Deep in a chapter of future kings — the kind of passage a reader's eye slides over, name begetting name for two pages — this sentence goes past:
"Sanjaya's son will be Maya and Maya's son will be Shakya. Shakya's son will be Shuddhoda and his son will be Langala."1
Śākya. Śuddhodana. The Buddha's clan and the Buddha's father, named in sequence, in a Śaiva Purāṇa's roll of the solar dynasty, with no comment of any kind.
And the son — the one the names exist to point at — is simply not there. The list steps from Śuddhodana to Laṅgala and carries on.
⛔ Reported throughout as this text's own handling of a lineage, never as history. Nothing on this page is a claim about the Śākya clan, about Buddhism, or about anything outside these chapters.
The passage is at ch 353-6(39), inside a prophetic roll of the Ikṣvāku line, and the surrounding names give the register:1
*"His son will be Kritanjaya and Kritanjaya's son will be Sanjaya. Sanjaya's son will be Maya and Maya's son will be Shakya. Shakya's son will be Shuddhoda and his son will be Langala. Langala's son will be Prasenjit and his son will be known as Shudraka. His son will be Runaka and Runaka's son will be known as Suratha. His son will be Sumitra and the lineage will end with him."*
⚑ Two features of the handling are worth naming.
It is absorption by cataloguing. The names are not disputed, refuted, or explained. They are filed — placed in a sequence of kings the text already owns, between a Sañjaya and a Laṅgala, in the grammar of every other entry.
And the absorption is total precisely because it is unremarkable. A polemic against a rival tradition announces itself and can be argued with. A name in a genealogy cannot: there is nothing to disagree with in "Shakya's son will be Shuddhoda." It is either in the list or it is not, and it is.
A whole-file sweep settles what happens to the name itself.
"Buddha" appears exactly once in the entire volume — body and endnotes together — and the occurrence is a translator's note.2
Ch 453-8.2(31), the volume's longest chapter, sets out a five-āvaraṇa worship scheme and lists the avatāras to be worshipped in the third enclosure: "Matsya, Kurma, Varaha, Narasimha, Vamana, Rama, the other Rama, you Krishna and Ashvamukha."
That is nine. And endnote 2053 supplies the reckoning:
"Parashurama. This is a listing of Vishnu's ten avataras, omitting Buddha. Ashvamukha means Kalki."2
⚑ So the volume's single mention of the name is the translator observing an absence — a ten-item list delivered as nine, with the missing item identified.
⚑⚑ Which means the omission happens twice, at two levels, in two Saṃhitās. The Umā Saṃhitā puts the father in the royal genealogy and passes over the son. The Vāyavīya Uttara runs a daśāvatāra list and leaves out the entry the tradition ordinarily supplies. Neither is flagged in the body; one is flagged in a note; and the note is the only place the name occurs.
⛔ [n=2 OBSERVED], stated at that strength. Two omissions is not a policy, and the second is a translator's observation about a count rather than a statement by the text.
The finding sharpens against Volume 2, which handles the same tradition by the opposite method.
Volume 2 folds Buddha into the avatāra list at ch 186, as part of a broader operation the Volume 2 build catalogued as heterodoxy manufactured to order — a composite heretic assembled from Jain iconography, "Buddhist agama texts" cited by name, and Cārvāka materialism, with a damnation of future believers attached.3
So the corpus runs two incompatible strategies for the same tradition, in two volumes:
| Volume 2 | Volume 3 | |
|---|---|---|
| Method | inclusion as subordination — Buddha named as an avatāra of Viṣṇu, inside a hymn | absorption by cataloguing — his clan and father placed in the royal genealogy, and he is passed over |
| What is claimed | that he is ours, at a rank we assign | that his line is ours, and he is not a datum |
| Visibility | announced, polemical, arguable | invisible, unremarkable, unarguable |
⚑ And the two cannot both be operative. If Buddha is an avatāra of Viṣṇu he belongs in a daśāvatāra list; ch 453's list, in the same corpus, does not have him, and note 2053 says so. Neither volume acknowledges the other's handling.
⛔ What this is not. It is not evidence of a change of doctrine, a change of period, or a change of hand. The volumes are separately compiled sections of a compilation whose sections demonstrably run incompatible positions on tattva counts, on hell's location and on who a story is about.4 Add this to that list rather than to a theory of development.
The genealogy does not merely absorb. It also closes, and then reopens:
*"His son will be Sumitra and the lineage will end with him. The Ikshvaku lineage will end with Sumitra… Having reached Sumitra, when kali yuga commences, the auspicious lineage will merge with outsiders and will again revive in krita yuga."*1
⚑ "Merge with outsiders" is doing a great deal of work in one clause. A line whose survival runs through absorption into people outside it — stated as the mechanism of its continuity rather than as its failure. And the revival is placed past the end of the current age, which puts it beyond any possible check.
⚑ Set beside the Śākya entry, the chapter is doing the same operation in both directions in the space of a page: taking in a lineage that was not its own, and forecasting its own dissolution into lineages that are not its own.
And two entries earlier the chapter installs a restorer, in the present tense:
*"Shighra's son was named Marut, and he obtained siddhi through yoga. Even now, the lord is alive in the place known as Kalapagrama. When kali yuga is over, he will revive the sages and in particular, surya vamsha, which has been destroyed."*5
A named king, alive at the time of narration, in a named place, waiting for the age to turn, whose function is to restore the very lineage the chapter has just said will end.
⚑ So the chapter states three things about one dynasty within two pages: it ends with Sumitra, it merges with outsiders and revives in kṛta yuga, and a specific man is currently alive and will restore it. ⛔ These are not obviously incompatible — a reader can sequence them — but the text does not sequence them, and it is worth noting that the restorer is introduced before the ending is announced.
Stated plainly, because the material invites overreach in every direction.
It is not evidence about the historical Śākyas, or about Buddhism, or about relations between traditions. It is evidence about what a sentence in this text does.
It is not evidence of suppression. Nothing is denied. A genealogy that omits a figure omits thousands of figures; the observation is only interesting because these two names in this order point unmistakably at a third, and because the corpus elsewhere has a great deal to say about him.
And it is not evidence of intent. ⚑ The absorption may be deliberate, or may be a lineage-list doing what lineage-lists do, which is take up whatever names are locally available. The page reports the two omissions and the two strategies and does not adjudicate, because nothing in the volume supplies a means to.
You are reading an institution's account of its own history and want to see how it has handled a rival.
Read the lists, not the arguments. The arguments are where a rival is engaged and therefore where it survives as a rival. The lists are where it is disposed of — a name in a sequence, a row in a table, an entry in an org chart. ⚑ Nobody argues with a list, which is exactly what makes it the more effective instrument.
Then check the absences with a sweep, not by reading. The Śākya finding is a presence and it was found by reading. The daśāvatāra finding is an absence and no amount of reading would have produced it — it took a whole-file search for one word, which returned a single hit, in a footnote, saying the thing was missing. Absences are cheap to check and impossible to notice.
Look for two incompatible handlings before assuming one policy. Here one section subordinates by inclusion and another absorbs by omission, and neither acknowledges the other. A corpus with two strategies has no strategy — and the useful finding is the incompatibility, not either half.
And separate the claim about the line from the claim about the person. The genealogy takes the clan and skips the individual. That is a different operation from either endorsement or denial, and it has no obvious name in ordinary usage — which is probably why it works.
Evidence. Ch 353-6(39) read in full in place, with the Śākya–Śuddhodana entry, the Sumitra ending, the "merge with outsiders" clause and the Marut passage all quoted directly from the body;15 a whole-file sweep for "Buddha", "Shakya", "Shuddhoda" and "Kalapagrama" returning one hit each, with the Buddha hit read at its own line;2 ch 453-8.2(31)'s avatāra list counted at nine.2
Tension 1 — the Volume 2 comparison rests on that build's catalogue rather than on a fresh read. ⚑ The ch 186 material is recorded there as a hymn to Viṣṇu containing the avatāra sequence, and this page uses it at that strength. The two-strategies claim would be stronger with both passages re-read side by side, and that has not been done here.
Tension 2 — ch 353 was covered by instrument sweep before it was read. An earlier pass classified ch 343–353 as creation-and-lineage material and swept it with doctrine-tuned instruments, which returned zero. ⚑ Instruments tuned to doctrine find nothing in a genealogy, and this chapter is the proof: it yielded the Śākya entry, the Marut passage, the lineage-ending, and several other candidates, none of which any sweep would have surfaced.
Tension 3 — the daśāvatāra omission may be conventional. Lists of ten avatāras vary, and substituting another figure for Buddha is not unknown. ⚑ What the page rests on is narrower and is the translator's own: this particular list runs to nine and note 2053 identifies the missing tenth by name.
Open question. Does "the auspicious lineage will merge with outsiders" have a referent anywhere else in the corpus — or is it, like the Kalāpagrāma king, a clause with nothing downstream of it? A sweep for the outsider-merger motif across all three volumes has not been run.
Against The Sectarian Exclusions: the contrast in method is the point. That page's material is loud — exclusions issued as curses, heresies manufactured with named texts, believers damned in advance. This is the same corpus's quiet register: two names in a king-list and an omission a reader could not detect without a search tool. ⚑ A tradition's most complete absorptions leave the least argumentative surface, and this page and that one are the two ends of the range.
Against Only What a Character Asks: ch 186 and ch 453 sit in different volumes and different Saṃhitās, so no listener is present at both avatāra lists. The incompatibility is unavailable to anybody inside the corpus — and note 2053 is the only place in three volumes where the two are close enough to touch, in a footnote, without the touching being remarked.
Against The Apparatus Is the Translator's: this is a case where the apparatus supplies something genuinely load-bearing and does it correctly. The volume's single occurrence of the name is a note observing a count. ⚑ Which is also a reminder of that page's thesis from the other side: a finding that exists only in the endnote layer is a finding about a book, and any vault page using it must say so.
The Founder Problem in Historical Perspective — that page's territory is how traditions acquire, assign and manage origins, and how much work an attribution does that the underlying material never asked for.
The Śākya entry supplies the inverse operation, and it is the one that page's frame does not naturally reach: not a founder acquired, but a founder's lineage acquired and the founder left out.
The usual moves with a rival founder are recognisable and both are documented in this corpus — subordinate him (Volume 2 makes Buddha an avatāra, at a rank the host assigns) or denounce him. Both engage the person. Both leave a surface to argue with, and both keep him in the frame as somebody about whom a claim is being made.
⚑ The genealogical move does neither. It takes the clan and the father into the host's own royal line and treats the son as not a datum. The result is that a reader of the roll acquires the family as ancestral property without ever encountering a proposition about the man — there is no claim, so there is nothing to dispute, so nothing has to be defended in perpetuity.
That is a distinct and cheaper technique than either of the standard two, and it generalises well beyond religion: the acquired company folded into the corporate history with its founders absent from the timeline, the prior art listed in a patent's ancestry with the inventor uncited, the school of thought entered in a discipline's family tree with its originator appearing only as somebody's teacher. The tell is always the same — the lineage is present and the person is a gap in it. And the reason it is so durable is the reason it is so easy to miss: a gap has no advocates.
Sharpest implication. In a Śaiva Purāṇa's roll of the solar dynasty, between a Sañjaya and a Laṅgala, the text says "Sanjaya's son will be Maya and Maya's son will be Shakya. Shakya's son will be Shuddhoda and his son will be Langala" — the Buddha's clan and the Buddha's father, filed in sequence, with the son simply absent and no comment of any kind. Absorption by cataloguing rather than by argument: there is nothing in a genealogical entry to disagree with. And a whole-file sweep shows the omission happening twice, at two levels: the word "Buddha" occurs exactly once in the entire volume, body and notes together, and the occurrence is endnote 2053 observing that ch 453's daśāvatāra list runs to nine — "This is a listing of Vishnu's ten avataras, omitting Buddha." ⚑ Which sits against Volume 2, where the same corpus handles the same tradition by the opposite method — folding Buddha into the avatāra list as part of a manufactured heterodoxy. Two incompatible strategies, two volumes, neither acknowledging the other. ⛔ Reported throughout as this text's own handling, never as history.
Generative questions.