⁂ Concepts · Knowledge · No. 41
Pramāṇa

Rules of evidence: enumerate the instruments, then say when each one fails.

Pramāṇa

A reliable source of knowledge - and the Indian tradition's question of how many there are.

Theme
Knowledge, truth and reason
Tradition
Indian
Term
प्रमाण · pramāṇa
Term fixed
c. 200

In plain terms

European epistemology usually asks what knowledge is. Indian epistemology asks a different question first: what are the ways of getting it?

The answers are enumerated and argued over. Perception, obviously. Inference, since you can know there is fire from smoke without seeing the fire. Testimony, since almost everything anyone knows came from someone else. Comparison, analogy, presumption, and the peculiar case of knowing an absence.

What makes this productive is that the list is contestable in a way that abstract definitions are not. A school that accepts only perception has to explain how anyone knows anything general. A school that accepts testimony has to say when it fails. The argument is about a finite list, so it can actually be conducted.

As a philosopher would put it

Nyāya accepts four pramāṇas: pratyakṣa (perception), anumāna (inference), upamāna (comparison) and śabda (testimony). Sāṃkhya accepts three, dropping comparison. Buddhist epistemology after Dignāga accepts two, perception and inference, and analyses testimony as a species of inference. Mīmāṃsā accepts six, adding presumption and non-apprehension - the last being the means by which one knows that something is absent. Cārvāka accepts perception alone, which is why the school is the tradition's standard opponent.

The deeper dispute is about self-validation. Are cognitions intrinsically valid until defeated, as Mīmāṃsā holds, or do they require external confirmation, as Nyāya holds? The first makes ordinary belief innocent until proven guilty and grounds scriptural authority; the second makes every cognition answerable.

Nāgārjuna's attack in the Vigrahavyāvartanī is the sceptical challenge: if the pramāṇas establish objects, what establishes the pramāṇas? Another pramāṇa begins a regress; self-establishment abandons the principle that things need establishing. The Nyāya reply, that a lamp illuminates itself and other things, is famous and not universally accepted.

Where you meet it

Every legal system enumerates what may count as evidence, and the resemblance to the pramāṇa debate is not superficial.

English and Welsh criminal procedure recognises direct testimony, documents, real evidence, expert opinion and admissions, and excludes hearsay except under specified gateways. Each category has conditions attached: expert opinion requires a recognised body of knowledge and a qualified witness; identification evidence carries a mandatory warning to the jury about the particular ways honest witnesses get it wrong.

What the rules encode is a considered position on which sources yield reliable belief and under what conditions, arrived at through centuries of finding out where each one fails. The hearsay rule is a judgement that second-hand testimony is a weaker instrument. The Turnbull warning is a judgement that perception under stress is weaker still. Nobody drafting those rules was reading Nyāya, and the shape of the enterprise is the same: enumerate the instruments, and specify the conditions under which each one delivers.

What is actually disputed

How many pramāṇas there are, and whether the framework can support itself.

The school-by-school count is not a scholastic quibble: accepting testimony as an independent instrument is what makes scriptural authority epistemically respectable, and denying it - as the Buddhists effectively do by reducing it to inference - removes that ground. Cārvāka's single-instrument position is the reductio the others argue against, since it appears to make even the belief that fire burns tomorrow unavailable.

Nāgārjuna's regress objection is unresolved and is one of the sharpest sceptical arguments produced anywhere. The lamp analogy answers it only if self-illumination is coherent, which is exactly what is in dispute - and the parallel with the Western regress problem is close enough that neither tradition can claim to have solved something the other missed.

Commonly confused with

  • epistemology in general Indian epistemology asks first about sources and second about analysis; European epistemology has usually done the reverse. The difference in starting point produces different literatures.
  • religious authority Śabda covers reliable testimony generally, including an expert telling you the way to a village. Its extension to scripture is a further and contested step.

Defined against

  • Ignorance The instruments are what remove it, and in the Indian traditions removing it is the point of the whole exercise

Philosophers

  • Vasubandhu The Buddhist reduction of the instruments to two
  • Nagarjuna The regress objection, and the challenge to the whole framework
  • Shankara Where the framework is used to argue for a non-dual conclusion
  • Kapila The Sāṃkhya count of three
  • Ramanujacharya Where the instruments are turned against Advaita

Works

Schools

Related concepts

Knowledge c. 369 BC · Perception c. 330 BC · Inference c. 330 BC · Testimony 1764 · Justification 1641

The Nyāya Sūtras enumerate four; the schools then disagree about the number for a millennium, and the disagreement is itself the tradition's central epistemological method.