Nyāya — the science of right reasoning

Nyāya is the tradition's toolkit for argument. Its subject is not what is true but how truth is established — and because every other school had to argue, every other school had to use Nyāya's apparatus, including those that rejected its conclusions.

Sixteen things, and one promise

Gautama's first aphorism is a single compound listing sixteen items, followed by the claim that knowing them truly brings the highest good. It is worth reading slowly, because the list is odd: it mixes the furniture of reality (the means of knowledge, the knowable) with the furniture of a debating hall (wrangling, cavil, quibbling, the grounds on which a debater is declared defeated).

That mixture is the school's whole character. Nyāya does not treat reasoning as a private mental activity that happens to be checkable by others. It treats it as a public procedure — a debate before a jury, with stated rules, recognised fouls and a definition of what counts as losing. Logic, for Gautama, is the codified etiquette of a conversation in which someone can be shown to be wrong.

The chain that liberation runs backwards along

The second sūtra explains how a treatise on debating can end in liberation. It names five links, each producing the next, and says that removing any link removes the one that followed it.

Read the chain in the direction of production and it runs the other way: false knowledge produces the defects (attraction, aversion, delusion); the defects produce activity; activity produces birth; birth produces suffering. Now notice where the chain is cut. Not at suffering — you cannot simply decide to stop hurting. Not at activity — restraint alone leaves the defect intact. It is cut at the far end, at mithyā-jñāna, false knowledge. That single move is why a school of logic is also a school of salvation: the disease is an epistemic one, so the cure has to be epistemic, and the study of what makes a cognition valid becomes the study of what makes a life free.

The four pramāṇas

The number four is itself a position, argued against rivals. Cārvāka allows only perception; Vaiśeṣika allows only perception and inference, folding the other two into them; Sāṅkhya allows three; Mīmāṁsā and Advaita allow six, adding presumption and non-apprehension. Every school has to say why the ones it rejects reduce to the ones it keeps — and Nyāya spends the whole first chapter doing exactly that.

Perception — four qualifiers, each doing work

Each qualifier excludes a specific rival case, and the commentators name them. Arising from sense-object contact excludes memory and inference, which have no such contact. Avyapadeśya — 'not to be designated' — marks the bare first moment of seeing, before the name 'pot' attaches; this is where the long Indian debate over nirvikalpaka (non-conceptual) and savikalpaka (conceptual) perception begins. Avyabhicāri — non-deviating — excludes the mirage and the shell mistaken for silver, which are also produced by sense contact but are false. And vyavasāyātmaka — determinate — excludes doubt: 'is that a post or a man?' is a genuine product of sense contact and is not perception, because it settles nothing.

Inference, and its three kinds

Tat-pūrvakam — 'preceded by that' — is the load-bearing phrase. No inference can float free of perception: you must at some point have perceived smoke together with fire, and you must now perceive smoke on the hill. Inference from the antecedent reasons forward from cause to effect (gathering clouds, therefore rain); from the consequent, backward from effect to cause (a swollen river upstream, therefore rain has fallen); and the third kind reasons where no such causal link is available at all — the sun is seen in different places at different times, and though nobody has ever perceived the sun moving, motion is inferred from the general observation that changed position implies motion. That third kind is what lets Nyāya reason about the imperceptible, including the self.

Testimony, and who counts as trustworthy

The definition is quietly radical. An āpta is defined by the commentators as one who has directly known the thing and reports it without motive to deceive — a qualification about the speaker's competence and honesty, not about his caste, his office or which book he carries. On Nyāya's account the Veda is authoritative because its speaker is trustworthy; it is not trustworthy because it is the Veda. Mīmāṁsā rejects this root and branch: for Jaimini the Veda has no speaker at all, and its authority is intrinsic precisely because no fallible person stands behind it. Two orthodox schools, opposite accounts of why scripture counts.

The five-membered inference

Nyāya's syllogism has five steps, and unlike the Aristotelian form it insists on an example — a demand that the general rule be shown to hold somewhere in experience.

1. Pratijñā — the thesis: 'the hill has fire'.
2. Hetu — the reason: 'because it has smoke'.
3. Udāharaṇa — the general rule with an example: 'wherever there is smoke there is fire, as in a kitchen'.
4. Upanaya — the application: 'the hill has smoke of that kind'.
5. Nigamana — the conclusion: 'therefore the hill has fire'.

Why the example matters

The requirement of a positive example, and often a negative one ('where there is no fire there is no smoke, as in a lake'), is a demand for empirical grounding built into the logic itself. An Aristotelian syllogism is valid whether or not any of its terms is ever instantiated; a Nyāya demonstration is not. The general rule must be shown holding in a case both parties have actually met — the kitchen — and, ideally, shown failing nowhere, in a case both parties agree lacks the conclusion: the lake. The technical names are sapakṣa, the class of instances where the property to be proved is known to occur, and vipakṣa, the class where it is known not to occur.

The Buddhist logicians Dignāga and Dharmakīrti compressed this into the trairūpya, the three marks a good reason must bear: it must be present in the subject, present in at least one similar case, and absent from every dissimilar case. They then argued that once you have the three marks, the five-membered form is padding — thesis and conclusion merely repeat each other, and two members suffice. Nyāya's answer was that the extra members are not logical redundancy but procedural necessity: a demonstration is delivered to a listener, and the listener needs the claim stated before it is argued and restated after. The disagreement is exactly the modern one between a proof and a presentation of a proof.

The problem underneath: how do you ever get vyāpti?

Everything rests on vyāpti — the invariable concomitance of smoke with fire. But how is a universal relation established? Not by perception, which reaches only the cases observed. Not by inference, which would presuppose another vyāpti and start a regress. This is the Indian form of the problem of induction, and it was posed with full force by the Cārvākas in order to destroy inference altogether, and by the Buddhists in order to restrict it.

Nyāya's replies are worth knowing because they are not evasions. First, repeated observation without a single counter-instance, together with tarka — a hypothetical reasoning that shows the denial of the relation leads to absurdity. Second, and more technically, Navya-Nyāya proposed sāmānya-lakṣaṇa-pratyāsatti: a special perceptual contact by which, in grasping one smoke as smoke, one grasps smoke-in-general, so that the universal is given rather than assembled. Whether that is a solution or a name for the problem is still argued; but the school saw the difficulty clearly and did not pretend it away.

Five ways a reason can fail

The list is not a schoolroom exercise. Because a debate had a formal loser, and because the nigraha-sthānas — grounds of defeat — were codified, an opponent who advanced a savyabhicāra reason could be shown to have lost by a public rule rather than by rhetorical advantage. This is a genuinely unusual achievement: a culture that agreed in advance on what would count as being refuted.

Three ways to argue, and only one of them is honest

Vādaवाद — discussion

Both sides advance a thesis and its contrary, supported by the recognised means of knowledge and by tarka, consistent with established doctrine and set out in the five members. The aim is the truth, and either party may be convinced. This is the only form the tradition regards as fully honourable, and it is the form in which the great commentaries are written.

Jalpaजल्प — wrangling

The same, but with victory as the aim; the debater may use quibbles, futile rejoinders and any ground of defeat available. Permitted, the commentators say, to a teacher defending a true doctrine against a skilled bad-faith opponent — as a farmer fences a growing crop with thorns.

Vitaṇḍāवितण्डा — cavil

Destructive criticism with no thesis of one's own: refute everything, assert nothing. Nyāya regards this as a defective form, and the standard charge against the Mādhyamika Buddhist is precisely that he is a vaitaṇḍika. Śrīharṣa, an Advaitin, later embraced the charge and turned it into a method.

Gautama even defines it, and the definition specifies the exact epistemic situation in which enquiry is licensed: you must see a property common to two things, or properties peculiar to each, and have no further mark that decides between them. Doubt in this sense is not a mood. It is a describable state of evidence.

Nyāya's God — and the sūtras that argue about Him

Nyāya is the one darśana that mounts a formal proof of Īśvara, and it is honest enough to record the objection immediately afterwards. Three consecutive sūtras stage the whole debate.

The next sūtra answers: no — for where human action is absent, no fruit arises at all, so the fruit is plainly the work of human action. And the third replies that this very fact is no objection, since human action is itself brought about by Him. Commentators divide over which of the three is Gautama's own view, and the tradition has never fully settled it. What is not in dispute is the later development: Udayana, in the eleventh century, assembled in his Nyāya-kusumāñjali a battery of arguments — from the world being an effect and so needing a maker, from the transmission of the Veda requiring a competent author, from the need for a supervisor to allot karmic results — specifically to answer the Buddhist denial. Nyāya's God is an intelligent efficient cause who arranges eternal atoms according to the accumulated deserts of souls. He does not make the material, and He does not make the moral law; He administers both.

Four theories of error — the sharpest disagreement in Indian philosophy

Everyone agrees that a man sometimes sees a shell and takes it for silver. What exactly is happening at that moment turned out to divide the schools more cleanly than any question about God, because each answer commits you to a metaphysics.

Follow the consequences. If Prabhākara is right that every cognition as such is valid, then the Veda needs no external certification and Mīmāṁsā's doctrine of intrinsic validity stands. If the Advaitin is right that something can be neither real nor unreal, then māyā is not a contradiction in terms and the world can be exactly that. And if Nyāya is right that the silver was real somewhere, then realism survives the existence of illusion — which is precisely what Nyāya wanted, since the whole school stands or falls on the world being as it appears.

Navya-Nyāya: logic without a subject matter

From Gaṅgeśa's Tattva-cintāmaṇi in the fourteenth century, the school turned its attention almost entirely to the analysis of cognition itself and built a technical language to do it. Its instruments are relational: a property is a qualifier, what it qualifies is the qualificand, and every relation is specified by a limitor — an avacchedaka — that fixes precisely under which description it holds. 'Fire on the hill' becomes a fully explicit structure naming the locus, the located, the relation, and the limiting descriptions of each. Compound expressions could be nested indefinitely, and the resulting sentences are famously unreadable to anyone not trained in them.

The strongest attack Nyāya ever faced

It came from an Advaitin. Śrīharṣa, in the twelfth century, wrote the Khaṇḍana-khaṇḍa-khādya — 'the sweetmeat of refutation' — and used Nyāya's own standards against it. His method: take each Nyāya definition, and show it is either too wide, too narrow, or circular. Define perception by sense-object contact, and you must define contact; define contact and you will use terms that themselves need defining, or you will loop back. Śrīharṣa advances no counter-theory. His conclusion is that the apparatus of definition cannot be made to work at all, and therefore cannot be used to establish the plurality of real things it was built to establish.

The Naiyāyika answer, from Gaṅgeśa onward, was not to abandon definition but to make it far more careful — which is precisely how Navya-Nyāya's technical machinery came into being. The attack produced the discipline's most rigorous phase. That is the pattern of this whole tradition: pressure from an opponent, absorbed as method.

॥ ❖ ॥

Last revised 2026-08-16. Corrections are welcome — see how this site is written.