Even the century or centuries in which Nagarjuna lived cannot be confidently identified. Scholarly references to Nagarjuna frequently take it as given that he lived in the second century, sometimes specifying the latter half of it.(2) On the basis of the archaeological evidence reported by K. R. Subramanian, whose 1932 publication is cited above, Venkata Ramanan drew the conclusion that Nagarjuna flourished within the period A.D. 50-120.(3) A recent account of Buddhist philosophy, by D. Kalupahana, similarly takes Subramanian as an authority, yet (citing Nakamura) assigns Nagarjuna to A.D. 150-250.(4) Shohei Ichimura claims that Nagarjuna lived from about A.D. 50 to about 150.(5) He says: "The earliest dates ever postulated for the times of Nagarjuna were proposed by Prof. H. Ui as A.D. 113 and 213"(6) but as we have seen this is wrong.

Substantially later dates also find their protagonists. For example. E. Lamotte was influenced by Chinese evidence, interpreted by him to entail that Nagarjuna's dates were A.D. 243-c. 300.(7) Prima facie, Chinese sources appear more useful than Indian ones (although as we shall see the results for the biography of Nagarjuna are disappointing), and this chronology, supported also by J. May,(8) should not be ignored.

Chronologies proposed vary freely between the first and third centuries.(9) One scholar, indeed, has claimed that "we do not think it an impossibility or absurdity to assign to him an age of about two hundred years," and assigns Nagarjuna's life to a period from the early first century B.C. to the early second A.D.(10) Nagarjuna's career has been variously located in widely different parts of India, often in the south (the late Satavahana stronghold along the Krishna, or further north in "Daksinakosala"), but sometimes in the west, the northwest, or the northeast. Again, for some, there is only one Nagarjuna, while for others there are two, three, or four.(11)

There is no need, however, to engage in any extended survey of the modern scholarship. It is clear that what has proved "baffling to the most brainy" is the quest for knowledge about Nagarjuna's life. Is the present state of confusion caused by any defect in scholarly method or rigor, or is it caused by the sheer inadequacy of the evidence to support any one convincing interpretation? This question needs to be confronted squarely. If (as I believe) the second answer is the correct one, a demonstration of the limits of the evidence can serve a valuable purpose. The common habit of assigning Nagarjuna to the second century is not properly justified; it is a self-validating majority vote, or a median possibility, rather than a demonstrable probability. I am inclined to believe that Nagarjuna may have lived later, but this cannot be proved; the purpose of the following survey is chiefly to show that claims commonly made are not well founded.

This is not, of course, the only sort of purpose that might be served by a study of the sources bearing on Nagarjuna's life. An important set of questions, not addressed here, concerns the role of hagiography in religious history, a topic that has been attracting increasing attention. The treatment of Nagarjuna as a supernatural figure is itself a phenomenon to be investigated.(12) However, my interest is not in those sources which present the fantastic but in those most likely to furnish historical evidence.

In what follows, a positive methodological contribution is also intended. It concerns our treatment of the many later works, mostly more or less tantric, which are traditionally credited to "Nagarjuna" but are unlikely to be by the original madhyamika philosopher. The later works in question were written by real people, but it is perhaps an error to suppose that they must all be attributable to some unique individual (or even two or three distinguishable individuals) who happened, confusingly, to have also the name Nagarjuna. Instead of a very small number of authentic Nagarjunas, perhaps we should seek rather a large number of people who played parts in the transmission of a legend. This point will be amplified below.

Finally, the present study is also informed by a survey of the results of archaeological research in the sites along the Krishna, a category of evidence which needs to be more closely integrated into the study of Nagarjuna's life than hitherto.

For the purpose of this discussion, unless otherwise specified, the name Nagarjuna shall designate the author, whoever he may be, of the Mulamadhyamakakarikas and of whatever other works can be agreed to come from the same writer. The interest here is in the philosopher of Madhyamaka; works of other sorts (tantric or proto-scientific) will be noticed only insofar as they have been held by some writers to share authorship with the Nagarjuna here identified. Where it is necessary to speak of more than one individual possibly bearing the same name, they can be distinguished as Nagarjuna I (the madhyamika), Nagarjuna II, and so forth.

Various sorts of evidence will here be given short shrift, on the assumption that there is little to be gained from them. Chronologies based on traditional lineages of teachers in Chinese or Tibetan do not appear promising avenues of enquiry. Tibetan sources in general seem too far removed from the age of Nagarjuna to furnish significant independent evidence. The study of the internal evidence of some philosophical literature (Nyaya, for example), is of a different order, and may raise more questions than it answers; here it is left on one side.(13)

A table of all the works traditionally accredited to Nagarjuna would need to include a very large number of texts extant in Tibetan and Chinese, as well as those in Sanskrit, but many of these are not taken seriously and need not concern us. There is no space here for a review of the case for including or excluding particular works in a bibliography of Nagarjuna; we can only note that the attributions of modern scholars all differ, as can be seen from a comparison of the conclusions of a number of writers: Winternitz,(14) Robinson,(15) T. R.V. Murti,(16) D. Seyfort Ruegg,(17) C. Lindtner,(18) and P. L. Vaidya.(19)

There is perhaps only a small core of the philosophical texts that clearly belong with the Mulamadhyamakakarikas as the work of one author. Of the works credited to Nagarjuna in the Tibetan tradition, six (the Mulamadhyamakakarikah, Sunyatasaptati, Vigrahavyavartani, Vaidalyaprakarana, Vyavaharasiddhi and Yuktisastika) are listed together as works of Nagarjuna by Buston (thirteenth to fourteenth centuries); these are specifically philosophical works. Elsewhere, as has sometimes been overlooked, he mentions others - the Ratnavali, Sutrasamuccaya, Bodhisambhara[ka], Suhrllekha, and the Svapnacintamaniparikatha.(20) The credentials of the Sunyatasaptati, the Vigrahavydvartani and the Yuktisastika are rarely disputed. It is difficult to study the Vigrahavyavartani, for example, without recognizing the essential identity of thought behind it and the Karikas. The Catuhstava, also commonly attributed to Nagarjuna I, is a special case which will be noticed further below; it is a set of four devotional hymns, thus quite different in genre from the others mentioned, and there is doubt about which particular four hymns constitute it. Also significant is the attribution of the Sutrasamuccaya to Nagarjuna by Santideva in his Bodhicaryavatara,(21) which constitutes a relatively early attribution of one core text.

Some of the works traditionally attributed to Nagarjuna (and accepted as his works by C. Lindtner(22)), such as the Bodhicittavivarana and the Vaidalyaprakarana, have had their authenticity seriously questioned.(23) The attribution to Nagarjuna of the Suhrllekha, a manual largely of ethical teachings in the form of a letter to a friend, a king, has been contested.(24) As for the Ratnavali, T. Vetter's analysis of its metrical characteristics and word frequency casts doubt upon it too.(25)

Among the miscellaneous works of chemistry and medicine traditionally ascribed to Nagarjuna, it should be noted that the Yogasataka, a medical text, has been accepted by Filliozat,(26) but the attribution is arguably implausible.(27)

The Mahaprajnaparamitopadesa has long been treated as an authentic work of Nagarjuna, but in fact is unlikely to be by him despite the weight of tradition; Lamotte originally accepted the tradition but changed his mind before he edited the third volume of his study and translation.(28)

The only conclusion that can be advanced here is that relatively few works can be treated with any confidence as authentic creations of the master. These few are written in an abstract, anonymous style. Thus, when we turn to internal evidence for biographical information, there are no very firm conclusions to be drawn, chiefly because the best-attested works are the least chatty. The best sort of source for biographical materials must be sought in other works which refer to Nagarjuna. Let us first consider originally Indian sources (some of which are no longer extant in Sanskrit).

The Lankavatara, known in Chinese and Tibetan translations, dates from as early as the fifth century. Chapter X was added to it, presumably between 443 and 513 A.D., the dates of the translations of Gunabhadra and Bodhiruci, respectively.(29) It contains an apocryphal prophecy of the Buddha, saying: daksinapatha-vedalyam bhiksuh sriman mahayasaskah nagahvayah sa namna tu sadasatpaksadarakah 165(30)

In Vedali [Vidarbha?], in the southern region, there will be an illustrious monk of great renown bearing, indeed, the appellation Naga (or by name Nagahvaya [?]); he will destroy the views of existence and non-existence.(31) The following verse celebrates this "Nagahvaya" as the declarer of Mahayana.

Various other sources associate Nagarjuna with Vidarbha, and some have taken the "Vedali" here to refer to it, but this has been disputed.(32) Tibetan tradition picked up the association of Nagarjuna with Vidarbha, but the origins of the tradition are obscure. The present passage is weak evidence. In the Tibetan, the name appears as Vedahi,(33) and P. S. Sastri has argued with abundant reference to local toponyms that a name like Vedali is likely to place Nagarjuna's birthplace in the Andhra country.(34) It has been seriously questioned whether "Nagahvaya" (whether a name in its own right, or a manner of allusion: "called Naga") really identifies our Nagarjuna. Nagahvayah sa namna, of course, may imply that the term ahvaya, "appellation," is part of the name, and from Sumpa and Taranatha we have references to a Nagahvaya who is apparently a disciple of the original Nagarjuna or who is otherwise known as Tathagatabhadra.(35) The doubters may all be wrong; given the fame of Nagarjuna, he rather than anybody else is likely to be given the credit for assailing the concepts of existence and non-existence; and there is some precedent for seeing "called Naga" as a way of referring to him.(36) However, it is clear that the Lankavatara does not offer any potential biographical information beyond the problematic allusion to "Vedali."

Also relatively early, but evidence of quite a different sort, is an inscription on a stele originally found on the northwest side of the Jaggayyapeta stupa. This site is on the Krishna. The inscription is on the base of the panel containing a high-relief standing Buddha image. It reads:

svasti bhadanta nagarjunacaryyasya sisya[syo] jayaprabhacaryya[h] tacchisyena ca[ndra]prabhena karapitam satu[tya?]-sugata-gataprasada-visesa-visista-samsare devamanu[ja]vibhutipurvvakam buddhattva-prapti-nimittam buddha-pratimam pratistha[stha]pitam anumodana [pakse?] kurvvantu sarvve saugaty-agrya[?]nyo pi(37)

The somewhat problematic text here incorporates the readings suggested by R. G. Bhandarkar and cited by J. Burgess; the sense of it can be rendered approximately:

Salute! The disciple of the Teacher Nagarjuna was the Teacher Jayaprabha. May all beings, even though themselves denied the perfection of Buddhahood, rejoice in sympathy at the Buddha image caused to be made by (Jayaprabha's) disciple Candraprabha, inaugurated for the purpose of the attainment of Buddhahood following upon prosperity in the world of gods and men in the course of a round of births distinguished by the great favors of the real Buddha.

Thus the inscription records an endowment of a Buddha image by an individual whose teacher's teacher was "Bhadanta Nagarjunacarya." This would be exciting if it could easily be made to fit plausibly with other evidence of the individual we are seeking. However, on paleographical grounds the inscription has been dated to about 600 by Burgess, to 450-500 A.D. by Ramachandran.(38) A fourth-century date for Nagarjuna, the earliest that could be managed on this evidence, has not been persuasively argued so far, and the allusion in this inscription must remain a mystery for the time being, though there will be occasion to revert to it briefly at the end of this study.

Bana's Harsacarita (seventh century) is the first Sanskrit text to make the link, widespread in the literature, between Nagarjuna and his friend "Satavahana, the lord of the three oceans."(39) The relevant passage tells of the fate of a string of pearls magically made by Vasuki, Lord of Serpents, from the tears wept by the moon:

samatikramati ca kiyatyapi kale kadacit tam ekavalim tasman nagarajan nagarjuno nama nagair evanitah patalatalam bhiksur abhiksata lebhe ca / nirgatya rasatalat trisamudradhipataye satavahananamne narendraya suhrde sa dadau tam / sa casmakam kalena sisyaparamparaya katham api hastam upagata.(40)

As time passed by, eventually a monk called Nagarjuna was brought to Patalatala [the abode of the nagas among the underworlds] by the nagas, begged this string from the Lord of the Nagas, and received it. Upon his departure from the infernal regions, he gave it to his friend the king called Satavahana, Lord of the Three Oceans, and in the course of time it came to our hand through the lineage of disciples.

The link with the Satavahanas, obscure though it is, nevertheless constitutes one of our best pieces of evidence about the life of Nagarjuna; it will be taken up again below. The rhetorical character of the Harsacarita, though, makes it unwise to place weight upon its fine detail.

Another text of equivocal value is the Mahameghasutra, extant in Chinese and Tibetan versions, though the line of descent of these versions is unclear and there is doubt how far they share a single origin.(41) According to Demieville, this text, which was translated into Chinese by the fourth and fifth centuries, was probably the same as one referred to by Candrakirti in the Madhyamakavatara, where he calls it the Aryadvadasasahasramahamegha, attributing to it the prophecy that Nagarjuna would be born in the future and teach the Dharma.(42)

Now, the Chinese text contains a number of spurious prophecies attributed to the Buddha; according to one, the Buddha prophesied to emissaries from the south, coming from the Black Mountain (possibly a reference to the Andhra country and the hills around Nagarjunakonda) that, twelve hundred years after his Nirvana, there would be a great king called ???, soto-p'o-ho-na, EMC sa-ta-ba-xa-[na.sup.h] (reconstructions of Early Middle Chinese follow Pulleyblank), clearly representing "Satavahana." Further, also in south India, a certain "Nagaraja" would be born in a place called Kusumamala or Sumanamala, near the river Supaya, in Surastra, and give his life for the Dharma.(43) This prophecy is of doubtful relevance because it does not specify the name Nagarjuna and offers biographical details not found elsewhere (though the name of the river does roughly correspond to the Tibetan version); the tenuous case for a connection must be built upon Candrakirti's use of the text, if indeed the prophecy in the version of it available to us accurately represents what Candrakirti's source said.

According to another prophecy in the Chinese version, the Buddha foretold that seven hundred years after his Nirvana, in the southern kingdom called [Chinese Text Omitted], wu-ming ('no-light', i.e., 'dark', = andha, translating andhra), on the south bank of the river [Chinese Text Omitted], hei-an ('black', i.e., Krsna), there would be a town called [Chinese Text Omitted], shou-ku ('ripe-grain', representing Dhanyakataka), and here would reign the king [Chinese Text Omitted] teng-ch'eng ('Equal-ride', = Sadvahana, = Satavahana); he would have a daughter called [Chinese Text Omitted], tseng-chang, 'Increase Prosperity', who would become ruler and receive homage from all the world.(44) The Tibetan version, which is evidently based on a variant text, declares that four hundred years after the Nirvana, on the north bank of Sundarabhuti in a country called Rishila in south India, there would be born as a Licchavi, in the time of King Vipatticikitsaka, a monk called Naga who would give his life for the Drarma. Tucci thought that this "Naga" could not refer to Nagarjuna.(45) The Mahameghasutra therefore offers us a "Naga" and a "Nagaraja," named in proximity to a prophecy about a Satavahana ruler at Dhanyakataka. Most authorities have not been impressed by the text as evidence, but Candrakirti's allusion to what appears to be the same work does suggest that, though garbled, it may provide early evidence of a tradition, strong in later centuries, that linked Nagarjuna, the teacher of Madhyamaka, with Satavahana rulers in the south.

The reference in the Tibetan version to a woman later becoming a ruler has been seen as a reflection of the importance of royal women in the culture to which the original text belonged, and Levi points to the prominence of women as donors to Buddhism at Nagarjunakonda, where the Iksvaku dynasty ruled in the third century.(46) Indian sources offer little from this time until about the eleventh century, when their distance from the origins of traditions about Nagarjuna severely limits their usefulness. The Kathasaritsagara, firstly, tells the story of King Long-lived, endowed with all gifts, whose wise minister, Nagarjuna, of bodhisattva origin, was possessed of compassion, generosity and good conduct:

cirayur namni nagare cirayur nama bhupatih purvam cirayur evasit ketanam sarvasampadam tasya nagarjuno nama bodhisattvamsasambhavah dayalur danasilas ca mantri vijnanavan abhut.(47)

According to this story (which became popular in later writings, whatever its origin) Nagarjuna had discovered the secret of immortality and shared it with his royal master; the two of them lived for many centuries, but ultimately a crown prince, greedy for the throne, tricked Nagarjuna into placing himself in the prince's power and submitting to decapitation, and the grief-stricken king died soon after.(48) Here we observe Nagarjuna the magician, the hero of many a legend in the Tibetan and Chinese stories of medieval times, and there is little here of historical significance besides a memory of the fame of Nagarjuna.(49)

Kalhana (twelfth century) offers a reference to Nagarjuna as lord of the earth (bhumisvaro) in Kashmir at the time of Huska, Juska, and Kaniska, living in the grove of the six arhants (sadarhadvana):

athabhavan svanamankapuratrayavidhayinah huskajuskakaniskakhyas trayas tatraiva parthivah 168 sa viharasya nirmata jusko juskapurasya yah jayasvamipurasyapi suddhadhih samvidhayakah 169 te turuskanvayodbhuta api punyasraya nrpah suskaletradidesesu mathacaityadi cakrire 170 prajye rajyaksane tesam prayah kasmiramandalam bhojyam aste sma bauddhanam pravrajyorjitatejasam 171 tada bragavatah sakyasimhasya paranirvrteh asmin mahilokadhatau sardham varsasatam hy agat 172 bodhisattvas ca dese 'sminn eko bhumisvaro 'bhavat sa ca nagarjunah sriman sadarhadvanasamsrayi 173(50)

This tells us that Huska, Juska and Kaniska were three rulers over the three cities that bore their names; the pure-souled Juska founded a vihara as well as Juskapura and Jayasvamipura; these kings, although of Turkish descent, gave themselves to pious deeds, building in Suskaletra and other places foundations such as monasteries and shrines; during their glorious reign Kashmir was mostly dedicated to the Buddhists, whose mendicant lifestyle enhanced their fame; at this period one hundred fifty years had passed since the Parinirvana of the Buddha; and in this country a bodhisattva, Nagarjuna, was the sole sovereign of the earth (eko bhumisvaro), dwelling in the Grove of the Six Arhats.

There has been some discussion of the meaning of this lordship,(51) which need not concern us. Nagarjuna did not rule as a lord anywhere; possibly the allusion can be taken to refer to his spiritual eminence, but the character of the passage as a whole should warn us against treating it as historical chronicle based on secure information. If (as we must) we reject the information that at the time in question the Buddha had been dead for a hundred and fifty years, why should we cling to the contemporaneity of a "lord of the earth" Nagarjuna with the Saka kings named?

It is possible, indeed, to find historical referents for the kings named; Kaniska could be the rather shadowy Kaniska II, supposed to have been ruling forty-one years after the accession of the dynastic founder and patron of Buddhism, Kaniska I, and the other two kings named could be the former's contemporaries Huviska and Vajeska.(52) Scholars have debated the chronology of Nagarjuna's life in relation to that of Kaniska I, whose reign in the northwest possibly began in A.D. 78;(53) however, there is little point in seeking to secure Nagarjuna's chronology to another which is itself so insecure; the serious doubts about the historical career of any Kaniska II, and about the date of the reign of Kaniska I, make it unwise to rely heavily on the links suggested in Kalhana's poem written a millennium later.

Two works which have the character of classical Mahayana texts, but may be spurious, are the Mahabherisutra and the Mahamdyasutra, both of which mention Nagarjuna, but their historical value is dubious.(54)

A later source, in impure Sanskrit, much used by Tibetan writers in their compilations of Buddhist lore, was the Manjusrimulakalpa. It contains a spurious prophecy by the Buddha:

bodhim prapsyati sarvajnim uttamartham acintiyam caturthe varsasate prapte nirvrte mayi tathagate nagahvayo nama sa u bhiksuh sasane 'smin hite ratah muditam bhumilabdhas tu jived varsasatani sat mayuri namato vidya siddha tasya mahatmanah nanasastrartkadhatvartham nihsvabhavarthatattvavit sukhavatyam copapadyeta yadasau tyaktakalebarah so 'nupurvena buddhatvam niyatam samprapatsyate(55)

Four hundred years after the Buddha's Nirvana, there would live a monk named Naga (nagahvaya) who would obtain the muditabhumi and would live six hundred years; he would acquire the Mahamayurividya and know the truth of non-substantiality (nihsvabhavarthatattva). He would go to the heaven Sukhavati on his death and in due course would certainly attain Buddhahood. The Tibetan version represents "Naga" as klu-shes-de-hbod, the equivalent of "Nagahvaya," 'called Naga'; according to Walleser the reference is definitely to Nagarjuna, and this, he thinks, helps to validate the expression nagahvaya as a reference to Nagarjuna where it also occurs in the Lankavatara.(56) It certainly appears plausible, given the reference to this Naga's knowledge of the emptiness of things (nihsvabhavarthatattvavit), to take the passage as an allusion to Nagarjuna the madhyamika.

An eleventh-century work based on Indian travels, but not by an Indian, is the account of Al-Biruni (Alberuni), who mentions a famous alchemist named Nagarjuna who lived a century before (i.e., in the tenth century); he came from Daihak near Somnath and was an expert in the science of rasayana and used plants to promote health, longevity, and enhanced sexuality.(57) Reference to the context, which contains various stories of miracles that are repeated uncritically, does not inspire confidence in its historical value. V. W. Karambelkar attaches weight to this evidence, although other authorities do not.(58) What we observe here is an early stage of the myth of Nagarjuna the practitioner of wonderful arts. Thus the evidence from India is disappointing. The stronger sense of history as a record of objective circumstances in China might be thought a more hopeful field to explore.

Kumarajiva, firstly, made translations of Buddhist texts very early in the fifth century. From him (whether as author or as translator) or from his disciples comes a short biography of Nagarjuna, the Long-shu p'u-sa chuan, the substance of which is largely mythical, but does offer a chronology: "Since Nagarjuna left the world up to the present, more than a hundred years have passed."(59) However, it is not clear whether these should be seen as words of Kumarajiva, or of an earlier work which he passed on, or of his disciples who edited the material he bequeathed them(60)

One disciple of Kumarajiva was Seng-jui, who wrote a preface to the Satyasiddhisastra. This preface is now lost, but according to a passage quoted by Chitsang, Asvaghosa was born 350 years after the Buddha's Nirvana and Nagarjuna was born "in the year 530"; it appears that Nagarjuna's year is dated from Asvaghosa's and thus comes 880 years after the Nirvana.(61) What weight should be attached to the date thus derived (a pedantic calculation shows that it should be A.D. 244, not 243, as is usually advanced) is not clear, for we have no ground for believing that the Chinese possessed Indian records capable of supporting such a precise chronology.(62) As Lamotte acknowledges: "On n'echappe pas a l'impression que toutes ces datations relevent de vues theoriques sur les etapes successives de la Bonne Loi et que, en chronologie absolue, leur valeur est plutot faible."(63)

Perhaps the earliest solid piece of biographical information of value comes in the sixth century from the translator Paramartha, who attests the connection between Nagarjuna and a Satavahana ruler that in later writings was to be repeated over and over again.(64) Hsuan-tsang, the famous seventh-century pilgrim, comes next. He visited a place which he identified as the capital of Kosala (i.e., Daksinakosala), the location of which has not been identified with certainty (it might have been in the area of Wairagarh or Bhandak(65)) and claimed that not far to the south of the city is an old sangharama, by the side of which is a stupa that was built by Asokaraja. . . . Afterwards Nagarjuna Bodhisattva (Lungmen-p'u-sa) dwelled in the sangharama.(66)

Nagarjuna had the friendship and patronage of the Satavahana king, who provided him with a residence. Nagarjuna concocted medicines for longevity, through which he and the king lived for centuries until Nagarjuna was decapitated; Hsuan-tsang offers a version of the story of Nagarjuna's end mentioned above. This king is called So-to-p'o-ho, clearly intending the dynastic name Satavahana, as we saw earlier.(67)

Traveling three hundred li to the south-west, Hsuantsang came to Po-lo-mo-lo-ki-li (Bhramaragiri), the site of a great vihara made by the king for Nagarjuna by hollowing out the rock in five tiers. The exact location of this vihara cannot easily be identified,(68) but the same recognizable monastery had already been described by Fa-hsien in the fifth century,(69) and in the early eighth century a Korean pilgrim, Hui Ch'ao, passed the same way.(70) Different versions of the Chinese text call the monastery [Chinese Text Omitted], hei feng ('black peak') and [Chinese Text Omitted], also pronounced hei feng, 'black bee', which as has been argued could represent the Sanskrit bhramara, epithet of the Saivite goddess.(71) Some have thought that this monastery bestowed upon Nagarjuna must have been at Nagarjunakonda, but strictly speaking the distances and directions stated in the text do not allow us to place it so far south.(72)

The narrative of Hsuan-tsang's travels resumes with a record of his journey southwards to the Andhra country.(73) It is in this general area, of course, that in the second and third centuries, first, the Satavahanas at Dhanyakataka and, then, the Iksvakus at Vijayapuri ruled at capital cities which could have been centers of patronage for celebrated Buddhist teachers. It is generally accepted that the site of Dhanyakataka was adjacent to the ancient stupa at Amaravati (where according to the much later Tibetan tradition Nagarjuna made benefactions). There are problems in the reconstruction of the history of this area, which have been discussed elsewhere.(74) Elsewhere Hsuan-tsang refers to Nagarjuna as one of "four suns which illumined the world," along with Asvaghosa, Aryadeva, and Kumarabdha,(75) but it is not likely that this allusion can furnish any useful chronological information: the four suns did not shine at the same time.(76)

Here it is necessary to revert briefly to the question of Kaniska's relationship with Nagarjuna, mentioned above, for Shohei Ichimura has recently appealed to it and to the "four suns" allusion in making his case that Nagarjuna lived in western India from about A.D. 50 to about A.D. 150.(77) The case for this chronology rests essentially upon the evidence of Hsuan-tsang that Nagarjuna was a contemporary of Asvaghosa, the assumption that Asvaghosa was a contemporary of Kaniska I, and the assumption that this Kaniska's reign was late in the first century.(78) The chronology of Kaniska, as noticed before, is notoriously problematic; so is the relationship of Nagarjuna to Asvaghosa. If we accept the poetic allusion to four suns as serious evidence that Nagarjuna and Asvaghosa were contemporaries, we must find better reasons for heeding it than the earlier evidence of Chi-tsang, according to which, of course, the two cannot be contemporary.

One other Chinese pilgrim, the late seventh-century traveler I-ching, shows an interest in Nagarjuna, telling us for example of the master's reputed alchemical wizardry (prolonging his life by breathing water), and linking him with Asvaghosa, Asanga, and Vasubandhu.(79) Further, he quotes the Suhrllekha at some length and identifies the friend to whom it was written as [Chinese Text Omitted], shih-yin-te-chia, EMC [zi.sup.h]-jin-tak-kia, titled So-to-p'o-han-na (spelled as above, but in this text with han represented by [Chinese Text Omitted]).(80) There has been much discussion of the proper Indian restoration of these names;(81) the latter is clearly Satavahana, while the former (variously taken to represent Santaka, Jantaka, Jayandhra, etc.) could possibly be based on the common Satavahana name Satakarni. We may note here that various forms of the name Satakarni (used by many Satavahana rulers) occur in puranic sources, including notably Santikarni, along with such variants as Sata, Sati, Sada and Sataka.(82)

These are the main relevant Chinese sources; Tibetan works constitute another category, but despite their importance for many aspects of Buddhist history they must be largely omitted from consideration here for reasons of space. They all stand at a substantial distance in time from the Nagarjuna I whose biography is here in view, and the evidence that they contain must be very indirect. A more substantial review would need to take account of the writings of Buston (1290-1364),(83) 'Goslo-tsa-ba gZon-du-dpal,(84) Taranatha,(85) and Sumpa.(86) Deserving of brief comment is a work by Manluns, who lived in the thirteenth century and described a magnificent stupa in Dhanyakataka where the Buddha was supposed to have preached. This picked up a tradition given currency in Tibet by the Kalacakra Tantra; Manluns took the further step of associating Nagarjuna with construction work at Dhanyakataka: temples to the northwest and southwest of the stupa are attributed to the benefaction of Nagarjuna. This would be important if the stupa described by Man-luffs could be identified with one that actually existed (such as the stupa at Amaravati), particularly since the author claims to have been to India and seen what he describes. Nevertheless, the detailed description that he gives appears to be a fanciful exercise, partly inspired by the Kalacakra Tantra, and cannot with confidence be attached to any known historical site. There are serious objections to any identification, including any along the Krishna River.(87)

This summary omits the many references to Nagarjuna as a mahasiddha who accumulated an impressive array of spells and powers, and who figures in many a list of past tantric masters, from which it would probably be vain to attempt any serious reconstruction of his biography. Most of the colorful and entertaining parts of the Nagarjuna evidence, unfortunately, must be omitted from this study. In many respects, the later traditions about a tantric Nagarjuna seem to represent a different person, and many have thought that the historical basis for the tantric legend must have lain in some second individual who lived long after the madhyamika master. For a number of scholars, the solution has seemed to be that two (or more) people called Nagarjuna, one of them a tantric mahasiddha, have become confused in Buddhist tradition.(88)

A variety of alchemical and medical treatises - not just tantric works - have been attributed to Nagarjuna, and some have been ready to attribute them to the madhyamika philosopher,(89) while others have considered that perhaps as many as four historical individuals, living at different times, came to be confused.(90) The argument for two (or more) Nagarjunas is basically that the aeuvre and the traditions credited to Nagarjuna point to a madhyamika acarya some time in the first three centuries A.D. and to a tantric mahasiddha, an adept in charms and spells, who would probably have lived later, when tantra flourished.(91) The tantric Nagarjuna can be identified as the founder of the Guhyasamaja system, as the sixteenth mahasiddha, and as the disciple of Saraha or Rahula;(92) it has been pointed out that the Guhyasamaja concepts do not date from before Dharmakirti and have a basis in Yogacara, not Madhyamaka, theory.(93)

There are difficulties with the case for plural Nagarjunas; we have to decide which of the facts alleged by tradition about Nagarjuna belong to one, or to another, or to both, or all by sheer coincidence; and in many cases the decision must be arbitrary. Nevertheless, most scholars have accepted that there must have been more than one Nagarjuna behind the traditions. The argument of Jan Yun-hua that, after all, there was only one Nagarjuna(94) is therefore all the more striking.

Jan Yun-hua offers several types of evidence which he says suggest that there was no later Nagarjuna. These are directed against specific claims for the existence of a tantric (or a medical) Nagarjuna in some particular later century.(95) However, Jan does not discuss the actual authorship of most of the various later texts attributed by tradition to Nagarjuna - a discussion which is required if the case for the existence of no more than one Nagarjuna is to be sustained.

Jan Yun-hua's arguments are salutary in casting doubt on the too-facile theories which seek to identify some particular Nagarjuna II (or III, or IV) in some particular later century. What we need to recognize here is that there are different grades of existence. We should heed the lessons of Madhyamaka logic: that a later Nagarjuna should exist, and that a later Nagarjuna should not exist, are not the only possibilities. There are several grades worth distinguishing:

(a) being Nagarjuna originally;
(b) taking the name Nagarjuna upon re-ordination;
(c) coming to be seen as a reincarnation of Nagarjuna;
(d) writing a book which contributes to a tradition that looks back to Nagarjuna, and which subsequently comes to be attributed to Nagarjuna;
(e) not being Nagarjuna in any sense at all.

Later Nagarjunas could indeed come into existence, in some sense. After all, there could be two Asvaghosas, two Sarahas or Rahulabhadras, even two Aryadevas and two Vimuktisenas. Tucci points out that, particularly in the tantric tradition, the repetition of names was specially liable to occur. Many siddhas might be seen as incarnations of a single person; further, many Tibetans would take new names upon initiation into new or different schools. This implies that some masters of the Siddha-sampradaya considered themselves or were considered by their disciples as the manifestation (Tib. rnam a'prul) of the first acaryas . . . and were given the same name. This fact explains quite well the contamination which we may trace between the biographical accounts of the older masters as given in the Chinese sources and those preserved in the Tibetan tradition.(96)

This much, of course, concerns Tibetan teachers, not Indian. But there is no reason why it should not apply to the later, tantric, forms of Indian Buddhism as well.

Few of the Sanskrit sources of Tibetan traditions are preserved, or we would be in a much better position to seek pathways by which genuine historical knowledge was transmitted. But a short Sanskrit work recording alleged teacher lineages, found in Nepal by Tucci, constitutes good evidence that, in the Sanskrit tradition also, teachers in a lineage could be regarded as reincarnations of earlier masters. The work in question(97) shows that one master, Nagarjuna, was held to be reembodied in the persons of a number of later siddhas - Damodara, Advayavajra, and Ratnamati. Further, it shows that teachers could take new names on being initiated in different schools.

The tendency here attested may have begun quite early, and it is not surprising that there should be confusing and discrepant traditions, more or less worthless for the purposes of true chronology, describing the place of Nagarjuna in one instruction lineage or another.

There was very possibly, then, one original Nagarjuna, but to him was added a legend which ramified. This legend, in turn, inspired the adoption of the name of Nagarjuna by many later texts written at different times, and likely also by some later teachers in the tantric tradition. In seeking the historical reality of these texts and teachers, therefore, we are not looking for some particular individual who was "the tantric Nagarjuna," or perhaps "the medical Nagarjuna," or "the alchemical Nagarjuna," who happened to have the same name. We are looking for the multiform particular manifestations of a single legend. These later manifestations need not be embodied in any new authentic Nagarjuna. They may be embodied in different ways of using the name.

It is necessary to recognize the importance of this perspective. There are implications here for our understanding of Indian views of authorship and tradition, of truth and its expression. It may be appropriate, as a comment on the world view that is here elicited, to cite some comments by Arthur Waley about authorship which, though they concern not India but China, may bear pondering nevertheless: Thus people in early China were used to regarding books as records of traditions. . . . When real authorship began writers should give their books the appearance of being records of ancient things, rather than present their ideas as new and personal discoveries. This was as natural and as inevitable as that the first railway carriages should imitate stage coaches. These early products of authorship were not, strictly speaking, what Western bibliographers call pseudepigraphs. No pretence was made that the books in question were written by the Ancients (though this was often believed in after ages by people who could only think in terms of modern authorship). It was merely pretended that what was now set down had once been taught by such or such an Ancient. Had this method not been adopted the people could not have been induced to read the books, any more than travellers could have been persuaded to enter a railway carriage if it had not looked something like a stage coach.(98)

There had to be a single historical figure at the origin of the whole process, and he, we need not doubt, was the author of the Malamadhyamakakarikas. What evidence about him emerges? The link with the Satavahana dynasty is probably the most striking feature. De Jong suggests that Paramartha's reference to the connection, in the sixth century, may be the first occurrence.(99) Hsuan-tsang's So-to-p'o-ho and I-ching's So-to-p'o-han-na evidently represent it phonetically, and the former's, Yin-cheng, 'Leading-right', conveys the sense of Sad-vaha. (The actual etymology of the dynastic name, which is Sanskritized from a local language and has been argued by some to represent a trace of a horse cult, does not matter; it is the way the name Satavahana struck its hearers that counts.) The Tibetan bDe.Spyod can be interpreted similarly, though the interpretation is conjectural: the first element means 'happy, good' (= sat), and the second 'walking, conducting oneself' (= vahana).

Which Satavahana ruler, if any, is most likely to have been Nagarjuna's patron? Nagarjuna's relationship to early Mahayana literature requires that we should look at the first three centuries A.D. In the course of this period, Satavahana kings lost their early dominions in the west and finished as lords of the lower Krishna region. The chronology of the dynasty, which must be reconstructed from fragmentary epigraphic records and implausible puranic lists, is far from clear, and different authorities differ widely in the dates they suggest for the reigns of rulers; there would be little point in assigning precise dates here.

Gautamiputra Satakarni is likely to have reigned around the beginning of the second century;(100) it is said in a Nasik inscription that his horses drank from the three oceans,(101) which recalls the allusion to the Satavahana friend of Nagarjuna in the Harsacarita (noticed above) as "lord of the three oceans," an epithet which has been thought appropriate to a ruler who had an empire in the west as well as the east, if we see in it an exaggerated claim to control the eastern, southern and western coasts of the Deccan. P.S. Sastri has argued that the Ratnavali (assuming it was written by Nagarjuna) contains references to the vilification of Mahayana and appears to be written to a monarch who had reverted to brahmanism; since Gautamiputra Satakarni is referred to in a Nasik inscription as ekabrahmana, this king could be he.(102) (This argument, of course, depends upon the conjunction of several hypotheses.)

Several other authorities have argued for Pulumavi II (late second century) or Yajna Sri Satakarni (variously dated in the later second century or at the turn of the third(103)), but the latter of these is the more favored candidate.(104)

Vijaya Sri Satakarni, at the beginning of the third century or in its early part, is not frequently given any preference,(105) but it deserves to be noticed that Vijaya is known from an inscription on a limestone pillar recording a Buddhist endowment near Vijayapuri (as it was to become), in the Nagarjunakonda valley.(106) H. Sarkar was led by his archaeological researches at Amaravati and Nagarjunakonda to propose a revision of the late Satavahana chronology: the inscriptions of the last three rulers, Vijaya Satakarni, Candasri, and Pulumavi III, are in different places, and they may have ruled concurrently from different capitals; Vijaya (possibly r. 203-9) could have been the real founder of Vijayapuri, which after his defeat became the Iksvaku capital.(107) (His chronology, however, must be acknowledged to be speculative.)

This leads naturally to a consideration of the connection between Nagarjuna and Sri Parvata - a connection widely enough claimed for him in the later sources (such as the Tibetan writers and the Rasaratnakara(108)).

The epigraphy of Nagarjunakonda identifies a Sri Parvata at that site, consisting at least of an eminence on a spur projecting into the valley (and possibly more generally of the surrounding heights);(109) some have considered that the name designated the whole Nallamallai range, the hills which surround the Nagarjunakonda valley and extend upstream, to include Sri Sailam (itself sometimes later called Sri Parvata(110)). I. K. Sarma has made a study of the toponym Sri Parvata, and according to him it is clear that in Iksvaku times at least it designated specifically the hill in the Nagarjunakonda valley, though it may subsequently have come to be used more generally.(111)

So, on the evidence reviewed above, it is difficult to choose between the region called "Daksinakosala" and the Krishna valley heartland of the latter Satavahana kingdom (supported by the references to Sri Parvata and the Satavahanas). Perhaps we need not choose. The two areas are close enough together for us to conceive of a famous Buddhist teacher moving freely between them; perhaps, for example, he gained fame in Daksinakosala and traveled to the Krishna area (perhaps to Nagarjunakonda) to receive patronage from a ruler there.

There is certainly evidence of Buddhist settlement at Nagarjunakonda predating the establishment there of the Iksvaku capital, and it is probably here that we should look for the home of the Aparasaila sect for whose benefit, in the reign of the Iksvaku monarch Mathariputra Virapurusadatta, there was erected the mahacaitya, a dhatugarbha, which must have contributed to the celebrity of the site in the Buddhist world, while the Purvasaila sect had long been established in places further east, primarily at Amaravati. Now, it is possible to indicate a chain of circumstantial connections fastening Nagarjuna to the Nagarjunakonda-Amaravati region in late Satavahana (or possibly Iksvaku) times; it needs to be emphasized at once, though, that the point of this exercise is to show, not that such a theory is correct, but that no others are better. The links are as follows. Despite the austere rationalism of the Mulamadhyamakakarikas and the other philosophical works, Nagarjuna's Dharmadhatustava, if correctly attributed to him, represents a devotional strain of Buddha worship, in which, as Ruegg argues, one can see elements of the tathagatagarbha doctrine.(112) The tathagatagarbha doctrine was emerging in association with devotional Buddhism. Elements of it have been discerned in the Srimalasutra, which has been tentatively attributed to the Buddhists of Iksvaku Nagarjunakonda, partly because of the active role of women as Buddhist donors there and the importance of the ideal woman portrayed in the Srimalasutra.(113) Some of the monasteries of Nagarjunakonda, according to H. Sarkar, represent the early influence of devotional religion with the incorporation of stupas inside vihara enclosures.(114) Nagarjuna is widely held to have resided at Sri Parvata for at least a part of his career, and there was a Sri Parvata at Nagarjunakonda. The Purvasaila sect flourished in the Satavahana dominions in Andhra, particularly at Amaravati, and Candrakirti in his Madhyamakavatara cites verses described as "following the Purvasailas," which indicate the influence of Prajnaparamita ideas and have been associated by La Vallee Poussin with the emergence in the south of the dharmadhatugarbha doctrine.(115) The association of these elements is obviously inconclusive; but it makes as much sense as the other theories that have been advanced about Nagarjuna. At this point we can return to the question who was the "Satavahana" ruler who was Nagarjuna's friend and patron. In the first place, we must not overlook the possibility that the name "Satavahana," so familiar in the Nagarjuna legend, came to be attached by posterity to some shadowy successor, not an actual Satavahana, who ruled in a part of the area whose history was made glorious chiefly by its association with the imperial Satavahanas. It is upon this supposition that one could cling to the Nagarjunacarya of the Jaggayyapeta inscription, which has not so far been allowed to date before the fifth century and points to a fourth- or fifth-century date for Nagarjuna. The supposition would also help us to deal with the claim made in the biography of Nagarjuna attributed (albeit on grounds which we have seen to be insecure) to Kumarajiva: "Since Nagarjuna left the world up to the present, more than a hundred years have passed" - which would point to a late third or possibly early fourth century sponsor for Nagarjuna's activities.

Alternatively, we should look for this sponsor among the later Satavahanas, preferably a very late one if we are persuaded that the relevant developments in Buddhist doctrine are to be found among the caitya sects (primarily the Purvasailas and Aparasailas) in the late second and third centuries. Vijaya Sri Satakarni deserves to be considered more favorably than he has been in the past, considering that he was responsible for Buddhist endowments in the Nagarjunakonda valley and has been taken to have established his capital there early in the third century; indeed, there is no reason why Nagarjuna should not have begun his career as a royal protege under Vijaya and subsequently lived under the Iksvakus.(116) The fact that Vijaya reigned for only a short time, while the royal patron of legend lived for several centuries, is of course irrelevant; we are not looking for a ruler who lived for several centuries. We are looking only for a ruler who might have been described by so many different sources in later times as a Satavahana (whether he was one or not).

What this enquiry shows is that, however inadequate the evidence for an original madhyamika Nagarjuna I may be, it enshrines memories of a real historical person.

The story is quite different when we seek facts about any subsequent Nagarjuna. There is an important methodological point to be made here. If we assume that some particular later Nagarjuna existed, about whom some fact is treated as given (for example, that he was an alchemist), we can treat certain writings as giving information about him; however, if we do not make such an assumption, the writings are not independently capable of constituting good evidence for his existence. That is, the object of our quest may itself be an artifact of the quest (maya or gandharvanagara, so to speak). We must give proper weight to the default hypothesis that the association of the name Nagarjuna with a profusion of tantric and quasi-scientific texts is a demonstration of the absorptive power of the legend originating in a single historical Nagarjuna, the author of Madhyamaka.

1 K. R. Subramanian, Buddhist Remains in Andhra (Madras: Diocesan Press, 1932), 62.

2 For example, F. J. Streng, in Encyclopaedia Britannica, s.v. Nagarjuna; and in Encyclopedia of Religion, ed. M. Eliade (N.Y: Macmillan, 1987), s.v, Nagarjuna; T. R.V. Murti, The Central Philosophy of Buddhism (London: Allen and Unwin, 1955), 87 f. Murti also states (p. 88) that Nagarjuna was "probably a Brahman from the south who came to Nalanda" (which did not in fact become a center of Buddhist learning until the Gupta period: see A.D. Sankalia, The Nalanda University [Delhi: Oriental Publishers, 1972], 48-59).

3 K. Venkata Ramanan, Nagarjuna's Philosophy, as Presented in the Maha-Prajnaparamita-Sastra (Rutland, Vt.: Tuttle, 1966), 30.

4 David J. Kalupahana, A History of Buddhist Philosophy (Honolulu: University of Hawaii Press, 1992), 160; H. Nakamura, Indian Buddhism (Delhi: Motilal Banarsidass, 1989), 235. This chronology has been accepted by most Japanese scholars (ibid., 236, n. 4).

5 Shohei Ichimura, "Re-examining the Period of Nagarjuna: Western India, A.D. 50-150," Journal of Indian and Buddhist Studies 40.2 (1992): 8-14 (= 1079-73).

6 Ibid., 8; cf. the citation, ibid., 14, n. 3; and H. Nakamura, Indian Buddhism, 236, n. 4.

7 E. Lamotte, L'Enseignement de Vimalakirti (Louvain, 1962), 74-77.

8 J. May, 'Chugan', Hobogirin 5 (1979): 473a, 478b.

9 See D. Seyfort Ruegg, The Literature of the Madhyamaka School of Philosophy in India, History of Indian Literature, vol. VII, fasc. 1 (Wiesbaden: Harrassowitz, 1981), 4f.

10 L. Joshi, "Life and Times of the Madhyamika Philosopher Nagarjuna," Mahabodhi 73.1-2 (1965): 46f.

11 See for example V. W. Karambelkar, "The Problem of Nagarjuna" Journal of Indian History 30.1 (1952): 21-33; N. Dutt, "Nagarjunikonda and Nagarjuna," Indian Historical Quarterly 7 (1931): 636-69.

12 The legendary material has been examined by M. Walleser, "The Life of Nagarjuna from Tibetan and Chinese Sources," in Asia Major: Hirth Anniversary Volume, ed. B. Schindler (rpt. Delhi, 1979), 421-55; the dynamics of hagiography in relation to Nagarjuna are discussed by Jan Yun-hua, "Nagarjuna, One or More? A New Interpretation of Buddhist Hagiography," History of Religions 10 (1970): 139-53.

13 The importance of Nyaya principles in the elucidation of Nagarjuna's Vigrahavyavartani has even now perhaps not received all the recognition it deserves. On the relative chronology of Nagarjuna and Nyaya, see J. Bronkhorst, "Nagarjuna and the Naiyayikas" Journal of Indian Philosophy 13 (1985): 107-32; on the dialogue between Nyaya and Madhyamaka see P.S. Sastri, "Nagarjuna and Aryadeva," Indian Historical Quarterly 31.3 (1955): 193-202.

14 M. Winternitz, History of Indian Literature, vol. II (Calcutta: Calcutta U.P., 1933), 346f.

15 R. Robinson, Early Madhyamika in India and China (Delhi: Motilal Banarsidass, 1976), 26f.

16 T. R. V. Murti, The Central Philosophy of Buddhism, 88-91.

17 D. Ruegg, The Literature of the Madhyamaka School of Philosophy in India, 8-29.

18 C. Lindtner, Nagarjuniana: Studies in the Writings and Philosophy of Nagarjuna (Copenhagen: Akademisk Forlag, 1982), 9-18, esp. p. 11. His list of thirteen includes all the most probable works of our Nagarjuna I, but also includes several that other scholars would be unwilling to accept.

19 P. L. Vaidya, ed., The Madhyamakasastra of Nagarjuna with the Commentary Prasannapada by Candrakirti (Darbhanga: Mithila Institute of Postgraduate Studies, 1960), xv.

20 J. W. de Jong, Review of E. Lamotte, Le Traite de la Grande Vertu de Sagesse de Nagarjuna, III, in Asia Major 17 (1971): 109; Buston, History of Buddhism, tr. E. Obermiller (Delhi: Satguru Publications, 1986), I: 50-51, II: 125; cf. T R. V. Murti, The Central Philosophy of Buddhism, 88, 91; D. Ruegg, The Literature of the Madhyamaka School of Philosophy in India, 8-29.

21 Bodhicaryavatara 5.106.

22 Lindtner, Nagarjuniana.

23 See T. Vetter, Asiatische Studien 46.1 (1992): 393.

24 Nagarjuna's Letter to King Gautamiputra, with Explanatory Notes based on Tibetan Commentaries, tr. the Ven. Lozang Jamspal et al. (Delhi: Motilal Banarsidass, 1978); S. Dietz, ("The Author of the Suhrllekha" in Contributions on Tibetan and Buddhist Religion and Philosophy, ed. E. Steinkellner and H. Tauscher [Wien: Arbeitskreis fur tibetische und buddhistische Studien, 1983], 59-72) considers the metre of the work, the nature of references to Mahayana, and the doctrinal content, and casts doubt on its connection with Nagarjuna.

25 C. Lindtner, Nagarjuniana, 163-69; T. Vetter, op. cit.

26 J. Filliozat, Yogasataka: Texte medical attribude a Nagarjuna (Pondicherry, 1979).

27 K. Zysk, Asceticism and Healing in Ancient India: Medicine in the Buddhist Monastery (N.Y.: O.U.P., 1991), 64, n. 3.

28 E. Lamotte, ed., Traite de la Grande Vertu de Sagesse de Nagarjuna (Mahaprajnaparamitasastra), vol. III (chs. XXXI-XLII) (Louvain: Institut Orientaliste de Louvain, 1970); idem, L'Enseignement de Vimalakirti (Louvain: Publications Universitaires, 1962), 74-77; and "Der Verfasser des Upadesa und seine Quellen," Nachrichten der Akademie der Wissenschaften in Gottingen, phil.-hist. Kl., 2 (1973): 3-5; J. W. de Jong, review of Le Traite de la Grande Vertu de Sagesse, 105-12.

29 M. Walleser, "The Life of Nagarjuna from Tibetan and Chinese Sources," 437; D. T. Suzuki, ed., The Lankavatarasutra (London: Routledge & Kegan Paul, 1932), 226-95.

30 For the text, see V. Karambelkar, "The Problem of Nagarjuna," 23.

31 See also D. T. Suzuki, Lankavatarasutra, 239f.

32 On Vedali and Vidarbha, see P.S. Sastri, "Nagarjuna and Aryadeva?" 193-202; N. Dutt, "Notes on the Nagarjunikonda

Inscriptions, I: Nagajunikonda and Nagarjunam," Indian Historical Quarterly 7 (1931): 635, n. 6.

33 N. Dutt, "Nagarjunikonda and Nagarjuna," 635, n. 6.

34 P.S. Sastri, "Nagarjuna and Aryadeva," 193-96.

35 Taranatha, History of Buddhism in India, tr. Lama Chimpa and Alaka Chattopadhyaya (Simla: Indian Institute of Advanced Study, 1970), 123, 126; D. Ruegg, "Le Dharmadhatustava de Nagarjuna," Etudes tibetaines dediees a la memoire de Marcel Lalou (Paris: Librairie d'Amerique et d'Orient, 1976), 449, n. 8.

36 M. Walleser, "The Life of Nagarjuna from Tibetan and Chinese sources," 439f.

37 J. Burgess, Notes on the Amaravati Stupa (Madras: Archaeological Survey of Southern India, 1882), 112 (cf. also, p. 57); T. N. Ramachandran, Nagarjunakonda, Memoirs of the Archaeological Survey of India, no. 71. (Delhi: Archaeological Survey of India, 1953),

28f; G. Tucci, Minor Buddhist Texts (Kyoto: Rinsen, 1978), 284.

38 See J. Burgess, The Buddhist Stupas of Amaravati and Jagayyapeta in the Krishna District, Madras Presidency (reprint, Varanasi: Indological Book House, 1970).

39 Bana, Harsacarita, tr. E. B. Cowell and F. W. Thomas (London: Royal Asiatic Society, 1897), 251f.

40 Bana, The Harsacarita of Banabhatta, with Commentary of Sankara, ed. Kasinath Pandurang Parab and Sastri Dhondo Paraguram Vaze (Bombay: Tukaram Javaji, 1892), 282.

41 D. Ruegg, "Le Dharmadhatustava de Nagarjuna," 450; G. Tucci, "Animadversiones Indicae" Journal of the Asiatic Society of Bengal, n.s., 26 (1930): 144-47.

42 P. Demieville, "Sur un passage du Mahameghasutra," appendix 2 of "Les versions chinoises du Milindapanha," Bulletin de l'Ecole francaise d'Extreme-Orient 24 (1924): 218-30.

43 P. Demieville, Mahameghasutra, 227. "Supaya" in the text.

44 Demieville, Mahameghasutra, 229; cf. K. Satchidananda Murty, Nagarjuna (New Delhi: National Book Trust, 1971-79), 43-44; S. Levi, "Kaniska et Satavahana," Journal asiatique (1936): 117; S. K. Pathak, "Life of Nagarjuna (from the Pag-Sam-Jon-Zang)," Indian Historical Quarterly 30 (1954): 93-95.

45 Demiville, Mahameghasutra, 118; Tucci, "Animadversiones" 146-47.

46 Levi, "Kaniska" 118-19.

47 Kathasaritsagara 41.9-10, ed. Pandits Durgaprasad and Kasinath Pandurang Parab, 3rd ed., rev. by W. L. S. Pansikar (Bombay: Tukaram Javaji, 1915), 188.

48 "Thus Nagarjuna went to his fate, prevented by the gods from destroying death": evam nagarjunarabdham martyanam mrtyunasanam / na sodhum devatair yavat so 'pi mrtyuvasam gatah, ibid., 189 (vs. 59).

49 Karambelkar, "Problem," 27; cf. J. J. Speyer, Studies in the Kathasaritsagara (Wiesbaden: Sandig, 1968), 163.

50 Rajatarangini, ed. Pandeya Ramtej Shastri (Benares: Pandit Pustakalaya, 1960), 1: 168-73.

51 Reviewed by S. Levi, "Kaniska," 119.

52 K. Venkata Ramanan, Nagarjuna's Philosophy, 28.

53 See A. L. Basham, ed., Papers on the Date of Kaniska (Leiden: Brill, 1968); D. Shackleton Bailey, Satapancasatka of Matrceta (Cambridge: C. U. P., 1951); and see other sources cited by D. Seyfort Ruegg, The Literature of the Madhyamaka School, 5, n. 11.

54 Mahabherisutra, cited by Buston, History of Buddhism, tr. E. Obermiller (Heidelberg, 1931-32), 129f.; T. Watters, On Yuan Chwang's Travels in India, II (Delhi: Munshiram Manoharlal, 1961),

204, n. 3; cf. Joshi, "Life and Times," 46; Murty, Nagarjuna, 45 (these two texts, not mentioned in Santideva's Siksasamuccaya, are possibly spurious); Ruegg, "Le Dharmadhatustava," 450.

55 Aryamanjusrimulakalpa, part III, ed. T. Ganapati Sastri, Trivandrum Sanskrit Series, 84 (Trivandrum: Government Press, 1925), 616f.

56 Walleser, "Life of Nagarjuna," 439f.; however, Walleser emphasizes (p. 440) that the text is a late one and unreliable.

57 Al-Biruni, Alberuni's India: An Account, tr. E. Sachau (Delhi: Chand, 1964), I: 188f.

58 Karambelkar, "Problem," 31-32; P. Kumar, tr., Nagarjuna's Yogaratnamala (Delhi: Nag Publishers, 1980), 18.

59 T 2047, 185b 2-3, 186b, cited by E. Lamotte, L'Enseignement de Vimalakirti, 76; cf. Walleser, "Life of Nagarjuna," 444.

60 J. W. de Jong, review of Le Traite de la Grande vertu de sagesse de Nagarjuna, 105-6; on the evidence of Hui-yuan see Robinson, Early Madhyamika, 22; cf. Murty, Nagarjuna, 47.

61 De Jong, review of Traite, 106.

62 Ibid., 110f.; cf. Jan Yun-hua, "Nagarjuna, One or More?" 148-49. The whole question of the date of the mahaparinirvana is again wide open: see, for example, H. Bechert, "Die Lebenszeit des Buddha: Das alteste feststehende Datum der indischen Geschichte?" Nachrichten der Akademie der Wissenschaften in Gottingen, phil.-hist. Kl., 4 (1986): 129-84; and Bechert, ed., The Dating of the Historical Buddha, part 1 (Gottingen: Vandenhoeck & Ruprecht), 1991.

63 Lamotte, Traite, III: liii.

64 J. W. de Jong, review of J. Hopkins and Lati Rimpoche, trs., The Precious Garland and the Song of the Four Mindfulnesses (London, 1975), in Indo-Iranian Journal 20 (1978): 137.

65 It is not clear where the capital of this Kosala would have been. See J. Fergusson, "On Hiouen-Thsang's Journey from Patna to Bhallabhi," Journal of the Royal Asiatic Society, n.s., 6 (1875):

260; Dutt, "Nagarjunikonda and Nagarjuna," 639; S. Beal, Hsuan-tsang, Si-Yu-Ki, Buddhist Records of the Western World, 2 vols. (Boston: Osgood, 1885), II: 209n.

66 S. Beal, Si-Yu-Ki, II: 210. On the Chinese transliteration of the name Nagarjuna, see Watters, On Yuan Chwang's Travels, 203.

67 Beal, Si-Yu-Ki, 210-12; see p. 210 n. 71; cf. T Watters, On Yuan Chwang's Travels, II: 206-7, referring to the rendering of Sadvaha (= Satavahana) as Yin-cheng, 'Leading-right'.

68 Beal, Si-Yu-Ki, 214, n. 80; cf. Karambelkar, "The Problem of Nagarjuna," 23-24; J. Burgess, The Buddhist Stupas of Amaravati,

6-7; Jan Yun-hua, "Nagarjunakonda: Note on a New Reference from Chinese Source," Journal of Indian History 48.2 (1970): 415-26.

69 Fa-hsien, The Travels of Fa-hsien (399-414 A.D.), tr. H. A. Giles (London: Routledge and Kegan Paul, 1923), 62: "There is a country called Deccan, in which there is a monastery dedicated to Kasyapa Buddha, made by hollowing out a great rock." From the bottom up, these storeys were shaped like an elephant and with 500 chambers, like a lion and with 400, like a horse and with 300, like an ox and with 200, like a pigeon and with 100; a spring of water flowed from the top and was channeled around all the chambers on the way down; there were windows on all storeys. The whole area was now waste and uninhabited, not frequented by Buddhists; cf. Watters, On Yuan Chwang's Travels, II: 207-8, Beal, Si-Yu-Ki, 214, n. 80.

70 Jan Yun-hua, "Nagarjuna, One or More?" 144 (citing W. Fuchs, 'Huei-ch'ao's Pilgerreise durch Nordwest-Indien und Zentral Asien', Sitzungsberichten der preussischen Alcademie der Wissenschaften, phil.-hist. Kl., 30 [1938]: 437ff, which was not accessible for this study).

71 Watters, On Yuan Chwang's Travels, II: 208; Beal, Si-Yu-Ki, 2: 214, n. 80.

72 On the discussion of this "Black Peak" monastery, see I. W. Mabbett, "Dhanyakataka," South Asia 16.2 (1993): 33f.

73 Beal, Si-Yu-Ki, 217. Dhanyakataka in ancient times was probably at Amaravati, but it is a puzzle why, in this case, Hsuan-tsang - usually so interested in the description of ancient and impressive Buddhist monuments - should omit any mention of the Amaravati stupa. It is possible that, by the seventh century, the name Dhanyakataka should have attached itself to another place, for the whole region was now a peripheral area, far from centers of power. See I. W. Mabbett, "Dhanya kataka."

74 Ibid.

75 Beal, Si-Yu-Ki, 97f, 302f; T Watters, On Yuan Chwang's Travels, II: 104.

76 S. Beal, tr., The Life of Shaman Hwui Li (Delhi: Academica Asiatica, 1973), 199; cf. M. Winternitz, History of Indian

Literature, II: 342.

77 Shohei Ichimura, "Reexamining The Period of Nagarjuna" 9 (= 1078).

78 Ichimura does not refer to the passage in the Rajatarangini; he discusses links with Satavahana rulers, and suggests that Nagarjuna wrote the Suhrllekha for a Satavahana king and the Ratnavali for a Mahayana-supporting Saka. (P. S. Sastri, on the other hand, has found reasons for supposing that Nagarjuna wrote the Ratnavali for Gautamiputra Satavahana, supposed to be brahmanical: "Nagarjuna and Aryadeva," 201f.).

79 I'-ching, A Record of the Buddhist Religion as Practised in India and the Malay Archipelago (A.D. 671-695), tr. J. Takakusu (Oxford: O.U.P., 1896), 181.

80 I-ching, A Record, 158-62.

81 Restored as "Jantaka" by Beal, but see de Jong, review of The Precious Garland, 136-39 (particularly n. 18). See also Watters, On Yuan Chwang's Travels, II: 207; Levi, "Kaniska et Satavahana" 107 (reading 'Jantaka'); S. Beal, "Some Remarks on the Suhrllekha or Friendly Communication of Nagarjuna Bodhisattva to King Shatopohanna," Indian Antiquary 16 (1887): 169-72; Joshi, "Life and times," 17; M. Walleser, "Die Lebenszeit des Nagarjuna," Zeitschrift fur Buddhismus 6 (1924-25): 95-103, 237-42.

82 K. Gopalachari, Early History of the Andhra Country (Madras: Madras U.P.), 38, 47.

83 Buston, History of Buddhism.

84 G. N. Roerich, The Blue Annals (Delhi: Motilal Banarsidass, 1979).

85 Taranatha, History of Buddhism in India; The Seven Instruction Lineages, tr. D. Templeman (Dharamsala: Library of Tibetan Works and Archives, 1983).

86 S. K. Pathak, "Life of Nagarjuna from the Pag Sam Jon Zang," Indian Historical Quarterly 30 (1954): 93-95; cf. Walleser, "Life of Nagarjuna," 422.

87 See A. Macdonald, "Le Dhanyakataka de Man-Luns guru," Bulletin de l'Ecole francaise d'Extreme-Orient 57 (1970): 169213, who suggests (p. 187) that the description of the stupa is at least in part influenced not by what was there to be seen but by the Kalacakra Tantra's versions of mandalas of Vajradhatu, Dharmadhatuvagisvara, etc.

88 Joshi, "Life and Times," 15-16; Karambelkar, "Problem," 28-29; Dutt, "Nagarjunikonda and Nagarjuna," 636-69.

89 The Yogaratnamala details experiments in chemistry; its translator, Kumar, argues contra Roy that there is evidence of chemical knowledge at the time of the original Nagarjuna: P. Kumar, ed., Nagarjuna's Yogaratnamala, 18-19. Murty, Nagarjuna, 58, thinks that the madhyamika philosopher might have been also the author of such works: be "dabbled (in alchemy), just as the philosopher Berkeley messed about with tar-water and wrote on its virtues."

90 Winternitz, History, II: 343-44, n. 2; cf. Karambelkar, "Problem"; Tucci, "Animadversiones," 143. The centuries of life attributed to him and the succession of careers in different parts of India could, suggests Ruegg, represent the conflation of several different historical individuals: Ruegg, "Dharmadhatustava," 452-53.

91 Pathak, "Life," 93-95; Joshi, "Life and Times," 13; K. R. Subramanian, Buddhist Remains in Andhra (Madras: Diocesan Press, 1932), 57, n. 2.

92 Joshi, "Life and Times," 19.

93 Tucci, "Animadversiones" 143f. The date of Dharmakirti is itself not certain.

94 Jan Yun-hua, "Nagarjuna, One or More?"

95 For example, (1) the failure of Hsuan-tsang to mention a second, tantric, Nagarjuna flourishing in India at the time of his own visit (as postulated for a Nagarjuna II by N. Dutt); (2) the existence of medical treatises in China attributed to Bodhisattva Nagarjuna (Lung-shu p'u-sa) referred to in the seventh century by the Sui shu (too early for the alleged medical Nagarjuna in the eighth century or any later); and (3) the eighth-century master Amoghavajra, who claimed that his instruction lineage passed through Vajrasattva, Bodhisattva Nagarjuna, Acarya Nagabodhi, and Acarya Vajrabodhi, who was Amoghavajra's own teacher (which would imply, since all these were supposed to live for centuries, that the tantric Nagarjuna existed in the early centuries A.D. and thus might not be different from Nagarjuna I).

96 Tucci, "Animadversiones," 140.

97 Tucci, "Animadversiones" ("A Sanskrit Biography of the Siddhas and some Questions Connected with Nagarjuna"), 138-55; text at pp. 148-55.

98 A. Waley, tr., The Way and its Power (New York: The Grove Press, 1958), 102-3.

99 De Jong, review of Precious Garland, 137.

100 The last quarter of the first century is another possibility; see K. Venkata Ramanan, Nagarjuna's Philosophy, 27.

101 Epigraphia Indica 1905-6: 61; cf. H. N. Sastri, The Philosophy of Nagarjuna as Contained in the Ratnavali (Calcutta: Saraswat Library, 1977), 11-12.

102 P.S. Sastri, "Nagarjuna and Aryadeva," 193-202.

103 Murty, Nagarjuna, 67.

104 For Winternitz (History, II: 347), influenced by Taranatha's

claim that Nagarjuna was born in the time of Kaniska, Yajna Sri (whose reign Winternitz dates to 166-96) seems right. J. Burgess favors Sri Yajna partly on the strength of his reading of I-ching's Shih-yin-te-chia, though adding that Pulumavi III (dated by him to 215-20 A.D.) is just as likely: Burgess, The Buddhist Stupas, 8; cf. Kumar, Yogaratnamala, 23-24.

105 M. Walleser, "Die Lebenszeit des Nagarjuna," 100-103, is led by his (in fact implausible) restoration of Jayandhra from I'-ching's Shih-yin-te-chia to favor Vijaya.

106 H. Sarkar and B. N. Misra, Nagarjunakonda (Calcutta: Krishna Murthy, 1966), 74.

107 H. Sarkar, "The Nagarjunakonda Phase of the Lower Krishna Valley Art: A Study based on Epigraphical Data," in Indian Epigraphy: Its Bearing on the History of Art, ed. F. Asher and G. S. Gai (New Delhi, 1985), 30.

108 Joshi, "Life and Times," 16-17; Bailey, Satapancasatka, 7; Walleser, "Life of Nagarjuna," 430; Levi, "Kaniska et Satavahana," 106.

109 J. Vogel, "Prakrit Inscriptions from a Buddhist Site at Nagarjunikonda," Epigraphia Indica 20 (1929): 22.

110 Murty, Nagarjuna, 62.

111 I. K. Sarma, Sri Parvata (unpublished MS, 1989).

112 Ruegg, "Le Dharmadhatustava," 448-71.

113 A. Wayman and H. Wayman, trs., The Lion's Roar of Queen Srimala: A Buddhist Scripture on the Tathagatagarbha Theory (New York: Columbia U.P., 1974); see the introduction.

114 H. Sarkar, Studies in Early Buddhist Architecture of India (Delhi: Archaeological Survey of India, 1966), 78.

115 L. de La Vallee Poussin, "Notes de bibliographie bouddhique," Melanges chinois et bouddhiques 1 (1931): 402.

116 Some scholars have been inclined to accept that Nagarjuna may have lived under the Iksvakus, whether or not they thought he began his career under an earlier dynasty. Longhurst half-whimsically questioned whether the broken-off head of a Nagarjunakonda statue, portraying a venerable capped figure, might not represent the master: A. H. Longhurst, "Excavations at Nagarjunakonda," Archaeological Survey of India Annual Report, 1927-28, 113-21; K. S. Murty, Nagarjuna, 69, refers to the discovery at Nagarjunakonda in 1938 of a purmakumbha containing only two small teeth, which some considered to be relics of the great Acarya. However intriguing, these are both guesses.