Classic and Contemporary Buddhist Works

The Buddhist Philosophy of the Middle - Selections

Essays on Indian and Tibetan Madhyamaka

Chapter 2: Towards a Chronology of the Madhyamaka school*

In investigating the literary, religious, and philosophical history of a school of thought it is necessary to attempt to establish its chronology. Three chronological frames basically come into question for this purpose: (1) the internal chronology of the masters and texts of the school relative to each other, (2) the external chronology of these masters and texts relative to those of other schools, allied and opposed, and (3) their absolute chronology. The determination of an absolute chronology will normally involve both the fixing of the relative chronology, in either or both of the above-mentioned forms, of authors and texts and the establishment of correlations with events known otherwise that can serve as independent historical landmarks with which these authors and texts can be brought into either a synchronic or a diachronic relation (terminus ante quem or post quem). In the case of the Madhyamaka literature, as with so much of Indian literature, it is, however, only occasionally that we are in the fortunate position of being able to bring external, historically verifiable data actually to bear on an author or text with a view to establishing a date with any high degree of precision and certainty. Frequently the historian has to satisfy himself with approximations, and datings to within the nearest century or two are often the best that can be achieved with the available documentation.

For Nāgārjuna, the source of the Madhyamaka school, we have certain traditionally accepted synchronisms. Thus his contemporaneousness, and even the existence of a personal relationship, with a Sātavāhana monarch are reported by authorities belonging to the seventh century. These are the Harṣacarita of Bāṇa (chapter VIII), the Hsi-yü-chi of Hsüan-tsang, and the Nan-hai chikuei nei-fa ch’uan of I-ching. The possibility of such a royal link is indeed suggested by two works ascribed to Nāgārjuna: the Suhrllekha, a paraenetic epistle addressed to a king, and the Ratnāvalī, an ethico-philosophical treatise also addressed to a king (rājaparikathā).1 A later commentary on the Suhrllekha by Mahāmati and one on the Ratnāvalī by Ajitamitra, both of which are extant in Tibetan versions, mention Bde spyod as the name of the king with whom Nāgārjuna was in contact.2

On the basis of this information modern scholars have variously identified the Sātavāhana monarch in question as Hāla, Gautamīputra Śātakarṇi, Vāsiṣṭhīputra Pulumāyi II, and Yajñaśrī Śātakarṇi. The fact that several different kings whose reigns cover a period of about a century have thus been suggested shows, however, how unhelpful such a synchronism is by itself for the purpose of obtaining a precise dating.

A further synchronism between Nāgārjuna and (a) Kaniṣka has been mentioned by Kalhaṇa, the twelfth-century historian of Kaśmīr, in his Rājataraṅgiṇī. There (i. 170) the Kaniṣka in question is described as a Turuṣka king, alongside Huṣka and Juṣka;3 and the context refers to the Ṣaḍarhadvana (Harvan). The difficulties here are not easier to solve than in the case of the synchronism with the Sātavāhana; for not only do we not know precisely which Kaniṣka should be considered, but modern scholars remain divided concerning the date of Kaniṣka I and the beginning of the Kuṣāṇa dynasty, which they have placed between the year 78 and the third century of the Christian era.

In one passage of his Hsi-yü-chi, Hsüan-tsang seems furthermore to have made Nāgārjuna a contemporary of aśvaghoṣa and Kumāralāta.4 But in other Chinese records of the Indian masters of Buddhism, aśvaghoṣa, the so-called twelfth patriarch who is frequently linked with Kaniṣka I, is placed before Nāgārjuna, who is counted as the fourteenth patriarch.5

The dates of Kumārajīva (344–413 or 350–409?), the founder of the Chinese Madhyamaka (San-lun) school who was of Serindian origin but studied as a boy in Kaśmīr (Chi-pin), also hardly allow any precise chronological determination; they can provide us only with a terminus ante quem for Nāgārjuna and his followers Āryadeva and Rāhulabhadra.6 It has furthermore to be noted that the commentary on the Daśabhūmikasūtra attributed to Nāgārjuna, and now extant only in Kumārajīva’s Chinese translation (T. 1521), is reported to have been translated earlier, in the second half of the third century, by Dharmarakṣa, who also translated the sūtra itself;7 this translator arrived in Ch’ang-an in 265 and worked there until about 313. But the ascription of this commentary to Nāgārjuna has been the subject of debate; and the relevance of this report, even if accurate, to the date of Nāgārjuna is therefore uncertain.8

The name Bhadanta nāgārjunācārya is found mentioned in an inscription discovered a century ago near the Jaggayyapeṭa Stūpa in South India.9 But because of the relative lateness of its script this reference unfortunately cannot help much in establishing the date of Nāgārjuna, the author of the Mūlamadhyamakakārikās etc.10

In sum, it is likely that Nāgārjuna flourished at some time when the Sātavāhana and Kuṣāṇa dynasties were holding sway in India. But of just which king(s) of either of these dynasties he was the contemporary, and whether he had personal links with any of them, it is hardly possible to determine either by means of criteria internal to the works ascribed to him or on the basis of the Indian and Chinese reports mentioned above. On balance it seems appropriate to place Nāgārjuna ca. 150–200 c.e. (with E. Frauwallner and a number of other scholars). The dating proposed by É. Lamotte and J. May, who place his birth precisely in 243 c.e.,11 even if not intrinsically impossible, is based on an inference drawn from the complex, and rather hypothetical, chronological calculations originating in Kumārajīva’s school.

Recently it has also been suggested by É. Lamotte that a later Indian author—perhaps a second Nāgārjuna originally trained in the Sarvāstivādin abhidharma and living in northwestern India somewhat later than the author of the Mūlamadhyamakakārikās—was the author of the Ta-chih-tu-lun (T 1509). This encyclopaedic commentary on the large Prajñāpāramitāsūtra traditionally ascribed to Nāgārjuna was translated into Chinese by Kumārajīva at the beginning of the fifth century. Lamotte’s suggestion is based in part on the fact that both Āryadeva’s Catuḥśataka and Rāhulabhadra’s Prajñāpāramitāstotra are cited in the Ta-chih-tu-lun. Lamotte observes, moreover, that such an extensive and encyclopaedic work referring to some thirty Mahāyāna sūtras cannot be ascribed to Nāgārjuna I and hence placed at the very beginning of “Madhyamaka Buddhism.”12 Now this enormous treatise appears to be quite unknown to the Indo-Tibetan traditions; and it seems in any case to be generally agreed among specialists on the subject that the extant text of the Ta-chih-tu-lun is in large part a composite work which is not even of exclusively Indian origin.13 Could it then be that its text as we now have it was the product of Serindian scholarship,14 and that it was subjected to a final process of Sinification at the time of its translation by Kumārajīva and his Chinese colleagues?15 The work could then have been ascribed to Nāgārjuna because its basic ideas—and perhaps even its core—go back to this Indian source of the Madhyamaka school without his having been the author, in the strict sense, of this work. Thus, although Lamotte’s hypothesis concerning the authorship of the Ta-chih-tu-lun is certainly not in itself impossible, it seems at least as simple—and certainly more economical than postulating an early Deutero-Nāgārjuna living quite soon after Nāgārjuna I—to account for references to Mādhyamikas later than Nāgārjuna I, as well as for other apparently later elements in this work, by attributing them to the activity of Serindian scholars and Chinese redactors.

The date of Āryadeva is to be determined in relation to that of Nāgārjuna I, whose direct disciple he is universally considered to have been. The relationship between Rāhulabhadra and Nāgārjuna will be discussed below (§ II).

Bhā(va)viveka (Bhavya), the founder of what later came to be known as the Svātantrika-Madhyamaka school, wrote a major commentary on Nāgārjuna’s Mūlamadhyamakakārikās, the Prajñāpradīpa, which was translated into Chinese in the fourth decade of the seventh century by Prabhākaramitra (T 1566). Later in the same century his “Jewel in hand treatise” was translated into Chinese by Hsüan-tsang (Ta-ch’eng chang-chen lun, T 1578). Bhā(va)viveka has been shown to be the elder contemporary of both Sthiramati (ca. 510–70) and Dharmapāla (ca. 530–61).16

Buddhapālita, the master of what later came to be known as the Prāsaṅgika branch of the Madhyamaka and the target of Bhā(va)viveka’s criticism, as well as Candrakīrti, the main authority of this Prāsaṅgika branch and Bhāvaviveka’s great opponent, are both unknown to the classical tradition of the Chinese San-lun school. and external historical evidence of the abovementioned kind making it possible to determine precisely their absolute chronology seems to be lacking.

The historical records of Tibet, on the contrary, permit reasonably accurate datings of several Indian Mādhyamikas living during the early propagation (snga dar) of Buddhism in that country. Thus they indicate that Śāntarakṣita, the great apostle of Tibet during the reign of Khri Srong lde btsan, arrived there for the first time ca. 763, and that he resided there a second time from about 775 to his death in about 788. He was closely associated with the foundation of Bsam yas (ca. 775 or 779?) and was in fact the first abbot (upādhyāya) of this great Tibetan monastic centre. On the death of Śāntarakṣita his disciple and commentator Kamalaśīla was invited to Tibet to continue and consolidate his master’s work there. He was in Tibet during at least part of the “Council of Tibet” (or “Council of Bsam yas”) in about 792–94, and he died there in about 795.17

The Tibetan records also make it possible to date with a fair degree of accuracy several important later Indian Mādhyamikas living during the earliest period of the later propagation (phyi dar) of the Dharma in Tibet. Dīpaṃkaraśrījñāna (ca. 982–1054) was the contemporary of Rin chen bzang po (958–1055), the senior monk-scholar of the kingdom of Western Tibet and the teacher of ’Brom ston pa (1005–64), the master from Central Tibet who founded the Bka’ gdams pa school. Jayānanda, a Kāśmīrī Mādhyamika and the commentator on Candrakīrti’s Madhyamakāvatāra, was closely associated with the Tibetan translator Pa tshab nyi ma grags (born in 1055) in the establishment on a solid basis in Tibet of the Prāsaṅgika branch of the Madhyamaka. And Abhayākaragupta, a later authority who developed a synthesis of the Madhyamaka with the Yogācāra, Prajñāpāramitā, and Vajrayāna, composed his Munimatālaṃkāra in the thirtieth year of the reign of the Pāla king Rāmapāla (r. ca. 1077–1130 or 1072–1126) according to the colophon of its Tibetan translation contained in the Bstan ’gyur.

Having surveyed some external historical evidence for the absolute chronology of the Mādhyamikas, let us now turn to evidence to be derived from internal analysis of their works for the relative chronology of their school.

a serious difficulty in the chronology of the texts and authors of the Madhyamaka arises, to begin with, from the fact that a large number of works of varied, and quite distinct, contents are traditionally ascribed to Nāgārjuna. Some of them are independent scholastic treatises (śāstra) such as the Mūlamadhyamakakārikās, the Vigrahavyāvartanī and the Yuktiṣaṣṭikā. Others are commentaries on sūtras, for example the Vibhāṣā on the Daśabhūmikasūtra (T 1521) and the above-mentioned commentary on the large Prajñāpāramitāsūtra (the Ta-chih-tu-lun, T 1509). Still others are epistles and homilies, such as the Suhrllekha and Ratnāvalī, and hymns (stava). A classification of Nāgārjuna’s works traditionally adopted in Tibet has divided them into a scholastic corpus (rigs tshogs), a hymnic corpus (bstod tshogs), and a homiletic corpus (gtam tshogs). But despite the differences it recognized between them not only as to literary genre but also, to a considerable degree, in contents, this tradition has nevertheless ascribed all the works included in these three categories to one and the same author.18

In such circumstances, one means of establishing authorship in the case of works of uncertain provenance, and hence also a relative chronology of texts and their authors, would be to attempt to apply to them terminological and philosophic criteria of the kind employed by Paul Hacker and his successors for determining the authenticity of works ascribed to Śaṃkara. In the case of Nāgārjuna, however, there are special difficulties. In the first place, most of the works in question are now available to us only in Chinese and/or Tibetan translations, so that the use of the usual literary, stylistic, and terminological criteria for establishing authorship becomes either problematic or simply impossible. no less important is the consideration that even when we do possess indubitable evidence of differences in style, terminology, and ideas, it is by no means certain that works that so differ have necessarily to be by different authors. It can in fact be just as legitimately supposed that they belong to different periods in the development of the thinking of a single author, or even that they represent one author’s complementary (rather than opposed) approaches to certain difficult philosophical problems. an example could be the treatment of reality in an apophatic manner (e.g., in the major portion of the Mūlamadhyamakakārikās and the related treatises ascribed to Nāgārjuna), in a cataphatic manner (e.g., in several of the hymns ascribed to him), and by the philosopher’s silence (āryatūṣṇībhāva, e.g., in a few passages of the Mūlamadhyamakakārikās and of the hymns). When seeking to establish authorship in such cases, then, it would clearly be begging the question were one to argue that, because of differences of style or terminology, such texts must necessarily be by different authors. Indeed, the question is precisely to define the full range of an author’s religious and philosophical thought; and this can surely be done only by taking into account all his works, without any text ascribed to him being excluded a priori on purely stylistic or terminological grounds. We have, in other words, to determine what styles an author has employed and what religious and philosophical ideas he has accepted on the basis of the entire corpus of his works; and to proceed the other way round can all too easily involve prejudgements and circular reasoning. In sum, we must be prepared to recognize the full range of historical, philological, and philosophical problems that arise in discussing the authenticity of the works ascribed to Nāgārjuna.

Even when our sources clearly refer to authors as famous in the history of the Madhyamaka as Nāgārjuna, Bhā(va)viveka (Bhavya), and Candrakīrti, further difficulties can arise owing to the fact that these names have been borne by two (or sometimes even more) different masters of the school who lived at quite different times. as a consequence, otherwise valuable sources have on occasion attributed improbably long lifespans to a single master, placing the same name/person at what are evidently quite different points in the history of the school. The resulting uncertainty as to whether such names are to be understood as designating one or more persons constitutes an additional major difficulty standing in the way of establishing a secure chronology of the Madhyamaka.

This problem is encountered in an acute form in the case of the relation between Nāgārjuna and Rāhulabhadra, whose names appear linked together at what clearly appear to be widely separated periods in the history of their school. and the difficulty is compounded by the fact that Rāhulabhadra is sometimes described in our sources as the successor and disciple of Nāgārjuna, whereas at other times he is stated to be his predecessor and master. Modern scholars have arrived at diametrically opposed solutions to this difficulty, which thus presents the historian with an intriguing and instructive problem.19

Scrutiny of the evidence seems to indicate that Nāgārjuna I was the source and originator of the Madhyamaka as a distinct school of thought and that (a) Rāhulabhadra was his successor. This Rāhulabhadra may have been the direct disciple of both Nāgārjuna and Āryadeva or of Āryadeva alone: the evidence is not clear on this point. (as already mentioned, both these masters are cited in the Ta-chih-tu-lun ascribed to Nāgārjuna and translated by Kumārajīva at the beginning of the fifth century.)

The statements that make Rāhulabhadra the master rather than the follower of Nāgārjuna can, however, be satisfactorily explained as referring to the Siddha Rāhulabhadra (= Sarahapāda), the master of the Deutero-Nāgārjuna who we have very good reason to believe lived no later than the seventh century.20 This Nāgārjunapāda, also one of the Siddhas, played a leading role in the development of the synthesis of the Madhyamaka and Vajrayāna which came into prominence at that time; and he is presumably to be identified with the source of the Ārya tradition of the Guhyasamājatantra. Supposing this Siddha Nāgārjuna to have been the disciple of Rāhulabhadra = Sarahapāda while Nāgārjuna I, the author of the Mūlamadhyamakakārikās etc., was on the contrary the predecessor of the Rāhulabhadra known to the Ta-chihtu-lun—as well as to asaṅga’s commentary on the Madhyamakaśāstra (T 1565) and *Sāramati’s *Mahāyānāvatāraśāstra (T 1634)21—we are in a position to resolve an apparent conflict in the evidence concerning the relationship between these two names and the persons they designate.

This solution does not of course dispose of all our chronological or other historical problems. and the question of the precise identity of Nāgārjuna in relation to “nāga” (Nāgāhvaya) and perhaps also nāgabodhi continues to raise questions. Sometimes we find mention of a “nāga” who appears to be identical with Nāgārjuna; but at other times “nāga” figures as a distinct person. His precise identity thus remains uncertain, but since no works are known to be by him he ceases to be a subject of major concern to the historian of Madhyamaka literature.22 as for nāgabodhi, being reported to be the teacher of Vajrabodhi (ca. 671–741?) and amoghavajra (ca. 705–74), the Indian masters who helped introduce the Vajrayāna into China and whose dates can be determined from Chinese sources,23 he has to be placed much later. (Whether nāgabodhi and the Vajrayānist Nāgārjunapāda mentioned above were identical it is difficult to say; the dates could support such an identification, but their identity is not established.)24

So that the long period of time separating Nāgārjuna, the author of basic treatises of the Madhyamaka school, and the Siddha Nāgārjuna(pāda), who lived approximately half a millennium later, might be bridged, those sources which do not distinguish between these two persons have ascribed to their single Nāgārjuna an unusually long lifespan (of six hundred years).

In the main line of Madhyamaka masters, Nāgārjuna I and his disciple Āryadeva were followed by Buddhapālita. as already noted, in the present state of our knowledge it is not possible to establish precisely the date of this master of what was later known as the Prāsaṅgika branch of the Madhyamaka; but the evidence points reasonably clearly to the period about 500 as the time when Buddhapālita flourished.25

The next major figure in the history of the Madhyamaka, Bhā(va)viveka or Bhavya, has already been mentioned above in connexion with the Chinese translations of two of his works. He found fault with Buddhapālita’s philosophical procedure founded on the exclusive use of prasaṅga-type reasoning to reveal the inapplicability of any thesis presupposing the existence of some kind of entity, and he sought to formulate independent inferences (svatantrānumāna) and syllogisms (prayoga) for this same purpose. Bhā(va)viveka was in his turn severely criticized by Candrakīrti, who came to Buddhapālita’s defence. as already noted also, Bhā(va)viveka was the elder contemporary of Sthiramati, the master of the Valabhī school of the Vijñānavāda who criticised him and wrote a commentary on Nāgārjuna’s Mūlamadhyamakakārikās (T 1567), and of Dharmapāla, the master of the nālandā school of the Vijñānavāda who also criticised him in his commentary on Āryadeva’s Catuḥśataka (T 1571). On the basis of these interrelationships Bhā(va)viveka can be placed ca. 500–570, a dating compatible with what has been noted above concerning the external historical evidence provided by the Chinese translations of two of his works, the first of which is dated to 630–32.26

For lack of external historical evidence, Candrakīrti’s date has to be fixed relatively to that of his predecessor Bhā(va)viveka, whom he criticised by name in his Prasannapadā Madhyamakavrttiḥ and whose own dates thus provide a terminus a quo, and to later developments in the Madhyamaka school.

Since Candrakīrti’s works were not translated into Chinese, the external evidence from that source, which has so often proved invaluable in other cases, is unavailable to us; an argument from silence is provided by the fact that Bhā(va)viveka, but not Candrakīrti, has been mentioned by I-ching, who left India about 685. Candrakīrti has then presumably to be placed some time in the seventh century.27 This is about the same time as Dharmakīrti (ca. 600– 660), to whom Candrakīrti does not, however, seem to refer when criticizing some doctrines of the Buddhist logico-epistemological school founded by Dignāga (cf. Prasannapadā i, pp. 59f ).

Here a problem arises owing to the fact that the Madhyamakaratnapradīpa, a treatise ascribed to Bhavya, refers to Candrakīrti. Because of this reference Candrakīrti’s date has even been put back by a recent writer to ca. 500–560, the time of Bhā(va)viveka/Bhavya.28 But this earlier dating appears in fact to be quite unnecessary on the grounds adduced by its proponent, and intrinsically it is quite improbable. In the first place, the Madhyamakaratnapradīpa refers also to Dharmakīrti, whose date would for the same reason have to be pushed back to the same time when he would have been an only slightly younger contemporary of Dignāga (ca. 480– 540); but such a dating for Dharmakīrti is most improbable. It is true that the Madhyamakaratnapradīpa has itself stated that its author also wrote the Tarkajvālā, the commentary on the Madhyamakahrdayakārikās ascribed, like the basic text, to Bhā(va)viveka. We thus seem to have to choose between regarding this statement as incorrect and ascribing the Tarkajvālā to a second, later Bhā(va)viveka/Bhavya. However this may be, in view of the reference to Dharmakīrti in the Madhyamakaratnapradīpa, it seems impossible to ascribe this text as a whole to Bhā(va)viveka I, the author of the Prajñāpradīpa, who has to be placed in the sixth century since he was a contemporary of Sthiramati and Dharmapāla. and there is every reason to suppose that the Madhyamakaratnapradīpa was composed by a second Bhavya who lived later than the middle of the seventh century.29 There is therefore no need to make the author of the Madhyamakāvatāra and Prasannapadā a contemporary of Bhā(va)viveka I.

In the case of Candrakīrti, just as in that of Nāgārjuna and Bhavya, it appears that we have, moreover, to reckon with more than one important master bearing this illustrious name. Thus the author of the Pradīpoddyotana, a commentary on the Guhyasamājatantra, has presumably to be distinguished from the author of the Madhyamakāvatāra and Prasannapadā. and still another Candrakīrti lived in the eleventh century, when he translated his *Madhyamakaprajñāvatāra (or *Madhyamakāvatāra-prajñā?) into Tibetan together with the lo tsā ba ’Gos khug pa Lha(s) btsas.

The relative position of Śāntideva in the history of the Madhyamaka can be determined from the fact that he is quoted by Śāntarakṣita; but he is not mentioned by I-ching. His contribution to the Madhyamaka can be regarded as a continuation of Candrakīrti’s, and he is usually counted as an exponent of the latter’s Prāsaṅgika school. It is thus likely that Śāntideva flourished about 700.30 His Bodhicaryāvatāra was already translated into Tibetan by Dpal brtsegs and his Śikṣāsamuccaya by Ye shes sde (ca. 800).

Śāntarakṣita has been credited in some sources with almost as long a lifespan as Nāgārjuna. But in the case of Śāntarakṣita the reason lies in the fact that he has been held by some doxographers to be not only the main representative but also the founder (shing rta’i srol ’byed “way-opener” or “pathfinder”) of the Yogācāra-(Svātantrika-)Madhyamaka school, while Ārya Vimuktisena was reckoned to be a member of this same school. now Ārya Vimuktisena, a commentator on the Pañcaviṃśatisāhasrikā Prajñāpāramitā following the system of the Abhisamayālaṃkāra, is stated to have been a younger contemporary of Buddhapālita (fl. ca. 500) and a disciple of Vasubandhu (ca. 400–480) or of Dignāga (480–540) and Bhavya (ca. 500–70). But, as the putative founder of the school to which Ārya Vimuktisena belonged, Śāntarakṣita had for historical-doxographical reasons to be regarded as the latter’s predecessor, and his birth had therefore to be put some time before that of his supposed disciple; he thus came to be credited with a very long life beginning in the fifth/sixth century and extending almost to the end of the eighth century.31 accordingly, whereas Nāgārjuna was assigned an extraordinarily long lifespan because (at least) two distinct persons of special importance in the Madhyamaka school bore his name, the reason for crediting Śāntarakṣita with almost as long a life was essentially doxographical in origin. It should however be noted that this relative chronology (mentioned, e.g., by Mkhas grub Dge legs dpal bzang in his Rgyud sde spyi rnam, fol. 26a) has not been accepted by the majority of the Tibetan historians and doxographers.

The dates of Śāntarakṣita’s grand-teacher Śrīgupta, who is counted sometimes as a Svātantrika-Mādhyamika and sometimes as a YogācāraMādhyamika, and of Jñānagarbha can also be determined relatively to his own. Śāntarakṣita’s teacher Jñānagarbha, a disciple of Śrīgupta, apparently has to be distinguished from the Jñānagarbha who lived a little later, about 800, and translated into Tibetan several Madhyamaka texts (including the commentary on Jñānagarbha’s Satyadvayavibhaṅga ascribed to Śāntarakṣita) with the Tibetan translators Dpal brtsegs, Klu’i rgyal mtshan, and Ye shes sde.

In several other cases of importance for the history of the Madhyamaka also the same name has been borne by more than one master. Thus the name Jitāri/Jetāri designates a master living ca. 800 and another living at the end of the tenth century. Dharmamitra, the Mādhyamika author of a commentary on Haribhadra’s Abhisamayālaṃkārakārikā-śāstravrtti (the “Sphuṭārthā”) who lived ca. 800, has to be distinguished from an earlier Dharmamitra who commented on Guṇaprabha’s Vinayasūtra. Buddhajñāna(pāda), a representative of the Yogācāra-Madhyamaka school and a disciple of Haribhadra, has to be distinguished from Buddhaśrījñāna, the author of the Jinamārgāvatāra who lived ca. 1200. Dānaśīla was the name of at least three different masters, the first of whom lived at the end of the eighth century, the second at the end of the tenth century, and the third in the twelfth century. Still other examples could be given of a single name designating two or more masters; in some cases only a minute study of the works attributed to the masters bearing such names will allow us to assign them correctly, whereas in others the date of a translation and the identification of the co-translator(s) of an authortranslator, if accurately known, make possible the correct attribution (examples are the cases of Jñānagarbha and Candrakīrti mentioned above).

The chronological results summed up above prompt an observation concerning the question of the reliability of Tāranātha’s History of the Dharma in India (Rgya gar chos ’byung). Since this work was first published and translated in Europe in the 1860s, attempts have been made to exploit for chronological purposes the information that this work contains, and which Tāranātha collected from both his own extensive reading and from the Indian paṇḍits with whom he was in contact.32 These attempts have, however, been often frustrated by a failure to understand how Tāranātha proceeded in recording his information and by our consequent inability to interpret this information; and chronological results based on Tāranātha’s works have consequently been impugned by scholars who have called into question all of Tāranātha’s statements. Tāranātha’s alleged unreliability has then been contrasted with the reliability of the Chinese pilgrims to India, especially Hsüan-tsang and I-ching. It has, however, to be recalled that the Chinese sources too are sometimes in serious error. Thus I-ching’s dating of Bhartrhari’s death to ca. 65133 brought considerable confusion into the history of Indian philosophy until it was finally corrected as late as the 1950s.34 Moreover, once the chronology of the Mādhyamikas has been examined on the basis of both internal and external evidence, some of the information that Tāranātha has provided falls into place, and it turns out to be quite compatible with results reached on the basis of other evidence. Thus, with regard to Tāranātha’s statements that Rāhulabhadra was a disciple35 of Nāgārjuna and his master,36 we now have reason to think that both statements are true provided that they are understood to refer to two different sets of masters bearing the names Nāgārjuna and Rāhulabhadra. The difficulty arose because Tāranātha has not clearly distinguished between the two sets of names and persons, so that the true historical situation could hardly be reconstructed from his statements on the subject alone. But even though his two statements require interpretation in the light of data from other sources, it nevertheless turns out that, once put in context, they are not contradictory as they at first appeared to be.37

Beside the absolute chronology discussed in § I and the internal relative chronology of the Madhyamaka discussed in § II above, it will be of interest also to consider briefly the external relative chronology of the Madhyamaka school vis-à-vis the great treatises and masters of other schools of Indian thought, non-Buddhist as well as Buddhist.

Nāgārjuna’s Mūlamadhyamakakārikās are in large part a critique of philosophical ideas current among the Buddhist Ābhidhārmikas. and the Ta-chih-tu-lun seems to represent an attempt to construct what might be called a Mahāyānist abhidharma in place of the Sarvāstivādin abhidharma. Nāgārjuna’s Ratnāvalī (I.61) alludes to the Sāṃkhya, Aulūkya (i.e., Vaiśeṣika?),38 and nirgrantha (i.e., Jaina) schools of thought; but unfortunately it is not possible to determine precisely to what period of these schools such a passing reference pertains. In addition to Kapila, the sage of the Sāṃkhya, and ulūka, the Vaidalyaprakaraṇa (§ 8) mentions Maṭhara.39 This treatise enumerates and refutes point by point the sixteen categories (padārtha) of Nyāyasūtra I.i.1; and it also mentions (§ 19) the twelve factors of I.i.9. Of special importance for a comparison with the Nyāyasūtras is Nāgārjuna’s Vigrahavyāvartanī, which, unlike the Vaidalyaprakaraṇa, is extant in Sanskrit. The editors and translators of this work have noted connections between VV 31f. and NS II.i.16–19, between VV 64 and NS II.i.13, and between VV 20 and 69 and NS II.i.8–12. In his study of ancient Indian eristic (vāda), G. Oberhammer has argued that the view expressed in NS V.i.10 is rejected in VV 31 f., which is then answered in NS II.i.19, and that the view expressed in NS V.i.18–20 is mentioned in VV 20 and 69.40 Recently, however, K. Bhattacharya has maintained that VV 31 f. indeed refer to NS II.i.19, and that VV 20 and 69 refer to nS II.i.12 f.41 a close historical proximity between Nāgārjuna’s discussion of the pramāṇas and several Nyāyasūtras seems in any case to be established, and G. Tucci has convincingly spoken of interdependence.42 More precise chronological results are difficult to obtain because of the paucity and ambiguity of the documents.

Similar problems arise concerning the question of the precise relationship of Āryadeva to the nyāya and Vaiśeṣika, discussed by G. Tucci,43 and to the Sāṃkhya, discussed by E. Frauwallner44 and M. Honda.45

Even the relatively abundant materials to be found in the Ta-chih-tu-lun do not easily yield clear results for chronology and the question of the historical relationship between the early Mādhyamikas and other schools. a great deal of this material has now been made available and annotated in E. Lamotte’s monumental translation of this encyclopaedic treatise (see in particular Le traité, III, pp. xiv–xxvi), and some of it has also been discussed in K. Venkata Ramanan’s Nāgārjuna’s Philosophy as Presented in the Mahā-PrajñāpāramitāŚāstra (Tokyo, 1966). On the Buddhist side, the author(s) of this treatise have made much use of the Sarvāstivādin abhidharma as well as of a large number of Mahāyānist sūtras.46 As for non-Buddhist schools, although this work argues against the Sāṃkhya and Vaiśeṣika as well as against the theists, it is hardly possible to identify precisely the targets of its criticism.47

Buddhapālita’s commentary on the Mūlamadhyamakakārikās has still to be studied in detail from the present point of view.

The historical position of Bhā(va)viveka is somewhat clearer. Notably in his Madhyamakahrdayakārikās, and in the accompanying Tarkajvālā, the doctrines of the Sāṃkhya, Vaiśeṣika, and Mīmāṃsā, as well as those of the Śrāvakas and Yogācāras, have been discussed in some detail. Even a reference to the Persian Magi is to be found in chapter IX of the Tarkajvālā. Bhā(va)viveka seems, moreover, to have been the first Mādhyamika to discuss the Vedānta, which he criticises in chapter VIII of his Madhyamakahrdayakārikās.48 But at the same time, in chapter III of the same work, Bhā(va)viveka has virtually assimilated the supreme brahman and the dharmakāya (verses 278–83), which he describes as cessation of discursive development (prapañca) and as inaccessible to those who engage in hypothetical reasoning (tārkika, 280); and he goes on to say that this brahman corresponds to the supreme reality of which the Muni (i.e., the Buddha) spoke (283), and that Sages (ārya) such as avalokiteśa, Maitreya, and the rest “approach” (upās-) it precisely through the mode of non-worship (anupāsanayogena, 284).49

In formulating his independent inferences and syllogisms to establish the Mādhyamika’s doctrines, Bhā(va)viveka was influenced by the Buddhist logico-epistemological school going back to Dignāga.50 He strongly opposed the Vijñānavāda of Guṇamati, the master of the Valabhī school; and he was in his turn criticized by Guṇamati’s disciple Sthiramati, as well as by Dharmapāla, a master of the nālandā school of the Vijñānavāda. Bhā(va)viveka approved the doctrine of Devaśarman, a commentator on the Mūlamadhyamakakārikās whose work is no longer extant.51 The commentary on the Prajñāpradīpa by avalokitavrata also provides much interesting information on the history of Indian philosophy which awaits analysis.

Candrakīrti criticizes the Sāṃkhya52 and Vaiśeṣika and mentions the Vedānta in his Madhyamakāvatāra. He followed and developed Buddhapālita’s Prāsaṅgika interpretation of Nāgārjuna’s philosophy; and he criticized both Bhā(va)viveka’s logical innovations in the Madhyamaka and Dignāga’s school. He does not seem to refer to Dharmakīrti, but it is not certain what chronological significance this silence has.53 In the course of his critique of his opponents’ epistemology, Candrakīrti has also quoted (Prasannapadā i, p.65) the Madhyāntavibhāga (i.9), one of the fundamental treatises of the Vijñānavāda school traditionally ascribed to Maitreyanātha and commented on by Vasubandhu.

In the following centuries explicit or implicit allusions to and discussions of opposed schools, Buddhist and non-Buddhist, become abundant in the literature of the Madhyamaka. In particular, Śāntarakṣita’s Tattvasaṃgraha and Kamalaśīla’s Pañjikā on it provide much information about Indian philosophy and its masters in the seventh and eighth centuries. Kamalaśīla’s Pañjikā (328), for example, contains what appears to be the earliest mention in a Buddhist treatise of the advaitadarśana; this mention in a work composed in the second half of the eighth century is consistent with the dating of Śaṃkarācārya in the early eighth century. The Nibandhana by Bodhibhadra (tenth century) on Āryadeva’s Jñānasārasamuccaya also supplies information of value on Indian philosophy of the late classical period. as for the mention of the upaniṣadvādins in the Pañjikā by Prajñākaramati (ca. 1000?) on Śāntideva’s Bodhicaryāvatāra (ix.6o), it refers back to Śāntarakṣita’s Tattvasaṃgraha (328–29). From the eleventh century we have a mention of the Vedāntin Bhāskara in advayavajra’s Tattvaratnāvalī (p. 19).54

Turning now to some of the great classics of Brāhmanical thought, we find in Kumārila’s Ślokavārttika a chapter entitled Śūnyavāda that follows immediately on another entitled nirālambanavāda. Kumārila was evidently a contemporary of Dharmakīrti, whose doctrine he criticises, and he is dated to ca. 600–660; thus he is in all likelihood also a contemporary of Candrakīrti, but he does not seem to refer to this Madhyamaka master anywhere. Contrary to what one might expect of a text called Śūnyavāda, it is not devoted primarily to a discussion of the Madhyamaka school. In the Śūnyavāda chapter (verse 259), just as in the nirālambanavāda chapter (e.g., verse 3), the term śūnya refers not to the Madhyamaka theory but to the absence of any separately real objective correlate (artha) in a cognition; and these two chapters of the Ślokavārttika are in fact concerned mostly with a criticism of Dharmakīrti’s logico-epistemological doctrine. It is true that Kumārila’s criticism in the nirālambanavāda (5–10) of the Buddhist theory of the two satyas—the saṃvrtisatya and the paramārthasatya—and of the idea that the saṃvrti is really false (mrṣā, mithyā) can be applied to Madhyamaka doctrine also; and when commenting on verse 5 of this chapter, uṃveka (eight-century) and Pārthasārathimiśra in fact quote Mūlamadhyamakakārikā xxiv.8 (dve satye samupāśritya buddhānāṃ dharmadeśanā/ loke [sic] saṃvrtisatyaṃ ca satyaṃ ca paramārthataḥ//).55 But it is at the same time clear that Kumārila had in mind the doctrine of the unreality of the external object, which—contrary to what has sometimes been maintained—is not a tenet of the pure Madhyamaka schools (in contradistinction to the synthesising Yogācāra-Madhyamaka). In verse 14 Kumārila distinguishes between the Yogācāras and the Mādhyamikavādins, explaining that the first accept a cognition empty of an object (arthaśūnya vijñāna), whereas the latter maintain the non-existence of even such a vijñāna.56 Kumārila’s doctrines were subsequently discussed and criticized in Śāntarakṣita’s Tattvasaṃgraha and Kamalaśīla’s Pañjikā on it.

The relationship between the Vedānta and the Madhyamaka poses a number of interesting questions which can be only briefly touched on here.57 Reference has already been made above to Bhā(va)viveka’s critique of the Vedānta and to his very remarkable comparison of the supreme brahman with the Buddhist idea of ultimate reality.

On the Brāhmaṇical side, Gauḍapāda—who is dated to ca. 500 but is sometimes considered to have been Śaṃkarācārya’s teacher’s teacher— shows clear and unmistakable links with Buddhist thought in chapter IV of his Māṇḍūkyakārikās, in particular with the Vijñānavāda and also with the Madhyamaka. However, in the other chapters of this work, starting with the third and culminating in the first, there is to be found a progressive movement away from Buddhist ideas and towards the Vedāntic sources.58

In his attitude towards what he terms Śūnyavāda (which in this case is unquestionably the Madhyamaka), Śaṃkarācārya has taken up a negative stance in his Bhāṣya on the Gauḍapādīyakārikās (see ii.32; cf. i.7 and iv.28).59 His position on the subject is especially clear in his Bhāṣyas on the Brahmasūtra and the Brhadāraṇyakopaniṣad, where he has simply rejected the Śūnyavāda, without discussion or analysis, as contrary to all means of correct knowledge (sarvapramāṇavipratiṣiddha) and as therefore unworthy of serious attention and refutation;60 according to the great advaitin, it is simply the teaching of nihilists (vaināśika) and thus like a sandy, dried-up well.61 It is indeed strange to find the Madhyamaka being dismissed out of hand for teaching something it has rejected. In fact, according to the Mādhyamika, nihilism (ucchedānta) is just as unacceptable a philosophical extreme as eternalism (śāśvatānta); and from the soteriological point of view it is perhaps an even more pernicious extreme than its opposite. Beginning with Nāgārjuna, the Mādhyamikas have furthermore accepted saṃvrtisatya and lokavyavahāra as necessary to philosophical thought and as duly recognized in the Buddha’s teaching.62 For Śaṃkarācārya to characterise the Madhyamaka as being in opposition to all means of philosophical knowledge is then at least an oversimplification demonstrating his lack of familiarity with the Madhyamaka as set out in the very large body of literature that it had produced by his time. How this is to be accounted for it is now difficult to say. It may have been the result of inveterate hostility—perhaps more theological and sociological than philosophical—to Buddhist thought. Moreover, Śaṃkarācārya may have felt himself obliged to distance himself from Buddhist thought as much as possible, not least because of the links that seem to have connected certain earlier Vedāntic ideas and Buddhist thought.63 a contributing factor may, finally, have been real unfamiliarity with Madhyamaka literature and philosophy due to a scarcity of this school’s books, at least in the places Śaṃkarācārya was living.

Quite different in his attitude to the Madhyamaka was Śrīharṣa, the twelfthcentury Vedāntin author of the Khaṇḍanakhaṇḍakhādya, who employed a method he termed deconstructive reasoning (khaṇḍanayukti). This form of reasoning has been recognized to be closely related to the Mādhyamika’s dialectic and to his prasaṅga-type reasoning. Śrīharṣa has moreover accepted the vitaṇḍā,64 which is certainly to be understood in this case not as referring to the form of argument employed by an unprincipled opponent merely to win an argument at any cost, but as philosophical criticism and reduction to the impossible.65 Rather than a cavil, then, vitaṇḍā is here an acceptable form of discussion which may be required by a true philosopher when he confronts the onto-logically indeterminate relative.66

In philosophical works of the Jainas, including doxographical texts like Haribhadrasūri’s Ṣaḍdarśanasamuccaya, as well as in later Hindu doxographical works such as Mādhava’s Sarvadarśanasaṃgraha, we find more or less condensed accounts of some form of Madhyamaka, and in particular of its epistemology and gnoseology. no comprehensive analysis of the Madhyamaka by a classical Indian authority of an opposed school seems to exist, however.

Many of the above-mentioned sources have not hitherto been systematically studied in detail from a comparative viewpoint for the purpose of tracing historical influences operating between the Mādhyamikas and the other philosophers of India.67 Once it has been carried out, such study can be expected to yield a rich harvest for our understanding of the history of Indian thought in its diverse forms; and it will at the same time no doubt resolve many a thorny chronological problem.

In conclusion, it is possible to state that the internal relative chronology of the Indian Madhyamaka school is clear in its broad outlines. But the exact place in this overall scheme of a few masters still remains doubtful; and it has not been possible to attribute a number of texts to specific authors in certain cases where the same name has been borne by more than one master. The absolute chronology of several masters also has not been settled beyond doubt, especially in the earlier period. Finally, the external relative chronology continues to pose problems, and in some cases it has not yet been possible to connect allusions made anonymously and without particulars with specific masters and texts known to us. Further detailed and patient research may be able to clear up many of the outstanding obscurities.


Postscript to pages 23–24
In a further article in Acta Orientalia 41 (1980), 27–37, in support of his dating of Bhavya’s Madhyamakaratnapradīpa to 570 (p. 33) and of Candrakīrti to the sixth century (see above, note 28), Christian Lindtner has now put back the date of Dharmakīrti also to ca. 530–600. He describes (pp. 31–32) Hsüan-tsang’s silence as “the apparently only objection to an earlier date of Dharmakīrti than ca. 600–660”; and in support of his earlier dating he refers to E. Frauwallner’s article on Kumārila’s lost Brhaṭṭīkā, where it is argued that this treatise must have been written about 630 as a reply to Dharmakīrti’s theory of inference and the necessary connection between cause and effect. But in his article (WZKSO 6 [1962], p. 89) Frauwallner sought in particular to show that Dharmakīrti’s work on this point is to be placed between Kumārila’s Ślokavārttika, which does not know of it and still treats the relation in question in terms of vyāpti, and the Brhaṭṭikā, the niyama theory of which must have been a response to Dharmakīrti’s theory of inference and the necessary connection between cause and effect. according to Frauwallner, then, Dharmakīrti and Kumārila must have been contemporaries; and Lindtner’s new dating of Dharmakīrti would as a consequence entail putting Kumārila’s date back to the sixth century too. (It might of course be supposed that Dharmakīrti’s contribution came to Kumārila’s attention only after he had finished his Ślokavārttika although Dharmakīrti preceded him by a long period of time; but this line of argument has not been followed by Lindtner.) Whether the (relatively minor) difficulty of the authorship and date of the Madhyamakaratnapradīpa in fact necessitates and justifies the chronological revisions involved by Christian Lindtner’s new datings will require further investigation and discussion which cannot be undertaken until he has presented in full the evidence announced in his last article.

* First published in Indological and Buddhist Studies, Volume in Honour of Professor J. W. de Jong. Ed. L. Hercus et al., pp. 505–30. Canberra: Faculty of asian Studies (australian national university), 1982.
1. See S. Beal, Si-yu-ki: Buddhist Records of the Western World, II (London, 1884), pp. 210–16 (cf. T. Watters, On Yuan Chwang’s Travels, II [London, 1905], pp. 200f ); and J. Takakusu, A Record of the Buddhist Religion As Practised in India and the Malay Archipelago by I-tsing (Oxford, 1896), pp. 159 f. See also the title of the Chinese translation by Paramārtha (500–69) of the Ratnāvalī (Pao-hsing wang, T 1656); cf. J. W. de Jong, IIJ 20 (1978), p. 137. The reference in Hsüan-tsang and I-ching as well as Bāṇa is to a Nāgārjuna who was also a physician. In Hsüan-tsang he is in addition an alchemist (Beal, op. cit., II, p. 216; cf. S. Lévi, JA 1936, pp. 105–6; J. Filliozat, Yogaśataka [Pondichéry, 1979], pp. xiii f ). On other evidence from I-ching see below, note 20.
For the Jaina tradition see Kutūhala, Līlāvaī, verse 1008; cf. a. K. Warder, Indian Kāvya Literature, II (Delhi, 1974), pp. 183, 213–14.
2. These two commentaries are contained in the Tibetan Bstan ’gyur. In verse 14 of the Tibetan translation of the Suhrllekha itself we find the name Bde byed. The Sanskrit original of spyod (bzang po) poses a problem. In his translation of Tāranātha’s Rgya gar chos ’byung, a. Schiefner has given udayana (Târanâtha’s Geschichte des Buddhismus in Indien [St. Petersburg, 1869], p. 2n2); and in the index to the Peking Bstan ’gyur we find udayi(bhadra) (see P. Cordier, Catalogue du fonds tibétain de la Bibliothèque Nationale, III [Paris, 1915], Mdo ’grel, gi, no. 32 [Suhrllekha] and nge, no. 27 [id.] and no. 35 [Suhrllekhaṭīkā]). Cf. S. Lévi, JA 1936, pp. 103–10, and É. Lamotte, Le Traité de la Grande rtu de Sagesse, III (Louvain, 1970), pp. liii–liv (referring to both Udayana and *Jantaka). But it appears that Bde spyod corresponds properly to Sātavāhana (Sāta). another Tibetan equivalent seems to be Mthar ’gro zhon, to which correspond Śālivāhana (Mahāvyutpatti 3654) and *antīvāhana (? which is found in Bu ston’s Chos ’byung, fol. 100b [Obermiller, II, p. 127]); here mtha’ “end” may be intended to render sāta understood as meaning “ended” (root sā-/so“to end, finish, destroy”?). Recently J. Filliozat has suggested that *Jantaka represents *Sā(ṃ)taka(ṇi) ~ Śātakarṇi (Yogaśataka, p. xviii, referring to I-ching).
3. On Kaniṣka and the Turuṣkas see, e.g., Buddha Prakash, Studies in Indo-Asian Art and Culture, 1 (ed. Perala Ratnam, Śatapiṭaka Series 95, new Delhi, 1972), pp. 32–33.
4. See S. Beal, Si-yu-ki, II, p. 302 (cf. pp. 97–101); T. Watters, On Yuan Chwang’s Travels, I (London, 1904), p. 245. See also J. Takakusu, Record of the Buddhist Religion, p. 181.—This synchronism has been dismissed by É. Lamotte, Le traité, III, p. 1n1.
5. Cf. É. Lamotte, L’enseignement de Vimalakīrti (Louvain, 1962), pp. 73–74; Le traité, III, pp. xlvi, li f.
6. On the basis of information originating in Kumārajīva’s school, É. Lamotte and J. May have, however, proposed 243 c.e. as the precise date of Nāgārjuna’s birth (see below).
7. See R. Hikata, Suvikrāntavikrāmi-Pariprcchā Prajñāpāramitā-Sūtra (Fukuoka, 1958), p. lii; É. Lamotte, L’enseignement de Vima ˚ kīrti, p. 76, and Le traité, III, pp. xl, li f.
8. See a. Hirakawa, IBK 5.2 (1957), pp. 504 f. But cf. R. Hikata, op. cit., pp. lxxii f; S. Yamaguchi, EB 1 (1966), pp. 45–47.
9. See J. Burgess, Notes on the Amarāvatī Stūpa (Madras, 1882), p. 57, and The Buddhist Stupas of Amaravati and Jaggayyapeta (archaeological Survey of Southern India, London, 1887), pp. 111–12.
10. Burgess dated the writing of this inscription to ca. 600; but G. Tucci has suggested the earlier dating of 450–500 (Minor Buddhist Texts, II [Rome, 1958], p. 284). T. Ν. Ramachandran thinks the inscription refers to the tantric Siddha Nāgārjuna (see Nagarjunakonda 1938, Memoirs of the archaeological Survey of India 71 [Delhi, 1953], pp. 28–29), on whom see below.
11. É. Lamotte, L’enseignement de Vimalakīrti, pp. 74–76; J. May, Hōbōgirin V, pp. 473a, 478b.
12. É. Lamotte, Le traité, III, pp. xiv f; IV, pp. xiii, xv.
13. See R. Hikata, op. cit., pp. lii f; R. H. Robinson, Early Mādhyamika in India and China (Madison, 1967), pp. 34–39. Cf. also P. Demiéville, JA 1950, pp. 383n, 386, and L’Inde classique, ed. L. Renou and J. Filliozat, II (Paris, 1954), § 2130. In Le traité, III, p. lv, n. 2, Lamotte has, however, declined to express an opinion on this subject (but see Le traité, III, pp. xlix–l).
14. This could have been the “Indian” version referred to by Kumārajīva’s disciple Seng-jui (352–436), since some Serindian scholars at least might well have been capable of writing in Sanskrit.
On this original of the Ta-chih-tu-lun, see Le traité, III, pp. xlvii–xlviii; cf. P. Demiéville, JA 1950, pp. 387 f.
15. While Lamotte accepts that parts of the Ta-chih-tu-lun are by Kumārajīva (Le traité, III, pp. xlix–l), he considers that neither a Chinese nor even a Serindian author could have been as well informed about things Indian as certain sections of this treatise show its author to have been (p. xxvi). (On this see also R. Hikata, op. cit., p. lxvi.) But given the importance of Indian culture in Central asia and, in addition, the fact that Kumārajīva had family links with Kaśmīr and studied there as a boy, it does not seem impossible to ascribe even such sections to a Serindian like him.
16. On Bhā(va)viveka’s relation to Dharmapāla, see S. Beal, op. cit., II, p. 223; cf. T. Watters, op. cit., II, p. 215, and Y. Kajiyama, WZKSO 12–13 (1968/69), pp. 194 f, 199 f.
17. Cf. E. Frauwallner, WZKSO 5 (1961), pp. 141–43; G. Tucci, Minor Buddhist Texts, II, pp. 28–31.
18. On this subject, see our article in Études tibétaines dédiées à la mémoire de Marcelle Lalou (Paris, 1971), pp. 448–71, reprinted as chapter 4 of the present volume.
19. See G. Tucci, JPASB 26 (1930), p. 141 (= Opera minora, II, p. 211 f ); R. Hikata, op. cit., pp. lxxi–lxxii; H. nakamura, Acta asiatica 1 (1960), p. 63; É. Lamotte, L’enseignement de Vimalakīrti, pp. 76–77, and Le traité, III, pp. xl, 1373–74; J. W. de Jong, AM 17 (1971), pp. 107–8.
20. I-ching (635–713), who left Tāmralipti about 685, already mentions Nāgārjuna’s having studied the Vidyādharapiṭaka. See E. Chavannes, Mémoire composé à l’époque de la grande dynastie T’ang sur les religieux éminents qui allèrent chercher la loi dans les pays d’Occident (Paris, 1894), p. 102. In I-ching’s Record (translated by Takakusu, p. 35), Nāgārjuna appears only as a physician and, perhaps, alchemist.
21. See É. Lamotte, Le traité, III, p. 1374.
22. The authorship of the *Trikāyastotra/stava or *Kāyatrayastotra/stava ascribed to nāga by Tāranātha (Rgya gar chos ’byung, p. 68) is uncertain. a Kāyatrayastotra is ascribed to (a) Nāgārjuna in the Bstan ’gyur; but it is doubtful that a hymn concerning the three kāyas can be attributed to Nāgārjuna I. In addition, the Kāyatrayāvatāramukha is ascribed to nāgamitra in the Bstan ’gyur.
23. See G. Tucci, JPASB 26 (1930), p. 142 (= Opera minora, I, p. 212); P. Demiéville, in L’Inde classique, II §§ 2092–93; Chou Yi-liang, HJAS 8 (1944/45), pp. 281 f, 313–14.
24. It is probably preferable to regard nāgabodhi as a follower of Ārya Nāgārjunapāda, the tantric Siddha.
25. Ārya Vimuktisena may have been a younger contemporary of Buddhapālita; cf. D. Seyfort Ruegg, WZKSO 12–13 (1968/69), p. 306.
26. See Y. Kajiyama, WZKSO 12–13 (1968/69), pp. 193–203, for the dates of Bhā(va)viveka and Sthiramati.
27. Candrakīrti was, however, placed in the sixth century by M. Winternitz, History of Indian Literature, II (Calcutta, 1933), p. 363, because he held him to be the pupil of Dharmapāla: but he gives no source for this statement (this may be Tāranātha’s Rgya gar chos ’byung, p. 133.18, which is, however, altogether obscure).
28. See C. Lindtner, AO 40 (1979), p. 91.
29. It has to be noted that the Madhyamakaratnapradīpa refers to Vajrayānist ideas and to the Vidyādharapiṭaka.
30. Cf. J. W. de Jong, IIJ 16 (1975), pp. 179–80.
31. For references see D. Seyfort Ruegg, WZKSO 12–13 (1968/69), p. 307n18.
32. Cf. Paṇ chen Blo bzang dpal ldan ye shes, Shambha la’i lam yig, fol. 18a–b; G. Tucci, IHQ 7 (1931), pp. 683–702 (= Opera, minora, II, pp. 305–19); Tibetan Painted Scrolls (Rome, 1949), pp. 137 and 164.
33. See J. Takakusu, A Record of the Buddhist Religion, pp. 179–80. (I-ching has also made Bhartrhari a contemporary of Dharmapāla!)
34. See H. R. Rangaswami Iyengar, JBBRAS 26 (1951), pp. 147–49; H. nakamura, Studies in Indology and Buddhology (Festschrift S. Yamaguchi, Kyoto, 1955), pp. 122–36; D. Seyfort Ruegg, Contributions à l’histoire de la philosophie linguistique indienne (Paris, 1959), p. 59; E. Frauwallner, WZKSO 3 (1959), pp. 107 f, and 5 (1961), pp. 134–35. Cf. also J. Brough, BSOAS 36 (1973), pp. 248–60.
35. Rgya gar chos ’byung (ed. a. Schiefner), pp. 68, 73. See also Bu ston, Chos ’byung, fol. 94a1 (Obermiller, II, p. 109, reads Bāhula, which is presumably a misprint for Rāhula).
36. Rgya gar chos ’byung, pp. 53, 83; Tāranātha, Bka’ babs bdun ldan gyi brgyud pa’i rnam thar, fols. 181 f. See also Bu ston, Chos ’byung, fol. 99a (Obermiller, II, p. 123
37. Sometimes, of course, Tāranātha does explicitly distinguish between two masters having the same name. In addition to Ārya Vimuktisena and Bhadanta Vimuktisena, a case in point is that of Dharmamitra, the Mādhyamika interpreter of the Prajñāpāramitā (ca. 800), whom Tāranātha distinguishes from the Vaibhāṣika and Vinaya master of the same name who lived earlier (see above). Tāranātha distinguishes also between two Mātrceṭas (p. 75; cf. p. 152).
38. Aulūka is also mentioned in lists of Sāṃkhya teachers, however; cf. P. Chakravarti, Origin and Development of the Sāṃkhya System of Thought (Calcutta, 1951), p. 131. In the Vaidalyaprakaraṇa (§ 8) his name appears beside those of Kapila, Maṭhara, Vyāsa, et al.
39. The Mātharav rtti itself is of course dated to a much later time. On the Māṭharas in a treatise ascribed t ˚ Āryadeva (and extant only in a Chinese translation by Bodhiruci [T. 1640], see H. nakamura, HJAS 18 (1955), p. 99, who also notes an identification with Vyāsa. See also P. Chakravarti, op. cit., p. 159. as already noted, in the Vaidalyaprakaraṇa (§ 8), Maṭhara and Vyāsa are listed separately, beside Aulūka et al.
40. WZKSO 7 (1963), pp. 64–70. See also G. Tucci, Pre-Diṅnāga Buddhist Texts on Logic from Chinese Sources (Baroda, 1929), notes on the Vigrahavyāvartanī, pp. 36 f, who considers that NS II.i contains replies to Nāgārjuna. according to Vācaspatimiśra’s Tātparyaṭīkā, the opponent in II.i.8 is a Mādhyamika.
41. Journal of Indo-European Studies 5 (1977), pp. 265–73. B. K. Matilal, Nyāya-Vaiśeṣika (A History of Indian Literature, Otto Harrassowitz, VI,2, Wiesbaden, 1977), p. 79, has suggested that NS IV.i.38 may be a reply to Nāgārjuna’s critique of svabhāva. But, contrary to W. Ruben’s opinion (Die Nyāyasūtra’s, Leipzig, 1928, p. 106), itaretarābhāva in the preceding sūtra can hardly apply to Nāgārjuna’s theory; and the Laṅkāvatārasūtra (p. 75) indeed condemns itaretaraśūnyatā as the most inferior emptiness.
42. Pre-Diṅnāga Buddhist Texts, p. xxvii. The above-mentioned discussions of the Vigrahavyāvartanī have mostly been conducted without reference to the Vaidalya“Sūtra” and Prakaraṇa, both of which are ascribed to Nāgārjuna and are therefore of the highest relevance.
43. Pre-Diṅnāga Buddhist Texts, p. xxvii. 44. WZKSO 2 (1958), p. 131.
45. IBK 23.1 (1974), pp. 7–12. On the later Sāṃkhyasūtra and the Madhyamaka see R. Garbe, Die Sâṃkhya-Philosophie (Leipzig, 1917), p. 146n265.
46. Cf. Le traité, III, pp. xiv–xxv, xxxii–xxxvii; Le traité, IV, pp. x–xii.
47. Cf. Le traité, III, pp. xxv–xxvi; R. H. Robinson, Early Mādhyamika, pp. 68 f, 90, 94. R. Hikata, Suvikrāntavikrāmipariprcchā, pp. lxiv f, places the Nyāyasūtra later than the Ta-chih-tu-lun.
48. This older form of Vedānta has been discussed by H. nakamura, Shoki no Vedānta tetsugaku I (Tokyo, 1950), pp. 238–332 (cf. HJAS 18 [1955], pp. 103–4) and V. V. Gokhale, IIJ 2 (1958), pp. 165 f.
49. See V. V. Gokhale, IIJ 5 (1961/62), pp. 271 f. On Bhā(va)viveka and Maitreya and avalokiteśvara, compare Hsüan-tsang’s remarks in his Hsi-yü-chi (trans. S. Beal, Si-yu-ki, II, pp. 223–26). and on his closeness to the “Sāṃkhya” see ibid., p. 223, and T. Watters, On Yuan Chwang’s Travels, II, pp. 215, 223.
50. Cf. Y. Kajiyama, WZKSO 7–8 (1963–64); 12–13 (1968/69), p. 198.
51. The information that Tāranātha has given on Devaśarman in his Rgya gar chos ’byung (p. 133.18) is hardly intelligible.
52. See R. Garbe, Die Sâṃkhya-Philosophie, pp. 391–92 (where most of Candrakīrti’s information has been described as untraceable in the Sāṃkhya sources).
53. According to I-ching (see J. Takakusu, A Record of the Buddhist Religion, p. 186), Dignāga’s treatises were still in use in his time as basic works on logic.
54. See H. nakamura, ABORI 48–49 (1968), pp. 119–22.
55. Cf. also Bhāsarvajña, Nyāyabhūṣaṇa (Ṣaḍdarśanaprakāśanagranthamālā, Banaras, 1968), pp. 518 f on saṃvrti.
56. For references to the Mādhyamikas see uṃveka, Ślokavārttika-Tātparyaṭīkā on Śūnyavāda 1 and 245, as well as on nirālambanavāda 14–18. Kumārila mentions Śūnyavādins in his nirālambanavāda 129.
57. On the early Madhyamaka of Nāgārjuna and Āryadeva and the Vedānta see H. nakamura, HJAS 18 (1955), pp. 89–102; K. Venkataramanan, Nāgārjuna’s Philosophy as Presented in the Mahā-Prajñāpāramitā-Śāstra (Tokyo, 1966), pp. 319–21.
58. For a recent discussion and analysis of the Gauḍapadīyaor Māṇdūkya-Kārikās see T. Vetter, WZKS 22 (1978), 95–131. Vetter, who places this work ca. 500, thinks (p. 95n1) that Bhā(va)viveka’s Madhyamakahrdayakārikā viii.12 may presuppose GK iii.5. at the same time he recognizes the Budd ˚ t background to chapter IV of GK, considering it to be the earliest of the four chapters; and he suggests that the other chapters of the GK reflect Gauḍapāda’s progressive development away from the Mahāyāna, the first chapter of the GK being then the one least indebted to Buddhist thought and closest to the upaniṣadic source (the Māṇḍūkyopaniṣad).
59. See T. Vetter, Studien zur Lehre und Entwicklung Śaṅkaras (Vienna, 1979), pp. 50, 53 f, 73–74.
60. Brahmasūtrabhāṣya II.ii.31 and Brhadāraṇyakopaniṣad-bhāṣya IV.iii.7.
61. Brahmasūtrabhāṣya II.ii.32. Cf. Brhadāraṇyakopaniṣad-bhāṣya I.ii.1 where, in a discussion of causality and the reality of kāraṇa in relation to kārya, a Śūnyavādin holding  the tenet sarvasyaivâbhāvo ’stu is supposed to be the opponent. But this passage in fact tells us nothing pertinent to the Madhyamaka as we know it from its own sources.
62. See, e.g., Mūlamadhyamakakārikā, chapter XXIV, especially verses 8–10. It is true that Candrakīrti has not invested vyavahāra and saṃvrtisatya with ultimate reality and “truth,” for he considers that it cannot resist ontologic ˚ analysis and that any attempt to investigate it in such a way is fraught with paradoxes and insoluble logical antinomies. But in itself this theory need not imply either indulgence in paralogisms and nihilism, or the taking up of an anti-philosophical position.
63. D. Ingalls has suggested that even if, metaphysically, the distance between “idealistic” Buddhism and Śaṃkara was but little, the psychological and historical chasm separating them was nevertheless deep (PEW 3 [1953], pp. 291 f ). See also T. Vetter, WZKSO 12–13 (1968/69), p. 409, and Studien, pp. 53 f, on Śaṃkara’s differences with Buddhism.
64. Kāśī Sanskrit Series no. 197 (Banaras, 1970), pp. 126–27 (= Chowkhamba Sanskrit Series ed. § 168). Cf. pp. 7, 12.
65. In the Nyāyasūtras, vitaṇḍā is mentioned together with jalpa (IV.ii.49) and also hetvābhāsa and other faults in debate (I.i.1). In I.ii.3 it has been defined as a jalpa that lacks the assertion of a counter-thesis (pratipakṣasthāpanāhīna).K. Bhattacharya, JA 1975, pp. 99–102, has denied that Śrīharṣa ever described himself as a vaitaṇḍika (which seems to be true); but he does not mention the above-mentioned passages of the Khaṇḍanakhaṇḍakhādya, where Śrīharṣa speaks favourably of the employment of vitaṇḍā. It seems necessary, then, to distinguish between vitaṇḍā as a cavil used to gain a debating point by demolishing the opponent’s thesis and vitaṇḍā as a form of reasoning and argument where a serious philosopher refrains from asserting a counter-thesis. (See also NS IV.ii.48: pratipakṣahīnam api vā prayojanārtham arthitve.) While he would no doubt reject a vitaṇḍā used in the first way, Śrīharṣa appears to be ready to accept the second use since, with the sole exception of brahman as vijñāna, everything is different from both existence and non-existence (vijñānavyatiriktaṃ punar idaṃ viśvaṃ sadasadbhyāṃ vilakṣaṇaṃ brahmavādinaḥ saṃgirante), much as is stated in the Laṅkāvatārasūtra. In their comments on NS I.ii.3, uddyotakara and Vācaspatimiśra discuss in addition the question whether the vitaṇḍā, defined as a jalpa that is pratipakṣasthāpanāhīna, can constitute proof by residue (pāriśeṣyāt pakṣasiddhiḥ) arrived at by excluding all opposed theses. In this case the Vaitaṇḍika would indeed have a thesis, even if it is established exclusively by refuting other theses.
66. It should be recalled here that the Vaidalyaprakaraṇa (§ 56) has rejected vitaṇḍā together with the other fifteen topics enumerated in Nyāyasūtra I.i.1. Though he uses prasaṅga-arguments, its author does not, therefore, identify himself with the Vaitaṇḍika who uses the vitaṇḍā in the first way mentioned in the previous note.
67. T. R. V. Murti, Central Philosophy of Buddhism (London, 1955), provides an interesting philosophical comparison between what he calls the two traditions of Indian philosophy embodied respectively in Buddhism and Vedānta. But Murti was not concerned with chronological sequences and interrelationships. nor did he make use of the Chinese and Tibetan sources that are now, owing to the loss of the Sanskrit originals of so many works of Indian philosophy, our principal documents not only for the history of the Madhyamaka but also for so much of classical Indian thought.


How to cite this document:
© David Seyfort Ruegg, The Buddhist Philosophy of the Middle (Wisdom Publications, 2010)

Creative Commons License
This selection from The Buddhist Philosophy of the Middle by David Seyfort Ruegg is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Based on a work at
Permissions beyond the scope of this license may be available at