Chapter One: Dignāga—Foundations of Buddhist Logic and Epistemology¶

Parts V, VI, and VII traced Indian Mahāyāna philosophy along the axes of ontology and practice. Madhyamaka declared everything empty; consciousness-only reconstructed epistemology on the premise of emptiness; tathāgatagarbha established a positive ground in buddha-nature from another direction. These three discourses were essentially internal developments within Indian Buddhism whose principal interlocutors were other Buddhist schools.

Beginning in the fifth and sixth centuries, another current developed within Indian Buddhism: systematic epistemology and logic, known as pramāṇa theory (pramāṇa-śāstra) and the science of reasons (hetu-vidyā). Its principal founder was Dignāga (c. 480–540). His system transformed not only the conduct of philosophical controversy within Buddhism but the direction of epistemological debate throughout Indian intellectual history.

I. The Philosophical Soil from Which Pramāṇa Theory Arose¶

Why at this time?

By the fifth century, the Indian tradition of interschool debate (vāda) had become highly developed. Vaiśeṣika and Nyāya had constructed precise theories of categories and logic. Mīmāṃsā had built an epistemology around the authority and interpretation of the Vedas. Sāṃkhya and Vedānta each possessed their own metaphysical and epistemological frameworks.

Within such an arena, a philosophical school without a systematic epistemology could scarcely maintain the rigor of its position in debate. Early Buddhist Abhidharma had already produced fine-grained analyses of mind and consciousness. Madhyamaka reductio was powerful, but Madhyamaka established no positive epistemology. Consciousness-only engaged epistemological questions, but tended toward psychological analysis rather than logic.

Before Dignāga, Buddhist logic essentially borrowed the Nyāya five-member argument: thesis, reason, example, application, and conclusion. The form contained imprecisions and had been built within a Nyāya philosophical framework. To use it was, to some extent, to accept Nyāya presuppositions.

Dignāga set out to construct a distinctively Buddhist system of epistemology and logic, independent of the Nyāya framework and capable of challenging Nyāya philosophically on its own ground.¹

The Epistemological Background in Consciousness-Only

Philosophically, Dignāga belonged to the Yogācāra tradition, as Part VI noted in the chronology in 6.3. He stood in a teacher–disciple relationship to Vasubandhu. The consciousness-only analysis of cognitive structure, especially its distinction among image aspect, perceiving aspect, and self-cognizing aspect, was refined by Dignāga into a rigorous epistemological theory. Consciousness-only asks, “What is the internal structure of consciousness?” Dignāga asks, “Under what conditions is cognition valid?” Their questions differ, but both share a realist commitment to the unique particular (svalakṣaṇa).

II. Dignāga’s Life and Works¶

Life

Dignāga, transliterated in Chinese as Chenna or translated as “Great Nāga of the Regions,” was active around the late fifth and early sixth century, approximately 480–540. Tradition says that he was born into a Brahmin family in South India, later converted to Buddhism, first studied Nyāya, and then became Vasubandhu’s student, moving toward a synthesis of consciousness-only and pramāṇa theory. What can be established is that his works demonstrate deep knowledge of Nyāya, Mīmāṃsā, and Vaiśeṣika epistemology and develop systematic criticisms of all three traditions.

Principal Works

The Pramāṇasamuccaya is Dignāga’s most important work. Its Sanskrit original is lost, and modern research reconstructs it chiefly from the Tibetan translation and Dharmakīrti’s later commentarial work, the Pramāṇavārttika.²

III. Two Valid Means of Cognition: Perception and Inference¶

Dignāga’s first major epistemological innovation was to reduce the sources of valid cognition (pramāṇa) from the four accepted by Nyāya to two: perception (pratyakṣa) and inference (anumāna).

Nyāya recognized perception, inference, comparison (upamāna), and authoritative testimony (śabda). Dignāga argued that comparison can be reduced to inference, since analogical reasoning is a form of inference; the validity of authoritative testimony ultimately depends on inference, since we judge through reasoning whether an authority is trustworthy. Thus only two independent means of valid cognition remain.¹

The Definition of Perception (Pratyakṣa)

Dignāga’s definition became classic: perception is cognition that is free from conceptual construction (kalpanāpoḍha) and nonerroneous (abhrānta).

Both conditions are indispensable:

• Free from conceptual construction: perception must be nonconceptual. It is only the direct presentation occurring when a sense faculty contacts an object, without conceptual superimposition. “Seeing a red thing” is perception; “seeing an apple” has already introduced the concept “apple” and is not pure perception. Strictly speaking, perception cannot be expressed in language.
• Nonerroneous: this excludes illusion, dreams, and other deceptive experiences.

Dignāga further distinguishes four kinds of perception:

The Definition of Inference (Anumāna)

Inference is cognition that produces new knowledge from known facts through reasoning. Dignāga’s major contribution was to impose strict conditions upon it: the “three characteristics of a reason” (trirūpa-hetu), discussed in Section VII.

IV. Unique Particulars and Universals: The Realist Basis of the Epistemology¶

Dignāga’s epistemology rests upon an explicit ontological commitment: only unique particulars (svalakṣaṇa) are real; universals (sāmānyalakṣaṇa) are conceptual constructions.

A unique particular is the individual character of each concrete thing or event: the exact combination of weight, color, and taste possessed by this apple at this moment, which can never be completely identical to another thing. Unique particulars are the objects of perception, what cognition contacts at its most immediate level.

A universal is a common characteristic abstracted from many unique particulars, such as “appleness,” “redness,” or “roundness.” A universal is not an independently existing entity, against Nyāya realism about universals. It is constructed when the mind conceptually integrates several unique particulars. Universals are the objects of inference.

The distinction has an important consequence. Language and propositions deal only with universals; universals are conceptual constructions, not immediate realities. Therefore, no linguistic proposition can directly capture the reality known through perception. Language handles abstract universals, while perception knows singular particulars. A principled gulf separates them.

This position fits deeply with the consciousness-only theory of “manifestation by consciousness.” What actually exists—the unique particular—is the immediate content of consciousness. What language and concepts handle is the universal abstracted from those contents, a construction rather than immediate reality.³

A doctrinal comparison, without implying historical influence, shows that Dignāga confronts the same “problem of universals” found in Western philosophy. Nyāya realism about universals structurally resembles Platonic realism: universals such as “cowness” and “roundness” are real, independently existing entities, and words operate by positively referring to these universal properties. Dignāga’s position approaches nominalism: only the concrete features of individual things—the unique particulars—are real; universals are concepts constructed from them and have no independent reality. The structure closely resembles Ockham’s fourteenth-century claim that only particulars are real and universals are merely names or concepts. Both confront the same question: if universals are unreal, how can a universal word such as “cow” apply to every cow?

For accuracy, Aristotle did not follow Plato in holding that universals transcend particulars as independently existing entities. His moderate realism locates universals within particulars—in re—and stands between Platonism and nominalism. In the Indian pramāṇa debate, Dignāga’s Nyāya opponents are therefore closer to Plato, Dignāga to nominalism, and no figure corresponds directly to Aristotle’s route.

Dignāga’s originality goes further. Ockham and Western nominalists generally retain the effective use of language while denying that universals are independent entities. Dignāga asks in addition how language is possible and gives a purely negative answer through the theory of exclusion (apoha): a word determines meaning by excluding what does not belong to its category, not by positive reference. This has no direct counterpart in Western nominalism and is closer to discussions of negative semantics after Frege. It is Dignāga’s distinctive response to this shared cross-cultural problem.

V. Self-Cognition: Cognition’s Apprehension of Itself¶

Dignāga classifies self-cognition (svasaṃvedana) as a kind of perception—inner perception—directly inheriting and deepening consciousness-only epistemology.

The basic claim is that when an act of cognition occurs, it not only cognizes an external object—or an object on the level of the image aspect—but simultaneously cognizes itself. Cognitive activity is “luminous,” or manifest to itself, and requires no further cognition to know it.

From the theoretical background of consciousness-only discussed in 6.3, this solves the question “How is cognition itself known?” If every cognition required a second cognition to cognize it, an infinite regress would result. Self-cognition stops the regress by holding that cognitive activity is self-illuminating.

Dignāga argues from memory. We can remember a past experience: “I perceived a tree.” The memory includes both the cognitive content of that perception—the tree—and the perception itself—the act of seeing. If the perceptual act had not somehow been known and registered, how could it later be remembered? The perception must therefore have cognized itself in some fashion when it occurred; without self-cognition, memory could not arise.²

A comparison clarifies another possible approach. Theravāda Abhidhamma explains the same phenomenon without self-cognition. Memory does not require that “perception knew itself when it occurred.” It requires only that the aggregate of perception (saññā; Sanskrit saṃjñā), a mental factor, automatically “mark” or apprehend the sign of an object at the time of perception, leaving an imprint of its characteristics. Later mental consciousness (mano-viññāṇa), taking a mental object as its object, can make the object—or its imprint—from a prior moment the content of present cognition. This is the mechanism of memory. Present cognition need not cognize itself while occurring; it need only take a past mental event as its object, made available by perception’s marking function.

Phenomenologically, this explanation has no obvious contradiction. It can explain memory without presupposing the “self-luminosity” of cognitive activity. A natural question follows: given this alternative, why did Dignāga and the entire consciousness-only–pramāṇa tradition, including Dharmakīrti, repeatedly return to self-cognition and treat it as a central epistemological question?

The reason is that the investigation concerns more than “How is memory possible?” It points toward a deeper question: what knows that cognition itself is occurring? The Theravāda Abhidhamma account—perception marks and later consciousness retrieves—says that after a cognition occurs, a subsequent cognition takes its content as an object. But an implicit question remains: when that act of retrieval or remembering occurs, how is it itself known? The specter of infinite regress has not disappeared; it has merely been postponed. More fundamentally, does cognition possess at the very moment it occurs an inward quality of “being experienced,” spontaneously present before any later recognition? This closely resembles contemporary debates in philosophy of consciousness concerning immediate phenomenal character. It is a structural problem that no tradition, whether or not it possesses a concept of self-cognition, can avoid. That is why the issue continued to develop in Indian philosophy, recurring from Dignāga through Dharmakīrti to Candrakīrti as a focal point of debate.

Candrakīrti’s criticism in 5.5 targets this argument: a knife cannot cut itself, so consciousness cannot cognize itself. The Dignāga lineage replies that “self-cognition” is not a cognitive relation between subject and object. It is the intrinsic self-manifestation of cognitive activity, the mode in which cognition exists. Section 6.5 already presented this dispute between Madhyamaka and consciousness-only.

VI. Exclusion (Apoha): How Words Refer¶

The theory of exclusion (apoha-vāda) is one of Dignāga’s most original philosophical contributions. It addresses a central question in philosophy of language: How do words refer to things?

Background to the Problem

Nyāya and Mīmāṃsā accept realism about universals. A word can refer to a class of things—“cow,” for example, can refer to every cow—because those things share a real common property, “cowness” (gotva). The word refers directly to this common property.

Dignāga rejects universals as real, since only unique particulars are real. But if universals do not exist, how can words refer? If “cow” does not refer to a universal property called “cowness,” why can it be used for every cow but not for other things?

Dignāga’s Answer

Words do not operate by positively referring to a shared property. They refer by excluding (vyāvṛtti) what does not belong to the word’s category. “Cow” does not refer to an entity called “cowness”; it determines a range of application by excluding everything that is not a cow. This is apoha, semantic determination through exclusion rather than positive denotation.⁴

A detail of cultural history deserves mention. From Pāṇini onward, Indian philosophy and linguistics overwhelmingly used the “cow” (go, gavī) as the primary noun in philosophical examples. “Cowness” (gotva) was the preferred example in Vaiśeṣika and Nyāya discussions of universals, and Dignāga likewise demonstrates exclusion through “cow” and “non-cow.” This plainly relates to the elevated status of cattle in Indian culture, where cows are sacred in Hindu traditions and among the civilization’s most important animals. The medieval European debate over universals gave an equivalent role to “horseness” (equinitas). Porphyry’s Isagoge used the horse in formulating the problem, Avicenna systematized that usage, and European scholastic philosophy continued it. Horses were medieval Europe’s most important animals in war, agriculture, and transport, mirroring the place of cattle in India. Two civilizations used their most culturally prominent animal to illustrate the same philosophical problem—a small anthropological observation at the center of a cross-cultural intellectual difficulty.

In modern terms, this closely resembles a negative semantic theory. The meaning of a word lies not in a positive entity to which it refers but in the negative set it excludes. “Cow” means “not non-cow”: the domain left after non-cows have been excluded.

The theory’s advantage is that it explains the universal application of words within a class without accepting real universals.

Its problem is circularity. If the meaning of “cow” is established by excluding non-cows, how is the concept “non-cow” established? The Mīmāṃsaka Kumārila Bhaṭṭa famously argued that determining the range of “non-cow” appears to presuppose that we already know what a cow is; otherwise, we could not know what to exclude. Dharmakīrti, in the next chapter, further revises and defends exclusion by introducing causal efficacy to explain the connection between words and things. Whether this escapes the circle or merely transfers it to the classification of effects is addressed directly in its sixth section.⁴

A doctrinal comparison, again without implying historical influence, reveals a striking parallel in Western philosophy of language: Ferdinand de Saussure’s 1916 Course in General Linguistics. Saussure’s central claim is that linguistic signs have no positive content, only differences. The character of a concept arises not from positive content but from its relation to other elements in the system. In his famous formulation, “in language there are only differences without positive terms.” This claim that meaning is purely differential almost exactly parallels apoha’s assertion that word meaning is established through exclusion rather than positive reference. Both reject direct reference to a positive universal entity, ground meaning in relation to others—difference or exclusion—and rebel against realism about universals. Their starting points differ: Dignāga begins with ontological nominalism, Saussure with linguistic structuralism. Kumārila’s circularity objection—“delineating non-cows already presupposes knowledge of cows”—arises equally for Saussure: if every sign is defined through difference, how does the differential system obtain its first anchor? This remains an unresolved problem shared by both theories.

VII. The Three Characteristics of a Reason: Reforming Logic¶

Dignāga’s other major contribution to Indian logic was the “three characteristics of a reason” (trirūpa-hetu), which imposed strict formal conditions upon valid inference.

Traditional Nyāya used the five-member argument: thesis, reason, example, application, and conclusion. Dignāga reduced it to three members and gave the “reason” three necessary conditions:

Take the classic inference “There is fire on the mountain” because “there is smoke on the mountain.”

• First, presence in the subject: smoke is a property of the mountain, providing the starting point of the inference.
• Second, presence in similar cases: wherever there is fire, as in a kitchen, smoke is also present; smoke correlates positively with fire.
• Third, absence from dissimilar cases: wherever there is no fire, as in a lake, smoke is absent; there is no counterexample in which smoke exists without fire.

Only a reason satisfying all three characteristics can support a valid inference.⁵

The system is comparable to an analysis of sufficient and necessary conditions in modern logic. The second characteristic ensures sufficiency—the reason accompanies the property to be proved—while the third ensures necessity—where the property is absent, the reason is absent. Dignāga is accordingly regarded as the first thinker in Indian logical history to formalize conditions for the validity of an argument.

Two parallels with Western logic help clarify the structure, again without implying historical influence.

The first is Aristotle’s syllogism, developed around the fourth century BCE. Aristotle likewise formalized conditions for valid inference, in the classic form: all human beings are mortal; Socrates is human; therefore Socrates is mortal. It derives a conclusion from a known universal proposition and an individual premise. But the structural difference is fundamental. Aristotle’s syllogism is purely deductive, moving downward from a universal rule to a particular case. Dignāga’s three characteristics test inductive inference, moving from observed concomitance to determine the validity of empirical relations such as smoke and fire. They address different levels of reasoning.

The second and structurally closer parallel is John Stuart Mill’s inductive logic in the 1843 System of Logic. Two of Mill’s famous Five Methods are central. Under the Method of Agreement, when a result occurs and a factor is always present, that factor is a candidate cause. Under the Method of Difference, when the result does not occur and that factor is absent, the factor is further confirmed as causal. These correspond nearly point for point to Dignāga’s second characteristic—presence of the reason in cases possessing the conclusion—and third—absence of the reason in cases lacking it. “Where there is smoke there is fire; where there is no fire there is no smoke” is a form of causal confirmation in Mill’s sense. Mill and Dignāga, separated by over two millennia without historical contact, formalized the same logical structure in strikingly parallel ways. Perhaps the double test of positive and negative correlation is the most natural method humans discover when confirming the reliability of empirical inference.

Dignāga also developed the “wheel of reasons” (hetucakra), systematically arranging all possible reasons—valid, invalid, and indeterminate—in a diagram resembling a truth table. It is among the closest Indian approaches to formal logic.

VIII. The Challenge to Non-Buddhist Epistemology¶

Dignāga’s pramāṇa theory developed from the outset in competition with Brahmanical schools.

Debate with Nyāya

Nyāya treated logic as its special domain, accepted four valid means of cognition, and grounded epistemology in realism about universals. Dignāga’s central criticisms were that universals are not independent realities and only unique particulars are real; that four means of cognition can be reduced to two; and that Nyāya reasons do not meet the genuine conditions of logical validity represented by the three characteristics.

The principal Nyāya respondent was the late fifth-century thinker Uddyotakara (c. 550–610), who wrote the Nyāyavārttika to refute Dignāga point by point. Their confrontation became one of the most celebrated interschool debates in Indian intellectual history and directly stimulated Dharmakīrti’s further developments, discussed next.

Debate with Mīmāṃsā

The Mīmāṃsaka Kumārila Bhaṭṭa (c. 600–700) was the principal critic of Dignāga’s theory of exclusion. He held that exclusion cannot explain how the range of “non-cow” is identified, since doing so appears to presuppose knowledge of what a cow is and thus becomes circular. Dharmakīrti’s revision of exclusion, introducing causal relations to explain the connection between words and things, responds to criticism of this kind.

IX. Historical Influence¶

In India

Dignāga inaugurated a pramāṇa tradition that developed continuously within Indian Buddhism for more than two centuries. Dharmakīrti (c. 600–660), his most important successor, constructed a still more precise system of epistemology and logic upon this foundation, as the next chapter explains. The lineage is conventionally called the “Dignāga–Dharmakīrti tradition” or “Buddhist pramāṇa tradition.” At Nālandā it became a major intellectual school alongside Madhyamaka and consciousness-only.

In China

Xuanzang brought Dignāga’s Nyāyamukha from India and translated it. His disciple Kuiji wrote a commentary on an introductory treatise on Buddhist logic, laying the foundations of its Chinese study. Its influence in China nevertheless remained limited. Buddhist logic was learned as a specialized technique but never deeply integrated with mainstream philosophies such as Madhyamaka or Tiantai. The central Chinese currents were tathāgatagarbha and Chan; the precise controversies of pramāṇa theory found no suitable place in that intellectual soil.

In Tibet

Dharmakīrti’s principal works, especially the Pramāṇavārttika, were systematically translated and studied and became a core subject of Tibetan monastic education. Geluk combined pramāṇa theory or logic with Madhyamaka view, treating them as work at different levels of one path. Buddhist logic addresses the conditions of valid conventional cognition; Madhyamaka addresses emptiness on the ultimate level. Through this conjunction, the Dignāga–Dharmakīrti tradition acquired the status of a “logical toolkit” within Tibetan Buddhism. Practitioners defending Madhyamaka in debate required pramāṇa theory to supply the instruments.

References¶