Modem scholarly acquaintance with the "sutra of the burden-bearer" (Bharaharasatra) goes as far back as the Western academic "discovery" of Indian Buddhism, for it is none other than Bumouf himself who first drew attention to a fragmentary text which, according to him, "directly established [...] [t]he existence of a thinking subject." Burnouf's source was Yasomitra's partial quotation of the sutra in the Abhidharmakosavyakhya (AKVy). (1) A little less than sixty years later (1901, and especially 1902), La Vallee Poussin provided new source materials to supplement the AKVy fragment (the Sanskrit of which had been published by Minayeff in 18942), alluded to the recently published (1899) Pali recension of the sutra, and criticized Hardy's interpretation of the text.3 He showed clear awareness of the fact that the sutra belonged to the scriptures most consistently resorted to by the Buddhist Pudgalavadins ("Personalists"). (4) Half a century later (1956), Frauwallner provided one of the two Chinese versions of the sutra (that in T. 99) with a German translation, and emphasized the importance of this "well-known" document for the "polemics of the later schools" (5) --thus echoing a similar statement by Bareau. (6) As is evident from this brief history, the first modern glimpse of the sutra came from a Sanskrit version, and ample materials for a comprehensive evaluation of this text have long been available. In addition, in view of the fact that allusions to and/ or discussions of the Bharaharasutra recur in nearly all documents reflecting the pudgala ("person") controversy, a fundamental doctrinal issue both in intra-Buddhist polemics and in Buddhist debates with their non-Buddhist opponents, one can only express surprise that no study has yet been dedicated to either its non-Pali recensions (for it is one or more Sanskrit recension[s] of the sutra that underlie[s] the intersectarian polemics) or its various scholastic interpretations. The present essay is meant as a step in this direction.
This study is comprised of two parts: first, a presentation of what appears to be a/the (Mula)Sarvastivada version of the sutra, and second, an evaluation of the uses to which the sutra was put in both Buddhist inter-sectarian conflicts and in discussions with non-Buddhists.
The first portion below presents an attempt at reconstructing one Sanskrit recension of this small but influential text found now in the Samyuktagama. (7) Four complete versions of the Bhara(hara)sutra have come down to us:
the Pali of SN HI.25-26
the Chinese of T. 99 (H) 19a15--b1
the Chinese of T. 125 (H) 631c11-632a6
the Tibetan version of Samathadeva's in extenso quotation in the Upayika Abhi-dharmakosatika (hereafter Upayika Tika; D nu 85b4-86a4/P thu 132a7-b7).
Of these four recensions, the closest ones are certainly T. 99 and Samathadeva's quotation, both of which differ significantly from SN I11.25-26 and T 125 (which, again, exhibit important mutual divergencies) and are likely to go back to a homogeneous (Sanskrit?) prototype. The present Sanskrit reconstruction of the portions not quoted mainly relies on the Tibetan of the Upayika Tika and T 99. Like most sutras the Bharahara is composed of otherwise well-attested building blocks reorganized in order to fit a new didactic structure and purpose (What is the burden? What is the taking up of the burden? What is the laying down of the burden'? What is the bearer of the burden?). Of these building blocks, some are of a material/didactic nature and find exact parallels in sutras such as the Dharmacakrupravartana (a Sanskrit version of which is incorporated in the Catusparisatsutra [CPSu] and the Sramanyaphala (substantial fragments of which have been preserved in the Sanghabhedavastu [SBhV]; others are of a formal/rhetorical nature (nidana or "setting," prose-verses transition formula, etc.), and find exact parallels in other Sanskrit sutras from the Samyuktagama (and elsewhere).
The stanzas that are meant to summarize the (prose) teaching of the sutra recur in slightly different forms in Sanskrit in the Udanavarga (Uv), the CPSu, and the SBhV. However, the backbone of the Bharaharasutra has no known parallel in Buddhist sutra literature. Fortunately, several fragments (identified in the Turfan materials, the Abhidharmakosabhaya [AKBh], the AKVy, and the Tattvasangrahapanjika [TSP]) make its organizing structure easily recoverable. Although we cannot be sure, what seems most likely is that the Sanskrit recension that I propose to reconstruct here belonged to a (Mu1a)Sarvastivada milieu. This conclusion has been reached based on the sectarian identifications offered for T. 99 and the Upayika Tika, in the first place, and secondarily because it seems highly likely that the Turfan fragments also belong to this lineage. This conclusion is certainly not contradicted by the presence of quotations in the AKBh and AKVy. The present study is not primarily concerned with either the early history of the text or its Pali recension (or the very close T 125, both of whose readings are only marginally taken into consideration in the suggested reconstruction of the Sanskrit text). Thus, I am not in a position to explain when, how and why the important structural/compositional discrepancies observable between the Pali and the Sanskrit recensions came about.
The second part of this essay bears not on the sutra itself, but on its uses and interpretations in the framework of the scholastic controversy that opposed the Pudgalavadins (mainly affiliated with the Vatsiputriya and Sammitiya sects) and the "mainstream" Buddhist intellectuals (representatives of the Madhyamika, Yogacara, Sautrantika, and Vaibha[section]ika schools).(8) For whatever may have been the original intention of its "author(s)," and especially the meaning to be given to the words sycld vacaniyam in [section]2b4, (9) the sutra was liable to be, and indeed was, interpreted as claiming the existence of a person over and above the five constituents (and this seems to have been, mutatis mutandis, the interpretation of Burnouf, La Vallee Poussin, and Frauwallner). The mainstream Buddhist scholiasts evolved several strategies in order to dismiss such a literal understanding of the text. These strategies can be divided into three. Some (including the author of the Mahayanasautra1ankrabhasya [MSABh] and Sthiramati) held the teaching of the pudgala to be undergirded by pragmatic/linguistic presuppositions: in short, the word pudgala had been resorted to by the Buddha in order to refer to the various psychological, cognitive, and soteriological dispositions of a given mental stream. This is the "typological" interpretation. Others (including Candrakirti) opted for a "skill in means" interpretation: as the alayavijnana and the tathagatagarbha, the pudgala had been taught as a proselytic device in a purely provisional way. Finally, yet others (including Vasubandhu and Kamalasila) engaged in frontal exegetical polemics against the Pudgalavadins and emphasized the purely conventional nature of the pudgala referred to by the Buddha.
A SANSKRIT BHARAHARASUTRA
1.1. Tibetan version
Any attempt at reconstructing a/the Sanskrit Bharaharasutra is bound to start from T 99 and the Tibetan version of Samathadeva's quotation of the entire sutra in his Upayika Tika For in spite of discrepancies in the nidana and most "stock phrases" (which may or not reflect different sectarian recensions of the text), these versions of the sutra most certainly translate a fairly homogeneous Sanskrit text, a conclusion that is supported by the fact that these versions faithfully reflect the wording of the available Sanskrit fragments. Below is the sutra in the Tibetan version of Samathadeva's Upayika Tika (D nu 85b4-86a4/ P thu 132a7-b7):
[section]1. glen (a) gzi mnan du yod pa na 'o II
[section]2a1. dge slon dag khur dan I khur len pa dan / khur spon ba dan I khur khyer ba dag (b) bstan par bya ste /
[section]2a2. de legs par non la yid la zun zig (c) dan bsad par bya '0 II
[section]2b1. khur gan ze na I ne bar len pa 'i phun pa lna rnams so II lna gan ze na I gzugs ne bar len pa 'i phun po dan I tshor ba dan I 'du ses dan I 'du byed dan I rnam par ses pa ne bar len pa'i phun po'o II
[section]2b2. khur len pa gan ze na I dga' ba 'i 'dod chags dan lhan cig par gyur pa 'i sred pa yan srid par kun 'byun ba de dan de la (d) mnon par dga' ba 'o II
[section]2b3. khur spon ba gan ze na I gan gi tshe dga' ba 'i 'dod chags dan lhan cig par gyur pa 'i sred pa yan srid par kun 'byun ba de dan de la (e) mnon par dga' ba 'di nid ma lus par spans sin so sor 'gags pa dan I nub pa dan I zad pa dan I 'dod chags dan bral ba dan I 'gags pa dan I ne bar ze ba dan I mi snan ba 'a II
[section]2b4. khur khyer ba gan ze na I gan zag ces brjod par bya ste I tshe dan ldan pa min ni 'di lta bu I rus ni 'di lta bu I rigs ni 'di lta bu I kha zas ni 'di Ita bu I bde ba dan sdug bsnal ni 'di Ita bu dag myon zin tshe rin pa 'am thun nu 'di tsam zig tu gnas te I tshe 'i mtha' 'di tsam du gyur zes bya 'o (f) II
[section]2c. dge slon dag khur daft khur len pa dan khur spoil ba dan khur khyer ba dag bstan par bya 'o zes nas ji skad du bstan pa de rgyas par rab tu bsad do II
[section]3a. bcom ldan 'das kyis de skad ces bka' stsal cin I bde bar gs'egs pas de skad gsuns nas I ston pas gzan yah 'di skad ces bka' stsal to II
[section]3b1. sin tu lci ba 'i khur spans nas II slar yan khur dag len mi 'gyur II khur len pa ni sdug bsnal te I khur spans pa ni bde ba yin II
[section]3b2. sred pa thams cad rab spans sin II kun sbyor thams cad zad pa yis (g) / blan bya yons su ses pa na (h) II yan srid dag tu 'gro mi 'gyur II zes gsuits so II.(10)
1.2. Sanskrit reconstruction
On the basis of the testimony presented below ([section]1.4), a corresponding Sanskrit version of the sutra can be reconstructed as follows:
[section]1.(A) sravastyam nidanam / (A)
[section]2a1. (B, C) bharam ca vo bhikyavo...