A man hears a story about a friend and cannot make the arithmetic work.
The story is that Kṛṣṇa killed Naraka, freed the women held in his fort, and married all of them — sixteen thousand — and that he is now living with every single one of them, in her own house, at the same time.
Nārada does not object to it. He does not call it false. He does what a certain kind of person always does with a claim like that: he gets on the road to go and look.
The text says so in the plainest words it has:
"Narada heard about Naraka's death and that Krishna had alone married several women. He wished to see him. It was extraordinary that though he possessed a single body, he simultaneously and separately married sixteen thousand women. Though alone, he dwelt in all their houses. Eager to witness this, the devarshi went to Dvaravati."1
⚑ Read what that sentence is doing. It states the source of the information (heard about), the claim (sixteen thousand, simultaneously, separately), the element that makes the claim implausible (one body), and the reason for the journey (eager to witness). That is the shape of a verification, not the shape of a pilgrimage.
He goes house to house. Not to one house. Not to the palace. To the houses.
And in each one the volume gives him something different to see.1 In one, Kṛṣṇa is fondling an infant. In another, offering oblations into fire. In another, eating the remains of a meal. In another, practising with a sword. Elsewhere: consulting with his ministers, arranging his children's marriages, having wells dug, sitting with brahmanas, being fanned, hearing petitions.
The audit runs long enough that the list stops being a list and starts being a working week.
And then the clause that changes what kind of scene this is:
"In some places, without any signs of who he was… roaming around in the inner quarters and homes as an ordinary person, wishing to know what people were thinking."1
⚑ The inspector walks in on the inspected running an inspection. Nārada has come unannounced to find out whether the reports about Kṛṣṇa are true. He finds Kṛṣṇa moving through his own city in plain clothes, unrecognised, to find out what his own people are thinking.
Two audits, nested, and the volume does not remark on the doubling.
Before the houses, the chapter takes an inventory. Not a rhapsody — an inventory.
"There were nine hundred thousand palaces, made out of crystal and silver… The city was laid out with main roads, smaller roads, quadrangles, markets, assembly halls and shrines for the gods… It was ornamented with sixteen thousand residences, for each of Shouri's wives."2
Count what is in that sentence. Housing stock. Road hierarchy — main roads and smaller roads. Quadrangles. Markets. Assembly halls. Shrines. And then the sixteen thousand residences, listed as one line item among the markets and the meeting halls.
The wives are infrastructure here. Not romance, not spoils — a housing programme with a number attached.
⚑ That register is consistent for this whole stretch of the volume. Dvārakā is founded as a planned settlement laid out "in conformity with principles of vastu", with treasuries, stores and upper-floor shrines, "full of people from the four varnas." Its jewel is described by its output — gold per day, and the abolition of famine and disease within range. Its wars are described as a problem of army numbers. The Dvārakā material is administrative writing, and this chapter is its census.
Here is the part that a devotional reading tends to skip.
Nārada came to check an implausible claim. He checked it thoroughly, in every household, over a stated period. And the audit returns nothing anomalous. Kṛṣṇa is complete in every house. There is no seam, no thinned-out version in the four-thousandth mansion, no wife who got a lesser copy.
A finding of "everything is exactly as reported" is a strange thing for a story to bother with. It has no plot value. Nobody is caught, nobody is exposed, nothing turns.
⚑ It has evidentiary value, and that is the only thing it has. The chapter exists to put a sceptic's name on a report.
And the sceptic is not a neutral one. Nārada is the volume's most active information vector: he is the man who walks into Kaṃsa's court and tells him which child is which, triggering the entire kill-plan; he is the man who recites Kṛṣṇa's remaining career in order, fifty chapters early; he is the man whose physical description of Kṛṣṇa lets a foreign king identify him on a battlefield and disarm himself on principle. The volume's chief carrier of information is also the man it sends to test information.
And here is the thing the chapter withholds, which is easy to miss because the chapter is so busy.
Nārada watches. He does not interview. In sixteen thousand houses he never once turns to the man and says: how are you doing this?
That makes his method purely observational, and it separates him sharply from the volume's other two sceptics. Akrūra provokes — he delivers a sermon nobody requested and reads the king by what the sermon fails to do to him. Bhṛgu intervenes — he withholds a bow, raises a foot, and measures the response. Both of them do something to the subject and grade the reaction. Nārada does nothing to anybody. He looks.
⚑ Observation confirms that. Provocation is what gets you what kind. And the one question this chapter's whole premise raises — one body, sixteen thousand households, how — is the one question its investigator declines to put.
Which may be piety, and may be method. The volume elsewhere insists that the answer to how is unavailable in principle: its own scripture, personified, ends by conceding that it reaches him only by negating every assertion about him. An investigator who already believes the mechanism is undescribable would run exactly this audit — count the outputs, never ask about the engine.
You hear a claim about someone's capacity that does not add up. You do not argue with it and you do not accept it. You name the exact element that is implausible — not "that sounds like a lot" but one body, sixteen thousand households, simultaneously — and then you go and look at that element specifically.
You want to know what your operation is actually like. You do not ask the people running it. You walk through it yourself with nothing on that identifies you, at a time nobody expects, and you watch what happens when the person in front of you thinks you are nobody.
Somebody arrives to inspect you on the same afternoon you are inspecting yourself. Notice that neither of you finds this odd, and notice that only one of you writes it down.
You run the check, and the check comes back clean. You publish that anyway. A verification that confirms is worth exactly as much as one that catches something, and it is the only kind most people never bother to record.
Strip the theology out and look at the two operations side by side.
Nārada's operation. Trigger: a report received second-hand. Target: a specific factual claim with a named implausible component. Method: unannounced, exhaustive, household-by-household observation. Output: confirmation, reported.
Kṛṣṇa's operation. Trigger: unstated. Target: "what people were thinking." Method: appearing in the inner quarters and homes of his own city "without any signs of who he was." Output: never reported.
⚑ The second one has no result in the text at all. We are told he does it. We are never told what he learns. The volume gives its protagonist a standing intelligence practice and never once tells us what it returns.
That silence is worth sitting with, because the volume is not usually shy. It prints Balarāma calling Kṛṣṇa's conduct contemptible. It prints a gopī's three-count indictment of him drawn from three separate lifetimes. It prints Śiśupāla's procedural case that he has no varṇa, no āśrama, no lineage. A text this willing to print what people say about him to his face declines to tell us what he heard when they did not know he was listening.
Two readings are available and the volume supports neither over the other. Either the practice is a courtesy — a king who checks on his people — or it is exactly what it looks like, a ruler who has arranged to hear the unguarded version. The chapter reports the method and withholds the finding, which is the one combination that keeps both readings alive.
The pattern this chapter belongs to is the strongest structural thing in the volume, and it needs three separate people to see it.
Akrūra, at 10(49). Sent to Hastināpura "to ascertain his present conduct, whether it is virtuous or wicked." Lives there for months. Collects testimony from Kuntī and Vidura. Delivers a sermon the king did not ask for, watches the king agree with every word of it and change nothing, and reports back.
Nārada, here at 10(69). Sent by nobody — self-dispatched — against a specific implausible report, with the implausible element named in advance.
Bhṛgu, at 10(89). A sacrifice full of ṛṣis cannot settle which of the three lords is greatest, so they "sent Bhṛgu" to find out. He applies the same deliberate insult to all three subjects, returns, describes the trials in detail, and "their doubts were dispelled."
⚑ [n≥3 PATTERN] — this volume repeatedly stages verification by a delegated sceptic, and prints the protocol as well as the finding.
A fourth instance sits in the frame rather than the story. At 10(75) Parīkṣit does not ask an open question; he cites a version he already holds — "That is what we have heard" — and asks for the discrepancy to be explained. The listener is auditing the teller.
⚠ Stated at the register the sample carries: four instances in one volume, three of them narrated episodes and one of them a framing device. No claim is made about the other volumes of this Purāṇa or about the genre.
The strongest support is the motive clause. The chapter does not need it. A visit from Nārada would have run perfectly well as a visit. Instead the text supplies the hearsay, the claim, the implausible component and the word "witness" — four pieces of apparatus that exist only to make the trip a test.
A second support, from the other end. The house-by-house catalogue is longer than the story requires. If the point were he is everywhere, one household would carry it. The point is he was checked in each one, and length is how a survey demonstrates coverage.
The tension, preserved. A verification is only worth what the verifier's independence is worth, and Nārada is not independent. He is a devotee of long standing, he leaves praising, and he is the same person whose information starts the Kaṃsa massacre and puts an unarmed Kālayavana in front of Kṛṣṇa. The volume runs an audit and staffs it from inside the company. ⚠ Note this cuts both ways and the page does not resolve it: an inside auditor with everything to lose from a bad finding is compromised, and an inside auditor is also the only one who would know what to look for.
A second tension. The chapter's proof-of-omnipresence rests on Nārada being unable to catch a seam. That is an absence-based finding, and an absence-based finding from a sympathetic observer is the weakest form the volume's own verification apparatus produces — weaker than Bhṛgu's, which measures a response, and weaker than Akrūra's, which produces a self-incriminating speech.
Open question. The undercover practice is stated and never resulted. Does the corpus anywhere report what a ruler learned by going unrecognised among his own people? Not found in the range read (Skandhas 10–12 of this volume); no claim is made about Volumes 1–2.
Set against The Women Who Chose Him Themselves, this chapter is the accounting version of that one's emotional claim. There, sixteen thousand women free themselves from a fort, choose him, and are married in one muhūrta in separate mansions. Here, a sceptic counts the mansions.
⚑ The same fact is a wonder in one chapter and a data point in the next, and the volume changes register without signalling it.
Against The Devotee Who Started the Massacre, the identical figure is doing the identical thing to opposite effect. There he carries accurate information to a tyrant and the information gets babies killed. Here he carries scepticism to a friend and the scepticism produces a testimonial. Nārada is a transmission channel in both, and the volume never treats the channel as the problem.
Against Ten People Look at Him and All Ten Are Right, there is a real friction here worth keeping open. That page reports the arena, where a single body is perceived ten incompatible ways and each perception is called accurate — perception indexed to relationship. Here the multiplicity is not in the seeing but in the man, and Nārada's whole method assumes it can be checked from outside. If perception were relationship-indexed all the way down, there would be nothing for him to verify. The volume runs both models and does not reconcile them.
Secret Shop Your Own Business — Baseline Discovery — you cannot see your own operation from inside it, so you send someone through it as a stranger and watch what actually happens rather than what the training says should happen.
Both halves of this chapter are that move, run in opposite directions on the same afternoon. Nārada secret-shops Kṛṣṇa. Kṛṣṇa secret-shops Dvārakā. Neither announces himself; both are counting on the gap between the performed version and the unwatched one.
⚑ What neither field gives alone: the business version assumes the gap is the finding, and this chapter reports no gap. Hormozi's whole premise is that what you'll find is horrifying — the distance between the trained script and the live call is the payload. Here the distance is zero, in every one of sixteen thousand households, and the zero is the result being published.
That inverts the instrument's usual meaning. A secret shop that comes back clean is normally treated as a failed test — you didn't look hard enough, you got a good day, the vendor recognised your friend. The volume treats it as the strongest possible finding precisely because the sceptic went in expecting a seam and named the seam he expected before he went. Which is the one condition under which a null result carries weight, and it is the condition Hormozi's version never has to specify because nobody secret-shops their own business hoping to find nothing.
The Probe Attack as Intelligence — the Mongols fought an opening engagement whose only purpose was to return a reading on the opponent's strength and morale. Men died and nothing was taken; the battle's product was information.
Nārada's journey is the same operation with the cost removed. The probe is a visit. Nothing is destroyed and nobody is hurt, and the output is identical in kind: an operation undertaken for no reason except that it returns a reading.
⚑ The insight the pairing produces: the probe attack's expense is what makes its result trustworthy, and that is exactly what this chapter cannot supply. A commander who spends men to learn something is understood to want the answer more than he wants to be right — the price is the proof of good faith. Nārada spends nothing. He travels to a friendly city, is welcomed, sees what he hoped to see, and leaves. His finding is free, and free findings are the ones a reader should discount hardest.
Set that against the volume's other two sceptics and the ranking is clear on the volume's own terms. Bhṛgu risks a curse from three deities and takes a kick's worth of consequence. Akrūra spends months in a hostile court and provokes a king to his face. Nārada goes to look at something lovely. The volume's verification apparatus is real, and this is its cheapest instance.
Sharpest implication. A devotional text sends an auditor to check whether its own central miracle is true, names in advance the element it finds implausible, prints the method in full — and the auditor walks in on the subject running an unannounced audit of his own city. The result is a confirmation, from an insider, at no cost, in a chapter whose real business is a housing inventory.
Generative questions.