Chapter Two: Dharmakīrti’s Logic and Epistemology¶

Dignāga established the foundations of Buddhist pramāṇa theory: the two valid means of cognition, perception and inference; the distinction between unique particulars and universals; the theory of exclusion; and the three characteristics of a reason. Dharmakīrti (c. 600–660) inherited this framework and refined every crucial point—sometimes defending it, sometimes revising it, and sometimes expanding it considerably. He was the most influential epistemologist in the history of Indian Buddhism. His works occupied a central place in the scholastic system of late Indian Buddhism and became standard textbooks on logic and epistemology in Tibet. Section 8.1 focused on Dignāga. This chapter asks where Dharmakīrti went further, where his path led, and how his fundamental disagreement with Candrakīrti concerned the very foundation of epistemology.

I. Dharmakīrti’s Life and Works¶

Life

Reliable historical information about Dharmakīrti is likewise scarce. Tradition says that he was born into a Brahmin family in South India, studied Brahmanical treatises in his youth, converted to Buddhism, and entered Nālandā, where he studied under Dignāga’s disciple Īśvarasena before surpassing his teacher and becoming the most important developer of Dignāga’s system.

The lineage itself is noteworthy. Dharmakīrti did not study directly under Dignāga, but under his disciple; yet no one interpreted and criticized Dignāga’s works more deeply. Traditional accounts place him roughly in the time of Candrakīrti and describe a direct philosophical rivalry between them.¹

The Seven Treatises on Valid Cognition

Dharmakīrti’s central works are conventionally called the “seven treatises on valid cognition” (saptaprakaraṇa):

The four chapters of the Pramāṇavārttika discuss the establishment of valid cognition (pramāṇasiddhi), arguing that the Buddha himself is an embodiment of valid cognition; perception (pratyakṣa); inference for oneself (svārtha-anumāna); and inference for others (parārtha-anumāna). The first chapter is philosophically unusual. It reconstructs on epistemological grounds why the Buddha’s teaching should be trusted rather than appealing directly to faith as a premise.¹

II. Inheritance and Innovation: From Dignāga to Dharmakīrti¶

Dharmakīrti’s relation to Dignāga can be stated simply: he accepted the overall framework but treated each important problem more precisely and sometimes more radically.

What he accepted: the two valid means of cognition—perception and inference; the ontological distinction between unique particulars and universals; the three characteristics of valid inference; and the fundamental direction of the theory of exclusion.

What he developed or revised:

First, Dignāga’s explanation of “why a particular reason supports an argument” remained at the level of formal description through the three characteristics. Dharmakīrti asks a deeper question: under what conditions can those characteristics obtain? His answer is “essential connection” (svabhāvapratibandha): inference must be grounded in an intrinsically necessary connection among its objects. This is Dharmakīrti’s most important contribution.

Second, Dignāga’s formulation of exclusion faced Kumārila’s circularity objection. Dharmakīrti introduced “causal efficacy” (arthakriyā) to answer it.

Third, Dharmakīrti systematized perception more thoroughly, especially the relation between momentary perception and conceptual judgment and the epistemological status of yogic perception.

III. Essential Connection: The Basis of Reliable Inference¶

The Problem

Dignāga’s three characteristics describe the formal structure of a valid inference: the reason must be present in the subject, present in similar cases, and absent from dissimilar cases. But the description remains external and formal. It does not explain why the structure exists or when we are entitled to believe that it genuinely obtains.

Suppose I observe that “whenever I see fire, there is smoke nearby” and construct an inference satisfying the three characteristics. Why is it reliable? Not only because it formally satisfies them, but because smoke and fire have an intrinsically necessary relation. Smoke is produced by fire; the two do not merely happen to coincide.

Dharmakīrti answers with essential connection (svabhāvapratibandha, a connection grounded in nature). An inference is reliable if and only if the reason and the property to be proved are linked by an intrinsically necessary relation. Such necessity has only two sources:²

First, a connection of nature (svabhāva): B belongs to A’s nature, or A is a form of B. For example: “This is a tree, so it is a plant.” A tree’s nature includes being a plant. Their connection is necessary, not accidental.

Second, a causal connection (kārya-kāraṇa): B is the necessary effect of A, or A is the cause producing B. For example: “There is smoke, so there is fire.” Smoke is a necessary product of fire. Causation guarantees the connection between reason and predicate rather than mere empirical regularity.

Only inference grounded in one of these two forms of essential connection is genuinely reliable. An inference may happen to satisfy the three formal characteristics while the reason and predicate merely coincide accidentally; it is then formally correct but epistemologically unreliable.

The implication is profound. Every reliable inference requires some internal, necessary, ontological relation between reason and predicate. This presupposes that things possess some real nature (svabhāva) capable of supporting necessity. That is precisely the target of Candrakīrti’s criticism, discussed in Section IX.

IV. Two Kinds of Reliable Inference: Essential Reasons and Effect-Reasons¶

From essential connection, Dharmakīrti derives two kinds of reliable inference:³

Essential reason (svabhāva-hetu), an inference based on inclusion within a nature.

When B belongs to A’s nature, the form is:

• Thesis: this A is B—for example, this tree is a plant.
• Reason: because it is A—because it is a tree.
• Similar cases: whatever is a tree is a plant.

This inference involves no causation but rests purely on inclusion within a nature. A tree’s nature includes being a plant; once something is identified as a tree, one may infer that it is a plant without further empirical evidence.

Effect-reason (kārya-hetu), an inference based on causal connection.

When the reason is a necessary result of the property to be inferred—B is a necessary effect of A—the form is:

• Thesis: there is fire there—A.
• Reason: because smoke is there—B, the effect of A.
• Similar cases: wherever there is smoke, there is fire.

This inference depends on causal necessity. Because B is produced by A, observing B allows one reliably to infer that A is present or was present.

These are, for Dharmakīrti, the only fully reliable kinds of inference. Reasoning grounded solely in empirical regularity, analogy, authority, or anything outside the two is epistemologically incomplete.

Dignāga also mentioned a third form, the “reason from nonapprehension” (anupalabdhi-hetu), in which failure to perceive something establishes its absence—for example, “there is no smoke here, therefore there is no fire.” Dharmakīrti accepts it but treats it ultimately as a negative form of the effect-reason.²

V. Deepening the Theory of Perception¶

Dharmakīrti analyzes the internal structure of perception more precisely upon the basis of Dignāga’s four kinds.

Momentariness (Kṣaṇikatā) and Perception

Dharmakīrti explicitly maintains Buddhist momentariness: everything arises and ceases from moment to moment, and nothing persists beyond a single instant. This deeply affects the theory of perception. Genuine perception cognizes a momentary object—an instantaneous, unrepeatable, unique particular. When we believe “I see a tree,” the actual cognitive activity is a sequence of momentary perceptions whose object at each moment is a distinct unique particular. The concept “tree” is a later conceptual superimposition.

Determinate Judgment (Vyavasāya) and Perception

Dharmakīrti distinguishes two levels: pure sensation—nonconceptual perception—and the subsequent determinate judgment (vyavasāya), which is conceptual. In ordinary experience they are always mixed. We perceive and automatically judge, “This is a tree.” Epistemological analysis must nevertheless separate them. The former is genuine perception; the latter conceptual construction (kalpanā).

The practical implication is that the boundary of language lies precisely between perception and conceptual judgment. Perception cannot be directly expressed in language, since concepts can never capture the unique particular completely. Language handles conceptual judgment. The sentence “I see a tree” no longer describes pure perception but perception plus conceptual superimposition.

Kinds of Perceptual Error (Bhrānta)

Dharmakīrti systematically analyzes illusions (māyā), mirages (marīci), dreams (svapna), double vision of the moon (dvicandra-bhrānta, produced by pressure upon the eye), and other perceptual errors. Their common feature is that a sensory faculty operates outside normal conditions or that conceptual judgment exceeds what sensation actually presents. This is not merely empirical cataloguing; it precisely defines the range excluded by the requirement that perception be “nonerroneous.”

VI. Improving the Theory of Exclusion: Causal Efficacy¶

Dignāga’s theory of exclusion faced Kumārila’s circularity criticism: if “cow” means “the exclusion of non-cows,” how do we determine the range of non-cows? Doing so presupposes that we already know what a cow is.

Dharmakīrti introduces causal efficacy (arthakriyā), the capacity to produce effects, to provide exclusion with a firmer foundation.⁵

Different kinds of things exhibit systematic differences in causal efficacy. What a cow can do—produce milk, pull a plow, and so on—is not what a horse or tree can do. Such causal differences determine the objects to which we apply “cow.” The application of words is ultimately rooted in actual causal differences among things and does not require a common entity called “cowness.”

The revision means that exclusion is no longer purely negative—“cow” equals “excluding non-cows”—but obtains a positive foundation. A thing’s distinctive causal powers determine its cognitive classification. We call something “a cow” because it shares a certain profile of causal efficacy with other cows. This profile is real on the level of unique particulars, not merely conventional.

This connects with Dharmakīrti’s criterion of existence in the next section. What truly exists is precisely what has causal efficacy; linguistic categories track actual differences in that efficacy.

Does the Causal Proposal Escape Circularity? It must be acknowledged that this revision did not end the debate. Mīmāṃsā criticism reappears on a new level. To say that individual cows “share a certain causal efficacy”—all produce milk or bear loads—already presupposes a classification of effects. What makes “producing milk” one kind of effect? If the answer is a real universal “milk-producing nature,” realism about universals returns through the back door. If the answer is exclusion—“producing milk” is the exclusion of not producing milk—the circle has merely moved from words to effects. Dharmakīrti stops the regress at nonconceptual perception and disposition. Similar unique particulars cause similar perceptual judgments in a cognizer; that “production of similar judgments” is itself a causal fact requiring no further conceptual classification. The stability of classification is ultimately sustained jointly by beginningless conceptual dispositions (vāsanā) and actual causal regularities. This response admits that exclusion’s foundation is causal and psychological rather than logical. Whether it resolves the circle or redescribes it as an acceptable naturalistic fact remains disputed in Buddhist epistemology.⁵ This book’s fair assessment is that Dharmakīrti demoted a fatal logical refutation into a theoretical cost with which the theory can coexist—a substantive achievement in defense, but not a final victory.

VII. The Criterion of Existence: Capacity for Causal Efficacy (Arthakriyāsāmarthya)¶

Dharmakīrti offers an influential criterion of existence: what truly exists is what can produce an effect (arthakriyāsāmarthya, the capacity for causal efficacy).²

The criterion has several implications.

First, only unique particulars genuinely produce effects. Universals are conceptual integrations of several particulars. As products of integration, they produce no effects themselves. Concrete individual things produce effects; abstract commonalities do not. Ontologically, unique particulars genuinely “exist,” not universals.

Second, to exist is to be capable of producing effects—sat equals arthakriyāsamartha. Conversely, something incapable of producing any effect does not exist, or is merely a conceptual construction. This supplies an ontological basis for Buddhist momentariness. A persisting thing produces effects at every moment, but those effects differ from one moment to the next; otherwise nothing changes, contradicting the phenomenon of continuity. Different effects show that the “thing” producing them changes at each moment, so nothing truly persists beyond an instant. A permanent, unchanging thing cannot produce an effect, because production itself entails change.

Third, the criterion connects epistemology and ontology. The objects of perception are precisely the individual things—unique particulars—with causal efficacy in the present moment. Inference grasps relations among them. The reliability of cognition ultimately derives from the causal reality of its objects.

In a modern philosophical comparison intended only as an aid, Dharmakīrti’s criterion closely resembles pragmatist theories of truth. Something is real if it “makes a difference,” producing distinct consequences in some fashion. It also resembles one criterion of scientific realism: an entity is real if we can interact with it causally. Dharmakīrti’s motivation, however, is Buddhist. He seeks to unite momentariness with epistemology, not to develop pragmatism.

VIII. Yogic Perception: Connecting the Epistemology of Practice with Pramāṇa Theory¶

Dignāga listed yogic perception (yogi-pratyakṣa) as a kind of perception but did not explain in detail why it genuinely qualifies. Dharmakīrti offered an extensive defense, among the most controversial and religiously consequential portions of his work.

The problem: yogic perception is the direct insight into reality—such as the four truths, not-self, and emptiness—said to arise through deep meditation and absorption. But how can it count as “perception”? By Dharmakīrti’s own criterion, perception must be nonconceptual and must cognize a unique particular, an individual causally efficacious entity. Yet “suffering,” “not-self,” and other objects supposedly known in meditation appear abstract. Can they be objects of perception?

Dharmakīrti’s defense:⁴

Through prolonged meditative training, practitioners can transform what began as merely conceptual cognition—“impermanence is true”—into a vivid present experience resembling direct sensory perception. This does not make impermanence a color visible to the eye. It means that practice can disclose the truth of a proposition in a directly presented, noninferential way. A person may know intellectually that “being burned by fire is painful,” but only when actually burned does the person know it directly. Through meditation, a practitioner transforms propositional knowledge that “suffering is suffering” into a perceptual experience that “directly feels the nature of suffering.”

The philosophical problem is whether this stretches the criterion of perception—nonconceptual cognition of particulars—beyond what the concept of a unique particular can cover. Dharmakīrti recognized the tension. His position is closer to saying that yogic perception directly and with extraordinary clarity cognizes concrete instances of suffering, impermanence, and not-self rather than abstract propositions, and does so with such purity that conceptual judgment no longer intervenes.

The defense is a crucial link between the pramāṇa tradition and Buddhist soteriology. If yogic cognition does not count as perception, insight acquired through practice falls outside the epistemological system and becomes indefensible mysticism. By incorporating it into perception, Dharmakīrti attempts to grant religious cognition—insight in practice—epistemic legitimacy.

IX. The Fundamental Disagreement with Candrakīrti¶

Sections 5.5 and 6.5 presented the opposition from Candrakīrti’s side. We can now complete the picture from Dharmakīrti’s.

In Candrakīrti’s eyes, Dharmakīrti’s essential connection (svabhāvapratibandha) is a “residue of realism.” To say that things possess an internally necessary relation implies that they each have some fixed intrinsic nature (svabhāva). Otherwise, upon what could “necessary connection” rest? Candrakīrti’s Prāsaṅgika Madhyamaka accepts no intrinsic nature beyond linguistic convention and therefore no “necessary connection” in the strong sense. It reduces causation to conventionally established practice, as discussed in 5.5.

Although the two were roughly contemporaneous and the line of direct exchange is unclear, Dharmakīrti’s reply can be inferred from his position. Without a necessary relation grounded in nature, the reliability of inference cannot be guaranteed. “Smoke entails fire” is true not merely because conventional practice says so, but because smoke is in fact produced by fire. This fact is independent of convention and reflects the actual causal structure of things. If the causal structure itself is reduced to conventional designation, how can we know that any inference is reliable rather than merely customary?

Ultimately, Dharmakīrti’s epistemology requires some “causal reality of the object of cognition” as its foundation; otherwise epistemology itself becomes unstable. Candrakīrti holds that “conventions commonly accepted in the world” are sufficient and require no metaphysical ground beyond convention. The disagreement extends the Dignāga discussion in 8.1 into the clearest dispute within the whole epistemological tradition: does cognitive validity require a metaphysical foundation, or are conventional practices enough?¹

X. Historical Influence¶

In India

Dharmakīrti became a central figure in Nālandā’s scholastic tradition. His seven treatises became standard works for teaching pramāṇa theory at Nālandā and later Vikramaśīla. Important commentators followed. Dharmottara, around 750, wrote a commentary on the Nyāyabindu and explanatory works on the Pramāṇaviniścaya. Prajñākaragupta, around 800, composed the extensive Pramāṇavārttikālaṃkāra. Śākyabuddhi also commented upon the Pramāṇavārttika. Their interpretations differ in detail and became sources for later Tibetan commentarial lineages.

The Nyāya challenge to Dharmakīrti also continued for a long time. Successors of Uddyotakara, especially Maṇḍanamiśra and later Vedānta thinkers, maintained critical pressure upon the Buddhist pramāṇa tradition.

In Tibet

Dharmakīrti entered Tibetan education differently from Dignāga. Dignāga’s works were earlier, but after the eleventh century Dharmakīrti’s seven treatises—especially the Pramāṇavārttika—became central to Tibetan scholastic training and were read and commented upon far more widely than Dignāga’s own works. Two principal Tibetan commentarial traditions developed.

The first is the Sakya tradition. Its representative Sakya Paṇḍita (1182–1251) wrote the Treasury of Reasoning on Valid Cognition (Tshad ma rigs pa'i gter), a foundational Tibetan work grounded primarily in Dharmakīrti. The second is the Geluk tradition, which integrated Dharmakīrti’s pramāṇa theory with Madhyamaka philosophy into a framework of “valid cognition within conventional discourse” (tha snyad pa'i tshad ma). Pramāṇa theory treats the conditions of valid conventional cognition; Madhyamaka treats ultimate emptiness. Each performs its own task without contradiction.¹

The argument in the first chapter of the Pramāṇavārttika, “Establishing Valid Cognition,” that the Buddha is a source or embodiment of valid cognition became an epistemological defense of the path in Tibetan logic curricula: a philosophical explanation of why the Buddha’s awakening should be trusted. It directly links pure logical pramāṇa theory with the spiritual meaning of practice and is one reason Dharmakīrti’s work remained important in Tibetan Buddhism for so long.

References¶