Chapter 4. Buddhapālita and Bhāvaviveka: the Svātantrika–Prāsaṅgika distinction¶

The preceding chapter ended with a question. Should Madhyamaka refute opponents only through reductio, asserting nothing itself, or should it also advance independent arguments? In the fifth and sixth centuries, this apparently technical disagreement opened two paths within Madhyamaka through Buddhapālita and Bhāvaviveka. Candrakīrti's seventh-century judgment then shaped what later became known as Svātantrika and Prāsaṅgika. This chapter traces the dispute's origins and issues; the next develops Candrakīrti's approach in detail.

Following the book's usual practice, the names need a historical qualification. Svātantrika and Prāsaṅgika are labels organized by later Tibetan classifications of teachings. Buddhapālita, Bhāvaviveka, and Candrakīrti did not call themselves by these names. They identify a real Indian dispute, but the tidy division into two schools bears the mark of later, especially Tibetan, systematization, discussed in Essay 11.¹

1. Should Madhyamaka offer its own arguments?¶

The methodological question returns. Reductio (prasaṅga) takes the opponent's premise that something has intrinsic nature and follows it to absurdity, showing it untenable without advancing a Madhyamaka thesis. Is showing the opponent's contradiction sufficient? Should one proceed further, using the prevailing logic of Dignāga, with its three members of thesis, reason, and example, to state an independent inference establishing that phenomena lack intrinsic nature?

Buddhapālita says it is unnecessary; Bhāvaviveka says it is required. The disagreement begins here.

2. Buddhapālita and Bhāvaviveka¶

Biographical information about both is extremely limited. Almost nothing reliable survives about Buddhapālita; even his dates can only be placed roughly through his influence on Bhāvaviveka. His prominence in later Madhyamaka history comes chiefly from Candrakīrti's defense, which elevated his reductio method to orthodoxy. Without Candrakīrti, Buddhapālita would probably have remained a commentator rather than a representative of a distinct position. Bhāvaviveka is slightly better documented, with a few stories preserved in Tibetan sources, but these too are difficult to verify. Reliable historical details remain scarce.

Buddhapālita, around the fifth to sixth centuries, follows pure reductio in his commentary on the Mūlamadhyamakakārikā. To refute production from self, he argues that a thing arising from itself is already itself, making further production redundant and leading to an infinite regress. This absurdity follows from the claim of self-production; it is not a separate thesis he establishes. He exposes the contradiction within the opponent's presupposition and asserts nothing himself.

The state of the evidence needs explanation. Buddhapālita's commentary has no ancient Chinese translation. His words enter Chinese only through quotations by his critic. To refute him, Bhāvaviveka's Prajñāpradīpa first had to quote him and thereby preserved his words. Every statement attributed to Buddhapālita below comes through this channel. Quotation by an adversary has the additional advantage that both parties acknowledge the testimony. His refutation of production from other is preserved in one sentence: "Production by another is likewise untenable ... because it would entail everything arising everywhere." If production from other works, everything different from the effect counts as other, so anything could arise from anything. There is no thesis, reason, or example, only the opponent's proposition pushed to its absurd consequences. Another quotation is equally important: "There is action because self, other, and the many conditions depend on one another." Action is established conventionally through mutual dependence. The reductio tradition never intended to abolish everyday life. The two sides share more than they dispute.²

Bhāvaviveka, around the sixth century, objects on several grounds. First, merely exposing absurdity leaves the argument incomplete: "You now give no reasons or examples, only a thesis and a fault assigned to the other." Under Dignāga's logic, a proper refutation should give all three members, thesis, reason, and example. Second, Buddhapālita does not explain why or block possible replies. Third, pure reductio implies a contrary commitment. Saying that something is not produced from itself seems to allow production from another. Bhāvaviveka describes the danger as "reversal of what is established and what establishes it." Once reductio is read as establishing something, the opponent can reverse its thesis and reason and use them for his own side.² Better, then, to state one's own position positively.³

Bhāvaviveka therefore holds that Madhyamaka should offer independent arguments, or autonomous inferences. This has a cost. An argument accepted by both sides needs a subject and terms that both acknowledge as established. He consequently allows things to be established through their particular characteristics (svalakṣaṇa) conventionally, sufficiently to provide a common basis. Candrakīrti later concentrates his attack on this concession, which leaves things a small measure of self-establishment at the conventional level. Part II, Essay 19, Chapters 1 and 2 relocate the issue. The concession's cost concerns the supply of a common basis, not the scope of refutation. See the end of this chapter's Section 4 and 5.5, Section 3.

Bhāvaviveka has three principal works. The Prajñāpradīpa is his commentary on the Mūlamadhyamakakārikā and the main text of his criticism of Buddhapālita. The Tarkajvālā debates directly with non-Buddhists. Xuanzang translated the Karatalaratna into Chinese. The last gives a standard independent argument. Its thesis is that "with regard to true reality, internal and external phenomena ... have no intrinsic nature"; its reason, "because of their conditions"; and its similar example, "like illusory things." This is a standard three-member inference with its own thesis, reason, and example, demonstrating the positive argument he requires of Madhyamaka.

The opening two Chinese characters, "真性" (true reality), need separate attention. Bhāvaviveka explains, "Therefore the thesis is qualified by true reality," and "conditioned things are established as empty with regard to ultimate truth, not conventionally." The qualifier fixes the target as the reading in terms of real nature, leaving untouched the conditioned things, such as eyes, commonly known and accepted even by cowherds. Part II, Essay 19, Chapter 1 judges this procedure equivalent in function to the book's scope clause. Both exclude everyday and linguistic assertions from applicability. Their methods differ. The scope clause sets a boundary by type, while the qualifier marks each sentence. Chapter 6 of that essay further judges it the closest step to theorem form among the different readings of the two truths. The danger is reification. If true reality is promoted from a sentence qualifier to a place, as though an ultimate domain existed elsewhere, the rejected divided-domain reading returns. Bhāvaviveka's own usage remains a qualifier.⁴

Particular characteristic (svalakṣaṇa) must be explained in Bhāvaviveka's usage for Candrakīrti's criticism to be understood. In the epistemology of Dignāga, who lived in the fifth century, it is a thing's own individual characteristic and the immediate object of perception. It is a present, particular reality reached directly rather than through conceptual inference. Bhāvaviveka's requirement that both sides accept the subject as established implies some cognitive independence in the object, namely a particular characteristic. Candrakīrti holds that admitting such a characteristic, even with a conventional qualification, assigns the thing some inherent feature established by more than convention alone. This conflicts with emptiness as absence of intrinsic nature. What Bhāvaviveka sees as an expedient conventional characteristic, Candrakīrti sees as a small opening for intrinsic nature.

3. Candrakīrti's reply: reductio alone, establishing nothing¶

The seventh-century Candrakīrti defends Buddhapālita and criticizes Bhāvaviveka. The next chapter develops the argument; its decisive point is that Madhyamaka cannot offer independent arguments at all. Such arguments require the subject and terms to be established for both sides by valid cognition. Madhyamaka holds everything ultimately empty and recognizes nothing as established through particular characteristics, even conventionally. Bhāvaviveka's concession implicitly commits him to things establishing themselves and departs from emptiness. Madhyamaka can therefore only use reductio, borrowing premises the opponent accepts and deriving consequences the opponent cannot accept, while establishing nothing of its own.⁵

The main argument, chiefly in Candrakīrti's Prasannapadā commentary, can be arranged as follows. An independent argument requires both sides to accept something as established by valid cognition (pramāṇasiddha). For this, the subject's object needs epistemological self-establishment, some particular characteristic. Yet Madhyamaka allows no such characteristic, even conventionally. Convention is linguistic agreement, not an inherent feature of the object itself. An independent argument would thus require recognition of conventional particular characteristics, contrary to emptiness. Only reductio remains available.

The objection also carries a methodological consideration. Pure reductio need not endorse the opponent's epistemology; it only needs to derive a contradiction within it. Independent argument requires Madhyamaka itself to accept some means of cognition, exposing it to the infinite regress of how those means are established, discussed in 5.3, Section 4. Prāsaṅgika's strategy stays wholly within the opponent's framework and introduces no epistemological commitment of its own.

Refutation without establishing a position is thus carried to its limit, retaining no common basis on which an argument can stand. This takes Nāgārjuna's and Āryadeva's absence of a thesis in 5.1 and 5.2 to its furthest point.

4. What is the dispute really about?¶

The technical details concern one underlying question. Does emptiness leave any conventional basis?

On the surface, Bhāvaviveka's concern has force. Without any thesis or shared footing for argument, how can one communicate or establish anything? Candrakīrti's concern is equally serious. Admitting even a trace of particular characteristic for the sake of argument compromises emptiness. One preserves the practicability of debate; the other, the thoroughness of emptiness. Each protects something and loses something.

A deeper problem lies within their dispute, one Madhyamaka has never clearly explained. If ultimate truth analyzes and refutes intrinsic nature, what is conventional truth? Is it merely a label for phenomena continuing to function, or does it have epistemological status and content of its own? More difficult still, does ultimate analysis itself belong to conventional or ultimate truth? If conventional, its concepts, terms, and even its conclusion of emptiness hold only conventionally, close to Bhāvaviveka's demand for a shared basis. If it belongs to ultimate truth, what means are used to conduct the analysis, and what is its own basis? Neither Candrakīrti nor Bhāvaviveka invented this tension; it lies within the two-truth framework. Candrakīrti's later Prāsaṅgika system does appear more internally coherent, avoiding the conventional particular characteristics Bhāvaviveka leaves open for argument. Yet coherence means only that the internal deduction has no obvious gaps, not that the tension is dissolved. The statement that ultimate analysis reveals everything as lacking intrinsic nature must itself be made at some level. Neither Buddhapālita nor Bhāvaviveka directly explains the status of that footing.

Part II, Essay 18 settles this question. Chapter 6 cites Qingmu on the two truths: "The ultimate depends entirely on speech; speech is conventional. Therefore without relying on convention, the ultimate cannot be spoken of." Ultimate discourse begins conventionally; the linguistic usage and logic employed at the metalevel are conventional tools. This is the two-truth framework's design, not a defect. The basis is thus identified. Ultimate analysis takes place conventionally, while the constraint it reveals belongs to the ultimate level. This answer gives a place both to Bhāvaviveka's shared basis and Candrakīrti's establishment of nothing. See Part II, Essay 18, Chapters 5 and 6, and Essay 19, Chapter 7.⁶

5. Further divisions within Svātantrika¶

Svātantrika is itself diverse. Scholarship usually distinguishes two branches:

The key difference concerns convention. Bhāvaviveka's Sautrāntika Madhyamaka holds that external objects conventionally exist independently of cognition. Śāntarakṣita and other Yogācāra Mādhyamikas accept only consciousness and its representations conventionally, with no external object apart from consciousness. They then argue that consciousness itself lacks intrinsic nature.

Śāntarakṣita, around the eighth century, is the latter branch's most influential figure. His Madhyamakālaṃkāra sets out the synthesis. After his death, his disciple Kamalaśīla entered Tibet at Trisong Detsen's invitation and represented the Indian side in the famous Lhasa debate, around 792–794. The dispute with Chinese Chan monks concerned gradual practice and sudden awakening. Tibetan histories report an Indian victory, making Kamalaśīla's gradual path early Tibetan Buddhism's orthodoxy. This was a key event in its formation, discussed in Essay 11.¹

Tibetan Buddhism's initial orientation was therefore Yogācāra Svātantrika Madhyamaka, rather than Prāsaṅgika. Tsongkhapa's later ranking of Candrakīrti's Prāsaṅgika as highest explicitly reordered the earlier Svātantrika predominance.

6. The two traditions and their place¶

The Svātantrika–Prāsaṅgika distinction is outwardly a methodological dispute over independent arguments and more deeply a disagreement over whether emptiness allows a conventional footing. Svātantrika further divides into Sautrāntika and Yogācāra approaches, the latter dominating a century of Buddhism's early Tibetan transmission. The Gelug ranking of Prāsaṅgika as the highest view is Tsongkhapa's reordering of Madhyamaka history, discussed in Essay 11.

The book's recurring observation should be recorded again. The entire dispute concerns removing attachment through debate. How should it work, and can it leave no position behind? From the perspective repeatedly stated in Essay 3 and 5.3, this marks Madhyamaka's full development into a scholastic debating tradition. Even how to continue arguing without advancing any thesis becomes a specialized dispute. Compared with the silent, nonsystematic path of the original intent described in 3.5 and 3.6, this is another intellectual environment. The next chapter examines Candrakīrti's Prāsaṅgika Madhyamaka in more detail.

References¶