Pramāṇa: The Sources of Valid Knowledge

Every time you believe something, a question lurks underneath: how do you know? You saw it, someone told you, you reasoned it out, but which of these really counts as knowledge, and why? This question drives the Indian philosophical tradition of pramāṇa-śāstra, the science of valid cognition. Among Buddhist thinkers, no one asked it more rigorously than Dignāga (c. 480–540 CE) and his successor Dharmakīrti (c. 600–660 CE), whose work transformed not only Buddhist philosophy but the entire field of Indian logic and epistemology.

Pramāṇa is often translated as "source of knowledge" or "means of valid cognition." Classical Indian philosophers debated how many pramāṇas there are: the Nyāya school accepted four (perception, inference, comparison, and testimony); Mīmāṃsakas added more. Dignāga was radically restrictive. In his magnum opus, Pramāṇa-samuccaya (Compendium of Valid Cognition), he argued that there are exactly two pramāṇas: pratyakṣa (direct perception) and anumāna (inference). Everything else, testimony, analogy, tradition, either reduces to one of these two or is not genuinely knowledge-producing at all.

Why so restrictive? Because Dignāga is a nominalist. He holds that reality consists entirely of unique, momentary particulars (svalakṣaṇa). Each particular is radically individual, no two are identical. Perception alone can touch these particulars directly, because perception is non-conceptual: it just receives the bare sensory event before the mind starts sorting, labeling, and comparing. The moment you form a concept, "this is a cow", you have moved away from the particular into the realm of constructed universals, which for Dignāga are mental fictions, not features of reality itself.

Inference, the second pramāṇa, operates entirely in the conceptual realm. It moves from signs to conclusions through three-part syllogisms (trairūpya). Dignāga formalized this: a reason (hetu) is valid if and only if it satisfies three conditions, it is present in the subject, it is present in similar cases, and it is absent in all dissimilar cases. This became the standard logical framework in Indian Buddhist philosophy for centuries.

Dignāga defended the validity of only two pramāṇas (instruments of knowledge), perception and inference, in his magnum opus, the Pramāṇa-samuccaya. His theory does not make a radical distinction between epistemology and the psychological processes of cognition. Dignāga's theory of knowledge is strongly grounded in perception as an epistemic modality for establishing a cognitive event as knowledge.

— Cristian Coseru, summarizing Dignāga's epistemology, as discussed in Buddhist Logico-Epistemology scholarship

The deepest puzzle Dignāga's system generates is this: if real knowledge is perception of bare particulars, and concepts are mental constructions that don't directly refer to real things, then how does language work at all? When I say "cow" and you understand me, what exactly are we both grasping? Dignāga's answer is the apoha (exclusion) theory: a word doesn't positively pick out a universal; it negatively excludes everything that is not that kind of thing. "Cow" means "not-non-cow." This is not a trivial logical trick, it reflects Dignāga's deep commitment to the view that all conceptual content is constructed by the mind through differentiation, not discovered in reality.