Where there is smoke

This is the third of six posts in which I work through the six Indian philosophical schools one at a time, writing down what I have understood. Sāṃkhya described what the world is made of; Yoga gave a method for getting out of it. Nyāya asks the question both of them took for granted: how do you know any of this? Its answers became the common toolkit of Indian thought, so much so that, as one of my teachers puts it, when the Vedāntins want to refute a point of Nyāya they borrow Nyāya's own logic to do it.

A note on what this is. I am not a scholar of Nyāya; I am a compiler. This page is my learning, set down in the order it made sense to me, from the teachers and books listed at the end. Where I add something from the Nyāya Sūtra that my teachers did not give, or report a view as one teacher's rather than the school's, I say so in place. The diagrams are mine.

The word

That by which one is led

Everyone takes nyāya to mean justice, and the first question I had was whether the school is about how to give it. The answer starts from the root. Nyāya comes from nī, to lead: नीयते अनेन इति न्यायः, that by which one is led to a conclusion. The court sense is right but downstream. A verdict is just when it rests on right knowledge, and right knowledge comes from right reasoning, argument and evidence. So the school's other names are tarka-śāstra, the science of reasoning; hetu-vidyā, the science of reasons; and ānvīkṣikī, the science of inquiry. A naiyāyika is simply someone who decides systematically. Everyone decides all day. Few do it with a method.

Any darśana has three dimensions: what there is to know, how it can be known, and how to behave while knowing it. Nyāya has all three, but its strength is the second, and the example that convinced me it matters to someone who is not a philosopher is a housewife. She has been someone's daughter, is someone's wife, will be someone's mother, is already an aunt several ways over. Among all these, who is she? The moment she asks, that is ontology. By what reasoning could she find out? That is epistemology. And can she find out without wounding the husband for whom the question does not arise, for whom "you are my wife" settles it? That is axiology: ethics is not only right and wrong but not making someone who thinks you his own feel like a stranger. Why her, of all people? Because Nyāya's most famous example is यत्र यत्र धूमः तत्र तत्र वह्निः, where there is smoke there is fire, and she sees smoke rise the moment her tempering hits the pan. From that one kitchen observation the whole school built its arguments, up to and including whether God exists.

Where it sits

The road and the map

Start with the pairing that the tradition takes for granted. Gautama tried to see the truth and realised that you cannot search for anything without first knowing the road. So Nyāya found the roads: the pramāṇas, the means of valid knowledge. Then, as I was taught it, the companions of Nyāya took those roads and started asking what the world is made of, and Vaiśeṣika joined. One school supplies the method, the other the inventory, and together they stand as a single darśana, a realism: as things are in the world, we accept them and proceed. The next post in this series is Vaiśeṣika's inventory; this one is the road.

The founder is Gautama, also called Akṣapāda, "eyes in the feet." There are two folk versions of the name: that he reasoned so well he could walk with his eyes shut and his feet knew the road, or that he was so lost in reasoning about the world and God that he did not know where he walked and fitted a second pair of eyes to watch the ground. Two traditions place him in Mithilā, as a native or as a southerner who made it his home; there is a Gautam Kund in Darbhanga today, and Mithila University once had a famous Nyāya department. The text is the Nyāya Sūtra: five chapters, ten sections, 528 aphorisms. On it, as in mathematics, a chain of commentary: Vātsyāyana's bhāṣya, then commentaries on that, then commentaries on those. One of my teachers claims that no Indian school produced as much literature, and that "system" is the wrong word for something that never closed. Why, then, has no one heard of any of it? Patañjali has Baba Ramdev, Śaṅkara has his hymns, Dayananda has the Arya Samaj, and Nyāya, being purely theoretical, never had an institution.

Step 1

The four pramāṇas

Pramā is valid knowledge; a pramāṇa is a means of getting it; a prameya is something to be known. Nyāya admits four means; here is each, with the example I learned it by.

Pratyakṣa, perception, is knowledge from the contact of a sense with its object, and it comes first. But the example I remember it by is about its failure. We saw a black shape at dusk and told the children a bear was standing there, do not go near. The children were naughty and went to hit the bear with a stick. It was a bush shaped like a bear. So perception alone does not deliver knowledge of what exists; after the eyes, apply the intellect. The same point, pushed further, becomes the difference between Gautama and Plato. Plato saw that rail tracks seem to meet and distant things blur, and threw perception out in favour of reason alone. Gautama keeps it and corrects it. Repair the heart, as I heard it put, rather than fitting a pacemaker.

Anumāna, inference, is deriving what you cannot see from what you can: the smoke in the hills, the fire behind them, because where there is no fire there is no smoke. It gets its own step below. Upamāna, comparison, is the textbook gavaya: you know a cow; here is an animal whose hooves and horns are not quite a cow's; compare feature by feature and decide whether it is a kind of cow or something else. The homegrown version is the nilgai, which matches the cow in most respects and was accepted as one. Śabda, testimony, is the word of someone authorised to speak, who has seen the truth and can describe it. You cannot see the earth move, but you read Newton. You may never have been to Africa, but a man who lived there many years is proof enough. The condition is the one that step 7 turns on: a source answerable for what it says.

Step 2

Inference has a shape: five limbs on the hill

The standard form is best staged as a scene: smoke is coming from a hill and two people are standing there. One says, there is fire on that hill. Why? Wherever there is smoke there is fire. How do you know? I have seen it in the hearth, smoke never comes without fire. And here is smoke. So the hill is on fire. Five sentences, and each is a limb, an avayava, of the inference.

The rule in the middle, where there is smoke there is fire, is vyāpti, invariable concomitance: the unbroken relation of sign to thing signified. The hill is the pakṣa, the subject of the claim; the fire is the sādhya, what is to be proved; the smoke is the hetu, the reason; and the hearth is the dṛṣṭānta, the example where sign and thing were seen together. Why do the extra two limbs matter? The reading I found most interesting runs like this. Aristotle's form is pure deduction: correct premises give a correct conclusion, and a wrong assumption gives a wrong conclusion however good the logic. The case I was given: two children with their heads bowed were asked whether the ceiling fan was running. One said no, there is no sound. The other said no, there is no breeze. Look up; it is running. The logic was fine; the assumptions that every fan makes noise and every fan gives wind were not. Dalton's atoms and the electrons-as-watermelon-seeds model were the same shape of error. Gautama's form forces the arguer to produce a concrete case from the world where the rule has already held, which is induction, and then to apply it, which is deduction. The image that fixed it for me is Batman and Spider-Man in one: the planner who deduces from a premise and the reactor whose tingle is pure observation. One of my teachers, a mathematician who came to Nyāya late, has made the same thesis the argument of his book: the West taught that induction is the opposite of deduction, and Nyāya says they are two faces of one coin.

There is a rule about the example itself that I want to keep. It has to be simple enough for an unlettered person to follow, what corporate trainers call monkey-proof. The old verse on the student's virtues, a crow's effort, a heron's concentration, a dog's sleep, is the model: a child can imitate an example, where an abstract rule can be read many ways. It is why a ten-thousand-page chargesheet looks to me like the Sophists' method rather than Gautama's.

Step 3

The map of the book: four, twelve, sixteen

The first sūtra of the Nyāya Sūtra is a list: sixteen padārthas, topics, through the true knowledge of which one attains niḥśreyasa, the highest good. The formula I was given for remembering how the list is built: four are the means of knowing. Twelve are the things to be known, the prameyas. Sixteen are the topics, knowing which nothing more need be known. The twelve prameyas start with the self and end with liberation, and they are the Nyāya statement of what the first two posts in this series were about. The remaining topics are about argument: doubt, purpose, example, settled position, the limbs of inference, hypothetical reasoning, decision, and then the whole machinery of debate.

From the text. The sixteen topics are Nyāya Sūtra 1.1.1; the twelve objects of knowledge (self, body, senses, their objects, intellect, mind, activity, defects, rebirth, fruit, suffering, release) are 1.1.9. I was given the count of twelve but not the list, and the recitation I learned dropped two of the sixteen, so the wheel follows the text.

Step 4

Debate has a protocol

The classical form first. To challenge another position you first state it, the pūrva-pakṣa, well enough to prove you understand it. Then you refute it, khaṇḍana. Then you state and prove your own position, the uttara-pakṣa. The conclusion is the siddhānta. Then the social frame: two people debate, there is no third, and the two of them decide who has reached right knowledge. In the famous debate between Śaṅkara and Maṇḍana Miśra both agreed the aim was truth and not victory, and whoever reached it, the other would follow. The loser said: you are a greater knower than me. There was no defeat.

The other story I was told is harder. In Mithilā, at Janaka's court, Aṣṭāvakra's father lost a debate to a sage, and losing then meant walking into the river. The boy grew up, studied, went to the same court bent in eight places, and won. When his opponent announced that he would now drown himself, Aṣṭāvakra said no: in debate one reaches the truth, one does not take a life. The engineer's gloss I was given on why Gautama wrote so many rules: rules are boundary conditions, as in thermodynamics. They limit the space so that a discussion can reach a nirṇaya, a decision, in finite time. Open-ended talk never concludes, and the root of the word is reaching a conclusion.

Step 5

Three kinds of talk

A logical person never shouts. The force of the words does the work, until the other side says: he is right. There is a borrowed phrase for what India has become instead, the argumentative Indian, and we forgot what the argument was for and how it was to be done. The house rule I took from this for public argument: it needs both tathya, facts, and tarka, reasoning. Debate whether the petrol rise is justified. Do not raise a vitaṇḍā about it.

Step 6

How a losing opponent cheats

This is the demonstration everything above builds to, and I learned it on a live example: the proposition that war is bad for the world, with the Israel and Iran exchanges and the closing of the Strait of Hormuz as the evidence. Set the proposition, the pratijñā; name which means of knowledge support it; note the saṃśaya, the doubt, since the arms-industry shareholder watching his stock rise holds the opposite; give examples; and lay out the five limbs. Then play the opponent, who cannot answer the argument and does not want to concede.

Each stage has a name. When the honest answer is unavailable the opponent wrangles, jalpa. When that fails he raises vitaṇḍā, illustrated for me with the Hindi sentence whose meaning flips on a comma: "roko mat, jāne do", do not stop him, let him go, re-punctuated as "roko, mat jāne do", stop him, do not let him go. When that fails he attacks the premises: you never stated the minor premise, how did you conclude? Then he cries fallacy, chala: the whole thesis rests on a trick. Then comes the move Gautama called jāti, a rejoinder based on mere resemblance, and here is its strongest form, as I was given it: every nation on earth came out of a war; Bangladesh exists because of one; no war, no Bangladesh. It sounds half right, which is what makes it a jāti and not an argument. The reply is to make him prove it to the end: who died? Fewer Pakistanis, more Bangladeshis, all of them people. And when every move is spent, the opponent is at a nigrahasthāna, a point of defeat, and the conclusion, the nigamana, stands.

The single example of the same thing that stayed with me is a lawyer. His client was caught adding water to milk. Milk, the lawyer argued, is seventy percent water, so his client added milk to water, for which no punishment is written. Gautama would call it kutarka, bad reasoning, and vitaṇḍā; the teacher I had the story from thinks today's lawyers carry the Sophists' DNA.

Terms from the text. In the Nyāya Sūtra's own scheme the comma trick is a chala, specifically vāk-chala, a quibble on words; vitaṇḍā is defined as wrangling with no counter-thesis, which is how one of my teachers glosses it, while the other files the comma trick under it. The five hetvābhāsas, fallacious reasons, are the erratic, the contradictory, the counterbalanced, the unproved and the mistimed. None of my teachers lists the sub-kinds; the counts above are from the sūtra.

Step 7

The same four tools filter the feed

The real question, for me, is about phones. Everyone has one, narrative forces are strong, and educated people hand over their OTPs. Does Nyāya give a basic framework for whether a piece of information is true? One of my teachers says that people like him predicted at the start of the internet that its central problem would be telling the authentic from the fake, and that Gautama's answer was always the same: judge by the means of knowledge, not by hearsay. His claim is that with the four pramāṇas you can sort fake news from real in two minutes, and he runs them on a recent example: a party with an invented name that Western media reported had been formed in India, quoted by the BBC, spread on X and Instagram, so that an Indian in America would tell you it existed. A political party's first act is registration with the Election Commission. No registration, no party.

Perception here is not only the naked eye but testing a viral item against other evidence and finding its origin: who said it, when, why. A razor advertisement with a beautiful woman in it has a cause, and the cause is money, not the razor. Inference is vyāpti applied to motive: when one party alleged during an election that the former chief minister's wife held several foreign passports, the public did not buy it, because the timing told them why it was being said. Comparison is checking against institutions with a record: when two countries at war both lose aircraft, neither prints it, and a third country's press does. And testimony is the narrow sense from step 1, a source answerable for its statements, which is exactly what a forwarded message is not.

The same five limbs run a life. Should I stay in this job? Proposition: a clear statement of the problem. Reason: my mental health is suffering, or the package is better elsewhere. Example: when I, or my friend, was in a high-pressure place, I could not use my capacity. Application: this new workplace is just as pressured. Conclusion: it is not suitable. One of my teachers, choosing between a flat in Delhi and one in Gurgaon, wrote the advantages on the left and the disadvantages on the right, which is a debate with yourself. It does not guarantee the right decision, but it becomes a habit, and a person is a puppet of his habits. Another of my sources mapped his own exit from a corporate drinking culture onto the five: quit; to do something larger; Ronaldo gave up alcohol; I stopped for fifteen, twenty, thirty days and saw the change; the decision.

Step 8

Karma, body, sorrow, release

For this purpose I was taught Nyāya and Vaiśeṣika as one school, and the account is physical. Karma binds, and both kinds bind: do good and you must return to enjoy its fruit, do bad and you return to suffer. The thing that experiences pleasure and pain is the body; the self experiences them only through contact with one. So liberation is defined as आत्मा का शरीर के साथ आत्यन्तिक वियोग, the absolute separation of the self from the body, the state in which it never has to touch one again. Death is only the temporary version. The five elements that made the body are not destroyed at death; they reorganise, according to the account of your karma, around the same self: as a dog, as a pig, as a rat, or, in my teacher's phrase, as a Birla or a Tata. The tradition counts 84 lakh kinds of birth. When the fruit of all karma is exhausted and no new binding karma is earned, the relation breaks forever. Nyāya called that state mukti and Vaiśeṣika mokṣa, and the two words became synonyms.

The obvious objection came early: a self that neither suffers nor enjoys is a stone. The Nyāya reply, as I was taught it, is that the self is not the enjoyer; call it a stone if you like, its state cannot be described, but this much can be said: it is free of all sorrow, and the body was the cause of the sorrow. Nyāya also admitted liberation in this life, jīvan-mukti, as well as after it: a man whose span is fixed and who has freed himself lives out his years with neither pleasure nor pain, something comes, something does not. And the practical takeaway I was given is about karma rather than metaphysics. Since you may have to come back as a dog, control your actions, and since Arjuna's complaint to Kṛṣṇa was that he did not know which action was karma, which akarma and which vikarma, knowledge comes first. The other version I have is the first sūtra: liberation through true knowledge of the sixteen topics.

Step 9

God at the edge of logic

The question anyone would ask: sannyāsīs are Vedāntins, yogis do japa, no one calls himself a yogi of Nyāya. Does the school of logic keep God out? The answer I have from the mathematician among my teachers is in two halves, and he does not reconcile them. Ask, as a mathematics student would, whether the God ordinary people mean is in Gautama's 528 sūtras, and the answer is no; the word there is niḥśreyasa, the highest good. And yet India's logic was for liberation, and liberation is where God is, so the land of logic became the land of God. The bridge between the halves is Udayana, the tenth-century Naiyāyika from Mithilā who wrote the Nyāyakusumāñjali, a whole book proving God's existence by argument against the Buddhists.

The story I was told about him, as tradition tells it: Udayana set out with students for Puri to see Jagannātha, walked without thinking of food, and reached the temple at midnight. The priest told him God was sleeping; come back in the morning. So Udayana addressed the closed door. "You who are so intoxicated with yourself: do you not know that when the Buddhists say there is no God on this earth, it is I, Udayana, who establish your existence with my arguments? And you will not give me darśana?" The doors, it is said, opened by themselves. In Udayana's village today, I am told, people pour a pot of water on the Śiva-liṅga and half of it on Udayana.

The argument itself. My addition, from the books: my teachers name the work and the opponents but not the proof. Udayana's central inference is Nyāya's own five-limbed form applied to the world: the world is an effect, like a pot; effects have intelligent makers; therefore the world has one. The Nyāya Sūtra is nearly silent on Īśvara, as I was taught; the theistic turn is the work of later Naiyāyikas, Udayana above all.

Another of my teachers circles the same edge from the side of the self. His case that the self cannot be described, only shown, is built from Western philosophers rather than from Nyāya, and he says so: Russell on the logically simple that can be known by acquaintance but not by description; Descartes, who could doubt the tree and its greenness but not the doubter; Kant's unity that collects thirty years of sight and sound into one experience and is never itself experienced. The eye cannot see itself. The only judgment the self admits is "I am"; the moment you say "I know myself", the known half is an object and the knower has slipped behind it again. And on whether reason can carry you the whole way, he tells a story he attributes to Russell. A man stood on logic: until logic satisfies me, I will not move. He raised one foot and asked for proof that the ground ahead would hold. A geologist tested it. Who is this man, what are his qualifications? An engineer vouched for the geologist. And so on. Tarka is never satisfied, and the man dies with his foot in the air. Put the second foot down on faith; if you fall, we will see. The other version I have is tighter and, I think, closer to the school's own view: tarka assists but does not conclude; things are proved by pramāṇa. That is why courts ask for witnesses, and why two lawyers with good logic are not allowed to settle a case between themselves.

Step 10

The living tradition, and the machine question

Nyāya did not stop with Gautama. The chain, as it was sketched for me: Vātsyāyana's commentary, Jayanta Bhaṭṭa's Nyāyamañjarī in Kashmir, Vācaspati Miśra, Udayana, and then Navya-Nyāya, the "new logic" with its own library. The door recommended to me for a reader today is the Tarkasaṃgraha, a late handbook that teaches Nyāya and Vaiśeṣika together. One of my teachers also notes, with some bitterness, that a sixty-page paper on a single Nyāya argument earns a doctorate at a university in Hawaii while Indian syllabi ignore the subject.

The machine question was put to me as a challenge: in logic, modelling and computing the West is far ahead, so why write about Nyāya at all? The answer I have is a report rather than a result. One of my teachers was told by a senior from IIT Bombay, from a family of grammarians, that he had spent seven years on Navya-Nyāya for the sake of AI modelling, because two or three times in ten the current systems answer in a way that makes you ask why they did not understand, and an alternative model would need a different account of knowledge. That teacher's own claims for what Nyāya has that a Western-derived model lacks are four, and I give them as his, not mine: an inference that fuses induction and deduction in one form instead of treating them as opposites; a logic that keeps perception and repairs it instead of distrusting it; testimony, śabda, admitted as a means of knowledge where the West calls an appeal to a great person's word a fallacy, though journals and citations later became its own śabda; and the pramātā, the knower, kept inside the method, so that refining the person who reasons is part of reasoning well. He adds that even moving from a binary to a ternary would be an upgrade, and that India's systems are ternary by habit: vāta, pitta, kapha; sattva, rajas, tamas. The strongest opinion I heard goes further than his: that as machines grow people will ask whether they can be conscious, and that question will bring them back to the Upaniṣads and to Nyāya.

Where my sources stop. None of the AI claims is worked out beyond the sentences above, and no system, paper or model is named. The dates in this step (Vātsyāyana around the fourth to fifth century, Jayanta in ninth-century Kashmir, Udayana in tenth-century Mithilā, Navya-Nyāya from Gaṅgeśa's Tattvacintāmaṇi around the fourteenth century, the Tarkasaṃgraha of Annambhaṭṭa in the seventeenth) are my addition from the standard histories, not from my teachers.

In one piece

The argument, start to finish

1. You cannot search for truth without first knowing the road; Nyāya is the school of the road.nyāya, from nī: that by which one is led
2. There are four roads: perception, inference, comparison and the word of someone answerable for it.the pramāṇas
3. Perception is kept and corrected, never thrown out: the bear was a bush, so apply the intellect after the eyes.repair, do not replace
4. Inference has a fixed anatomy: a subject, a sign, a thing to prove, a rule, and a case where the rule already held.pakṣa, hetu, sādhya, vyāpti, dṛṣṭānta
5. Stated in full it has five limbs, two more than the syllogism, and the extra two are where argument touches experience.udāharaṇa and upanaya
6. Four means, twelve objects, sixteen topics; true knowledge of the sixteen is the highest good.the first sūtra
7. Two people argue, no third decides, and the one shown wrong follows the one shown right; the aim is truth, not victory.vāda
8. Argument aimed at victory wrangles; argument with nothing to defend only tears down.jalpa, vitaṇḍā
9. A losing side has a finite repertoire, quibbles, premise attacks, cries of fallacy, rejoinders from mere likeness, and each move has a name, down to the point of defeat.chala, jāti, nigrahasthāna
10. The same four roads sort true from false in whatever arrives on a screen, and the five limbs settle a job, a flat, a habit.the feed filter · the life decision
11. All of it is for the end of suffering: the self freed forever from the bodies karma keeps building round it.ātyantika viyoga · apavarga
12. At the edge, logic hands over: tarka assists, pramāṇa proves; God is argued for by a later master; the knower is never the known.Udayana · "I am"

Terms

Glossary

The school

न्याय nyāya

From nī, to lead: that by which one is led to a conclusion. Right reasoning; justice as a derived sense.

आन्वीक्षिकी ānvīkṣikī

The science of inquiry; an old name for the school.

तर्कशास्त्र · हेतुविद्या tarka-śāstra · hetu-vidyā

Science of reasoning; science of reasons. Two more names.

नैयायिक naiyāyika

A follower of Nyāya; "one who decides systematically."

अक्षपाद Akṣapāda

"Eyes in the feet": Gautama's other name.

न्यायसूत्र Nyāya Sūtra

Gautama's text: five chapters, ten sections, 528 aphorisms, with Vātsyāyana's commentary and a chain after it.

समानतन्त्र samāna-tantra

"Of one system": the traditional label for the Nyāya and Vaiśeṣika pair.

नव्यन्याय Navya-Nyāya

The "new logic" from Gaṅgeśa onward, with its own technical language.

Knowing

प्रमा · प्रमाण · प्रमेय pramā · pramāṇa · prameya

Valid knowledge; its means; its object.

प्रमाता pramātā

The knower, who must be refined for knowledge to be right.

प्रत्यक्ष pratyakṣa

Perception: knowledge from the contact of a sense with its object.

अनुमान anumāna

Inference: from the seen sign to the unseen thing.

उपमान upamāna

Comparison: knowing a new thing by its likeness to a known one. The gavaya; the nilgai.

शब्द śabda

Testimony: the word of a trustworthy, answerable source.

व्याप्ति vyāpti

Invariable concomitance: wherever smoke, fire.

पक्ष · साध्य · हेतु pakṣa · sādhya · hetu

The subject of an inference (the hill); what is to be proved (fire); the reason (smoke).

दृष्टान्त · उदाहरण dṛṣṭānta · udāharaṇa

The example where rule and case were seen together: the hearth.

अवयव avayava

A limb of the five-membered inference: pratijñā, hetu, udāharaṇa, upanaya, nigamana.

तर्क tarka

Reasoning, especially hypothetical; it assists proof but does not itself prove.

पदार्थ padārtha

Topic or category; sixteen in the first sūtra.

निःश्रेयस niḥśreyasa

The highest good; liberation, attained through true knowledge of the sixteen.

Arguing

पूर्वपक्ष · उत्तरपक्ष pūrva-pakṣa · uttara-pakṣa

The opponent's position, stated first; one's own, stated and proved after the refutation.

खण्डन · सिद्धान्त khaṇḍana · siddhānta

Refutation; the established conclusion.

संशय saṃśaya

Doubt, when thesis meets antithesis.

वाद · जल्प · वितण्डा vāda · jalpa · vitaṇḍā

Discussion for truth; wrangling to win; destructive criticism with no position of one's own.

हेत्वाभास hetvābhāsa

A reason that only looks like one; five kinds in the sūtra.

छल chala

Quibble; three kinds, the first on words, as in the comma trick.

जाति jāti

A futile rejoinder based on mere similarity or dissimilarity; twenty-four kinds.

निग्रहस्थान nigrahasthāna

A point of defeat; twenty-two kinds.

निर्णय nirṇaya

Decision; what the rules exist to make possible.

कुतर्क kutarka

Bad reasoning; the milk-in-water lawyer.

The end

आत्यन्तिक वियोग ātyantika viyoga

Absolute separation of the self from the body: Nyāya-Vaiśeṣika's liberation.

मुक्ति · मोक्ष · अपवर्ग mukti · mokṣa · apavarga

Release; the first two used as synonyms by the paired schools, the third the sūtra's own word.

जीवन्मुक्ति jīvan-mukti

Release while still alive.

कर्म · अकर्म · विकर्म karma · akarma · vikarma

Action, inaction, wrong action: Arjuna's question, and why knowledge comes first.

न्यायकुसुमाञ्जलि Nyāyakusumāñjali

Udayana's proof of God's existence by argument.

तर्कसंग्रह Tarkasaṃgraha

Annambhaṭṭa's handbook of Nyāya and Vaiśeṣika together; the recommended first book.