The catuskoti (tetralemma) is the Buddhist schema of four possible answers to any question: it is, it is not, it both is and is not, it neither is nor is not. In the Mulamadhyamakakarika, Nagarjuna (c. 150 to 250 CE) often rejects all four, to show that the thing asked about has no intrinsic nature (svabhava) for any answer to latch onto. Modern scholars read this either as two different kinds of negation (Westerhoff) or as a deliberate use of paraconsistent logic in which contradictions can be true (Garfield and Priest).
- Four corners: P, not-P, both, neither, the full menu of answers in early Buddhist debate.
- MMK 1.1 rejects all four ways a thing could arise: from itself, from another, from both, or from no cause.
- Westerhoff (2006): the denials are prasajya negations that cancel a shared assumption, so classical logic survives.
- Priest (2010, 2018): First Degree Entailment gives four truth values, and a fifth "empty" value handles rejecting all four.
- Not nihilism: MMK 24.18 ties emptiness to dependent origination, not to non-existence.
This is a post for readers who already know that Madhyamaka means "middle way" and that emptiness (sunyata) is Nagarjuna's central idea. The aim is to set out exactly what the four corners claim, why rejecting all of them looks like a logical crime, and what the main formal readings commit you to.
It treats the catuskoti as a piece of philosophy to be tested, not as a mystical slogan.
What are the four corners of the catuskoti?
For any proposition P, the catuskoti lists four positions:
- P ("it is");
- not-P ("it is not");
- both P and not-P;
- neither P nor not-P.
Early Buddhist texts use it as the full menu of answers a debater could give, and the menu was standard well before Nagarjuna. The classic case is the Aggivacchagotta Sutta (Majjhima Nikaya 72).
At Jeta's Grove in Savatthi, the wanderer Vacchagotta puts ten views to the Buddha one by one: is the cosmos eternal, is it finite, are soul and body the same, and does a Tathagata exist after death, not exist, both, or neither. To each, the Buddha says he does not hold it. Vacchagotta presses: where does a monk whose mind is released reappear? Each of the four options, the Buddha says, "doesn't apply".
Vacchagotta admits he is now confused. So the Buddha asks him about a fire burning in front of him, fed by grass and timber. When it goes out, which direction did it go, east, west, north or south? Vacchagotta sees the point at once: the question does not apply, the fire is simply "out". He ends the conversation by taking refuge as a lay follower.
Aggivacchagotta Sutta (MN 72), translated by Thanissaro Bhikkhu, Access to InsightSee the data as a table
| Number of questions (questions) | |
|---|---|
| Is the cosmos eternal? (2 options) | 2 |
| Is the cosmos finite? (2 options) | 2 |
| Are soul and body the same? (2 options) | 2 |
| Does a Tathagata exist after death? (4 options) | 4 |
Source: Aggivacchagotta Sutta, MN 72, trans. Thanissaro Bhikkhu · The Buddha sets aside all ten. Only the question about the Tathagata is posed as a full four-cornered set.
Why classical logic objects
Classical logic finds two of these corners awkward. Under the law of non-contradiction, corner 3 (both) is always false. Under the law of excluded middle, corner 4 (neither) is always false too.
By De Morgan's laws, "neither P nor not-P" is equivalent to "not-P and P", the same as corner 3. So a classical logician sees only two live options, and a rejection of all four looks like a rejection of P and of not-P together, which is itself a contradiction. That is the puzzle every modern reading has to solve.
Where does Nagarjuna use the catuskoti?
Nagarjuna uses the tetralemma in two directions in the Mulamadhyamakakarika (MMK): negatively, to deny all four positions about something, and at least once positively, to affirm all four as stages of teaching. The verse numbers below follow Stephen Batchelor's literal translation from the Tibetan (2000); Jay Garfield's 1995 translation (Oxford University Press, also from Tibetan) and Mark Siderits and Shoryu Katsura's 2013 translation from Sanskrit are the standard academic versions.
| Verse | Topic | How the four corners are used |
|---|---|---|
| MMK 1.1 | Arising | Nothing arises from itself, from another, from both, or without a cause. All four rejected. |
| MMK 18.8 | Teaching | Everything is real, not real, both, neither: offered as the Buddha's graded teaching. All four affirmed. |
| MMK 22.11 | The Tathagata | Do not say empty, not empty, both or neither; the terms are used only for conventional understanding. |
| MMK 25.17–18 | The Buddha after nirvana, and while alive | He cannot be grasped as existing, not existing, both or neither. |
| MMK 24.18 | Emptiness | Whatever is dependently related is explained as emptiness; this is the middle path. |
| MMK 27.30 | Closing homage | Homage to Gautama, who taught the dharma so that all views could be let go. |
Causation in MMK 1.1
The two uses tell you something. In 1.1 the target is causation. Later commentators matched the options to rival schools:
- from itself suits the Samkhya view that effects pre-exist in their causes;
- from another suits views on which cause and effect are distinct;
- from both and without a cause cover mixtures and pure chance.
Rejecting all four is an argument that causation cannot be explained in terms of things with svabhava. In 18.8 the four corners become a ladder for students at different stages, which no formal reading can treat as four simultaneous assertions about one fact.
The Stanford Encyclopedia of Philosophy entry on Nagarjuna, written by Jan Westerhoff and substantively revised in May 2026, makes the same point about causation: each of the four options is found wanting, and the conclusion is that cause and effect lack svabhava, translated as "intrinsic nature", "essence" or "substance".

Why does rejecting all four corners seem impossible?
If the four corners exhaust the options, denying all four looks like denying reality any shape at all, and in classical logic it produces a contradiction. Denying corner 1 gives not-P; denying corner 2 gives not-not-P, which is P; so you are left asserting P and not-P together, the very corner 3 you also denied.
Three ways out have been proposed, and they are not equivalent:
- Change the negation. The denials are not ordinary truth-functional "not". This is Westerhoff's route.
- Change the logic. Allow sentences to be both true and false, or neither, so the corners stop collapsing. This is Priest's route, with Garfield on the dialetheist side.
- Change the job. Treat the tetralemma as a therapeutic or soteriological device whose point is to stop the questioner grasping at any view, not to state a fact. Commentators from Robinson (1957) to Green (2025) lean this way in different forms.
How does Westerhoff read the negations?
Jan Westerhoff's 2006 paper "Nagarjuna's catuskoti" in the Journal of Indian Philosophy argues that the puzzle dissolves once you separate two negations that Indian grammarians already distinguished. Paryudasa (implicative) negation attaches to a term: "a non-brahmin" implies some other kind of person. Prasajya (non-implicative) negation attaches to the verb: "it is not the case that..." implies nothing about what is the case.
On Westerhoff's formulation, a paryudasa negation negates only the final member of a set of options, while a prasajya negation can negate that member and some or all of its predecessors. If Nagarjuna's rejection of all four corners is prasajya, it does not commit him to any replacement claim.
It cancels a presupposition shared by all four: that there is a thing with intrinsic nature for P to be true or false of. "The barren woman's son is tall, short, both, neither" fails in the same way because there is no son.
Christopher Rahlwes extended this in the same journal in 2022 ("Nagarjuna's Negation"), using Sanskrit and Tibetan commentaries and a grammatical study of na and a(n)- to show that the scope of each negation varies across MMK passages.
The presupposition-failure reading keeps classical logic intact; the price is that the catuskoti says nothing about reality except that one picture of it fails.
Argues that Nagarjuna's apparently contradictory use of the tetralemma becomes coherent once two Indian negations are distinguished: paryudasa, which negates a single option and implies an alternative, and prasajya, which can cancel a whole set of options without implying anything. Rejecting all four corners then denies a shared presupposition (intrinsic existence) rather than asserting a contradiction. Limit: the reading saves consistency by making the denials non-truth-functional, which critics say reads modern presupposition theory into the text.
Read the paper (repository record)How do Garfield and Priest model it with paraconsistent logic?
Graham Priest argues that the four corners make immediate sense in First Degree Entailment (FDE), a paraconsistent logic in which a sentence can be true only, false only, both, or neither. Read "true" and "false" as two independent questions, as in the grid above, and the four corners are just the four cells. Corner 3 is a truth-value glut and corner 4 a gap, and neither collapses into the other.
The fifth corner
In "The Logic of the Catuskoti" (Comparative Philosophy, 2010), Priest sets out this FDE reading for the early Buddhist schema, then notes that Nagarjuna seems to go further, allowing that none of the four may hold. To handle that, he adds a fifth status for the ineffable or empty.
In his 2018 book The Fifth Corner of Four (Oxford University Press) he develops it with FDE and plurivalent logic, a framework in which a sentence can take more than one value at once. On this reading, rejecting all four corners is not a contradiction; it is assigning the "fifth corner".
True contradictions at the limits of thought
Jay Garfield and Priest's earlier paper "Nagarjuna and the Limits of Thought" (Philosophy East and West, 2003) makes the bolder claim. They take Nagarjuna at his word when he says he asserts no thesis, and argue that he finds true contradictions at the limits of thought, especially at the boundary between conventional and ultimate truth.
The structure, they argue, is the same as paradoxes found by Western philosophers and mathematicians who also tried to describe the edge of what can be described. For a dialetheist, accepting such contradictions is a sign of rigour, not confusion.
Rejecting all four corners is either a contradiction, a cancelled assumption or a fifth value. Which one you pick decides what logic you think Nagarjuna used.
Shows that the early Buddhist four-cornered schema, which strains the laws of excluded middle and non-contradiction, has a natural semantics in First Degree Entailment, where sentences may be true, false, both or neither. Starts from the Majjhima Nikaya exchange with Vaccha about the saint after death, then turns to Nagarjuna, whose denial of all four options requires going beyond FDE. Limit: the formal model is Priest's construction; it shows the texts can be made consistent with a logic, not that Nagarjuna held that logic.
Read the paper (free)
Which reading is right, and where do scholars disagree?
No reading has won. Each one fits some verses well and strains on others, and each carries a philosophical price. The table sets out what you sign up for with each.
| Reading | Key text | Logic kept | What "reject all four" means | Main objection |
|---|---|---|---|---|
| Double refutation, middle path | Robinson 1957 | Classical | A dialectical device to block extremes | Leaves the logic informal |
| Two negations, presupposition failure | Westerhoff 2006; Rahlwes 2022 | Classical | The shared assumption of svabhava is false | Non-truth-functional negation may be anachronistic |
| Dialetheism at the limits of thought | Garfield & Priest 2003 | Paraconsistent | Some contradictions about the ultimate are true | Asks you to accept true contradictions |
| FDE plus a fifth value | Priest 2010, 2018 | FDE, plurivalent | The claim takes the "empty" or ineffable value | Adds machinery the texts never state |
| Soteriological, anti-formal | Green 2025 | Not the point | A practice for dropping all views (MMK 27.30) | Gives up on testing the arguments |
Two live debates are worth knowing:
- Plain denial (2024). A paper in Asian Philosophy by Wang and Wen-fang criticises the Garfield-Priest interpretation and argues that Nagarjuna's rejection of the four corners should be read as plain denial, not a positive assertion of a contradiction.
- Logic versus liberation (2025). Ronald Green's April 2025 study compares the non-classical reconstructions with East Asian Madhyamaka commentators such as Jizang and Zhiyi, and argues that adding a fifth truth value risks turning a practice for freeing the mind into a logic puzzle.
Which verses fit which reading
Westerhoff's account explains MMK 1.1 and 25.17 cleanly with the least machinery, because both are about whether a thing with fixed nature exists to be described.
Priest's account has more to say about MMK 22.11, where the word "empty" itself is put on the list of things not to assert. If emptiness is also empty, as MMK 24.18 implies by tying it to dependent designation, then some self-reference is unavoidable, and that is exactly where dialetheists say contradictions live.
A hermeneutic and comparative study of how Priest, Garfield and Adrian Kreutz reconstruct the tetralemma in non-classical logic, read against the MMK and East Asian commentators including Jizang, Zhiyi, Kuiji and Gomyo. Finds that logical reconstructions, especially a fifth value for ineffability, can reduce the catuskoti's soteriological aim to paradox, whereas East Asian traditions resisted adding logical dimensions. Limit: qualitative and interpretive; it does not test a formal model.
Read the paper (free)How does the catuskoti compare with Jain and Nyaya logic?
The three classical Indian traditions handle "maybe" very differently. Nyaya logic is broadly two-valued and builds proofs with a five-membered inference; Jain syadvada builds seven predications from "is", "is not" and "inexpressible", each qualified by "in some respect"; Madhyamaka uses four corners mainly to deny.
| Tradition | Basic values | Number of positions | Typical use |
|---|---|---|---|
| Nyaya | True, false | 2 | Prove a thesis by inference (see Nyaya's five-step syllogism) |
| Early Buddhist and Madhyamaka | Is, is not, both, neither | 4 (plus "none") | Map the options, then set them aside |
| Jain syadvada | Is, is not, inexpressible | 7 | Affirm each claim from a standpoint (see Jain syadvada) |
The contrast is useful. The Jain inexpressible value and Priest's fifth corner both make room for what cannot be said, but the Jain system affirms every standpoint conditionally, while Nagarjuna refuses to let any standpoint describe the ultimate. The SEP entry on logic in classical Indian philosophy lists the tetralemma and syadvada as the two forms of argument peculiar to Buddhists and Jains respectively.

Does the catuskoti mean Nagarjuna was a nihilist?
No, and the MMK says so directly. In chapter 24 an opponent objects that if everything is empty, the Four Noble Truths collapse; Nagarjuna answers (24.18 to 24.20) that only because things are empty, that is, dependently arisen, can anything change, and so the truths hold.
Emptiness denies intrinsic existence, not conventional existence. Dan Arnold's IEP entry on Madhyamaka frames this through the two truths: conventional truth is the ordinary level of persons and objects, and for Madhyamaka the set of ultimately existent things is itself empty, an emptiness that is also dependently originated.
This matters for the logic. If "empty" meant "non-existent", rejecting the four corners would be nihilism with extra steps. Because emptiness is identified with dependent origination, the denials target one specific picture, things that carry their nature on their own, and leave a world of dependent, conventional things standing.
That same move underlies the Buddhist answer to what gets reborn without a self, and it shapes how Buddhist thinkers treat time in the theory of momentariness. More essays on Indian thought and the mind are collected on Manthan.
What's new
- The Stanford Encyclopedia of Philosophy's Nagarjuna entry, by Jan Westerhoff, was substantively revised Stanford Encyclopedia of Philosophy, May 2026
- New study weighs fifth-truth-value readings of the catuskoti against East Asian Madhyamaka Smaratungga Journal, Apr 2025
- Critique of Garfield and Priest's reading of the catuskoti in the Mulamadhyamakakarika Asian Philosophy, 2024
Questions people ask
Is there a truth table for Buddhist four-fold logic?
Yes, if you use First Degree Entailment: treat "is true" and "is false" as separate yes-no questions and you get four values (true only, false only, both, neither), which match the four corners. Priest's 2010 paper sets this out. The catch is that Nagarjuna's rejection of all four needs something beyond the four-valued table, which is why Priest adds a fifth value.
If Nagarjuna rejects all four options, isn't that just scepticism?
Not quite. A sceptic suspends judgement because the evidence is unclear; Nagarjuna argues that the question presupposes something false, that the thing has intrinsic nature. The Philosophy Forum thread on this question reaches the same split modern scholarship does: some read it as a logical claim, others as a practice for dropping views.
Does the catuskoti violate the law of non-contradiction?
On a classical reading of corner 3, yes. Westerhoff avoids this by treating the denials as non-implicative negations, so nothing contradictory is asserted. Garfield and Priest accept that some contradictions at the limit of thought are true, which is dialetheism.
Which translation of the Mulamadhyamakakarika should I read?
For philosophy, Jay Garfield's The Fundamental Wisdom of the Middle Way (1995, from Tibetan, with commentary) or Siderits and Katsura's 2013 translation from Sanskrit. Stephen Batchelor's literal Tibetan-English draft for Verses from the Center (2000) is freely available online and is useful for checking verse numbers.
Sources
- Westerhoff, Stanford Encyclopedia of Philosophy: Nagarjuna (2010, revised 2026)
- Nagarjuna, Mulamadhyamakakarika, literal translation from Tibetan by Stephen Batchelor (draft of Verses from the Center) (2000)
- Aggivacchagotta Sutta (MN 72), translated by Thanissaro Bhikkhu
- Westerhoff: Nagarjuna's catuskoti, Journal of Indian Philosophy 34 (2006)
- Rahlwes: Nagarjuna's Negation, Journal of Indian Philosophy (2022)
- Priest: The Logic of the Catuskoti, Comparative Philosophy 1(2) (2010)
- Priest: The Fifth Corner of Four: An Essay on Buddhist Metaphysics and the Catuskoti, Oxford University Press (2018)
- Garfield & Priest: Nagarjuna and the Limits of Thought, Philosophy East and West 53(1) (2003)
- Green: Reinterpreting Catuskoti in Contemporary Philosophy, Smaratungga Journal 5(1) (2025)
- Wang & Wen-fang: On Garfield and Priest's interpretation of the use of the catuskoti in the Mulamadhyamakakarika, Asian Philosophy 34(3) (2024)
- Arnold, Internet Encyclopedia of Philosophy: Madhyamaka Buddhist Philosophy
- Stanford Encyclopedia of Philosophy: Logic in Classical Indian Philosophy (2011, revised 2023)
- The Philosophy Forum: The Catuskoti and Skepticism (2020)
