Here is the order of events in chapter 38, and it is worth writing out plainly because it sounds wrong the first time.
There is nothing. Then there is a formless absolute. Then the absolute makes itself a form. Then that form makes a goddess.
Then the two of them build a city.
Then, after the city, they decide someone should be created to handle the actual universe — the making of it, the running of it, the ending of it — so that they can get on with enjoying the place they have just built.1
The city comes first. Before the elements, before the worlds, before Vishnu, before anything the city could plausibly be in.
And it is not an allegory or a heaven with a poetic name. Debroy's note on the word is one line: Kashi.2 The city on the Ganga. The one you can catch a train to.
The sentence is compact and every clause in it is doing something:
The brahman assumed the form known as Kala and along with Shakti, created the kshetra known as Shivaloka. This is the supreme kshetra known as Kashika.1
Three things happen at once here.
Shivaloka and Kashi are declared the same object. Not corresponding, not reflecting — the same. The god's world and the town are one item under two names.
To do it, the brahman takes the form of Kala — time, and also death. Which is peculiar placement, given that this same chapter has just told you there is no night or day.3 Time is put on like a coat, specifically in order to build a place.
And it is a kshetra — a field, a sacred zone, the technical term the Vidyeshvara Samhita already spent a whole chapter grading and mapping.4 So the vocabulary of pilgrimage is being applied to something that predates the existence of ground.
There is no world yet. There is a city.
Then the claim that does the real work:
Even at the time of pralaya, that kshetra is never freed from Shiva and Shivaa. Therefore, it is known as Avimukta.1
Debroy glosses it: literally, never forsaken.5
Go back to the beginning of the same chapter. The dissolution passage removed the suns, the moon, day and night, fire, wind, earth, water, Pradhana, every sense-object, and finally direction itself. A darkness a needle cannot get into.3
One thing is now exempt.
That is an extraordinary carve-out and the text makes it in a subordinate clause. The universal solvent has an exception, the exception is a specific municipality, and nobody stops to remark on it.
Whatever else Kashi is, it is now the only object in this cosmology with a guarantee. Everything else in existence is on a lease that expires. The city is freehold.
The next sentence is stranger still, and it usually gets read past.
Earlier, the wielder of Pinaka gave it the name of Anandavana, since the kshetra bestows bliss. It came to be known as Avimukta later.1
Earlier. Later.
We are still before creation. There is no "later" available. And yet the place has a name-history — an original name, a reason for it, and a subsequent renaming with a different reason.
Two things follow from that.
First, the same time-leak as elsewhere in the chapter. A narrator telling a story has to say earlier and later; the grammar of telling brings sequence with it whether or not the content permits it.6
Second, and more interesting: the renaming is an argument. Anandavana — forest of bliss — says what the place gives you. Avimukta — never forsaken — says what the place is guaranteed. The volume elsewhere uses etymology as a way of installing doctrine, deriving Shiva from three syllables and licensing the whole method with a pun on a lion.7 Here it does it twice on the same city, and the upgrade runs from this is nice to this is permanent.
That is the direction a property claim strengthens in.
Set the devotional reading aside for one section and read the passage as a document with an interest.
The Vidyeshvara Samhita already built a full sacred geography — rivers graded by their number of mouths, which world each bathing place bestows, an astrological calendar for when to go.4 That is a pilgrimage economy written down: a ranked list of destinations, each with a stated return.
This chapter does something categorically different. It does not add Kashi to the top of that list. It removes Kashi from the list's logic entirely by dating it before the list's subject matter exists and exempting it from the list's eventual destruction.
Compare the two claims as an ordinary person would hear them.
"Bathing at this confluence gives you Brahma's world" is a good offer. It is also one of many, and the passage that makes it goes on to make several more.
"This city was here before the sun and will be here after it" is not an offer at all. It is a category change. It cannot be outbid, because everything that could outbid it is on the list of things that get dissolved.
And unlike the graded list, it costs nothing to maintain. No priest has to certify it, no calendar governs it, no rival tirtha can produce a better number. It is a single sentence in a subordinate clause and it settles the market permanently.
The most valuable line in the volume's entire sacred geography is the one that is not part of the sacred geography.
You are standing on the ghats at five in the morning because someone told you to see it at dawn.
It is not beautiful in the way you were led to expect. There is rubbish. There is a diesel generator running somewhere behind you. A man is brushing his teeth in the river. The famous burning ground is a working site with logs stacked and priced, and the smoke goes where the wind puts it, including over you.
And the thing you are supposed to be feeling is not arriving.
Now try the chapter's claim on, not as a belief but as a frame. This was here first. Not first among cities. First — before the elements, before the sun, before there was anywhere for it to be.
Nothing changes in front of you. The generator keeps running.
But the sequence inverts. You had been standing in a world, in a country, in a city, at a river. The chapter says: no, you are standing in the original site, and everything else — the country, the elements, the sun — was added around it afterwards to give it a setting.
That is what the passage is engineered to do, and it works on the ghats better than it does on the page. It does not make the place more impressive. It makes everything else provisional.
Then someone asks you for two hundred rupees and it wears off.
Say the uncomfortable thing plainly, since the vault's job is to hold what the source is doing and not only what it says.
Kashi is a real city with a real economy, and that economy runs on the belief that dying there is better than dying elsewhere. Priests, boatmen, wood-sellers, guesthouses, funeral specialists — a whole trade, sustained by a metaphysical claim about one location.
Chapter 38 is that claim in its strongest possible form, placed in the mouth of Brahma, inside a conversation with Narada, inside a bard's recitation, at the opening of the Rudra Samhita.6
This does not make the passage insincere. Devotional literature is written by people who live somewhere, and a text composed in or for a place will say that place matters. That is ordinary.
What is worth flagging is the structure of the argument, because it is unusually well built. The claim is placed before creation, which puts it beyond any evidence. It is exempted from dissolution, which puts it beyond any future. It is fused with Shivaloka, which means denying the city denies the heaven. And it is attached to the god's own naming, twice.
Four independent supports, none of which can be checked, all in one paragraph.
Whoever wrote this understood exactly what they were securing.
The volume is disputed in ways that make this passage more interesting rather than less.
Hazra treated the Śiva-purāṇa not as a Mahāpurāṇa at all but as a late sectarian Upapurāṇa raised by "a comparatively late sectarian zeal." Wilson suggested it might be identical with the Vāyu. Debroy declines to adjudicate any of it.8
Hold that beside chapter 38.
A text of contested canonical status, whose elevation is attributed by a leading scholar to sectarian enthusiasm, contains a passage asserting that one particular city predates the universe and survives its end.
That is precisely the kind of passage a "comparatively late sectarian zeal" would produce. It is also precisely the kind of passage a genuinely ancient tradition centred on that city would produce. The evidence is the same either way, which is exactly why Debroy will not decide and why this page must not either.
What can be said without choosing: any page in this vault that treats the Śiva Purana's claims about Kashi as equivalent in standing to, say, the Vishnu Purana's claims about anything is taking a side in a live dispute, and must say so.8
Strongest support is the order of chapter 38 itself — absolute, form, goddess, city, and only then a being made to produce the universe — together with the explicit pralaya exemption and Debroy's flat gloss of Kashika as Kashi.
Tension one, and it is a flat contradiction inside the volume. The Vidyeshvara Samhita builds a ten-fold multiplier ladder of sacred places — house, cowpen, riverbank, mountain summit — and then voids it: "the place where the mind is at peace is the best."9 Fifteen chapters later the Rudra Samhita makes the strongest geographical claim available in the entire literature. The ladder is released and then a single rung is made eternal. Preserve unresolved.
Tension two. "Earlier" and "later" are used of a period the chapter has established has no time. The renaming from Anandavana to Avimukta happens in a sequence that cannot exist.
Tension three. The brahman assumes the form of Kala, time, in order to build the city — nine lines after the passage abolishing time. The text does not connect the two.
Tension four. The pralaya exemption is stated in a subordinate clause and never revisited. The most consequential metaphysical carve-out in the volume is delivered as an aside.
Open question. Volumes 2 and 3 of this translation exist and are unread. Does the Kashi claim recur, get elaborated, or get quietly dropped? A claim this valuable that appears once would be a very different object from one the corpus repeats — and this is cheap to check when those volumes are read.
The vault's existing sacred-geography material and this passage are pulling in opposite directions, and the tension is productive rather than embarrassing.
The Place Multiplier and Its Collapse documents the Vidyeshvara Samhita voiding its own geography with a last-sentence escape clause — the volume's signature move, found eight times across two Samhitas.
The Bilva Tree goes further in the same direction: every tirtha on earth is present at the root of a common scrub tree, which abolishes the pilgrimage economy outright by distributing it.
And The Kshetra Map of Bharata holds the ranked destinations that both of those undercut.
Chapter 38 has no escape clause. No "but the place where the mind is at peace is best." No distribution. The one geographical claim in the volume that arrives without its own antidote is the one attached to a specific, named, monetisable city.
That is a pattern worth stating: the escape clauses appear where the claim costs the institution nothing, and vanish where it does. The volume releases you from the ladder, from fasting, from caste-assigned lingams, from ritual timing — and does not release you from Kashi.
Akhenaton's Shadowless City (cross-domain). A pharaoh set out to build a city engineered so completely with light that nothing inside it could cast a shadow — a whole municipality as a theological argument, dedicated to the sun-disk. The priests of the displaced gods killed him with the foundations still being poured, and the city stands in the record as a thing dreamed and abandoned.
Set that beside a city whose defining name means never forsaken.
Both are cities that exist to make a doctrinal claim rather than to house people conveniently. Both were produced by a sectarian movement asserting one deity against an established pantheon. And they are exact opposites in outcome: one was abandoned within a reign, and the other is named for the impossibility of being abandoned — a name the text derives specifically from its survival of the universe's own destruction.
The insight neither gives alone: a doctrinal city's real vulnerability is not military, it is that the doctrine can be outlived. Akhetaten fell because it was tied to one man's theology and one man's lifespan. Kashi's chapter solves precisely that problem, and you can watch it solving it — by dating the city before any founder could have founded it and exempting it from any end. This is a defence engineered against the exact failure mode the Egyptian case demonstrates. Which gives a general test for reading founded sacred sites: ask what the founding story does about the founder's mortality, because the ones that survive are usually the ones that wrote the founder out.
Mandala City — Urban Planning (cross-domain). That framework's core complaint is that modern cities are chaotic because they have no centre and no axis mundi — and its model city is built outward from a permanent central fire, with everything else arranged as neighbourhoods around it.
Chapter 38 is that principle taken to its limit. The centre is built first and the world is arranged around it afterwards — not as a design philosophy but as literal chronology. There is a plaza before there is a planet.
What the pair produces: a centred city and a decentred one are not two arrangements of the same material. They encode opposite claims about what is provisional. In a centred city the periphery is contingent and the centre is given; the further out you live, the more your location is a matter of circumstance. Chapter 38 pushes this until the entire universe is periphery. And that explains something the urban framework only asserts — why a centre has to be built first rather than designated later. A centre appointed after the fact is an administrative decision and can be revoked; a centre that was there before the map cannot. Which is a real constraint on anyone trying to give an existing chaotic city a centre now: the move is not available retroactively, and every tradition that has a genuine axis mundi got it by claiming priority, not by choosing well.
Sharpest implication. In chapter 38 the city is built before the universe — after the absolute, after Shakti, and before the being created to make and run the worlds — and it is identified flatly as Kashi. Then, in a subordinate clause, it is exempted from the dissolution that the same chapter's opening page removes literally everything else from, and named Avimukta, the never-forsaken. That is four independent, uncheckable supports for a single property claim, stacked in one paragraph: dated before evidence, exempted from the future, fused with Shivaloka so that denying the city denies the heaven, and named twice by the god personally. And it arrives in the same volume that had already released its readers from the ladder of sacred places with "the place where the mind is at peace is the best." The volume hands back every geography except the one with a working economy attached.
Generative questions.