The Milindapanha: The Questions of King Milinda

By Tim Yu, Founder & Editor · Published 27 August 2026 · Updated 8 September 2026

Checked against the primary texts · How we check · Report a correction

The Milindapanha (“Questions of Milinda”) is a revered Pali Buddhist text: a dialogue between the Indo-Greek king Menander (Pali: Milinda) and the learned monk Nagasena. Famous above all for the chariot analogy for not-self, it is a landmark of Buddhist philosophical argument — and a rare literary meeting of Greek and Buddhist worlds.

It holds an unusual, in-between place in Buddhist scripture: canonical in Burma, where it is counted among the books of the Khuddaka Nikaya, and para-canonical — revered, but outside the canon proper — in Sri Lanka and Thailand. This page gives you the dialogue itself.

What happens in the Milindapanha?

The Milindapanha stages one of the great conversations in Buddhist literature. On one side sits King Milinda — the historical Greek king Menander, who ruled in the northwest of India in the 2nd century BCE — a sharp and sceptical questioner. On the other sits Nagasena, a monk of profound learning. The king presses hard questions and apparent contradictions about the Buddha’s teaching; Nagasena answers each with reasoning and, above all, with memorable similes. The result is both a clear introduction to Buddhist thought and a genuine philosophical dialogue in the classical manner. Its most celebrated moment — the analysis of a chariot to demonstrate that there is no fixed “self” — has become one of the most-taught illustrations in all of Buddhism.

The king and the monk

The Milindapanha is unusual for being anchored to a real historical figure. King Milinda is Menander I, the most famous of the Indo-Greek kings who ruled the Bactrian and northwest-Indian regions in the centuries after Alexander the Great’s conquests. He reigned in the 2nd century BCE, and he is the only one of those Greek kings whom Indian tradition remembers by name — a sign of the impression he left.

The text presents him as a brilliant, restless mind who has bested every teacher he has questioned, until he meets Nagasena, who is more than his match. That such a book exists at all is a small marvel of history: a record, however stylised, of the meeting of Greek and Buddhist civilisations in the wake of Alexander — one of the most fascinating cultural contacts of the ancient world, part of the long story of how Buddhism spread.

The chariot: the argument for not-self

The Milindapanha’s most famous passage stands at the head of the dialogue proper (Mil 3.1.1). Asked his name, Nagasena answers that “Nāgasena” is “only a generally understood term, a designation in common use”, because “there is no permanent individuality (no soul) involved in the matter”. The king pounces: if no one is there, who gives the robes and the food and who enjoys them, who “lives a life of righteousness”, and if a man were to kill you, where is the murder? He then walks Nagasena through the thirty-two parts of the body and the five aggregates, is told no to each in turn, and springs the trap.

Two things are worth noticing before the famous reply. The first is that Milinda cross-examines the monk using Buddhism’s own inventories — the thirty-two parts of the body, then the five aggregates — so the trap is built entirely out of Nagasena’s materials. (Rhys Davids writes the aggregates as “Skandhas”, the Sanskrit form of the Pali khandhā.) The second is the register: a falsehood, an untruth. That is a courtroom charge, not a philosopher’s objection. Milinda is still arguing as a king.

Nagasena does not defend the claim. He asks how the king arrived, is told “I came in a carriage”, and puts the carriage through the identical cross-examination — pole, axle, wheels, framework, ropes, yoke, spokes, goad — before handing the king back his own sentence.

The simile only works on a made thing whose parts are separable and separately named, which an ancient chariot, itemised piece by piece like that, is. No single part is a smaller chariot. That is the assumption doing the work, and it is worth naming, because it is also the assumption a critic would attack. Notice, too, where this is happening. It is not a court scene: the king has travelled to the monk. The introductory narrative has Milinda, “attended by the five hundred Yonakas”, mount “his royal chariot” and proceed to the Saṅkheyya hermitage, where Nagasena was sitting with the brethren “in the open hall in front of the hermitage”. So the vehicle the king arrived in is the one turned against him — and it is those same Yonakas, his own Greek retinue, whom Nagasena calls to witness, and the five hundred of them “shouted their applause”. The argument is won in public, on the monk’s ground, in front of the king’s own men.

The king then answers his own question, so that the demonstration is his and not the monk’s.

“The generally understood term, the designation in common use” is Rhys Davids compressing a Pali pile-up — saṅkhā samaññā paññatti vohāro nāmamattaṁ — five near-synonyms for naming, which John Kelly’s 2023 translation opens out as “a concept, a designation, a label, an appellation, a mere name”. The same five-word string appears three times in the section: of the name “Nāgasena” in Nagasena’s first answer, of the chariot in the king’s answer here, and of Nāgasena again in the conclusion. The point is carried by repetition rather than by argument. Nagasena then turns the king’s answer back on him, citing as his warrant a verse he attributes to “our Sister Vajirā in the presence of the Blessed One” — the chariot verse of the Vajira Sutta (SN 5.10), which our page on not-self quotes in Ṭhānissaro Bhikkhu’s translation. That page explains what the argument establishes; what the Milindapanha adds is the setting, the cross-examination, and the fact that the king performs the conclusion himself.

Talking as a scholar, not as a king

Two sections later, at Mil 3.1.3, Milinda asks to carry on, and Nagasena sets a condition before he will. Scholars, he says, talk a matter over until “one or other is convicted of error, and he then acknowledges his mistake”; kings, contradicted, say “Inflict such and such a punishment upon that fellow!” So he will discuss with a scholar — Rhys Davids keeps the Pali in brackets, “as a scholar (paṇḍit)” — and not with a king. Milinda accepts the terms.

Rhys Davids heads the section an “(Interlude)”, and it is easy to read past, but it is the clause that makes the rest of the book possible: a homeless monk imposing terms on a monarch, who accepts them. Look at the ladder the king offers — brother, novice, lay disciple, servant — each rung lower than the last. He is not promising courtesy; he is waiving rank, and granting Nagasena the right to treat him as a nobody. And “Be not afraid!” concedes what everyone in the room already knows: that fear rather than argument is what a king normally has on his side.

How Nagasena answers: the similes

Nagasena’s method barely varies. He gives a definition in a word or two; the king asks for an analogy; and the answer arrives as a picture from ordinary life — a lamp brought into a dark house, two rams butting, the king’s store-keeper perceiving the colours of the goods in the storehouse, a city-superintendent at the crossroads seeing who comes from each of the four directions. Asked what distinguishes attention (manasikāra) from wisdom (paññā), he answers with a harvest.

The picture is doing exact work: reaping is two-handed, and nothing is cut until something has first been gathered and held. Attention, on this account, is not itself the insight — it is the grip that makes cutting possible, which is why the Pali gives attention the mark ūhana and wisdom the mark chedana, the pair Kelly renders “examination” and “severing”.

The English versions part company here. Kelly has “attention” and “spiritual aspirant” (yogāvacara in the Pali of this passage); Rhys Davids calls the same manasikāra “reasoning” and the same yogāvacara “the recluse”, and gives the two marks differently again: “Reasoning has always comprehension as its mark; but wisdom has cutting off.” The choice carries weight. A section earlier, that same Pali word is what Nagasena grants to animals while withholding wisdom from them: “Sheep and goats, oxen and buffaloes, camels and asses have reasoning, but wisdom they have not.”

Rebirth without a self

The chariot settles what a person is; it leaves open what carries on. The Milindapanha comes back to that in the chapter Kelly’s translation heads “Buddha Chapter” (Buddhavagga), and puts the hardest form of the question first — flatly, with no simile at all.

That is the whole section: question, answer, “Very good, Nāgasena!”, and on to the next. Both halves of the reply are load-bearing. “In the highest sense” translates paramatthena, the qualifier that keeps the denial at the level of ultimate analysis instead of making it a flat “there is nobody here” — Kelly renders the same word “According to ultimate reality”. And “soul” translates vedagū, which is not the ordinary Pali word for a self at all. The brevity has troubled editors: in Rhys Davids’ 1890 Clarendon printing the answer carries a footnote reading “Mr. Trenckner thinks there is a lacuna here”, adding that the Sinhalese follows the Pali word for word (p. 111).

The section that follows asks the question that actually worries the king, and it is a ruler’s question rather than a philosopher’s.

Read the answer slowly, because it is conditional rather than doctrinal. Nagasena does not say that a wrongdoer is carried forward; he says that release from evil deeds would follow if there were no rebirth, and does not follow because there is. Nothing passes from body to body, and on this account nothing needs to.

The illustration Nagasena gives next is the stolen mangoes, which our guide to rebirth vs reincarnation works through alongside the lamp lit from another lamp; this page does not repeat it. What is worth adding here is the word the answer turns on. Rhys Davids renders the nāmarūpa of that reply “name-and-form”, and so does Kelly’s 2023 SuttaCentral translation; Kelly’s 2005 Access to Insight version has “mind-and-body”. The 2005 English makes plain what the compound holds together: the mental and the physical sides of a person — the very ground the chariot cross-examination walked over, part by part and aggregate by aggregate, without turning up an owner of any of it.

How the translations differ

The soul question turns on a single Pali word, vedagū, and three careful English versions do not agree about it.

The middle line is the one our glossary entry on rebirth quotes; it stands here beside the other two because the disagreement between them is the point.

Vedagū was a Brahmanical word before it was a Buddhist one. Kelly’s note on the line calls it “originally a brahmanical term related to mastery of the Vedas”, says the Buddha appropriated it to mean “one who has attained highest knowledge”, and quotes the Pali Text Society dictionary calling the sense “soul” “a peculiar meaning” of the word, recorded at Miln 54 and 71. What turns on the choice is how wide the denial is. “Soul” lands in English inside a Christian and Cartesian frame the Pali does not carry, so the line can read as a sweeping metaphysical denial. “Experiencer” narrows it to what the word is reaching for: a permanent subject standing behind the senses, to whom experience happens. And the qualifier matters as much as the noun — paramatthena marks the denial as one made under final analysis, not on the level at which there is plainly a king in the room being answered by a monk.

An unusual place in the canon

Here precision matters, because the Milindapanha’s status is genuinely mixed. It is best described as para-canonical — a text of the highest respect that different Theravada traditions place differently:

• In the Burmese tradition, it is treated as canonical, included among the books of the Khuddaka Nikaya.
• In the Sri Lankan and Thai traditions, it is not part of the canon proper, though it is deeply revered as an authoritative non-canonical work.

So it would be wrong to call it simply “canonical” or simply “non-canonical.” It sits in a respected borderland — cherished everywhere, formally canonised only in some places.

Its date is likewise uncertain. Scholars think the core of the text (roughly its earlier portion, which has a parallel preserved in Chinese) may date to around the 1st century BCE, with substantial later additions made over the following centuries, much of it in Sri Lanka. So the historical king Menander belongs firmly to the 2nd century BCE, while the book about him grew up considerably later — the two should not be run together.

Why the Milindapanha matters

The Milindapanha matters for two reasons. As philosophy, it is one of the clearest and most engaging expositions of core Buddhist ideas — the chariot analogy alone has taught not-self to countless generations. As history, it is a precious trace of the encounter between the Greek and Buddhist worlds, a conversation across cultures preserved in scripture. Few texts are at once so useful to a beginner and so intriguing to a historian.

Read the whole dialogue

SuttaCentral carries the complete Milindapañha, the Pali text alongside Rhys Davids’ translation, section by section; Access to Insight has John Kelly’s selected passages, marked “(excerpt)”, with his notes on the Pali words. Start where the book starts, at the chariot (Mil 3.1.1). The 1890 Clarendon volume sets out the rest under its own headings: Book II, “The Distinguishing Characteristics of Ethical Qualities”; Book III, “The Removal of Difficulties”; and Book IV, “The Solving of Dilemmas”, which opens in Rhys Davids’ contents with “Milinda finds dilemmas in the Holy Writ”.

For the teaching at its heart, see not-self; for the canon it sits beside, the Pali Canon; for the world that produced it, how Buddhism spread; and for the wider library, Buddhist scriptures.