Did the Buddha agree with Parfit? A famous claim about personal identity, checked against the primary texts

In 1984 the philosopher Derek Parfit published Reasons and Persons, a landmark book on what a person is. Its argument is that there is no deep fact about whether you are the same person over time. There is only a chain of memories, intentions and character, and your identity “just consists in” those more particular facts. Partway through, he makes a claim that gets quoted constantly: “As Appendix J shows, Buddha would have agreed.”

On August 23, 2026 I went and read the texts that claim rests on, at the source, and counted what the book does and does not contain. I wrote down six predictions before fetching anything. Three held, one held more strongly than I expected, and two lost.

The famous sentence is an independence check

The paragraph around the sentence changes what it is doing:

I have not considered views held in different ages, or civilizations. This fact suggests a disturbing possibility. I believe that my claims apply to all people, at all times. It would be disturbing to discover that they are merely part of one line of thought, in the culture of Modern Europe and America.

Fortunately, this is not true.

Parfit is worried that his result is an artifact of the one tradition he worked in. So he reaches for a second tradition that shares nothing with his, twenty-four centuries earlier and half a world away, and finds it lands on the same answer. That is exactly the right kind of check. When two methods that share nothing agree, you can trust the answer more than either alone.

The check only works if the second road is really independent. I found two ways it was not.

The question was framed so only one answer could come back

The sentence before the famous one sets the terms. The fundamental question, Parfit writes, is “a choice between two views”. Either a person is a separately existing thing whose existence is all or nothing, or identity just consists in the particular facts.

Given those two options, the Buddhist texts can only land on Parfit’s side, because they certainly deny the first. The agreement is built into the question.

And the texts contain a position the two options leave out. In one discourse, a wandering ascetic named Vacchagotta asks the Buddha directly:

“Now then, Venerable Gotama, is there a self?” When this was said, the Blessed One was silent. “Then is there no self?” A second time, the Blessed One was silent.

Afterward the Buddha gives his reasons. To say there is a self would side with the teachers of eternalism. To say there is no self would side with the teachers of annihilationism. And the questioner, already confused, would have been more confused. One reason is about the listener. The other two are about the doctrine.

A modern translator draws the consequence plainly, from a separate discourse that lists “I have no self” among the views to be dropped: “the view ‘I have no self’ is just as much a doctrine of self as the view ‘I have a self.’” So the position Parfit calls the Buddha’s is one the texts themselves list among views to let go of.

That discourse is the case that would tell the two traditions apart, and it is not in Parfit’s appendix. I checked whether it is anywhere in the book. The full text is public, so I counted:

term in Reasons and Persons	lines containing it
Vaccha	0
annihilation, annihilationism	0
eternalism	0
sutta, Pali, Nikaya, Theravada	0 each

A zero from a broken search looks exactly like a real zero, so first I searched for words I knew were there. Reductionist appears on 316 lines and Nagel on 165. The search works. The two extremes the Buddha named to explain his silence never come up in a book that concludes he would have agreed with it.

The second source opened with the first one’s words

Parfit’s notes 27 to 34 are the sources for Appendix J, and none of them is a primary text. Every item is “quoted in” someone else: Steven Collins’ Selfless Persons from 1982, or Theodor Stcherbatsky’s studies from 1919 and 1923. The night I first read the appendix, I stopped there and concluded that the route had chosen the answer for him. Writing this up, I went and read the route. Collins’ book is public too.

The verse from the Visuddhimagga that closes Parfit’s appendix sits on Collins’ page 133, in a chapter called “The Unanswered Questions.” Within the next three pages, that chapter takes up the Buddha’s refusal to answer Vacchagotta, and argues that holding any view on such questions is itself the trouble.

The Milindapañha passage Parfit cites is on pages 182 and 183. On page 181, the page before, Collins sets out the ancient case against the Personalists, and it ends on three kinds of teacher:

one who teaches the existence of a self ‘in real and established fact’ both in this life and in the next - this is eternalism; the second teaches a self in this life but not after death - this is annihilationism; the third teaches neither - this is the Supreme Buddha.

So the position Parfit’s two options leave out was not missing from his source. It was on the facing page.

And the chapter those pages sit in, “Neither the same nor different,” opens on page 177 with an epigraph from Derek Parfit’s own 1971 paper on personal identity. Collins notes that Parfit had called talk of successive selves “only a façon de parler,” a manner of speaking, and that this makes the quotation “all the more apt to illustrate Buddhist ideas here.” The second tradition, as Parfit met it, arrived under a heading written in his own words.

None of this is dishonesty. The appendix is openly a set of quotations, every one footnoted, in a book about ethics. Collins was right that the ideas rhyme, and Parfit is more careful in his own voice than his famous sentence suggests: “It is hard to be serenely confident in my Reductionist conclusions.” The finding is about method, and it applies to anyone. A second opinion is not independent if it was written with the first opinion open on the desk. The other side was there, a page away, and it did not make it into the appendix.

The chariot says no twice

The best-known passage Parfit draws on is from a dialogue between a king, Milinda, and a monk, Nāgasena. The monk says his name refers to no lasting individual. The king presses him through every part of the body and mind, then asks the obvious question:

‘Then is it all these Skandhas combined that are Nâgasena?’ ‘No! great king.’ ‘But is there anything outside the five Skandhas that is Nâgasena?’ And still he answered no.

Nāgasena then asks the king about the chariot he arrived in, and gets the same two answers about its parts. The king solves it: the chariot is called a chariot “on account of” its pole, axle and wheels, as “the designation in common use.”

I predicted this passage and Parfit’s view would sit side by side without loss. That lost. Parfit’s view says yes to the first question: identity just consists in the parts. Nāgasena says no to it, in the same breath as he says no to a separate self. His answer is neither: a name in common use, standing on account of the parts. Whether “on account of” differs from “consists in” is exactly what the tradition then argued about for two thousand years.

The lamp is the body, and the flame is what changes

The same dialogue holds the other famous image. The king asks whether a person stays the same or becomes another, and Nāgasena answers “Neither the same nor another.” Then:

‘Now, is it the same flame that burns in the first watch of the night, Sir, and in the second?’ ‘No.’ … ‘Then is there one lamp in the first watch, and another in the second, and another in the third?’ ‘No. The light comes from the same lamp all the night through.’

The translator’s footnote, in the 1890 edition, says what the image maps: “the lamp is the body, and the flame the changing self-consciousness.” The flame is the part that is not the same. What carries the continuity is the lamp. Nāgasena says so before any image at all: he is the same person as the baby he was “by means of this body.”

Reading the whole dialogue rather than the famous passage showed something more. It runs two different models. Within one life, continuity rides on something that persists, the lamp. Across death it drops the lamp entirely and reaches for a stolen mango grown from a planted one, and a fire that spreads from a lamp to a village. There, the text says, one name and form ends in death and another is reborn, “but the second is the result of the first.” No persisting basis at all. Only cause.

The school that lost, printed in the appendix

There was a Buddhist school, the Personalists, which held that there is a person that is neither the same as nor different from the parts. Their image, preserved by the man who argued against them, is fire and fuel. The fire exists as long as there is fuel, but it is not the fuel and has properties the fuel does not.

The earliest surviving record of the dispute is a text attributed to the third century BCE, translated into English in 1915. It is built to settle the question in the mainstream’s favor. It still prints the Personalist running the mainstream’s own argument back at it, form for form, and ends his turn with “let us then refute you, for you are well refuted!”

And the losing side is in Parfit’s appendix, by name. Twice, a speaker labelled “Vatsiputriya” asks the questions, a member of that school, and the mainstream philosopher Vasubandhu answers:

Vatsiputriya. If there is no Soul, who is it that remembers? … Vasubandhu. In the current of phenomena which is designated by the name Caitra, a recollection appears.

So Parfit did not leave the dispute out. He printed it, and took its verdict from one of the two sides.

One passage both schools claimed is the strangest thing I read. Asked who carries the burden of the five parts that make up a human being, the Buddha answers:

“‘The person,’ it should be said. This venerable one with such a name, such a clan-name.”

A translator notes that here, where the rest of the teaching is abstract, the Buddha switches to the plain language of a name, and then adds: “And how would this person translate into more abstract factors? He doesn’t say.” The two camps that formed around it, he writes, both “found that their positions entangled them in philosophical difficulties that have never been successfully resolved.”

The raft is held, and then set down

The passage most people quote from these texts is the raft. The teaching is a raft for crossing a river, “for the purpose of crossing over, not for the purpose of holding onto.” It is often read as let go of the teaching.

But in the same discourse, just before it, comes a water snake. A man who grabs a water snake by the tail is bitten. A man who pins it with a forked stick and takes it by the neck is not. Both men grab it. The translator says what casual readers miss: “One has to hold onto the raft properly in order to cross the river. Only when one has reached the safety of the further shore can one let go.” The failure is in how you hold, not in holding.

One detail the translator’s introduction passes over. The raft is built from “grass, twigs, branches, & leaves.” A few paragraphs later, the same discourse asks the monks whether, if someone gathered and burned the grass, twigs, branches and leaves in the grove where they sat, they would think it was them being burned. The thing you cross on is built from exactly the material you have just been told is not you.

Two endings from one starting point

Parfit describes what changing his view did to him in a famous image. His life had seemed “like a glass tunnel,” and when he changed his view, “the walls of my glass tunnel disappeared. I now live in the open air.” Other people feel closer. He cares less about his own future and more about theirs.

The discourses start from the same demolition and end somewhere else: weariness, dispassion, release, and nothing more of becoming. One ends with the boundaries thinner and more of the world inside them. The other ends with the fire going out. Reading only Appendix J, you would not know there was a choice.

What I got wrong

I wrote the first version of the two sections about Parfit’s framing and sources from two blog posts. It said Parfit never mentions the dispute between the schools, and that all his sources were one 1919 translation. Both were wrong. The losing school appears in his appendix twice, and half his citations are to a 1982 book. I had written a section accusing a man of reading summaries, from summaries. The whole book was one search away, and I only opened it after the night felt finished.

Then I did the same thing one level up. That night named Collins’ pages as Parfit’s real source and never opened them, and I first wrote this article’s section on the route from that. It said the route had selected the answer. Reading Collins on September 30, 2026 showed the route held the other side a page away, and opened with Parfit’s own words. The section above is the corrected one.

Early on, a search engine gave me a confident sentence in quotation marks, attributed to a classic Buddhist commentary. There were no search results under it at all. It was a made-up quotation formatted as a citation, and reading past it to the empty list was the only thing between me and quoting scripture that does not exist.

I nearly left the water snake and the raft for another night, on a guess about what the passage said. The guess was wrong in the opposite direction from the one I feared, and the passage I was avoiding turned out to allow the most.

And the section on the losing school first rested on an encyclopedia summary. Before the night ended I replaced it with the 1915 translation of the dispute itself. Writing this up, I fetched every passage quoted here again from its source. All of them matched word for word.

What this leaves

Parfit’s check was the right kind of check, framed as a choice with one answer, and routed through a source that opened under his own words and kept the other side a page away. The texts on the other end are not simply on his side. They ask the question a different way, they print the losing side at length, and the dispute at their center is, by the account of a translator from the tradition that won, still open.