this school. The Saāstiādins ae e.g. mentioned on a pot fom Haḍḍa (Fussman 969). See also the suey of epigaphical souces gien by Lamotte (988: ). 9 The eidence of this manuscipt is now moe extensiely discussed in Stauch fothcoming, a.

16 60 HARRY FALK AND INGO STRAUCH with the obseations made on the Pātimokṣasūta we hae to assume that the composition of the Gāndhāī Vinaya texts attested in the Bajau Collection is pio to the emegence of standadized canonical texts which became exclusiely authoitatie within the boundaies of a cetain school (Stauch 2008: 7f.) Pātimokṣasūta Naiḥsagika Pācittiya 9 (BajC 3 ecto, Stauch 2008: 6f. and Stauch, fothcoming, a) Naiḥsagika Pācittiya 8 (BajC 3 eso, Stauch 2008: 6f. and Stauch, fothcoming, a) Kamaācanā śayanāsanagāhaka fomula (BajC 5 ecto, Stauch 2008: 7f.) aṣopagamana fomula (BajC 5 eso, Stauch 2008: 7f.) Vinaya-elated pose texts conesion of Sujātā and he family (RS 5 + 8, Allon 2007a: 3) Moggalāna tells the Buddha about Deadatta s wish to lead the Saṅgha, enumeation of fie kinds of teaches (RS A., Allon 2007a: 6 7) Anuuddha s wish fo odination (RS A.2. Allon 2007a: 7) stoy of Nāla (Skt. Nālada/Nālaka) and Eakapatta (Skt. Elāpatta) (RS 24., Allon 2007a:7) stoy of Tapussa (Skt. Tipuṣa/Tipusa) and Bhallika and the Buddha s bowl (RS 24.2, Allon 2007a: 7 8) Canonical Āgama Texts Vese texts and the Kṣudakāgama Vese texts like the Dhamapada now found in thee diffeent Gāndhāī esions (KhDhp: Bough 962, BL: Lenz 2003, SplitC: Falk 20), the Rhinoceos sūta (BL: Salomon 2000) o the Anaataptagāthā peseed in two esions (BL, RS: Salomon 2009) epesent a popula gene of Buddhist liteatue which is pat of at least thee majo collections of Gāndhāī manuscipts (BL, RS, SplitC). Its populaity is also confimed by the fact that some of the texts ae found in moe than one esion while othes fomed the basis of an extensie commentaial liteatue. In paticula the diffeent esions of the Dhamapada show the athe flexible natue of Buddhist Gāndhāī liteatue, which is fa fom being a static textual tadition. While Lenz (2003: 3) was tempted to egad the few fagments of the Dhamapada fom the BL collection as most likely being a second esion of essentially the same [i.e. Khotan DhP] text, the much moe extensie emains of the Dhamapada in the Split collection (SplitC 3, Falk 20) athe point to the opposite, being compaable with the Khotan Dhamapada in sequence and content only in ae cases. This new manuscipt offes ample scope fo compaison with the esion fom Khotan. It contains 87 stanzas fom at least seen agas. The othogaphy diffes consideably fom the one used at Khotan and whee both esions ae extant, the diction is often slightly diffeent. With egad to the sequence of chaptes and stanzas, the esion fom Subashi and the Udānaaga ae close than any othe paallel, without being identical. The inclusion of seeal stanzas known othewise only fom the Aṅguttaanikāya, the Saṃyuttanikāya and paticulaly the Majjhimanikāya shows how feely the compiles woked. Veses fom the Dhamapada ae also among the oot eses of Gāndhāī commentaies. 0 Othe commented ese texts include the Athapada/Athaaga (P. Aṭṭhaka- 0 A suey of commented eses found in the BL commentaies is poided by Baums (2009: 50).

17 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 6 agga), the Pāāyaṇaaga, the Udāna and the Itiṛttaka (P. Itiuttaka) which ae commented on in seeal of the commentaies of the Bitish Libay Collection (see below, suey). That these texts wee not only used fo commentaial puposes is shown by the independent manuscipts of this gene. Beside the Dhamapada manuscipts discussed aboe we now hae at ou disposal anothe independent text of this class, namely a ese collection among the fagments of the Split Collection which coesponds to a pat of the Athapada (P. Aṭṭhakaagga) (SplitC ). This text is aailable pesently in just one stip of bich bak. A ecent offe in the manuscipt maket seems to hae contained moe fagments of this manuscipt, thus aising the hope that some pats of the Split Collection will be eunited soone o late. The fagment measues.8 cm, being almost completely peseed in its width. Both sides show the full content o taces of fou lines, containing the full o patial text coesponding to stanzas and of the Aṭṭhakaagga diision of the Pāli Suttanipāta. The othogaphy used is diffeent fom othe hands, some lettes hae aiant foms as if some olde esions contibuted to this mixtue of shapes. This fagment appeas to be the oldest in the collection. A 4 C date is so fa not yet aailable. The eadings ae gien in Falk 20. The Āgama affiliation of some of these ese texts is a matte of continuous discussion. The majoity of them, howee, wee subsumed unde the categoy of the so-called Mino Texts. The status of this categoy within the canons of the diffeent schools is not quite clea. Wheeas in some taditions it is peceied as pat of the Sūtapiṭaka called eithe nikāya/āgama o piṭaka 2 othe taditions classify it as a sepaate piṭaka beside the Sūta- and Vinayapiṭaka. 3 Since the numbe and length of the texts which ae found in the aious lists of Mino Texts ay makedly, it can be suggested that many of the schools nee completed an authoitatie collection of this categoy (Lamotte 988: 62f.). Fo the Dhamaguptakas, howee the most pobable candidates fo ou canonical Gāndhāī liteatue we suely know fom thei Vinaya (T. 428, ch. 54, p. 968b) that they possessed a special collection of Mino Texts, called Kṣudakapiṭaka and foming pat of the Sūtapiṭaka, which included inte alia the Dhamapada, the Athapada and the Pāāyaṇa, which do seem, in fact, to hae fomed the oiginal coe of the mino texts (Lamotte 988: 60f.). The moe complicated issue of the Anaataptagāthā, which is not listed in any of the known canonical Āgama o Nikāya collections, was extensiely discussed by Richad Salomon, who concludes that we find a pepondeance of eidence, though no single conclusie poof, that in the Gandhāan tadition of the ealy centuies of the Chistian ea the A[naatapta]-G[āthā] was constued as a canonical sūta in the Kṣudakāgama class (2008: 8). Thus the pesence of independent ese texts like the Dhamapada, the Rhinoceos sūta and the Athapada as well as the extensie commentaies on eses fom this text type indicate the existence of a Kṣudakāgama/Kṣudakapiṭaka-like compilation in Buddhist Gandhāan liteatue (Salomon 999: 59 6) Āgama sūtas By fa the lagest amount of canonical sūtas in Gāndhāī liteatue is found in the Senio Collection. All its manuscipts wee witten by a single scibe, pobably accoding to a The authentic Gāndhāī tem which is appaently also used by the Dhamaguptakas can now be established as Athapada. Cf. the extensie discussion in Baums (2009: 38 44). 2 Accoding to Lamotte (988: 5), the Mahāsāṃghikas, Haimaatas, Mahīśāsakas and Dhamaguptakas used the designation Kṣudakapiṭaka, but included this collection in thei Sūtapiṭaka. 3 Such a diision with a Sūtapiṭaka consisting of fou Āgamas and an additional Kṣudakapiṭaka which is not pat of the Sūtapiṭaka is attested in aious ealy taditions including that of the Saāstiādins which nee had moe than fou Āgamas (Lamotte 988: 5).

18 62 HARRY FALK AND INGO STRAUCH peiously fixed plan (Allon 2007a: 3 25). Of the 4 suiing texts, 33 belong to the class of Āgama sūtas: 29 of them can be ascibed to the Saṃyuktāgama, fou to the Madhyamāgama and only one to the Dīghāgama. The chaacte of the emaining eight texts is difficult to establish. Six of them ae naatie texts with paallels in Vinaya liteatue, in one case in both Vinaya and Sūta liteatue. 4 Especially the aangement of the sūtas which can be attibuted to a Saṃyuktāgama indicates the existence of a text compilation which is patially paallel to the Pāli Saṃyuttanikāya and the Chinese Saṃyuktāgama (T 99) without being identical with any of them (Glass 2007: 26 50). This collection of canonical sūtas in the Senio Collection is supplemented by some isolated examples in othe collections. Thus the Bitish Libay Collection contains thee sūtas fom an Ekottaikāgama (Allon 200) and a futhe non-identified sūta text (BL 26+29, Salomon 999: 24). A piece of indiect eidence fo a Dīghagama text is poided by a Bitish Libay commentay on the Saṅgītisūta. Its attibution to a Dīghāgama is indicated not only by the Pāli canon, but also by the Chinese (Dhamaguptaka) Dīghāgama (T ), the Saāstiāda (Obelies 2003, n. 83 with efeences) and the newly discoeed Mūlasaāstiāda Dīghāgama fom Gilgit (Obelies 2003: 66, Hatmann 2004). A futhe Dīghāgama text is epesented by the Mahāpainiāṇasūta of the Schøyen Collection (Allon and Salomon 2000). The eidence fo Madhyamāgama texts is supplemented by fagment of the Bajau Collection. It contains the Gāndhāī esion of a sūta which is paallel to the Pāli Dakkhi- ṇāibhaṅgasutta gien as No. 42 of its Majjhimanikāya (MN III ). The Chinese tanslation of the Madhyamāgama (T 26) by Gautama Saṅghadea (tanslated ) lists this sūta as No. 80 (p. 72c2) unde the name Qutanmi jing = Skt. Gautamīsūta. The compaison of this Gāndhāī esion with its paallels, including two small fagments of Sanskit esions in the Tufan and Schøyen Collections and Śamathadea s quotations fom a pobably Mūlasaāstiādin Madhyamāgama, cannot help to settle the question of the school affiliation of this text. Since no Dhamaguptaka Madhyamāgama is known, it is, howee, possible that the Gāndhāī text epesents an extact fom a Madhyamāgama of this school (Stauch 2008: 8 9; see now also Stauch fothcoming, b). Dīghāgama Skt. Mahāpainiāṇa-sūta (MS 279, Allon and Salomon 2000) P. Sāmaññaphala-sutta (RS 2, Allon 2007a: 8) Skt. Saṅgīti-sūta C (BL 5, Salomon 999: 24, 38, 7 73) Madhyamāgama P. Cūḷagosiṅga-sutta (RS 2, Allon 2007a: ) P. Dakkhiṇāibhaṅga-sutta (BajC, Stauch 2008: 8f. and Stauch, fothcoming, b) P. Dhammacetiya-sutta (RS +3, Allon 2007a: 7) P. Saṅkhāupapatti-sutta (RS 0, Allon 2007a: ) Ch. Shìzhě jīng, Sūta on an attendant, T no. 26 pp. 47c 475a (RS 4A, see Allon supa) 4 In the pesent suey we will not include the eidence of the so-called index-scoll of the collection (RS 7 + 8), which enumeates the titles of seeal sūtas, but does not completely coespond to the contents of the collection. It is extensiely discussed in Allon (2007a: 8 2) and in his contibution to this olume.

19 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 63 Saṃyuktāgama P. Anattalakkhaṇa-sutta (RS 22.2, Allon 2007a: 5) P. Dāukkhandha-sutta (RS 9, Lee 2009) P. Dhanuggaha-sutta (?) (RS 22., Allon 2007a: 5) P. Dutiyacchiggaḷayuga-sutta (RS 22.3, Allon 2007a: 5 6) P. Gaddula-sutta (see Nadīsutta) P. Kulaputtenadukkhā-sutta alias Nibbidābahula-sutta (= G. Ṣadha-suta, Skt. Śaddhā-sūta, RS 5.3, Glass 2007) P. Mahāpaiḷāha-sutta (RS 20.2, Allon 2007a: 4) P. Nadī-sutta cum Gaddula-sutta (RS 7., Allon 2007a: 3) P. Natumhāka-sutta (G. Ṇatuṣpahu-suta, RS 5.2, Glass 2007) P. Nibbidābahula-sutta (see P. Kulaputtenadukkhā-sutta) P. Puppha-sutta (RS 22.6, Allon 2007a: 6) Skt. *Ṣaḍāyatana-saṃyukta, pobably a sūta theeof about the causes of happiness and suffeing (RS 20., Allon 2007a: 4) G. Ṣadha-suta (see Kulaputtenadukkhā-sutta) Skt. *Saṃjñā-sūta (G. aña-suta, RS 5, Glass 2007), pehaps fom a Saṃyuktāgama like the othe texts of this scoll P. Tissa-sutta (RS 7.2, Allon 2007a: 3) P. Uppāda-sutta (RS 22.4, Allon 2007a: 6) P. Vana-saṃyutta (4 suttas theeof, 9. 4) (RS, Allon 2007a: ) P. Vāsijaṭa-sutta (G. *Vasijaḍa-suta, RS 5.4, Glass 2007) P. Veḷudāeyya-sutta (RS 3, Allon 2007a: 2) Ekottaikāgama G. *Buddhabayaṇa-suta (Skt. Buddhaacana, BL 2 + 4, Allon 200) P. Dona-sutta (G. Dhoṇasuta, BL 2 + 4, Allon 200) P. Saṃaa-sutta (G. Pasaṇasuta, Skt. Padhāna, BL 2 + 4, Allon 200) Kṣudakāgama / Kṣudakapiṭaka Skt. Anaataptagāthā (BL /, RS 4, Salomon 2008) Skt. Athapada (P. Aṭṭhakaagga, SplitC 3, Falk 20) Skt. Athapada C (P. Aṭṭhaagga): BL ese commentay 2 (BL 7, 9, 8, 3 line 90, Baums 2009) Skt. Dhamapada (Khotan: Bough 962, BL /, Lenz 2003: 76, SplitC 3, Falk 20) Skt. Dhamapada C : BL ese commentay (BL 4/, Salomon 999: 27), BL ese commentay 2 (BL 7, 9, 8, 3 line 90, Baums 2009), BL ese commentay 3 (BL 3, line 9, Baums 2009: ) P. Itiuttaka C : BL ese commentay 2 (BL 7, 9, 8, 3 line 90, Baums 2009) P. Khaggaisāṇa-sutta (G. *Khagaiṣaṇa-suta, BL 5B, Salomon 2000) P. Pāāyaṇa C : BL ese commentay 2 (BL 7, 9, 8, 3 line 90,) Baums 2009) P. Udāna C : BL ese commentay (BL 4/, Salomon 999: 27), BL ese commentay 2 (BL 7, 9, 8, 3 line 90, Baums 2009) Unidentified Āgama type texts Sūta text concening the fou stages of meditatie tance (BL : Salomon 999: 24)

20 64 HARRY FALK AND INGO STRAUCH eses (mixed with pose?) esembling Jātaka no. 480 (IV ) (RS 22.5, Allon 2007a: 6) Both the Āgama sūta and ese texts and the Vinaya texts epesent liteay genes which wee ceated and deeloped outside Gandhāa east and south of it. They ae supplemented by a lage numbe of texts which seem to hae been ceated in Gandhāa itself Abhidhama liteatue One of these autochthonous liteay genes is the scholastic and commentaial liteatue (see the contibution by Cox in this olume) which is epesented by a lage numbe of texts in the Bitish Libay and the Bajau Collections. The canonical status of the scholastic and commentaial liteatue of this peiod is disputed. Due to the fact that the Abhidhamapiṭakas of the aious schools ae athe diese and contain heteogeneous mateial including commentaial texts like the Saṅgītipayāya o the Dhamaskandha of the Saāstiādins it cannot be excluded that texts of this type ae indeed pat of a (poto-)abhidhamapiṭaka. 5 On the othe hand, the inclusion of these texts in the aious Buddhist canons took place duing a athe late phase in the histoy of Buddhist liteatue. Thee is stong eidence that the pocess of the fomation of canonical Abhidhamapiṭakas was not yet completed at the time of ou manuscipts. 6 That this liteay gene epesents a quite ecent deelopment is also suppoted by the fact that the specimens of both collections display a distinctiely diffeent chaacte. This clealy speaks in faou of consideing them as poducts of local scholastic taditions which had not yet been subject to a pocess of hamonizing and standadization. Due to this somehow tansitional chaacte which places these texts in between canonical and paacanonical liteatue we will subsume them unde the categoy of Paacanonical texts, while beaing in mind the fact that some of them might hae fomed pat of an ealy, possibly poto-canonical Abhidhamapiṭaka-type collection Paacanonical texts Scholastic texts and commentaies (Abhidhama) In the BL Collection this categoy compises a significant pat of the classified texts. In geneal tems it can be diided into two majo goups: independent scholastic teatises of aious chaactes and commentaial texts. The actual state of eseach and the main featues of the BL texts ae coeed by Collett Cox (supa). These BL specimens of scholastic texts and commentaies ae supplemented by altogethe nine manuscipts fom the Bajau Collection which display, howee, a distinctly diffeent chaacte, being mostly independent scholastic teatises athe than commentaial texts. Most of them ae ey fagmentay and shot and hadly allow any fa-eaching conclusions about thei contents. One of these manuscipts, howee, is ey well peseed (BajC 9 eso). It is composed in the style of a polemical scholastic teatise citing diffeent authoities, who ae intoduced by phases like keci(d) aho / keyi aho (Skt. kecid āhuḥ) some say o apae aho (Skt. apaa āhuḥ) othes say (Stauch 2008: 9). The emaining scholastic texts of the Bajau Collection ae in a much wose state of peseation. Accoding to some shaed phases and teminological expessions six of them 5 Cf. the detailed discussion of the Saṅgītisūta commentay in the contibution by Cox, pp A simila conclusion is also possible with egad to the Kuṣāṇa peiod Abhidhama-type texts fom Tufan which can hadly be attibuted to any of the Abhidhama woks of the Saāstiādins known fom late manuscipts o tanslations (Dietz 2007: 67, efeing to Sande 99: 33 34).

21 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 65 can be attibuted to two diffeent text goups. Goup A consists of thee fagments (BajC 4, 6, ), of which two (BajC 4, ) ae possibly pat of the same scoll. The text is mainly concened with the discussion of diffeent kinds of happiness (Gāndhāī suha, Skt. sukha). Expessions like bodhimaṇḍa, gagaṇadialia amal(o)ǵadhadu = Skt. gaṅgānadīālukasama-lokadhātu and efeences to pañapaamida (4,2,) (Skt. pajñāpāamitā) and the six pāamitās (edeṣa ṣahi paamidehi) (,2,) might indicate the Mahāyāna chaacte of this text goup (Stauch 2008: 9). The second goup B (BajC 4, 6, 8) is een less well-peseed. Hee only single phases like yadi jie bhaeadi if he is meditating upon life and yadi dhama bhaeadi if he is meditating upon the dhama and few teminological paallels indicate thei association with the same text o text tadition. Due to thei biefness and bad state of peseation it is at pesent difficult to gie any eliable infomation about the contents of Bajau fagments 2 and 9, which seem, howee, to belong to the same gene of scholastic teatises. Commentaies ese commentay (BL 4/, Salomon 999: 27) ese commentay 2 (BL 7, 9, 8, 3 line 90, Baums 2009) ese commentay 3 (BL 3, line 9, Baums 2009: ) Saṅgītisūta commentay (BL 5, Salomon 999: 24) commentay on a text simila to the Dhātuibhaṅga-sutta (MN III ) (UW, Glass 2004: 4f.) Independent scholastic teatises Scholastic teatise (BL 0, Salomon 999: 47) Scholastic teatise (BL 7, Salomon 999: 49f.) Scholastic teatise (BL 20+23, Salomon 999: 50f.) Scholastic teatise (BL 28, Salomon 999: 30) Bajau scholastic text (BajC 9 eso, Stauch 2008: 9) Bajau scholastic text 2 (BajC 2, Stauch 2008: 9) Bajau scholastic text 3 (BajC 9, Stauch 2008: 9) Bajau scholastic texts goup A (BajC 4, 6,, Stauch 2008: 9) Bajau scholastic texts goup B (BajC 4, 6, 8, Stauch 2008: 9) Unspecified scholastic texts text fom some moe technical gene, whethe Sūta, Abhidhama, o commentay (PC 8, Salomon 998: 47 50) commentay on a sūta o an abhidhama text that utilizes canonical quotes (MS 279/42 = MS 42, Allon et al. 2006: 288) Aadāna/Pūayoga texts Aadānas and Pūayogas seem to fom anothe class of popula contempoay texts composed in Gāndhāī in opposition to the canonical texts which must hae been tanslated fom a Middle Indian oiginal. They ae epesented as well in the Bitish Libay as they ae in the Split Collection, but ae supisingly absent fom the othewise athe compehensie Bajau Collection. In many cases the stoies epoted in these shot texts ae based on local taditions containing local toponyms and pesonal names with occasional paallels to Aadāna taditions fom othe pats of India. Timothy Lenz,

22 66 HARRY FALK AND INGO STRAUCH who ecently published the second olume of his editions of the Bitish Libay Collection Aadāna texts, poides an excellent suey of Gāndhāī Aadāna liteatue which can be efeed to hee (200: 3 7, especially tables +2, pp. 8 2). In descibing the geneal featues of this gene Lenz says: The Gāndhāī aadānas contain only a single stoy, eithe one concening past actions (self-styled as pūayogas) o one about pesent actions (self-styled as aadānas). On the othe hand, the aadāna texts ae summaies, each giing the outline of a stoy that its autho o compile assumed would be ecognized by the eade and would be expanded quite easily into a naatie by that same peson (Lenz 200: 6). Lenz plausibly agues fo the existence of a goup of specialists aadānists who wee esponsible fo the composition of these texts which diffe consideably fom what we know fom late Aadāna taditions. A Gandhāan aadānist s palette was aied, including not only aadānas concened with kamic ineitability (...) and giing (...), as one would expect, but also with impemanence (...), the Buddha s powe beyond niāṇa (...), women in dhama (...), and the histoy of the Fist Council (...) that wee pesented in aious foms, including but not limited to a standad aadāna fomat compising a tale of the pesent, a tale of the past, and a conclusion identifying stoy chaactes (Lenz 200: 4). The eidence ealuated by Lenz based on the texts of a single scibe (200: 6) can now be supplemented by an Aadāna compilation fom the Split Collection which can theefoe help to poide a moe epesentatie pictue of this paticula liteay gene of Gandhāan liteatue. The text (SplitC 4) consists of about 300 fagments, few of them lage than 6 cm, most of them much smalle. The gene is clealy discenible though phases like NN aadano japati. In shot sentences seeal stoies ae outlined, coeing just a few lines, ending in a glyph esembling a wheel. The witing is emakably achaic, with peconsonantal epesented by a hook, not a backwad loop. Despite this achaism it is difficult to fully accept the outcome of a 4 C check, which yielded a possible ange fom 84 to 46 BC, with a peak of about 70 BC. With egad to the contents, it is inteesting to ead about Buddhist schools; the Dhamaguptakas and the Mahāsāṃghikas ae no supise in Gandhāa, but the Seiyaputas ae, as ae the Ājīikas, who execise a stong and negatie influence in one stoy on one king. The topic of anothe stoy is the well-known thief Aṅgulimāla. The phases aadano eo ṣuyadi o eo paiśao found in othe Aadāna o Pūayoga texts (Lenz 2003: 83) ae completely absent, pointing to a diffeent banch of stoy deelopes, if thee is not simply a beak in time, with the BL pieces mentioning kings and dynasties actie in the middle of the fist centuy AD, and ou collection being possibly somewhat olde. It is possible that thee of the pesumably naatie texts on some of the badly peseed fagments of the Pelliot Collection also belong to this gene (Salomon 998, fagments PC, 2, 3+6). The majoity of the Buddhist Mainsteam taditions include Aadānas in the list of aṅgas, o constituents of the Buddha s wods, and thus accod canonical status to them (Lamotte 988: 43 47). Due to thei explicit local chaacte and the absence of diect paallels in othe canonical taditions we neetheless decided to goup the peseed Gāndhāī examples unde the categoy of paacanonical texts, again being awae as in the case of the Abhidhama texts that they might hae enjoyed canonical status in a gien community.

23 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 67 Aadāna collection (BL /2, Lenz 200) Aadāna collection (BL 2, Lenz 200) Aadāna collection (BL 3A/2, Lenz 200) Aadāna collection (BL 4/2, Salomon 999: 35f.) Aadāna collection (BL 2+4/2, Salomon 999: 36f.) Pūayoga collection (BL 6+25/2, Lenz 2003) Aadāna collection (BL 2/2, Lenz 200) Aadāna collection (SplitC 4, Falk 20) 3 unspecified texts of Aadāna/Jātaka chaacte (PC, 2, 3+6, Salomon 998: 24 45) Rakṣāsūtas / Dhāaṇī This liteay gene, which gained quite a pominent status in the fist centuies of ou ea, is epesented by only two manuscipts fom the Bajau Collection. One of them (BajC ) is pooly peseed. It is witten in lage and caelessly outlined lettes on the back side of fagment which contains the Gāndhāī esion of the *Dakṣiṇāibhaṅga-sūta on its obese (see aboe). Only some chaacteistic wods and phases like migili pitili and efeences to yakṣas and nāgas allow this peliminay attibution. Moe telling and much bette peseed is the text of Bajau fagment 3. It is composed in the conentional sūta style and efes to the nāgaāja Manasin, who is exclusiely known fom nothen Buddhist souces. The manta which this nāga king pesents to the Buddha is called maṇasi-ṇagaaya-ija (Skt. manasi-nāgaāja-idyā) and contains names of poisons. Nealy all of them ae also pat of a compehensie list peseed in the Mahāmāyūī (ed. Takubo 972: 55, cf. ed. Oldenbug 899: ). Anothe close paallel to the manta is poided by the shot Tibetan text Phags pa klu i gyal po gzi can gyis žus pa žes bya ba i gzuṅs whose oiginal Skt. title has to be estoed as Āya-Manasināgaāja-paipṛcchā-nāma-dhāaṇī. 7 Accoding to its liteay style and stuctue the Gāndhāī text displays significant paallels with othe post-canonical woks of this gene, like the Pañcaakṣā texts Mahāmāyūī and Mahāsāhasapamadinī o the appendix (yākaaṇa) of the Cental Asian esions of the Nagaopamasūta (cf. ed. Bongad-Lein et al. 996: 30 37). Summing up this eidence, the Gāndhāī *Manasināgaāja-sūta can theefoe be elated to the post-canonical Pañcaakṣā collections, which accoding to Pete Skilling s classification fom one of the fou majo goups of akṣā liteatue (992: 3). At the same time it is the oldest manuscipt of any akṣā text. Skt. *Manasināgaāja-sūta (G. *Maṇasiṇagaaya-suta, BajC 3, Stauch 2008: 20f.; Stauch, fothcoming, c) unspecified dhāaṇī-like text (BajC eso, Stauch 2008: 20) Buddha paises One of the Bitish Libay fagments (BL 5C) contains eses paising the Buddha which esemble the stotas known fom the Niya documents 50 and 5 (Boye et al., 2: 84 87). It is composed in aious poetic metes such as Vasantatilakā, descibing him [the Buddha] with such epithets as gunehi guna-paami-pataṃ, who has attained though his itues the pefection of itue ; soma-sadiśa-[a]dana, whose face is like the moon, and saasatutamaṃ, supeme among all beings (Salomon 999: 39). Othe epesentaties of this gene ae pat of the Bajau Collection: fagment 8 contains fou stuctually paallel eses composed in the Śādūlaikīḍita mete. 7 Fo futhe details and efeences cf. Stauch 2008: 20f.

24 68 HARRY FALK AND INGO STRAUCH Bajau fagment 0 is a non-metical text about the paisewothy things (Gāndhāī pa aśa hana, Skt. pāśaṃsyasthāna, P. pāsaṃsaṭhāna) of a Buddha (Stauch 2008: 2). The Buddha is paised hee by conentional attibutes like nilinakileśa (Skt. nilīnakleśa) whose passions ae destoyed, iśudhipato (Skt. iśuddhipāpta) who has attained puity, o sadia iadeasabuda (Skt. smṛtyādhipateyasaṃṛta) contolled though the powe of mindfulness. The Split Collection pesees one moe text of this gene (SplitC 2). It is still 6 cm wide, haing lost about 5 cm on the ight side whee most stanzas stated. It seems that the text deals with paticula pominent eents in the life of the Buddha. Conspicuous is the mete, being pedominantly Vegaatī, a subaiety of the Vaitālīya well-known fom Pāli texts, but found hee fo the fist time in Gandhāa. stota in diffeent metes (BL 5C, Salomon 999: 39, 46) stota of fou eses in Śādūlaikīḍita (BajC 8, Stauch 2008: 2) non-metical text paising the Buddha (BajC 0, Stauch 2008: 2) paise of the Buddha in Vegaatī eses (SplitC 2, Falk 20) Miscellaneous/not detemined texts Thee is a small numbe of texts which cannot be safely put in any of the aboementioned categoies. Bitish Libay fagments BL 8,, 2/, 22, 24, 27 8 as well as some of the ey tiny Pelliot fagments esist identification o at least chaacteization, mainly due to thei bad state of peseation o insufficient extent. Othe texts can be chaacteized but do not match any of the known o accepted liteay categoies. Fou of them ae paacanonical ese texts of diffeent types. The shot metical text on BL 5A consists of possibly didactic eses (Salomon 999: 45, ed. Salomon 2000: ). Unclea is the chaacte of the ey pooly peseed eses of Bajau fagment 7. This heaily damaged manuscipt, which consists of only a few patially peseed lines of text, does not allow any eliable conclusions. A much bette peseed and paticulaly inteesting and impotant text of this class is epesented by Bajau fagment 5, which contains a collection of Buddhist eses aanged accoding to the Aapacana alphabet. The ight side of this scoll is almost completely peseed and thus allows fo the fist time the econstuction of the entie inentoy of the alphabet fom its 2nd lette a up to its last lette ḍha on the basis of a contempoay Khaoṣṭhī manuscipt. 9 Othe texts of this categoy ae composed in pose. The Libay of Congess Scoll appeas to consist of fomulaic accounts of the lies of fifteen Buddhas, fom Dīpaṅkaa to Maiteya, enumeating fo each Buddha the kalpa in which he lied, his life-span, his class (bāhmaṇa o kṣatiya) the size of his assembly (saṃnipāta), the duation of his dhama, etc. (Salomon and Baums 2007: 202). Accoding to Salomon and Baums, the text is paticulaly closely elated to some potions of the Bahubuddha-sūta, which is pat of the Mahāastu (2007: 202). A futhe hitheto unidentified pose text is found on a single palm-leaf fagment collected by Segey Oldenbug on his expedition to Easten Tukestan and kept today in the collection of the St. Petesbug Academy of Sciences unde the siglum SI O/ 0. Although it has been identified by Voob ea-desi atoskai a (2006: 48) as pat of the 8 Of these manuscipt fagments, 2 and 24 ae possibly witten by the same hand and contain the same text which is cautiously chaacteized by Salomon as possibly a ese text (999: 47). Similaly, fagments 22 and 27 ae pehaps pat of the same scoll (Salomon 999: 5f.). Its contents as well as that of fagment 8 ae, howee, undetemined. 9 Fo futhe details and othe souces fo the sequence and contents of the Aapacana alphabet see Stauch 2008: 2 23 and paticulaly Stauch 20.

25 The Bajau and Split Collections of Khaoṣṭhī Manuscipts 69 Mahāpainiāṇasūta, fom the ey same esion as the text fom the Schøyen Collection (Allon and Salomon 2000), thee is no incontoetible eidence to suppot this identification. 20 We would theefoe suggest that this text be tentatiely consideed as anothe, as yet unidentified paacanonical Buddhist text. Vese texts ese text aanged accoding to the Aapacana alphabet (Stauch 2008: 2 23, Stauch 20) unidentified pobably metical text (BL 5A, Salomon 2000: ) unidentified possibly metical text (BajC 7) Pose texts unidentified text elated to the Bahubuddha-sūta of the Mahāastu (LC, cf. Salomon and Baums 2007: 202) unidentified text (Oldenbug Collection fagment, Voob ea-desi atoskai a 2006) Unspecified fagments undetemined texts (BL 8,, 2/, 22, 24, 27, Salomon 999: 39, 46 52) emnants of small potions of text o lettes (PC 4, 5, 7, Salomon 998: 45f.) Mahāyāna liteatue Only ecently could a gowing numbe of Mahāyāna texts be identified among the Gāndhāī manuscipts. Thee of them belong to the Bamiyan fagments and can consequently be ascibed to a slightly late peiod than the bich-bak manuscipts contained in the Bitish Libay, Senio, Bajau and Split Collections. Radiocabon dating as well as the athe adanced stage of the Sanskitization of language and scipt suggest a date in the 3d, pehaps een ealy 4th c. AD (Allon et al. 2006: 289f.). The Mahāyāna texts identified so fa among the Bamiyan fagments can be attibuted to Gāndhāī esions of the Bhadakalpika-sūta, the Saapuṇyasamuccayasamādhi-sūta and the Bodhisattapiṭakasūta (Allon and Salomon 200: 6 9). While these manuscipts indicate the pesence of Mahāyāna liteatue in the late phase of Gāndhāī, thee is now clea eidence fo the ealie peiod as well. Two of these old Mahāyāna texts ae pats of the collections studied in Belin. The text easiest to ealuate calls itself Pajñāpāamitā in the colophon and has poed to be a esion on which the classical Aṣṭasāhasikāpajñāpāamitā (ASPP) is based. The manuscipt is pat of the Split Collection (SplitC 5, Falk 20). It is peseed to a length of about 90 cm, both sides hae suffeed fom wea to an equal extent (Figue 3). Figue 3: Fagment fom the Pajñāpāamitā manuscipt of the Split Collection (Photogaph: Hay Falk) 20 We would like to thank Klaus Wille, who took the touble to seach his database of Buddhist texts fo possible paallels of this text and the unidentified new Khaoṣṭhī fagment fom Peshawa pesented aboe (p. 56).