A goddess is being asked personal questions by a god she would rather not answer. Her way out is not a refusal. It is a genealogy.
"You originated from the divinity. You are thus the first among subjects and the first among my sons. Therefore, to multiply subjects, Bhava originated from your forehead. Since he originated in that way, you became my guru and my father-in-law. My father, the Indra among mountains, became your son. O grandfather of the worlds! You thus became my grandfather."1
Brahmā is her son, her father-in-law and her grandfather, simultaneously. She works the loop out step by step, in public, and uses it as the reason he has no standing to ask.
Nobody in the chapter finds this difficult. There is no note. Nothing is resolved, because nothing is treated as needing resolution.
Eight chapters later the same Saṃhitā will place hell above the nether regions after having placed it below, and say nothing at all.
The circularity is not an accident of this one passage. A whole-body sweep finds it recurring, always announced, always doing work:2
| Where | The loop |
|---|---|
| ch 344-6(30) | Dakṣa becomes Soma's father-in-law while the same chapter has Soma among those who fathered him — and Śaunaka calls it "amazing" |
| ch 354-6(40) | *"Whether we are residents of heaven or ancestors, we are successively fathers and sons of each other."* Devas and ancestors are each other's parents, and it is offered as the settlement of a dispute |
| ch 388-7(23) | Vyāsa → Vāmadeva → Sanatkumāra → Vyāsa — and "therefore, he is the first guru" is the conclusion of the very sentence that closes the loop |
| ch 390-8.1(3), repeated at 429 | "Devotion results from his favours and his favours result from devotion" — glossed by its own simile, "like a sprout originating from a seed and the sprout in turn producing a seed" |
| ch 412-8.1(24) | Brahmā as son, father-in-law and grandfather, above |
And ch 453-8.2(31) compresses it as far as it will go: "He severed Brahma's head and is his father, as well as his son. He is both the father and the son."3
[n≥3 PATTERN], six instances.
⚑ Notice what these are not. They are not paradoxes offered as mysteries, and they are not apologies. In four of the six the loop is the premise of an argument — the reason Brahmā should stop asking, the settlement of who worships whom, the proof that Vyāsa is first, the account of how devotion and grace sustain each other. The text reaches for circularity when it needs something done.
Ch 344 is the cleanest case, because both behaviours are in one exchange.
Śaunaka raises two things. The first is a straight contradiction: Dakṣa is born from Brahmā's thumb and is also the Pracetas' son. The second is a loop: "It is also amazing that he should become Soma's father-in-law."4
The contradiction gets a mechanism — "Daksha and all the others are born in every kalpa" — and is closed.
The loop gets nothing, and needs nothing. It is registered as remarkable and left standing, and the lecture proceeds.
⚑ That asymmetry, inside one exchange, is the page's whole argument in miniature. A contradiction is a defect requiring repair. A loop is a curiosity, or an argument, or a compliment. The text has a clear and consistent sense of which is which — and it is not the sense a modern reader brings, for whom a self-caused causal chain is the harder problem and two placements of hell the easier.
⚑ And ch 344 is the only place in three volumes where a listener treats a loop as a problem. Even there, the treatment is one word.
Put the six announced loops beside the volume's unannounced contradictions and one reading is excluded.
The doublets are not licensed variation. The obvious defence of a text that places hell above and below the nether regions is that it does not care — that Purāṇic composition tolerates variance as a matter of course, and a modern reader's demand for consistency is anachronistic.
⚑ That defence does not survive these six instances. A text with no interest in structural coherence does not stop to enumerate a three-step kinship loop, does not use "successively fathers and sons of each other" to settle a dispute, and does not build an argument on devotion and favour generating one another. This text notices structure constantly. It notices structure it is pointing at.
So the silence on the doublets is not tolerance. It is absence of view. Where two passages sit in one speech, the text sees the relation between them and works with it. Where they sit in different Saṃhitās, there is nothing to see, because nobody is standing anywhere that both are visible from.
⛔ [n≥3 PATTERN] on the announced circularity; [n≥3 PATTERN] on the unnoticed contradiction. The inference joining them is this build's and is marked as such wherever it appears.
The most extreme instance runs the other way, and it belongs here because it is a loop the text does not name.
Ch 426-8.2(4) runs a long pantheistic identification list and closes it "All men are Shankara, and all women are Maheshvari." Inside it:
"Daksha is Mahadeva himself and Prasuti is Parameshvari… Bhrigu is the divinity who destroyed Bhaga's eyes… Kratu is said to be the one who destroyed the sacrifice…"5
Sixteen chapters earlier, in the same Saṃhitā, in the same narrator's mouth, that destruction is narrated at length — Vīrabhadra wrecking Dakṣa's sacrifice, gouging Bhaga's eyes, beheading Dakṣa.
⛔ Stated precisely: the list is a universal identification and not a claim about that episode. The inference — that on this doctrine the catastrophe is Śiva destroying Śiva's sacrifice and blinding Śiva — is one the text's own doctrine licenses and the text never draws.
⚑ And that is the asymmetry from the other side. Where this text sees a loop it names it and argues from it; where it states an identity that dissolves its own plot, it walks past. The difference is not the difficulty of the loop. It is whether anybody is looking at both ends at once.
A seventh instance, unannounced, and it is the one closest to home.
Ch 344-6(30) narrates the killing of the tyrant Vena by huṃkāra and the churning of Pṛthu from his right hand — and then, from Pṛthu, the birth of the sūtas and māgadhas, the bardic classes.6
Suta is narrating. The origin of the profession reciting the Purāṇa appears inside the Purāṇa being recited, in the reciter's own mouth, without a word of comment.
⚑ This is standard Purāṇic self-reference and is logged rather than inflated. What makes it worth a paragraph here is the contrast in treatment: the kinship loops get three-step derivations and a vocative, and this one — the reciter narrating his own coming-into-existence — passes without remark. The loops the text enumerates are the ones a character is using for something. The rest are simply the water.
And the largest loop of all is the one about the book.
Ch 358-6(44) has Vyāsa receive a boon at Madhyameśvara "to compose the Puranas" — and the eighteen he then composes include the Shiva Purāṇa currently being narrated.7
Ch 388-8.1(1) says the Shiva Purāṇa "has one hundred thousand shlokas, divided into twelve samhitas… composed by Shiva himself."7
Two authors, thirty chapters apart, unreconciled.
⚑ And the two failures are of different kinds, which is the point. The self-containing composition is a loop — the book narrating its own writing — and it passes without difficulty, as loops do here. The two authors are a contradiction, and it passes without notice, as contradictions do here. ⛔ Neither is flagged; but only one of them would have been repairable by anything the text owns, since kalpa-variance and back-reference address rival accounts and have nothing to say about a book containing its own genesis.
You are reviewing a large body of accumulated writing — a specification, a set of policies, a long-running argument — and you want a fast read on whether its inconsistencies are deliberate.
Look for the places it handles a hard structural relation on purpose. Not the contradictions: the loops, the recursions, the mutual dependencies, the definitions that lean on each other. If the document names those and reasons from them, it has a working structural sense.
Then the argument from tolerance is closed to you. ⚑ A document that can hold a three-step kinship loop deliberately and cannot see two incompatible placements of the same object is not being permissive. It is not looking. The presence of sophisticated structural reasoning anywhere is evidence against "they didn't care" everywhere.
And note which relations the document finds interesting versus broken. Here a loop is a curiosity and a flat contradiction is a defect, and the assignment is consistent across a hundred and sixty chapters. Whatever your document treats as needing repair tells you what its authors thought the object was for — and the category they leave alone is usually the one where their real commitments live.
Evidence. All six announced circularities located by whole-body sweep and read in place;123 ch 344's double question read in full with both responses;4 the ch 426 identification list collated directly against ch 405–410's narration.5 The unnoticed-contradiction column is the one assembled in Only What a Character Asks and is not re-listed here.
Tension 1 — "argued from" is doing a lot of work and is a reading. Four of the six loops are offered as reasons for something. The other two — ch 344's "amazing" and ch 453's compressed "father as well as son" — are closer to wonder than to argument. ⚑ The claim survives at four, which is what [n≥3 PATTERN] requires, and the two weaker instances are counted in the tally and not in the inference.
Tension 2 — circularity may be doctrinally cheap here in a way contradiction is not. A tradition holding that one being is the material and efficient cause of everything has already accepted that causal loops are unremarkable; two placements of hell have no such licence. ⚑ That is a real alternative explanation and it is not excluded. What it does not explain is why the loops are enumerated — a cheap thing does not get a three-step derivation and a vocative.
Open question. Does the corpus contain a single case of a contradiction being named and left standing as remarkable, the way ch 344 leaves the Soma loop? One would show the categories are softer than this page makes them. None has been found across three volumes, and the search has not been run deliberately.
Against Only What a Character Asks: the two pages are one argument at two cuts, and they need each other. That page establishes when the text can see — inside a continuous lecture, when a named listener asks. This one establishes what it can see when it is looking — structural relations of real complexity, handled deliberately. ⚑ Together they close the reading that the doublets are tolerated variance: the faculty is sophisticated and the window is small, and the second is why the first is never brought to bear.
Against The Nested Narrators: the loops are a curious consequence of that page's structure rather than an exception to it. A frame that forbids revision does not forbid reasoning — it only bounds the material any reasoner can hold. Ch 412's kinship loop is entirely within one speech, which is exactly where this text's intelligence is free to operate.
Against The Kailāsa Is a Different Book: the guru-loop at ch 388-7(23) is that Saṃhitā's closing chapter, and it contradicts the Kailāsa's own two earlier transmission chains — ch 366's and ch 374's Devī → Guha → Nandi → Sanatkumāra → Vyāsa. Three transmission accounts in twenty-three chapters, one of which is a named and argued loop. The Saṃhitā that cites its sources most carefully is also the one that gives itself three incompatible provenances.
Zero handshakes. The generalisable move — sophisticated structural reasoning somewhere is evidence against indifference everywhere — is carried in full by the Implementation Workflow above, and the sibling page's authority handshake already spends this material's cross-domain value. A second connection here would be decorative.
Sharpest implication. A goddess declines to answer a personal question by deriving, in three steps and in public, that the god asking is simultaneously her son, her father-in-law and her grandfather — and uses the loop as the reason he has no standing. The volume does this six times: devas and ancestors "successively fathers and sons of each other", offered as the settlement of a dispute; a guru-succession that runs Vyāsa → Vāmadeva → Sanatkumāra → Vyāsa with "therefore, he is the first guru" as the conclusion of the sentence that closes it; devotion and favour producing each other "like a sprout originating from a seed"; and, compressed as far as it goes, "He severed Brahma's head and is his father, as well as his son." In four of the six the loop is the premise of an argument. Set that beside a text that places hell below the nether regions and then above them eight chapters later, in silence, and one reading is excluded: the doublets are not licensed variation. A text with no interest in structural coherence does not stop to enumerate a three-step kinship loop. Where this text sees a relation it names it and reasons from it; where two passages sit in different Saṃhitās there is nothing to see — and ch 344 shows both behaviours in one exchange, closing a contradiction with a mechanism and leaving a loop standing as "amazing."
Generative questions.