Resurrecting Candrakīrti

1. The Indian Discovery of Candrakīrti

+ –


1. The Indian Discovery of Candrakīrti

PROMINENT TIBETAN SCHOLARS of the fourteenth and fifteenth centuries uniformly speak of a Prāsaṅgika school of Madhyamaka (“Middle Way”), founded by the Indians Buddhapālita (c. 500)1 and Candrakīrti (c. 570–640),2 developed in India—in some accounts by a lineage of mostly unlettered disciples but always including such luminaries as Śāntideva (early eighth century) and Atiśa (c. 982–1054)—and later propagated in Tibet. The Tibetan systematizers likewise speak of clear differences between the Prāsaṅgika and Svātantrika interpretations of Madhyamaka and of the superiority of Prāsaṅgika in elucidating the “true thought” of Nāgārjuna (c. 200), the founder of Madhyamaka. In this vein, Candrakīrti is said to have “refuted” Bhāvaviveka (c. 500–570), the “founder” of the Svātantrika interpretation, and established the preeminence of Prāsaṅgika through writing his commentary on Nāgārjuna’s Fundamental Treatise on the Middle.3 While the precise nature of the Prāsaṅgika-Svātantrika division was debated in the Kagyu, Sakya, and Geluk schools throughout the fourteenth to eighteenth centuries and in the Nyingma school in the nineteenth century, there was no disagreement that just such a division accurately reflected Indian Buddhist developments of the sixth and seventh centuries.4 Tibetan scholarship on this distinction, from the fourteenth century into the present, has influenced a great deal of contemporary scholarship that continues to speak of two schools of Indian, and then Tibetan, Madhyamaka.

However, the Indian textual record presents a remarkably different view than fifteenth-century Tibetan scholars’ accounts. When we consider this record, we must conclude that Candrakīrti, rather than forming a school of Madhyamaka and triumphing over or refuting Bhāvaviveka, was in fact largely ignored in his day and for some three hundred years in both India and Tibet. Jayānanda (twelfth century)5 is the only known Indian commentator on the works of Candrakīrti,6 whereas there were eight Indian commentaries 18on Nāgārjuna’s Fundamental Treatise on the Middle and twenty-one Indian commentaries on Maitreya’s Ornament for Realization.7 The lineage of Indian Prāsaṅgika disciples stretching from Candrakīrti through Śāntideva and extending to Atiśa, the supposed progenitor of Prāsaṅgika in Tibet, varies widely in the Tibetan accounts and rarely includes figures known elsewhere.8 Furthermore, the argument from silence against Candrakīrti’s importance in India is bolstered by the fact that none of these shadowy figures is known by Tibetan scholars to have written on Madhyamaka (or anything else). This absence of any reported texts strongly suggests that, unlike the many volumes known to Tibetan scholars to have existed in India or Tibet in the past but no longer accessible to them or to us,9 no such texts by these figures ever existed. Rather, these figures would seem to represent Tibetan historians’ acknowledgment of great gaps in the Prāsaṅgika “lineage” and their attempts to fill in these holes with names, if not writings.

One can infer that the very survival of Candrakīrti’s writings down to the time of Jayānanda could only have been brought about by some kind of following, whether Candrakīrti’s writings were preserved in monastic libraries or transmitted in scribal families.10 Most strongly, we can imagine the existence of a marginal school of thought that did not champion Candrakīrti with new treatises (at least none that survived even until the time of Jayānanda) but studied and preserved his texts. It may have been this sense of a “lineage” that Tibetan authors imagined and attempted to enliven with names. A school, family, or library preserving Candrakīrti’s writings furthermore provides a more coherent picture of how his texts could later be popularized.

While the ongoing search for Sanskrit manuscripts could one day turn up a treatise from an early member of a putative Candrakīrti following, recent discoveries strengthen the case that Candrakīrti’s popularity arose long after his death. Studies of the recently recovered eighteen-folio Lakṣaṇaṭīkā show it to be a series of notes composed mostly in Sanskrit (with parts of four folios consisting of Tibetan notes) on three of Candrakīrti’s compositions.11 The colophons to the texts that the “Lakṣaṇaṭīkā” was bundled with lead Yonezawa tentatively to conclude that these comments stem from Abhayākaragupta (c. 1025–1125) through the pen of Nur Dharmadrak, who served as the scribe.12 While attributing these comments to Abhayākaragupta will require a great deal of further investigation into this manuscript and comparison with his known writings, the dating of the text seems secure, given Nur Dharmadrak’s role. A late eleventh- to early twelfth-century frame for these important notes on several of Candrakīrti’s major writings aligns well with the surviving evidence for Indian interest in his work. Thus, at present 19we can deduce that Candrakīrti’s writings did not spawn a literary tradition for many hundreds of years, with Jayānanda’s commentary and the “Lakṣaṇaṭīkā” representing the earliest known works that take Candrakīrti as their subject matter.

The silence of Candrakīrti’s supposed Middle Way adversaries rings even more tellingly. While Avalokitavrata (c. 700) in his subcommentary on Bhāvaviveka’s Lamp for Wisdom mentions Candrakīrti in a list of Indian scholars who wrote commentaries on Nāgārjuna’s Fundamental Treatise on the Middle,13 he says nothing about Candrakīrti’s lengthy criticisms of Bhāvaviveka. One can well assume that, in the Indian commentarial tradition, if Avalokitavrata deemed Candrakīrti’s attacks damaging, it would have been incumbent upon him to respond. His silence, in an otherwise extensive treatise (spanning three volumes in Tibetan translation), suggests that he viewed Candrakīrti’s criticisms as insignificant, not worthy of response, perhaps not even as serious philosophy.

Likewise, the important Mādhyamikas Śāntarakṣita (eighth century) and Kamalaśīla (c. 740–95) remained silent on Candrakīrti.14 Their extensive use of the Buddhist epistemological tradition, to an even greater degree than Bhāvaviveka, would require their responses to Candrakīrti’s attacks on that tradition, had they viewed his attacks to be damaging. Both authors, instead, were more concerned with Dharmapāla’s critique—from a Yogācāra viewpoint—of the feasibility of joining epistemology with Madhyamaka ontology.15 Furthermore, in Ichigō’s analysis, Kamalaśīla worked to refine Bhāvaviveka’s and Śāntarakṣita’s views, arguing against subtleties in their writings16 rather than concern himself with the widely divergent views of Candrakīrti. The wide success of Śāntarakṣita’s and Kamalaśīla’s Yogācāra-Madhyamaka interpretation, an interpretation well at odds with Candrakīrti’s own, suggests Candrakīrti’s insignificance during this time.17 In contradistinction to what fifteenth-century Tibetan authors state, the textual evidence leads one to conclude that Candrakīrti was a marginal figure in his day and uninfluential in India until the close of the first millennium.

Tibetan evidence—translations of Sanskrit Madhyamaka texts and native Tibetan commentaries and doxographies—from the establishment of Buddhism in Tibet until 1000 show a similar disinterest in Candrakīrti. Whereas a wealth of important Madhyamaka texts by Nāgārjuna, Āryadeva, Buddhapālita, Bhāvaviveka, and Śāntideva were translated during the “early diffusion” (snga dar) of Buddhism into Tibet, Candrakīrti’s major writings were not translated into Tibetan until the eleventh century. Only two of his commentaries, both on two of Nāgārjuna’s texts, Sixty Stanzas of Reasoning18 20and Seventy Stanzas on Emptiness,19 were translated in the “early diffusion.” These commentaries would be likely candidates for translation as they represent the only Indian commentaries on these important Nāgārjuna texts.20 In cases where Tibetans had a choice of commentarial tradition, for instance with Nāgārjuna’s Fundamental Treatise on the Middle, Candrakīrti was left out.

As is well known, Śāntarakṣita and Kamalaśīla, both later categorized as Svātantrika-Mādhyamikas,21 were instrumental in the early diffusion of Buddhism in Tibet, the former credited with creating the first monastery in Tibet at Samyé and ordaining the first Tibetan monks and the latter, his student, credited with establishing the orthodox “gradual path” at the purported Great Debate at Samyé. Their most important Madhyamaka texts were translated during the early diffusion, along with those by another key Indian author, Jñānagarbha, who blended components of Dharmakīrti’s epistemology with Madhyamaka thought.22 Not surprisingly then, the first Tibetan doxographies by Yeshé Dé (Ye shes sde) and Kawa Peltsek (Ka ba dpal brtsegs) in the eighth century esteem the Yogācāra-Madhyamaka synthesis created by Śāntarakṣita and Kamalaśīla as the highest Buddhist school of thought.23 Bhāvaviveka’s Sautrāntika-Madhyamaka is ranked second. Candrakīrti is not mentioned. No Prāsaṅgika school is identified nor do we see the appellation, “Svātantrika,” which—as discussed in chapter 2—is first employed only in the twelfth century, in contradistinction to Candrakīrti’s views. This earlier bifurcation of Madhyamaka into Yogācāra and Sautrāntika sub-streams, to the apparent exclusion of Candrakīrti’s views, appears also in Rongzom Chokyi Zangpo’s (Rong zom Chos kyi bzang po, eleventh century) three doxographical works, perhaps our earliest sources for the Madhyamaka of the “later diffusion” (phyi dar) of Buddhism in Tibet.24

The Indian and Tibetan evidence point to an eleventh-century resurrection of Candrakīrti’s writings in India and a twelfth-century birth of the Prāsaṅgika movement in Tibet. In addition to detailing Candrakīrti’s Indian rise, this chapter discusses the fragmented evidence that illuminates the philosophical and doctrinal issues (treated more fully in chapters 3 through 5) engendered by his writings that polarized Indian and Tibetan Buddhists in this period. The central issue around which Candrakīrti’s fame grew was his perceived denial of “valid cognition”25—the epistemological enterprise foundational to Indian thought from at least the sixth century. As will be seen, both Candrakīrti’s supporters and detractors saw his philosophy as denying the validity of ordinary human cognition in the project of reaching 21enlightening knowledge. This denial held far-reaching ramifications, extending from a low appraisal of the value of human intellect to the very nature of buddhahood.

Reviving Candrakīrti’s Critique of Ultimate Valid Cognition

As mentioned above, fifteenth century and later Tibetan authors frequently group Śāntideva’s writings with Candrakīrti’s as “Prāsaṅgika” and place him in a lineage stretching from Candrakīrti down to these authors themselves. Śāntideva is the one figure in these lineage lists prior to Atiśa about whom we have literary information. However, Śāntideva’s own writings make no reference to Candrakīrti nor to any other of the figures that Tibetan historians would place in a lineage between Candrakīrti and Śāntideva. Śāntideva’s surviving writings, consisting of poetry and comments interspersing his collection of sūtra fragments,26 allow a great deal of interpretive room. Several verses from the ninth chapter of Engaging in the Bodhisattva’s Practice echo sentiments found in Candrakīrti’s writings, especially the denial that ultimate truth is a referent of human intellect, the explanation of ultimate truth as “non-seeing,” and the refutation of self-cognizing consciousness.27

Despite this seeming harmony between Candrakīrti and Śāntideva, it is important to note that this text was commented upon from a decidedly non-Prāsaṅgika standpoint both in the early diffusion of Buddhism in Tibet and during the later spread. Saito points to two Indian commentaries, likely the earliest, that treat Śāntideva’s text from a Yogācāra-Madhyamaka perspective.28 From the later diffusion, we see commentaries to the text by Ngok the Translator and Chapa, both of whom opposed Candrakīrti’s views.29 Furthermore, Śāntideva’s text grew over hundreds of years: the version cataloged in the Den karma collection (c. 800) is 600 stanzas in length,30 while that preserved in the Dunhuang caves (dated to before 950) contains 701.5 stanzas.31 Both are far shorter than the present canonical version in 913 stanzas. The fact that Śāntideva’s stanzas were important to Yogācāra-Mādhyamikas and the growth of the text heightens our uncertainty as to which views we may ascribe to the eighth-century Śāntideva.

While we thus cannot with any certainty show a historical link between Śāntideva and Candrakīrti, over two hundred years after Śāntideva wrote, his commentator, Prajñākaramati (950–1030), ties Śāntideva’s views to Candrakīrti. Prajñākaramati cites Candrakīrti’s Entrance to the Middle repeatedly in his commentary to Śāntideva’s Engaging in the Bodhisattva’s Practice.32 Prajñākaramati relies particularly heavily on Candrakīrti in his comments 22on Śāntideva’s proclamation of the two truths (satyadvaya, stanza IX.2), citing Entrance to the Middle four times (stanzas VI.23, 25, 28, and 29),33 a lengthy sūtra passage found in Entrance to the Middle, and three sūtra passages found in Candrakīrti’s Clear Words. The stongest link we may establish is not between Candrakīrti and Śāntideva but between Candrakīrti and Prajñākaramati. The latter is the earliest Indian author that we know of to express overt enthusiasm for Candrakīrti’s writings.

Vibhūticandra (c. 1200) likewise employs stanzas from Entrance to the Middle in his comments on Śāntideva’s stanza IX.2.34 Vibhūticandra was part of the last entourage of Indian paṇḍitas to travel to Central Tibet in 1204; he traveled with Śākya Śrībhadra (1127/1145–1225/1243),35 with whom the extremely influential Tibetan scholar Sakya Paṇḍita (1182–1251) worked. As will be discussed in more detail in chapter 2, Sakya Paṇḍita was the first to adopt the Prāsaṅgika position in the Sakya school.36 Vibhūticandra, then, may have been partly responsible for this development within Sakya.

Śāntideva’s brief stanza, in which ultimate truth is declared outside the realm of human intellect, reads:37

           It is asserted that there are two truths—obscurational and ultimate.

           The ultimate is not a referent of awareness; awareness is said to be obscurational.

Classical Indian aesthetics valued poetic brevity; elaborating on meaning was left to a commentator. So while Śāntideva may have meant his stanza to echo an important theme in Candrakīrti’s writings that the ultimate “is just not an object of consciousness,”38 Prajñākaramati makes the first certain connection between the two and is the earliest commentator to explicate this theme in either Candrakīrti’s or Śāntideva’s works. Prajñākaramati elaborates at some length, writing that Śāntideva’s “awareness” means “all consciousness,” that the ultimate “surpasses the sphere of all consciousness,” and that it is “impossible to bring [the ultimate] within the sphere of awareness in any way.”39 He later notes, “All awareness, whether having an object or not having an object, has a nature of conceptuality and all conceptuality has a nature of ignorance.”40 Prajñākaramati also utilizes Candrakīrti’s comparison of objects of consciousness to the flickering hairs seen by those suffering from eye diseases, suggesting that just as all awareness is flawed with ignorance, so all objects of awareness are unreal.41

In linking Candrakīrti’s and Śāntideva’s texts on the radical separation of human consciousness and the ultimate, Prajñākaramati establishes an 23important tenet of his interpretation of Madhyamaka and shows a lengthy pedigree of the tenet, drawing upon a sūtra in which the Buddha himself makes such a proclamation;42 he establishes a tradition of exegesis, the beginnings of what would become a Prāsaṅgika interpretation of Buddhist philosophy. Furthermore, he posits a radical separation of ultimate truth and those things known by ordinary consciousness: what we call knowledge he calls ignorance. Such a vast divide between ordinary consciousness and knowledge of the ultimate runs directly counter to the epistemological project of establishing the valid foundations of all knowledge, conventional and ultimate, engaged in by Mādhyamikas since the sixth century. Indeed, Prajñākaramati concludes his discussion of ultimate truth by stating that only āryas—advanced bodhisattvas who realize emptiness directly—have valid cognition concerning the ultimate.43 This conclusion, too, is lifted directly from Candrakīrti’s Entrance to the Middle.44 However, unlike in Candrakīrti’s India, Prajñākaramati’s great divide separating conventional knowledge from knowledge of the ultimate sparked debate among Mādhyamikas on both sides of the Himalayas. Candrakīrti’s rejection of valid cognition in ultimate pursuits ran directly counter to the tenor of Indian philosophy that in his day was dominated by epistemological concerns across religious traditions. In such a religious climate, his separation could not be taken seriously. By Prajñākaramati’s time, Candrakīrti’s conservative bent could be recast as a unique and viable Madhyamaka interpretation.

As noted above, Atiśa (Dīpaṅkaraśrījñāna), a junior contemporary of Prajñākaramati, is commonly credited with establishing Prāsaṅgika in Tibet. His Introduction to the Two Truths twice praises Candrakīrti, once for Candrakīrti’s presentation of the two truths45 and once for Candrakīrti’s understanding of ultimate truth. The latter passage reads:46

           Candrakīrti is the disciple of Nāgārjuna

           Who saw the truth of the final nature.

           The truth of the final nature is to be realized

           According to the instructions of his lineage.

Leading up to this stanza, Atiśa denied the validity of both forms of valid cognition accepted by Buddhists—direct perception and inference—to realize the ultimate and further rejected the ability of conceptual and non-conceptual consciousness to realize it.47 Clearly, his praise of Candrakīrti’s understanding of ultimate truth references Candrakīrti’s denial that the ultimate can be known by human intellect.


In this same text, Atiśa refers favorably to Bhāvaviveka,48 the supposed founder of the Svātantrika interpretation of Madhyamaka, over whom Candrakīrti’s Prāsaṅgika supposedly triumphed. Atiśa also translated into Tibetan two of Bhāvaviveka’s most important texts, Heart of the Middle and its autocommentary, Blaze of Reasoning.49 These two texts were translated at the request of Ngok Lekpay Sherab,50 whose monastic institute staunchly opposed Candrakīrti’s views. Elsewhere, Atiśa lists Bhāvaviveka and Candrakīrti as authoritative interpreters of Madhyamaka, along with Nāgārjuna, Āryadeva, Śāntideva, and Atiśa’s own teacher, Bodhibhadra.51 We might conclude that Atiśa saw Candrakīrti and Bhāvaviveka as upholding equally valid positions. Ruegg notes that, “In Dīpaṅkaraśrījñāna’s time and circle, Bhavya’s and Candrakīrti’s schools of the Madhyamaka were apparently not clearly differentiated by distinct designations and they were evidently being studied side by side.”52 Extending Ruegg’s point, the very issues that would polarize Bhāvaviveka’s and Candrakīrti’s writings into separate schools of thought were only in Atiśa’s day coming to be elucidated.

However, Atiśa’s endorsement of Candrakīrti may be the earliest instance of an Indian author favoring Candrakīrti over Bhāvaviveka. When we look more closely at Atiśa’s Introduction to the Two Truths, we see that his praise of Bhāvaviveka was misplaced. He wrote “The Master scholar Bhavya stated that [the ultimate] is not realized by either conceptual nor nonconceptual consciousness.53 Lindtner adeptly identifies a very similar statement, to which Atiśa likely referred, in the Jewel Lamp of the Middle, a text attributed to Bhāvaviveka.54 The Bhāvaviveka who wrote the Jewel Lamp of the Middle, and who claims to be the author of Heart of the Middle and Blaze of Reasoning as well,55 refers favorably to Candrakīrti’s Entrance to the Middle, and so must post-date him.56 The Bhāvaviveka who wrote Heart of the Middle and Lamp for Wisdom—a section of which Candrakīrti critiqued in detail—preceded Candrakīrti, making a common attribution impossible. The statement that Atiśa endorses from the Jewel Lamp of the Middle shows much more affinity for Candrakīrti’s views—at least as presented by Atiśa—than those expressed in Heart of the Middle.

More tellingly, Atiśa criticizes another of Bhāvaviveka’s views (without mentioning Bhāvaviveka by name). He writes, “Ultimate truth is only one; others assert that it is two.”57 As we have seen, Candrakīrti and Prajñākaramati rejected the applicability of valid cognition to ultimate truth. Anticipating such criticism, Bhāvaviveka wrote of two kinds of ultimate consciousnesses, which realize two kinds of objects:


           The ultimate is of two types: one engages thoroughly effortlessly, passes beyond the world, is undefiled, and lacks proliferation; the second engages with thorough effort, accords with the collection of merit and wisdom, is called “pure worldly wisdom,” and possesses proliferations.58

Bhāvaviveka explains that the second kind of ultimate consciousness realizes ultimate truth inferentially; he explains how inference is utilized in knowing ultimate truth. In rejecting that ultimate truth is two, Atiśa rejects Bhāvaviveka’s solution to inference’s utility. Atiśa further states that, “The deluded whose vision is narrow say that the two [kinds of valid cognition, direct perception and inference] realize emptiness.”59 Atiśa’s critique of Bhāvaviveka’s use of valid cognition in general and inference in particular, seemingly inspired by Candrakīrti’s views, is perhaps the first implicit hierarchical ranking of Candrakīrti over Bhāvaviveka. The validity of human cognition in knowing emptiness is the litmus test in Atiśa’s ranking.

Apart from Atiśa’s attributions, the later Bhāvaviveka’s work itself reveals the impact of Candrakīrti’s views and represents a further mark of his growing influence. It may well be that the passage from Blaze of Reasoning (Tarkajvāla) cited just above does not anticipate Candrakīrti’s criticism but, in fact, responds to it. Ruegg utilizes the work of Ejima, who distinguishes an “Ur-Tarkajvāla” from a revised version, to suggest that many parts of Blaze of Reasoning were not written by the Bhāvaviveka who wrote Heart of the Middle (the stanzas upon which Blaze of Reasoning comments) but by the later Bhāvaviveka who wrote the Jewel Lamp of the Middle.60 Furthermore, the commonalities between the Jewel Lamp of the Middle, the Compendium of Meanings of the Middle (Madhyamakārthasaṃgraha), and certain parts of Blaze of Reasoning suggest that the same hand (Bhāvaviveka II) composed them.61 The passage I have cited from Blaze of Reasoning can likely be considered a later accretion as the Compendium of Meanings of the Middle likewise posits a two-fold ultimate truth, a figurative ultimate (paryāyaparamārtha) that can be expressed in language and grasped by conceptual thought and a non-figurative ultimate that is beyond expression and thought.62 This closely mirrors the two-fold ultimate explained in the Blaze of Reasoning passage...

Join Wisdom

Members of the Wisdom Experience will be able to read Wisdom books online as part of their membership. Please join our waiting list and be the first to find out when Wisdom Experience membership opens!

rotate left rotate right