Part I, Chapter 3: Candrakīrti and the Prāsaṅgika Path to Liberation¶

The previous chapter stopped at an unfinished project. Early Madhyamaka—Nāgārjuna and Āryadeva—offered a theory of the path with a clear direction but an incomplete sequence. It pointed toward breaking the view of intrinsic existence by seeing emptiness, but did not lay out that direction as a series of stages that could be followed level by level and tested step by step. Connecting the view of emptiness to a complete set of path stages, especially integrating the ten bodhisattva grounds of the Daśabhūmika-sūtra, ground by ground, with seeing emptiness and compassion, was accomplished gradually after Nāgārjuna and Āryadeva. This chapter addresses one of the most important successors in that project, Candrakīrti (c. seventh century), and the Prāsaṅgika path to liberation that he represents.

Following the preceding two chapters, this chapter presents the position in its strongest form without yet developing a critique. It will, however, mark which of the five questions his additions answer and which unresolved problems they sharpen. There are really two questions: how did later thinkers supply the systematic stages of the path that early Madhyamaka lacked? And did the system they supplied alter the core assumption identified in Chapter 1—the epistemological optimism that “seeing emptiness is liberation”?

The Madhyamakāvatāra's dual role: commentary and systematization

Within the Madhyamaka tradition, Candrakīrti is both a commentator and a systematizer. The Prasannapadā is the most important commentary on Nāgārjuna's Mūlamadhyamakakārikā, and the main battleground of his Prāsaṅgika methodological position. The Madhyamakāvatāra takes the ten bodhisattva grounds of the Daśabhūmika-sūtra as its framework and expands the Madhyamaka view of emptiness into a sequence of path stages.¹ For him, these are two aspects of a single undertaking. His methodological criticism of Bhāvaviveka clears the ground for his own theory of the path; that theory in turn is meant to demonstrate that the Prāsaṅgika position of “establishing no thesis of one's own” remains complete and practicable. Candrakīrti thus answers both “how should Madhyamaka argue?” and “how should Madhyamaka practice?”, insisting that the two answers must be consistent.

Systematization I: the ten grounds and six perfections as stages of the path

Here is the alignment that the previous chapter left “for the next chapter.” The Madhyamakāvatāra makes the six perfections (ṣaṭ-pāramitā) the backbone of its theory of the path, pairing them with the ten bodhisattva grounds. Each of the first six grounds centers on one perfection: giving, ethical discipline, patience, effort, meditative absorption, and wisdom. The sixth ground, the Ground of Presence, centers on directly seeing emptiness through the perfection of wisdom; the seventh and higher grounds deepen this further. Candrakīrti did not originate this framework of ten grounds and six perfections; it comes from the Daśabhūmika-sūtra and the prajñāpāramitā sūtras. His contribution is to bind the practice of the perfections closely to the deepening view of emptiness at every ground, specifying at each level which perfection is practiced, which principle is seen, and which obstruction is abandoned.

In terms of the five questions, this supplies an operational sequence for the truth of the path, the mechanism. In Chapter 1, Madhyamaka's mechanism was stated only in principle: seeing emptiness breaks the view of intrinsic existence. Early Madhyamaka (Chapter 2) expanded this into a path of counteraction and accumulation, but still lacked verifiable stages. Once Candrakīrti's line took up the project, the “construction drawing” that early Madhyamaka lacked acquired a relatively complete outline for the first time. Ground, perfection practiced, principle seen, and affliction abandoned were all arranged in sequence. In systematic completeness, these path stages could finally compare with the Abhidharma theory of the path that Madhyamaka had once criticized, such as the Sarvāstivāda's five paths (see Chapter 2). The critics were several centuries late, but at last they too supplied a drawing that was no longer sparse.

Systematization II: conventional valid cognition as the foundation of practice

Systematization must supply more than a sequence. It must also answer a methodological problem left by early Madhyamaka: if Madhyamaka “establishes no thesis of its own,” on what basis should practitioners follow its direction? Candrakīrti's answer emerges from his famous disagreement with Bhāvaviveka.

Bhāvaviveka holds that Madhyamaka thinkers must use autonomous inference (svatantra-anumāna) to break an opponent's view of intrinsic nature: a syllogism that stands independently under cognitive standards accepted by both parties. Merely negating the opponent through reductio—the prasaṅga used by Nāgārjuna and Candrakīrti—does not suffice to establish right view. Candrakīrti objects that an autonomous inference can hold only if both parties “see” the object of debate in the same way. Yet adherents of intrinsic nature and Madhyamaka thinkers do not see the same thing: one sees it as really existing by its own characteristic, the other as lacking intrinsic nature. No such common ground exists. Insisting that it must exist amounts to Madhyamaka first conceding an independently real epistemic foundation, the very thing it seeks to refute. Madhyamaka can therefore use only reductio: borrow the premises the opponent already accepts, derive a contradiction, and establish no separate thesis of its own. Chapter 2 of Essay 19 in Volume 1, Part II reconstructs this step as a dilemma, adding the pivot not mentioned in this section: Bhāvaviveka's qualifier. Bhāvaviveka himself writes, “Therefore, I qualify the establishment of the thesis by true nature.” He establishes the emptiness of the conditioned with respect to ultimate truth, not conventionally. The dilemma then runs as follows. If the inference strictly applies the qualifier, the “conditioned” in its thesis is already the conditioned qualified by true nature, a term the opponent does not recognize. If the qualifier is not applied, what is established is no longer one's own thesis. In this domain, the entry ticket for a positive inference cannot itself be issued.²

The key is that reductio borrows premises that are “commonly accepted in the world” (loka-prasiddha) at the conventional level: inferring fire on seeing smoke, and everyday judgments of color, sound, smell, and taste. These function normally and validly within conventional truth without any metaphysical guarantee. Tsongkhapa later calls this “conventional valid cognition” (Chapter 1): cognition valid at the conventional level without involving intrinsic nature.

This step has great significance for the path: it provides a foundation for practice within conventional truth. Practicing the six perfections, arousing bodhicitta, and responding compassionately to sentient beings' suffering are all activities at the conventional level. Their validity is guaranteed by what is commonly accepted in the world—conventional valid cognition—and they need not wait for the direct perceptual realization of emptiness to begin. Before directly realizing emptiness, practitioners can rely on a correct understanding of conventional truth to practice the six perfections and gather the two accumulations steadily. The question “Since Madhyamaka has no thesis of its own, why should practitioners follow it?” thus receives an initial answer: they follow a direction guaranteed by conventional valid cognition at the level of conventional truth, not some Madhyamaka “thesis” about ultimate truth. As to whether this foundation has itself been proved, Chapter 2 of Essay 19 in Volume 1, Part II records a difference. A preservation clause of the same form is a proved theorem in Essay 18; for Candrakīrti, compliance with that constraint is a methodological position, with the overall success of reductio indirectly taking the place of its proof. The criticism in Part II, Chapter 5 can therefore find a more precise target. The issue is not why conventional valid cognition is valid; it is that, in Candrakīrti's hands, it never had the form of a theorem. Whether this answer truly stands is precisely for that chapter, “The Tension Between Having No Thesis and Practice,” to examine. Here we first record the Prāsaṅgika answer.

The nonduality of the two truths: conventional truth is not an illusion to transcend

Conventional valid cognition can ground practice because of Candrakīrti's judgment about the relation between the two truths: conventional truth is the field in which practice actually occurs, not an inferior illusion that must be transcended to reach ultimate truth. Ultimate truth is likewise not a higher level of reality superimposed on conventional truth. It simply describes the empty nature of the same conventional things. Viewed conventionally, the same process of practice has steps, content, and effects. Viewed ultimately, nothing within it possesses a fixed intrinsic nature. These are two readings of the same thing, not two worlds.

This judgment has an important soteriological consequence: advancing in practice does not require repeatedly switching perspectives between the truths—“now look conventionally, now turn to the ultimate.” Two descriptions of the same experience coexist throughout. The more clearly bodhisattvas see the workings of conventional truth, the more deeply they understand its empty nature. The two directions support rather than exclude each other. This follows the same logic as the Madhyamaka truth of cessation identified in Chapter 1: since nirvāṇa too lacks intrinsic nature and saṃsāra and nirvāṇa are nondual, liberation is not in some ultimate world elsewhere. It lies in seeing through this present conventional world as it is.

Compassion and wisdom working together: great compassion as the pivot

The most distinctive feature of the emotional structure of Candrakīrti's theory of the path is the pivotal place of great compassion (mahā-karuṇā). The Madhyamakāvatāra opens by naming three causes of a bodhisattva—great compassion, nondual wisdom (advaya-jñāna), and bodhicitta—and places great compassion first. This arrangement has a purpose: the Mahāyāna path begins not with disenchantment with one's own suffering, but with a response to the suffering of all sentient beings. Compassion drives the whole course of the path; it is not an emotion appended to practice.

Great compassion is central at two levels. Epistemically, it is a seeing of sentient beings' suffering as it is, with an emotional quality: seeing suffering, its cause (ignorance), and beings wandering in ignorance. This seeing unites compassion and wisdom, the two “working together,” rather than observing philosophically from a cold distance. Soteriologically, bodhicitta arising from great compassion provides the structural motive that prevents practitioners from falling midway into śrāvaka-style extinction and keeps them turning toward sentient beings. Candrakīrti states explicitly that without great compassion, bodhisattvas who have seen emptiness might enter extinction directly out of disenchantment with suffering and cease turning back toward others. Great compassion is precisely what steadies the bodhisattva path in an altruistic direction.

This returns to a point foreshadowed in Chapter 1. Some later Madhyamaka classifications rank the śrāvaka arhat's attainment as “incomplete liberation” (realizing only the selflessness of persons, not the selflessness of dharmas), placing the śrāvaka path at an initial stage of the bodhisattva path. Candrakīrti gives a different explanation of why the bodhisattva path continues: when śrāvakas stop along the way and enter cessation, what they lack is the driving force of great compassion, not sufficient realization. Bodhisattvas' further progress is explained by compassion's continuing drive. In this respect, Candrakīrti not only completes the stages of the path; he also supplies a motivational explanation of why this path “must go all the way to Buddhahood.”

Recapitulation: the system of the path completed, the core assumption untouched

Taken together, these developments show that Candrakīrti's line supplied what early Madhyamaka most lacked: a complete system of practice that could be followed and tested. The ten grounds and six perfections provide a sequence practitioners can follow; conventional valid cognition supplies practice's conventional foundation; the nonduality of the two truths resolves the problem of perspectives in “how practice and contemplation of emptiness can proceed together”; and compassion and wisdom working together provide the motivational structure that sustains the bodhisattva direction. In theoretical precision and consistency, this is a far more complete scheme than early Madhyamaka's.

One thing must nevertheless be clear: everything Candrakīrti completes concerns the roads and procedures leading to liberation. The core assumption underlying them all remains wholly untouched. As Chapter 1 identified, Madhyamaka soteriology's entire success or failure rests on one epistemic act: seeing emptiness, breaking the two kinds of grasping, and thereby removing suffering's conditions. Everything Candrakīrti does makes the path toward this act more complete, more orderly, and easier to follow. He neither touches nor intends to touch the core assumption. Far from reexamining it, systematization places it at the very center of the edifice as a premise needing no proof.

In the Chinese figure, “epistemological optimism” describes the central assumption, “seeing emptiness is liberation.” The English figure places that label above conventional valid cognition in the upper-left building; it should be read with the central pillar. Conventional valid cognition supplies the ground for practice in the conventional world. The lower-left building says that practice and contemplation of emptiness continue together without interruption. The closing claim concerns this systematization: it did not re-examine the assumption but placed it at the center as a premise needing no proof.

For this very reason, systematization does not dissolve the two unresolved problems left by early Madhyamaka; it sharpens them. First, can conventional valid cognition alone truly reconcile “having no thesis of one's own” with asking practitioners to follow a direction? That is Part II, Chapter 5's subject. Second, is the pivotal “seeing emptiness” a direct perceptual realization or a right view reached through inference, and who determines that direct realization has occurred—the verification question from Chapter 1? Once Candrakīrti arranges the path of seeing within the ten grounds while relying heavily on argument, the threshold for certification can only become higher and more specialized. Part II, Chapter 6, “The Drift in Certifying Realization,” returns to this.

This chapter has still done no more than present systematized Madhyamaka's theory of the path in its strongest form, together with the core assumption it has never touched. Judgment belongs to the chapters that follow.

References¶