Yadu meets a brahmana wandering with nothing, and the question he asks is not about doctrine. It is about a discrepancy he can see.
You are capable, wise, accomplished, handsome, and your speech is like amṛta — and you want nothing. Everyone else is scorched in the fire of desire and avarice. You are not. You move around "like a dumb or crazy person, like a pishacha."1
⚑ The question is empirical. A man with every advantage is behaving like a man with none, and Yadu wants the mechanism.
The avadhūta answers by naming his teachers, and he opens with a count and a commitment:
"I have accepted twenty-four as my preceptors… My knowledge and conduct are derived from whatever I have learnt from them. O Yayāti's son! O tiger among men! I will tell you exactly what I have learnt from each of them. Listen to what I have to say."1
⚑ Exactly what I have learnt from each of them. That is a stated undertaking, and it is what makes the rest of this page a finding rather than a quibble.
Here is the enumeration as printed, because the arithmetic depends on it:
"the earth, the air, space, water, fire, the moon, the sun, the pigeon, the python, the ocean, the moth, the bee, the elephant, the gatherer of honey, the deer, the fish, Pingala, the osprey, the child, the maiden, the maker of arrows, the snake, the spider and the wasp."1
Twenty-four names. Five elements, two luminaries, eleven creatures, two human exempla, a craftsman, a maiden, and a bird.
Now follow the teaching.
Earth — patience under attack, and it subdivides at once: "as a student of the mountain and as a student of the tree." Air — contact without adhesion, bearing scents without being attached to them. Space — pervasion, untouched by clouds. Water — purity that acts outward, "friendly towards men and must purify them." Fire — austerity as radiance, "uses his stomach as a vessel."
And then Chapter 11(8) opens: "Like the python, a person should not make any efforts, but should eat whatever turns up on its own."2
⚑ Earth, air, space, water, fire — and then straight to the python.
The moon, the sun and the pigeon are never taught. Three consecutive entries, sitting between fire and the python in the enumeration, and the teaching passes over all three without a break, a transition, or a note.
⚠ Checked across the whole Skandha rather than the chapter. Moon occurs four times in the Eleventh Skandha and not one is a teaching passage — a simile about toenail-radiance at 11(14), the enumeration itself, an instruction about a meditation-lotus, and the vibhūti catalogue at 11(16). Sun, within 11(7)–11(9), occurs only inside the enumeration.
⚠ And the bounding is not optional. This is a claim about this EPUB as ingested. Three consecutive items dropped from a list is precisely the signature an omission at any production stage would leave, and it cannot be distinguished from a recension difference using this file. No claim is made about Debroy's Sanskrit or the print edition. ⚑ Debroy does not flag it — no endnote attaches to the gap.
The stranger fact is that the list also over-delivers.
The earth is explicitly subdivided — "as a student of the mountain and as a student of the tree" — so two teachers are taught under one entry.
And at the very end, after the twenty-fourth, the avadhūta adds one more, outside the enumeration entirely:
"Now hear about what my intelligence has learnt from my own body… my body is my preceptor."3
⚑ So the arithmetic is: twenty-four announced, twenty-one taught, and three further teachers supplied that were never counted.
The number is announced, undershot, exceeded, and never reconciled. Debroy flags neither direction.
Somebody gives you a count and then a list. Check the list against the count. It is the cheapest audit available and almost nobody runs it.
An enumeration promises completeness, and completeness is most of what it is for. Three items quietly missing does not reduce the authority of the list by three twenty-fourths — it puts the whole thing in question, which is why nobody checks.
You cite an example everybody knows. Ask yourself whether you could actually tell it. If not, you are borrowing a name, and the borrowing works precisely because the audience will not ask either.
The count is wrong in both directions at once. That is not sloppiness cancelling out. It means the number and the content were never checked against each other by anyone in the chain.
This volume does this twice, and the two instances are the same species.
At the front: the printed front-matter table lists Chapter 10(85): 59 shlokas and Chapter 10(86): 59 shlokas. Neither chapter is printed. The body runs 10(84) → 10(87) with no break and no note. And endnote 673 cross-references 10(86) as though it were present.
Here: three of twenty-four preceptors announced, promised individually, and never taught.
⚑ In both cases an enumeration is stated, the body does not deliver it, and the apparatus does not notice. One hundred and eighteen ślokas by the book's own count in the first case; three teachers in the second.
⚠ Both bounded identically to this EPUB as ingested. Both unflagged.
And a further pass over the volume's technical enumerations found them intact: the twelve yamas, the twelve niyamas, the eighteen siddhis as eight plus ten, the ten characteristics of a Purāṇa — all correct, item for item. ⚑ The failures are bibliographic and narrative; the doctrinal lists hold.
⚠ And the obvious alternative explanation is available and untested: doctrinal lists are short, closed and mnemonically drilled, while narrative counts are long, open and easy to corrupt. That would produce the same distribution with no claim about editorial care at all. The distribution is recorded. The explanation is not chosen.
There is one more thing about the pigeon specifically, and it is why the gap is worth a page rather than a footnote.
At 10(72), disguised brahmanas working on Jarāsandha deploy a seven-name exempla list: "Harishchandra, Rantideva, Unchhavritti, Shibi, Bali, the hunter, the pigeon" — invoked as figures the audience is assumed to know.4
No story is attached, there or here.
⚑ So this volume names the pigeon twice as a known teaching and supplies it neither time.
⚠ Bounded precisely: an observation about this volume. [n=2 OBSERVED]. No claim that the pigeon and the hunter are absent from Volumes 1–2 — that check is unrun, and the candidate that would state it is deliberately held unbuilt for exactly this reason. An absence-claim over unread material is not available.
The closing line of the whole discourse explains the design, and it makes the gap sharper rather than softer:
"Jnana obtained from a single preceptor is not stable and may not even be complete. Though the brahman is without a second, rishis have chanted about it in many ways."3
⚑ The number is the argument. Twenty-four is not decoration — it is a claim that knowledge from one source is unstable and must be assembled from many.
A doctrine whose whole point is the multiplicity of sources arrives with three of its sources missing. ⚠ The volume does not notice, and this page does not claim it intends the irony. What it prints is a stated principle about completeness, and an incomplete list.
The strongest support is the promise. I will tell you exactly what I have learnt from each of them is an undertaking in the text's own voice, which is what makes non-delivery checkable rather than interpretive.
A second support. The gap is contiguous — moon, sun, pigeon, in enumeration order, with the teaching resuming exactly at the next entry. Three scattered omissions would suggest carelessness; three consecutive ones suggest a block.
A third support. The Skandha-wide check. Moon and sun were traced across the whole Eleventh Skandha, not just the chapter, and no teaching passage exists for either.
The tension, preserved. ⚑ The count fails in both directions in one list, and the same volume's technical enumerations are all correct. Held as a distribution.
Open question. The volume's principle is that single-source knowledge is unstable, and its demonstration of multi-sourcing is missing three sources. ⚠ Is the number load-bearing, or is twenty-four itself conventional? Not determinable from this volume, and the answer would change what the gap means.
Set against He Was Terrified of Him Every Waking Second, the wasp — the twenty-fourth and final preceptor — is taught, and its page is built in the antagonist cluster because the volume uses the same figure there for something else entirely. ⚑ It counts as delivered in this page's arithmetic, and it is cited rather than rebuilt.
Against They Killed Him Through the One Thing About Him That Was Not Corrupt, the descending exempla list at Girivraja shares the pigeon and shares the assumption that the audience already knows it. Two lists, one shared unnarrated name, and the same reliance on common knowledge the volume never supplies.
Against The Narrator Stops the Story to Say That What He Just Narrated Cannot Be True, the charge the narrator lays at 10(77) is that transmitters "do not remember what they themselves have said." ⚑ Here is a passage where the text does not remember what it said forty lines earlier, and nobody interrupts.
Against The Forward Index, the contrast is the volume's own apparatus working. Nārada recites the remaining plot fifty chapters early and it holds, in order. The same book that keeps that appointment fails this one, and fails the one at 10(86).
Borrowed Credibility: The Naming Tactic — you name a recognised authority, client or reference, and the association does the work. No substance has to be supplied, because the name carries the credibility and audiences do not audit names.
The hunter, the pigeon is that tactic operating inside a sacred text. Both are invoked as known teachings. Neither is ever told. They function exactly as a borrowed name does: their authority comes from the audience's assumption that the story exists somewhere.
⚑ What neither field gives alone: the commercial version borrows from a real referent that could be checked, and here the referent may not exist in this text at all. A named client can be phoned. The pigeon can only be looked up in a corpus that, on this volume's evidence, does not contain it — and the volume names it twice, in two unrelated chapters, in two different speakers' mouths.
Which is what makes it more than an omission. ⚠ The page cannot say the exemplum is absent from the tradition — that check is unrun and asserting it would be fabrication. What it can say is that within this volume the name does all the work and the content never arrives, twice, and that this is the same structure as a borrowed credential whether or not anyone intended it as one.
Build an Inventory of Stories — the practitioner assembles a stock of examples in advance, indexed and ready, so that the right one can be produced on demand. The inventory is the asset; its completeness is what makes it reliable under pressure.
The avadhūta's twenty-four is exactly that: an indexed inventory, announced by count, produced on demand to a king who asked a hard question.
⚑ The insight the pairing produces: an inventory's authority comes from its completeness, and completeness is the one property nobody in the room can verify. Yadu cannot check. He has been given a number, a list, and a promise, and the teaching that follows is long, coherent and impressive. Nothing about the delivery signals that three entries went past unaddressed — the gap is only findable by holding the enumeration beside the exposition and counting.
And that is the general form of the failure, not a fact about this text. ⚠ A stated count is a promise about a set; an audience checks the items it hears and never the items it does not. The volume's other enumerations survive the same audit — twelve yamas, twelve niyamas, eighteen siddhis, ten Purāṇa characteristics — so the instrument is sound and the one that fails is the one nobody would have thought to check.
Sharpest implication. *The avadhūta names twenty-four preceptors, promises "I will tell you exactly what I have learnt from each of them", and teaches twenty-one — the moon, the sun and the pigeon are named and never taught, three consecutive entries, with the exposition resuming exactly at the next name.* He then adds three teachers nobody counted. The count is wrong in both directions in one list; the volume's technical enumerations are all correct; and the closing principle of the whole discourse is that knowledge from a single source is unstable and must be gathered from many.
Generative questions.