What is a text doing when it builds the strongest case against itself and then declines to answer it?
The Vol. 3 build produced 118 pages, and the thing that kept recurring across clusters was not a doctrine. It was a procedure: raise the objection at full strength, from a credible mouth, in detail — and then partition rather than rebut, or leave it standing and move on.
1. Objections are raised by insiders, not enemies. The narrator attacks his own verse. Balarāma, who is divine, rebukes Kṛṣṇa. Parīkṣit halts the recitation to demand a discrepancy be explained. ⚑ The prosecution is never outsourced to a straw man. — The Prosecution Is Inside the Devotion
2. Resolution is by partition, and the sets are left unranked. Four models of death. Four rankings of jñāna. Three criteria for one judgement. Two dharmas in one chariot. Five relativities of purity. ⚑ Every one is a set held without a priority rule. — Partition Is Not an Answer
3. Two bars, moving at different rates. Caste comes down and stays down; sex comes down and goes back up in the same Skandha. And a third position nobody was looking for: the proceeds are graded even where the door is open. — Two Bars, One Lever
4. Counts fail wherever a number is announced first, and never where one is not. Six instances in twelve chapters, all checkable on the page, two flagged and four silent. — The Number Was Never Counting
5. The expensive doctrine is routed through the rival deity. [n≥3] — and the volume's counter-instances (11(12), 11(14)) are Kṛṣṇa's own mouth, which cuts against it. — The Rival Deity Is the Credible Witness
6. Attainment is tested by what somebody failed to notice. [n≥3] — the arrow-maker, Mārkaṇḍeya, Parīkṣit. Never by self-report. — The Arrow-Maker Did Not Notice the King
7. The recommended age is the one the best man is evacuated from, and the reconciliation is a reader's patch. — The Age It Recommends Is the One It Evacuates
8. Explanation extends waiting. The women given a reason grieve for chapters; the women given nothing are said to have come out ahead. — The Ones Who Were Given Something to Wait For
This volume is not a text with contradictions in it. It is a text organised around refusing to rank incompatible positions — and the refusal is what lets it serve readers in incompatible situations without admitting it has more than one audience.
⚠ The disconfirming instance is inside the corpus and is not being hidden: 11(28)'s prohibition on praising and reprimanding runs simultaneously with the volume's own continuous praising and reprimanding, and nobody notices. ⚑ A deliberate refusal would not have missed that one.
A. The Vols 1–2 rate check. [NEWSLETTER] — Almost every [n≥3] in this build is bounded to Vol. 3. The count failures, the not-noticing test, the Śiva routing, the listener-as-auditor habit all need a rate across three volumes before any of them can be stated as a feature of the corpus.
B. Partition as a genre property or a local one. [WILDCARD] — If Purāṇic form generally resolves by partition, thread pattern 2 is a description of a genre and not a finding about this book. This is the single highest-leverage check on the list.
C. What a summary can hold. [NEWSLETTER] — 12(12) keeps the antagonists by name and reduces the volume's longest grief sequence to four words. — The Recap Keeps the Crimes and Drops the Grief
D. Reading contracts and false specifics. [NEWSLETTER] — the count material carried forward into the present. — The Numbers Nobody Was Counting
E. Hope as an inhibitor. [WILDCARD] — filed under tamas, deployed as a merchant's motive, and dissolved as Piṅgalā's liberation. — Hope Is Filed Under Tamas
Angle 1 — "The Numbers Nobody Was Counting." Pillar: how traditions transmit. Core argument: a false specific is only false if readers treat specifics as claims; the Purāṇa's broken counts survived because nobody was holding them to arithmetic. Unique angle: the AI-hallucination frame arrives late and as a consequence, not as the hook. Target reader: someone who works with AI output and has never thought about the reading contract as the other half of the problem. Type: argument essay. Opening hook: count the kings yourself; it takes forty seconds.
Angle 2 — "The Test Is What You Failed to Notice." Pillar: attention and practice. Core argument: this text's working criterion for attainment is always a third-party observation of something the person missed — which is exactly what self-report cannot reach. Unique angle: it closes on keep company with people who have it, which is the tradition's own answer to an unverifiable criterion. Target reader: anyone who logs, rates or tracks an inner state. Type: essay. Opening hook: the arrow-maker and the king.
Angle 3 — "What a Summary Can Hold." Pillar: narrative and memory. Core argument: recaps are made of verbs, so events survive and conditions do not — and what a tradition remembers about itself is filtered through that. Unique angle: the strongest counter-explanation is given room rather than buried. Target reader: anyone who has watched a long story get flattened by its own retelling. Type: essay. Opening hook: four words for the grief, an inventory for the killings.
Seven collisions filed, all status: speculative. ⚑ None is close to promotion, and the reason is the same for all seven: every one has a promotion criterion that requires Volumes 1–2, and Volumes 1–2 have not been read for these questions.
⚑ Read Volumes 1–2 against this list rather than as fresh ingests. Five of the seven collisions and three of the five sparks resolve or die on the same evidence, and gathering it once is cheaper than eight separate checks.
⚠ Held, not to be built: #329 (the-hunter-and-the-pigeon-are-cited-and-never-told). Building it before Vols 1–2 are read would be an absence claim over unread material.