Before any of this was written down, somebody had to carry it.
That job belonged to the sutas — bards, minstrels, raconteurs, professional rememberers.1 They held the genealogies of gods and kings, the ballads about famous men, the tales and songs and lore. Not as a hobby. As a livelihood and an office, "from remote times," with a duty attached.2
They were the library. There was no other copy.
And then, over some number of generations, the material stopped being theirs.
Pargiter states it in one clause, quoted approvingly in Debroy's Introduction, and the clause is remarkably direct:
It thus passed into the hands of brahmans, and their appropriation and development of it increased in the course of time, as the Purana grew into many Puranas, as Sanskrit learning became peculiarly the province of the brahmans, and as new and frankly sectarian Puranas were composed.3
Appropriation. Pargiter's word, in 1922, not a modern gloss.
The transfer isn't described as a seizure, and reading it that way misses the interesting part. It happened through the ordinary operation of teaching.
The chain given is specific. Vyasa compiles a Purana from the tales, songs and ancient lore that had come down through the ages, and entrusts it to his fifth disciple — Romaharshana, a suta. Romaharshana teaches it to his son Ugrasravas, who appears as the reciter in some surviving Puranas. So far the material is in suta hands and being transmitted father to son.
Then one more line: Romaharshana also taught it to six disciples, and at least five of them were brahmanas.4
That's the pivot. One teacher, one generation, a class ratio of five to one — and afterwards the material exists in both lineages, but only one of those lineages is on the rising side of a curve where "Sanskrit learning became peculiarly the province of the brahmans."
Nobody took anything. The sutas kept reciting; Pargiter says they "still retained the right to recite it for their livelihood." They just stopped being the only people who could, in a world increasingly organized so that literate authority sat elsewhere.
The word Pargiter reaches for after appropriation is development, and the pairing is exact.
Once the material is in the hands of a class with growing control over Sanskrit learning, new Puranas get composed. The corpus grows from one to many. The later ones are — Pargiter's phrase — "frankly sectarian."5
So the transfer isn't a single event that could be pointed at and dated. It's a compounding advantage. Each new composition enters the corpus from the new side. Over centuries the proportion of the literature that originated with bards shrinks, not because anything was removed, but because everything added came from somewhere else.
By the time the Shiva Purana consolidates around 1000–1400 CE, the corpus it joins has been growing that way for a very long time.6
Here's the part that makes this more than a historical footnote about custody.
The Shiva Purana is still narrated by a suta.
The frame is Suta reciting to the sages at Naimisha — and Debroy notes that Suta here is Romaharshana himself, the original disciple.7 Every "Suta replied" in this volume, hundreds of them, is a brahmana-composed text putting its words in a bard's mouth.
The material changed hands. The voice didn't.
That is a strange and telling arrangement. The class that acquired the literature kept, as a permanent formal feature, the fiction that it is being told by a member of the class it acquired it from.
Two readings are available and both are probably right.
The generous one: the frame is a memorial. It preserves, in the structure of every Purana, an accurate record of where the material came from, and it kept the sutas' name attached to a corpus long after they stopped controlling it.
The harder one: a bard's voice is the warrant. These are old tales, and their authority depends on being old — on having come down through the ages rather than having been composed last century by an interested party. Retaining the suta narrator retains the aura of inheritance for material that is, increasingly, new and sectarian composition.
The frame does both jobs at once, and there is no way to separate them.
You are a suta and your father was a suta.
You know the lineages — all of them, both dynasties, the solar and the lunar, and the ballads that go with the names. You know which king comes after which, and what he did, and the song about the thing his brother did. Nobody wrote any of this down. It is in you.
Your father's teacher taught six others alongside your family, and five of them were brahmanas, and that seemed fine at the time. It was fine. Teaching is what you do with something you know.
Forty years on you notice you are being invited to fewer places.
Nothing was announced. You still have the right to recite — nobody disputes it, and you still make your living at it. But the versions being recited in the larger halls have material in them you don't recognize, arranged around a god you were not brought up to put at the centre, and the men reciting them learned it from manuscripts.
And in those versions, at the very front, there is a bard called Suta narrating the whole thing to the sages in a forest.
You are inside the text and outside the room. Both are true and neither is a mistake.
There's a practical consequence for how the vault should read this corpus, and it applies well beyond this volume.
Debroy is explicit that the Puranas evolved and were added to, that texts have been "completely restructured hundreds of years later," and that within a single Purana there are earlier and later sections.8 Combine that with Pargiter's account of who was doing the adding, and a general caution follows.
Where a Purana is sectarian, self-promoting, or engaged in ranking itself against rivals, that is precisely the material most likely to belong to the later, appropriated, "frankly sectarian" layer — not to the old bardic inheritance whose antiquity gives the genre its authority.
The Shiva Purana's opening chapters are exactly that kind of material. And, per Debroy's endnote 1, some of them aren't even from this text.9
The old tales and the arguments about which text is best are not the same stratum, and the frame narrative encourages you to hear them in one voice.
It's worth being concrete about the inheritance, because "tales and songs" undersells it.
Pargiter's list is specific: the genealogies of gods, rishis and kings, and traditions and ballads about celebrated men.14 Genealogy is the core, and genealogy is not decoration — in a world where legitimacy runs through descent, the man who knows which king came after which is holding the record that settles claims.
Two of the five topics that formally define a Purana are genealogical: vamsha, the lineages of gods and sages, and vamshanucharita, the lineages of kings and their conduct.15
So the sutas' material sits at the definitional centre of the genre, not at its edges. Whatever else a Purana acquired later — cosmology, ritual manuals, sectarian argument — the thing that makes it a Purana at all is substantially what the bards were carrying.
One small thing about the quotation deserves flagging, because it affects how much weight the page can bear.
Pargiter wrote appropriation in 1922, in a book Debroy calls one of the best introductions to the corpus. Debroy quotes the passage in full, without comment, hedge or counter-argument, and follows it by saying that to explain the transmission process "one can do no better than to quote him."16
Neither man treats the word as inflammatory. It appears in a sentence otherwise concerned with ordinary historical development, alongside "and their development of it increased in the course of time."
That's the register to hold this page in. Not an accusation recovered by a modern reader against the grain of the sources — a description the sources themselves give plainly, which happens to sound sharper now than it did when it was written.
The evidence is Pargiter, at length, quoted approvingly. Debroy calls his 1922 book "still one of the best introductions to this corpus" and says that to explain composition and transmission "one can do no better than to quote him."10 So this is a translator endorsing a century-old reconstruction. [SCHOLARLY — translator's introduction, quoting Pargiter 1922]
Tension one — the source is old. A 1922 reconstruction of a transmission that supposedly happened millennia earlier is a chain of inference, not a documented event. Debroy's endorsement is real but he is a translator, not a historian of the corpus, and he cites no more recent work on the question. This page should be held as the account the vault's source endorses, not as established fact.
Tension two — Pargiter himself declines a key question. He writes that whether Vyasa composed the original Purana or merely superintended its compilation "is immaterial for the present purpose."11 That is a large thing to set aside, and it means the chain begins on an explicitly unresolved link.
Tension three, and it's unresolved in Debroy. He reports Pargiter's view that the Mahabharata was composed after the original Purana, notes this runs contrary to popular perception, and then says the linear picture "is too simplistic, since texts evolved in parallel, not necessarily sequentially."12 If they evolved in parallel, the clean custody sequence — sutas held it, then it passed — is also too simple, and Debroy does not follow that implication.
Open question. Did the sutas leave any trace of their own view of this? The account we have is entirely from the receiving side, preserved in brahmana-transmitted texts. What a suta lineage would have said about the transfer is not recoverable from anything in this volume.
The vault's eastern-spirituality corpus has essentially no material on who owned this literature, which makes this page load-bearing rather than decorative.
The closest existing point of contact is A Tantra Is a Text (Reclaim the Source), where Selvalingam argues that what circulates as "tantra" isn't in the tantras and that the remedy is to go back to the vetted corpus.
That argument assumes the corpus is a stable object with a settled provenance. Pargiter's account says the corpus is the output of a custody transfer and several centuries of interested composition afterwards. Going back to it is still better than not — but "the source" turns out to be a thing with a history, and part of that history is about who got to add to it.
There's also a quieter convergence worth naming. The vault holds a good deal of material on lineage transmission, and it is uniformly written from inside the tradition, where a lineage is a guarantee of fidelity. Here is the same structure viewed from outside, where a lineage is also a distribution channel — and channels can change hands.
The Founder Problem in Historical Perspective (history). The history corpus examines how movements construct a founder retrospectively, and how the constructed origin serves the institution doing the constructing rather than the historical record.
The parallel is Vyasa, and it is unusually clean. He is credited with compiling the original Purana; Debroy notes that "Vedavyasa isn't a proper name" but a title held twenty-eight times.13 A named individual founder who is actually an office is the founder problem in its most explicit possible form — the tradition is telling you, in the definition of the word, that this is a role rather than a man.
The insight the pairing produces, which neither gives alone: this is a founder-construction where the construction is disclosed and the authority survives anyway. The history pages generally treat the constructed founder as a story that works because it is believed literally. Here everyone knows Vyasa is a title. The material still gets its warrant from him. That suggests founder-authority may not depend on historicity at all — it may only need a stable name at the head of a chain, and the chain can be openly institutional without losing force. Worth testing against the vault's other founder cases.
The Deliberate Incompleteness (business). The Brunson ingest read a manual as engineered to leave its reader dependent, and the vault built that as an authorial act with a commercial function.
The mechanism here runs over centuries and has no author. Nobody decided that a bardic corpus should become a brahmana one. A teacher taught six students, five of whom happened to be from a class whose control of Sanskrit learning was rising, and the compounding did the rest.
What the pairing produces: the vault's commercial and behavioral pages assume a designer, and this is a clean case of an identical outcome arising from drift plus a ratio. Ownership of a body of knowledge transferred without a transfer event, through nothing but differential access to the medium it was moving into. That is a category the vault currently has no page for — capture by literacy rather than by intent — and it is likely to be the better model for a great deal of what the behavioral corpus reads as strategy.
Sharpest implication. A literature changed owners without anyone taking it. One teacher, six disciples, five of them from the class whose control of Sanskrit learning was rising — and afterwards every addition to the corpus entered from the new side, until the proportion originating with bards shrank to nothing. Pargiter's word for it is appropriation, and his companion word is development, and the two are the same process. And the acquiring class kept the bard's voice as the permanent narrative frame, so that the corpus still opens, a thousand years later, with a suta telling the story.
Generative questions.