A sage is farming. A feeble cow wanders into his grain. He shoos it away with a blade of grass — the gentlest thing in reach — and the cow drops dead on the spot.
That scene happens twice in this volume, four pages apart, with the same sage, the same grain, the same blade of grass and the same instant death. And the two tellings end in opposite places. In the first, Gautama is expelled, does penance, brings the Gaṅgā down to earth for everyone's benefit, and then talks the river into forgiving the men who framed him. In the second, he works out what was done to him and lays a hereditary curse on them that reads like a description of a living rival sect.
What makes this worth a page is not the repetition. It is that the text notices. The listeners interrupt and say they have heard a different version. The narrator agrees, names the reason, and tells the second one in full.
That is the volume doing, out loud, the thing it conspicuously does not do for its other repetitions.
The first version starts with a queue at a well.
Gautama's disciples go for water. The rishis' wives are already there and pull rank — "We are wives of rishis. We will take it first. Go far away." Gautama's wife Ahalyā comes down, takes the water herself, and gives it to her husband for his daily rites. The other wives reprimand her, go home, and "in front of their husbands, they reported the opposite of what had happened."1
The husbands, believing it, decide to have Gautama expelled. And their method is the interesting part: they worship Gaṇeśa, and ask him for it.
Gaṇeśa says no, at length, and the reasons he gives are correct:
"Anger when no offence has been committed leads to harm. It is not good to cause misery to those who have done a good turn in the past… Earlier, as a result of loss of food, you confronted hardships. It is maharshi Goutama who gave you this happiness. You now wish to inflict miseries on him… Deluded by the power of your wives, you are not acting in accordance with my words. … Crossing him is not right. Please ask for another boon."1
They answer with one sentence: "This is what you must do, nothing else."
And he yields: "I will indeed do what you have asked for. What is going to happen, will certainly happen thereafter. Saying this, he vanished."1
This belongs squarely with the corpus's boon material. The granter names the objection in full, makes the moral case, is overruled by petitioners with no argument at all, and grants it anyway. Not because he is persuaded, and not because he is bound by a prior promise — simply because that is what a boon does in this text.
Then the detail that turns this from a bad decision into something stranger.
"Ganesha assumed the form of a feeble cow and went there. As a result of the boon, he went there, with his body trembling, and started to eat the vrihi and barley."1
The god who has just argued that framing an innocent man leads to destruction goes and personally performs the frame-up. He walks into the field in a trembling body, eats the grain, and dies at the touch of a blade of grass while the conspirators hide nearby and shriek "What has Goutama done?"
Gautama's own reaction is not anger. It is bewilderment, and he reaches for the wrong explanation: *"O lady! What is this and how has it happened? Maheshvara must be angry."*1
He is stoned, expelled, told the fire and the ancestors will not accept his offerings, and moved a krośa away. He asks the rishis how to purify himself and they set the terms — circumambulate the earth three times proclaiming the sin aloud, or go round Brahmagiri a hundred and one times, or make a crore of pārthiva liṅgas and bathe in the Gaṅgā.1 He accepts the lightest option and performs it.
Note what has happened to responsibility. A god committed the fraud, on the instruction of men he told were wrong, and the penance falls on the man who swung the blade of grass.
Gautama's penance brings the Gaṅgā down at Brahmagiri, which is the whole point of the episode — Tryambaka and the river Gautamī both come from it.2
And then the framers come to bathe, and "as soon as she saw them, she instantly vanished."3
Gautama intercedes for them. Gaṅgā refuses, and her assessment is blunt: "They are evil-souled and ungrateful. They have injured their own lord. They are rogues and heretics. Their sight must always be avoided."3
Gautama's counter is the move this corpus almost never lets work — he quotes an authority back at her:
"O mother! Please listen to the great words Bhagavan uttered. You must make them come true. 'If a person does a good turn to a person who has injured him, I am purified by that.' These are the words Bhagavan spoke."3
She concedes. "O rishi Goutama! You have spoken the truth, and these are auspicious words." She sets a penance — a hundred and one circumambulations — the rishis perform it, beg his pardon, and are restored. Gautama makes them the Kuśāvarta tīrtha. "Goutama and the other rishis met with each other. They were ashamed because of their former ungratefulness."3
Version one ends in reconciliation, and the reconciliation is achieved by quoting Śiva's own teaching at a goddess who was refusing.
And then the sages listening to the story interrupt.
"O Suta! We have heard you tell us about a different account, how the angry Goutama cursed the brahmanas. Please enlighten us about this."4
Sūta does not deny it, does not choose between them, and does not treat the question as a challenge. He answers with the corpus's harmonisation instrument:
"That is also true. These differences occur because incidents have occurred in different kalpas. O ones excellent in vows! I will specifically tell you about that account too."4
And he narrates the second version in full.
Three things are worth separating here.
First, this is the second deployment of the kalpa device in the volume. The first is at ch 163, on Gaṇeśa's two birth-stories: "Because of differences in kalpas, the story of Ganesha's birth is spoken about in different ways." Together with the back-reference-and-decline at ch 239 and the announced re-telling at ch 286, that gives the volume four working mechanisms for disclosing repetition, used four times.
Second, and this is what makes the finding sharp: none of the four is used on the volume's actual unflagged doublets — Nṛsiṃha's containment at ch 216 against ch 239–241, or the Avadhūteśvara episode at ch 183 against ch 259. The machinery exists, is in active use, and was not applied where it was most needed. That is a much stronger statement about those doublets than "the text never notices repetition," which would have been a perfectly good excuse.
Third, a qualification this build owes itself. The listeners say they heard the other account from this narrator, which reads as internal. But the translator's note on that sentence locates the variant elsewhere: "For example, in the Brahma Purana."5 On that reading the kalpa device is being deployed on an inter-textual variant — a version circulating in another Purāṇa — rather than on an internal repetition. Both readings are available from the English, and the distinction matters, because a text harmonising a rival Purāṇa's account is doing something different from a text catching itself.
Two accounts of the same incident, and you are the one who has to say something.
The instinct is to pick. Sometimes you have to. But watch what this text does first, because it is a real option and it is usually not on the table when people think it should be.
It says "that is also true," names a frame in which both hold, and then tells the second one in full.
That move costs almost nothing and buys three things. Nobody is called a liar. Neither account has to be suppressed, which means neither acquires the energy that suppressed accounts acquire. And the person who raised the discrepancy is answered rather than managed.
The catch — and it is the whole catch — is that it only works if you actually narrate the second version. Sūta does. He gives the rival account at full length, including the parts that make the first version look naive. A harmonisation that says "both are true in their own way" and then only ever tells one of them is not a harmonisation. It is a dismissal with better manners, and everyone can tell.
So: if you are going to say both are true, publish both. Same length, same detail, same platform.
And then the second half, which is the diagnostic. When two accounts of one incident end differently, look for what differs in the middle, not at the end. People reach for motive — one of them is protecting someone — and usually the answer is duller and more useful. Here the two versions end in forgiveness and in a hereditary curse, and the reason is not that one narrator is kinder. It is that in one version a god committed the fraud and in the other the men did it themselves. Different offence, different ending. The endings are not arbitrary and they are not evidence of bias; they are consequences.
Find the divergence in the facts of the middle before you attribute anything to the teller.
⚑ Everything in this section is what this text says about a rival tradition. It is a Śaiva polemic, composed as an origin-myth, and it is reported here as such. Nothing in it describes Śrīvaiṣṇavism, and it must never be cited as though it did.
The second version keeps the drought, the water dispute and the cow, and changes two things.
The water comes from Varuṇa, not from Gaṅgā: Gautama performs austerities to him, is given an inexhaustible supply, sows grain and feeds the starving rishis.4 And the cow is not a god. "The brahmanas created an artificial cow. Crooked in their minds, they made it feed on the grain."4
Gautama strikes it with the blade of grass, it dies, the rishis raise the cry — and then: *"When he got to know the truth about the cow, Goutama was filled with rage."*4
The curse that follows is long and it is specific:
"Therefore, you will always turn away from the Vedas… From now, you will have faith in wicked paths… From now on, your foreheads will be smeared with mud. O brahmanas! With your foreheads smeared in mud, you will descend into hell. You will no longer have Shiva as your single divinity. Shiva is without a second, but you will treat him on par with other devas. You will no longer have any love for rites connected with Shiva's worship… Let it always apply to your descendants too… You will therefore be marked with hot coins."4
And the destination is named: *"Expelled from the dharma of being Shiva's devotees, they resided in Kanchi."* The translator's notes gloss the branding as being applied "as punishment, on the forehead" and identify Kāñcī as Kāñcīpuram.5
A mud-marked forehead, branding, hereditary transmission, refusal of Śiva's singularity, and a South Indian address. This is the third sectarian-origin polemic in the volume, after the manufacture of the Jain-and-Buddhist heresy at ch 174–175 and its damnation of future believers at ch 182, and the inclusion of Buddha in the avatāra list at ch 186.
Its value to this vault is as dating evidence, not as description. A polemic that names a specific mark, a specific punishment and a specific city is describing something its author could see, and that narrows the window at the late end. It belongs on the source stub's governance note, and it does.
Set the two versions side by side and the divergence is not moral temperament. It is culpability.
Version one: the rishis ask a god to arrange an expulsion. The god objects, is overruled, and performs the fraud himself. The men are guilty of a bad request; the deed is the god's. So the offence is ingratitude and bad judgement — and ingratitude is forgivable, on Śiva's own stated principle, which is exactly what Gautama quotes to get them restored.
Version two: there is no god in the mechanism at all. "The brahmanas created an artificial cow." They build the instrument, they place it, they raise the cry. The offence is a manufactured judicial frame-up executed with their own hands, and Gautama's rage begins at the moment "he got to know the truth about the cow."
Different act, different verdict. The two accounts are each internally coherent and they are coherent with each other's endings once you notice that the fraud has a different author in each.
That is worth stating because it cuts against the easy reading. These are not a kind version and a cruel version of one story. They are two stories that agree about the setup and disagree about who did the thing — and the sentences follow from that.
The strongest evidence is that the text raises the discrepancy in its own listeners' mouths and answers it with a named device, then narrates the variant in full.4 Nothing here requires the reader to detect anything.
Tension 1 — Gaṇeśa argues correctly, loses, and executes the fraud personally.1 The volume does not comment on a god performing an act he has just called destructive.
Tension 2 — the internal-versus-inter-textual reading of the kalpa deployment. The listeners' phrasing reads as internal; the translator's note locates the variant "for example, in the Brahma Purana."5 Both readings are available and the page does not decide.
Tension 3 — the disclosure machinery is used here and not on the volume's real doublets. Four mechanisms, four deployments, and none on ch 216/241 or ch 183/259.
Tension 4 — version one lets an argument win. Gautama moves Gaṅgā by quoting "the great words Bhagavan uttered."3 That is the same mechanism as the volume's two acknowledged winning arguments — a citation the opponent cannot repudiate — and it is worth flagging as a possible third instance rather than asserting it, since Gaṅgā is not refusing on the merits but on disgust.
Open question. Version one's Gaṅgā and version two's Varuṇa supply the same water for the same famine. Does the volume elsewhere swap the deity in a type-scene while keeping the scene intact — and if so, is that variation or transmission?
Against The Śarabha and the Hide: these two pages are the volume's control pair on disclosure. Here the repetition is raised by the audience, named as kalpa-variance, and the variant is told in full. There, an incompatible account of Nṛsiṃha's containment sits twenty-five chapters from its twin with no acknowledgement, in a chapter that uses a different disclosure device in its opening paragraph. Same volume, same narrator layer, opposite handling.
Against Gaṇeśa: Made, Beheaded, and Promoted: the ch 163 kalpa deployment is about Gaṇeśa's own birth, and here Gaṇeśa is the agent of the episode the second kalpa deployment covers. He is at both ends of the volume's harmonisation machinery, which is a coincidence worth noting and not worth building on.
Against Tripura, the Hardest Case: the ch 298 curse completes a pattern the Tripura chapters begin. There a heresy is manufactured by divine command and its future believers damned; here a rival sect is generated by curse out of the tradition's own brahmins, hereditarily. In both, the rival is not encountered — it is produced, from inside, by the text's own gods and sages.
Against In a Different Kalpa: that page holds the device. This chapter is its second and larger deployment, and supplies the point that matters for method — the device's non-use elsewhere is informative precisely because it is demonstrably available.
Christian Heresies and Their Buddhist Parallels — that page records an argument that early Christianity kept generating Buddhist-shaped doctrines from inside itself — self-responsibility, levels of truth, a God who suffers — and kept condemning them; two traditions drawing on one well, with only one permitted to say so. The heretic, on that reading, is not an invader. He is an internal product.
Chapter 298 is that structure stated as myth rather than reconstructed from records. The people who end up mud-marked at Kāñcī are not foreigners and not converts. They are brahmins, sages, colleagues of Gautama, fed through a famine by his austerities. Their heterodoxy is not something they brought or discovered. It is issued to them, hereditarily, as a punishment, by one of their own — and its content is specified in the curse itself: they will treat Śiva as one deity among others.
What the pairing produces is a distinction the historical case cannot easily make and the mythic case makes plainly. In the Christian material the internal origin of heresy has to be inferred, against a tradition insisting the heretics were corrupted from outside. Here the tradition asserts the internal origin itself, because asserting it is the more useful claim: a rival who was never other than you is a rival whose existence needs no explanation beyond your own moral history, and whose defining doctrine can be presented as a sentence rather than a position. The curse does not argue that treating Śiva as one among many is wrong. It makes it the punishment.
Carried back, that suggests reading origin-stories about rivals for their direction of causation. A tradition that says its rival came from outside is defending its own coherence. A tradition that says the rival was expelled from within is doing something more aggressive — it is denying the rival any independent starting point at all.
One handshake, not two. A second, on the kalpa device as a harmonisation technology, is worked as this page's own Implementation Workflow and would restate rather than add.
Sharpest implication. The volume's clearest act of self-awareness and its ugliest piece of polemic are the same passage. Sūta is asked about a discrepancy, concedes it without defensiveness, names a frame in which both accounts hold, and narrates the second at full length — the most methodologically honest thing the text does anywhere. And the account he then narrates is a hereditary curse that manufactures a rival sect out of the tradition's own brahmins, complete with forehead mark, branding and postcode. The candour and the polemic are not in tension; the candour is what makes room for the polemic, because a text that can hold two versions at once never has to withdraw either.
Generative questions.