⚑ Count note: this build's PRD described seven forward promises with a 6/7 in-volume completion rate. A recount of the endnote layer gives a different and more interesting picture, set out below. The corrected figures are used here.
The Bhāgavata Purāṇa makes forward promises. I will tell you about this later. A composite text of this size could easily leave them dangling, and a reader has no way to check without the whole corpus in front of them.
Debroy's endnotes resolve them, and the resolutions are specific — this is discharged in the Eighth Skandha, this in the Ninth, this in the Twelfth.
The finding is not that the system works. It is where it works.
Every Skandha cross-reference in the endnote layers of Skandhas 5 through 10, counted:
| Skandha | Cross-references | Forward, in this volume | Forward, beyond it | Backward |
|---|---|---|---|---|
| 5 | 7 | 5 | 1 | 1 |
| 6 | 4 | 3 | 0 | 1 |
| 7 | 3 | 1 | 0 | 2 |
| 8 | 0 | — | — | — |
| 9 | 0 | — | — | — |
| 10 | 0 | — | — | — |
| Total | 14 | 9 | 1 | 4 |
Fourteen cross-references. Nine point forward to a Skandha inside this volume; one points to the Twelfth, which is outside it; four point backward to Skandhas 1–3, which is Volume 1.
And every one of them is in the endnotes of Skandhas 5, 6 and 7. The endnote layers of Skandhas 8, 9 and 10 contain none at all.
The in-volume forward references are genuine navigational aids and they land:
Each names a destination the reader can reach without leaving the volume. The Vāmana note at 6(18) points to Skandha 8, where Vāmana duly appears. The elephant note in Skandha 7 points to 8(2)–8(4), where Gajendra duly appears. The turīya note sends a reader forward for a technical term.
That is a working apparatus, and it is worth saying plainly: the promises this volume makes to itself are kept.
The distribution is the finding, and it is not what a reader would predict.
Skandhas 8, 9 and 10 are not thin. Skandha 9 is the longest in the volume by shloka count and is almost entirely dynastic material — the exact kind of text that generates cross-references, since it is a dense web of lineages, and this build has repeatedly found it gesturing at stories told elsewhere. Skandha 10 opens the corpus's most famous section and is explicitly continued in the next volume.
Yet the annotation stops.
Three explanations are available and the volume does not choose between them.
The material stopped requiring it. Possible for Skandha 10, whose narrative is self-contained; much harder to argue for Skandha 9's genealogies.
The annotator's attention decayed across a long work. The commonest explanation for any apparatus that thins toward the end, and unfalsifiable from inside the text.
Or the front-loading is deliberate — the cross-references exist to orient a reader arriving at Volume 2 without Volume 1, and once oriented, they are no longer needed.
The third is the most generous and has one piece of evidence for it: all four backward references are also in Skandhas 5–7, pointing to the earlier volume. If the apparatus were about orientation, that is exactly the pattern it would show — heavy at the entry point, absent once the reader is settled.
Of the fourteen, one is genuinely open at the volume's end, and it is the one the PRD flagged for a different reason.
At 6(18) the text reads: "Through Ashana, Bali had one hundred sons and Bana was the eldest. Through excellent shlokas, we will later describe his praiseworthy character." The nearer referent is Bāṇa. The endnote resolves it to Bali — "Bali's character, in the Eighth Skandha."2
And Bali's character is described at length in Skandha 8. So the note's promise closes.
What does not close is Bāṇa. He appears in this volume only as a name in the battle roster at 8(10) and as a follower of Śiva at 6(18).3 If the pronoun refers to him, the promise is outstanding.
The pronoun is genuinely ambiguous and Debroy's reading is defensible — it is also, characteristically, the reading on which the apparatus has nothing left to do. This build reports the ambiguity and does not adjudicate it.
The four pointing to Skandhas 1–3 deserve separate treatment, because they are a different instrument.
"The Second Skandha." "In the Second and Third Skandhas." "There are units for measuring time and lava is one of these. The details have been given in the Third Skandha." And the fullest of them: "The sage's name wasn't Ajagara. From the third chapter of the First Skandha, we can deduce that this was Dattatreya."4
The last is not a cross-reference at all. It is an identification — the volume gives an unnamed sage, and Debroy uses material from Volume 1 to name him, and shows his working. This build's page on that figure treats him as deliberately unclassifiable, which is what the text does; the annotator supplies a name the text withholds.
That is worth flagging as a category the vault should watch. A backward reference that resolves an ambiguity the passage was maintaining is doing something more than navigation, and it can quietly undo a feature of the narrative — here, a sage whose entire discipline is being unidentifiable.
The other three are ordinary orientation: a technical term, a doctrine, a measure, each located in the earlier volume.
Check where an apparatus is dense before trusting its coverage. A cross-reference system that is thorough for three sections and absent for three others is not a system with gaps; it is a system with a scope, and the scope is usually undeclared. The same applies to indexes, glossaries and footnote layers generally.
Distinguish a promise kept from a promise resolvable. The nine forward references are kept within this volume. The one pointing to the Twelfth Skandha is resolvable only by someone holding a different book. For a vault working from one volume, those are different objects and should be cited differently.
Recount inherited figures. The PRD's seven promises, 6/7 discharged was a plausible summary that a recount does not support. Neither the numerator nor the denominator survived, and the actual distribution — front-loaded, then nothing — is a more interesting finding than the one it replaced.
And treat backward references as orientation, not as citation. The four pointing to Skandhas 1–3 are telling a reader where they have arrived from. They are not evidence that this volume depends on those Skandhas for its arguments.
The apparatus is front-loaded. Dense at the entry point, absent later. Orientation rather than coverage.
Forward and backward references cluster together. Both types stopping at the same point is the strongest evidence that the pattern is about the reader's position rather than about the material.
A promise's referent is ambiguous and the note resolves it conveniently. Not dishonest, and worth marking — the resolution that closes the loop is the one an annotator will reach for.
The densest material has the thinnest annotation. Skandha 9's genealogies. Where the section most in need of cross-referencing has none, the apparatus is not tracking need.
Strongest support. All fourteen references were extracted by pattern from the endnote layers of Skandhas 5–10 and classified by direction against the Skandha they appear in; the tally is reproducible and the nine forward in-volume references are quoted.
Tension the page preserves. The three available explanations for the stop — material, attention, design — are not separable from inside the text, and the page does not choose.
Absence-claim discipline, and it matters here. Zero cross-references in the Skandha 8, 9 and 10 endnote layers is a claim about a pattern search for the string Skandha across those files. A cross-reference phrased without the word — this is told later, see the account of — would not have been caught. The claim is none found by this method, not none exist.
Open question. If the front-loading is deliberate orientation, the same pattern should appear in Debroy's other Purāṇic volumes — heavy annotation at the start of each volume, thinning after. The vault holds the Śiva Purāṇa build and a concurrent Vol. 1 ingest; the test is cheap and has not been run.
The vault should hold this page and its predecessor together, because they are the two halves of one object: debroys-simplest-reading states the declared method, and this page and the contradiction page show the apparatus that method produced.
What emerges across the three is a translator who is rigorous about disclosure and uneven about coverage. He flags contradictions, admits hard passages, discloses supplied subjects and assigned punctuation, reverses one consensus reading and says so — and he annotates the first three Skandhas of the volume with a cross-reference system he then stops maintaining.
Those are not in tension. Disclosure is a habit of individual notes; coverage is a property of a project. A scholar can be scrupulous in every note he writes and still write fewer of them by Skandha 9.
The consequence for the vault is practical rather than critical: the endnote layer is reliable where it exists and is not evenly distributed, so an absence of annotation in Skandhas 8–10 is not evidence about the text.
No load-bearing cross-domain handshake. This page's subject is the coverage pattern of one translated volume's cross-reference apparatus, and its connections — to the declared method, to the disclosure layer, to the passages the promises point at — are all inside eastern-spirituality and are carried in Connected Concepts.
A handshake could be manufactured here to something on documentation decay or on editorial attention, and it would be decorative. The standing rule is that a connection which does not produce an insight neither domain generates alone should be cut, and this one would not.
Sharpest implication. Fourteen Skandha cross-references, nine of them forward and landing inside this volume — and every single one sits in the endnotes of Skandhas 5, 6 and 7. The apparatus stops dead before the longest and most cross-referable Skandha in the book.
Generative questions.