A tradition tells you two things about who it is for. It states rules, and it tells stories.
This volume's rules are restrictive and they are stated twice, in two different Saṃhitās, in the same words: "brahmanas, kshatriyas and vaishyas have a right to the worship." Women have no rights of their own and must be authorised by a husband, or failing that a son, or failing that a father. The initiation ladder starts at vaiśya, so the machinery for climbing is available only to people already above the bottom.
Then look at who the stories are actually about.
Demons. Three courtesans. A caṇḍālī. Two hunters, one of them described as a man who "caused violence to others." A declared adulterer who is "not touched by any karma." A tiger. A man who died in the form of a dog. And two people who changed sex.
Not one of them qualifies under the rules on the same volume's own pages. And they are not marginal cases smuggled in at the edges — they are the exemplars. They are who the text reaches for when it wants to show that this works.
Precisely, because the page turns on the rules being real rather than a strawman:1
Śūdras are excluded from Śaiva worship rights, twice, in two Saṃhitās. Ch 388-8.1(1): "following Shiva's dharma, men are divided into three varnas." Ch 430-8.2(8): "They understood that brahmanas, kshatriyas and vaishyas have a right to the worship." ⚑ Elsewhere in this corpus "three varṇas" consistently means the twice-born — ch 419's three bead-colours run "the three varnas, brahmanas and others, should respectively wear beads that are yellow, red and black" — so this is doctrine rather than a slip.
Women's rights are derivative and run through a complete chain of male permission. Ch 437-8.2(15): "On their own, women do not possess the rights to rites connected with Shiva's samskara. They can do so if instructed by the husband… a woman without a husband must seek the son's permission. With the father's permission, a maiden obtains the right."
The initiation ladder starts above the bottom. Ch 438-8.2(16) raises a śiṣya through the varṇas inside a single ceremony — vaiśya to kṣatriya to brahmana, and then "raised up to the status of Rudra" — and it begins at vaiśya. ⚑ So the promotion machinery is available precisely to those who already have access.
And the probation is varṇa-graded, inversely. "the guru must test him for a year, if he is a brahmana. A kshatriya must be tested for two years and a vaishya for three years." The lower the varṇa, the longer the wait.
⛔ Reported throughout as the text's own scheme, never as description of any historical practice, community or period.
A scripted census of the body returns 133 instances of the litany form — [Name] … obtained / became / was freed. The three principal litanies alone carry 85 named-figure-with-outcome sentences: ch 308-5(4) has 37, ch 316-6(2) has 37, ch 317-6(3) has 11.2
Who is in them:3
| Category | Named exemplars |
|---|---|
| Demons | Hiraṇyakaśipu, Virocana, Bali, Bāṇa, Vṛṣaparvan, Danu, Vidyutprabha, Śatamukha |
| Courtesans and puṃścalīs | Cañculā, Piṅgalā, Mahānandā |
| Caṇḍālī, outcaste, forest | Sauminī the caṇḍālī; Gurudruha the bhilla; Nala's bhilla former life |
| Hunters and killers | the hunter Śaṅkara; the hunter Mahākāla — "This hunter caused violence to others" |
| Declared sinners | Durjana — "a sinner… deceitful and lascivious, indulging in intercourse with many women" → "he was not touched by any karma" |
| Animals | the tiger made a gaṇeśvara; Vimarśana, who "in the form of a dog, died in front of a Shiva lingam"; the three deer of ch 311 |
| Sex-changed | Sudyumna, alternating month by month; the brahmana's son who "became a woman" |
| Women | Śāradā the widow, Sīmantinī, Kaikeyī, Anasūyā, Lopāmudrā, Ahalyā |
⚑ Notice the demons especially. The text is not making a grudging exception for them; it puts them first and says so plainly: "The daitya Hiranyakashipu, his younger brother and his sons constantly worshipped Shiva, as did Virochana and Bali. Bana, Hiranyaksha's son, is remembered as Shiva's great devotee."3
The class the cosmology exists to defeat is the class the devotional record celebrates.
The sharpest single line in the material is not in the body. It is in an endnote, and it is the moment the two halves touch.
Sudyumna appears in a devotee list at ch 308-5(4), among the saved. He is also, in the same volume, a man who was once a woman — the volume gives two incompatible accounts of how, one a curse and one a boon.4
Ch 350-6(36) then has him decline a kingdom in favour of Purūravas, on Vasiṣṭha's advice, and the body does not say why.
Endnote 1381 says why:
"Having been a woman earlier, Sudyumna was not entitled to a kingdom and Vasishtha advised that it should be given to Pururava instead."5
⚑ That is the whole architecture in one sentence. Qualified for liberation, disqualified from a throne. The devotional record admits him without comment; the institutional order does not, and the reason is a fact about his body that he did not choose in either version of the story.
⛔ Provenance matters here and is stated. The disqualification is supplied by the translator, not by the body, which gives only the outcome. What the body supplies is the outcome and the advisor; the reason is the apparatus reconstructing the rule from the result.5 That is a weaker citation than a body quotation and it is the hardest line of its kind in the volume — both facts belong on the page.
⚑ And it is the exact inverse of the litany's own logic. The litanies never ask what anybody was entitled to. They record what happened: he worshipped, he obtained. The rules ask entitlement and nothing else. The two are not in tension over any particular person — they are asking different questions, and the volume never notices that it is asking both.
The rules are not the only pole, and a page that suppressed the other one would be reporting half a corpus.
The volume's most inclusive sentences are as absolute as its most restrictive:6
Four data points at each pole, both stated in the register of doctrine, neither ever set against the other.
⚑ Which changes what the litanies are evidence of. Read against the restrictive pole alone they look like a contradiction. Read against both, they look like the corpus's own summary of the inclusive pole — the flat statements say anyone may approach, and the hundred and thirty-three names are the working out. ⛔ Preserved, not resolved. The volume states both and reconciles neither, and this page does not choose.
And there is a third position, which is neither flat inclusion nor flat exclusion, and it is the one an institution actually runs on.
Ch 437-8.2(15), immediately after the husband-son-father chain:
*"These paths are not recommended for mortal shudras… However, if such a person is full of natural devotion… Sins can be cleansed by offering padodaka."*1
Śūdras get the guru's foot-water instead of the rite. Not excluded, not admitted — given a lesser instrument that achieves a stated result.
⚑ That is a different kind of statement from either pole, and it is the most institutionally realistic thing in the volume: a graded offer, with a real benefit at the bottom and a real ceiling above it. [n≥3] on the variety of positions — flat inclusion, flat exclusion, graded access — and no claim about any one of them being the volume's view.
⚑ And ch 453-8.2(31) supplies a fourth axis that cuts across all three, because its exclusion list is not about birth at all: "let non-believers, deceivers, ingrates, those who follow darkness, heretics and great sinners maintain their distance."7 Conduct and belief, not varṇa — in the same chapter that names and accepts four Śaiva sects including the kāpālika mode. The volume has at least two independent axes of exclusion and never puts them in one sentence either.
One entry deserves its own paragraph, because it is where the two halves collide inside a single passage rather than across a hundred chapters.
Ch 437-8.2(15) states the husband-son-father chain, and then — four sentences later, in the same chapter — licenses a wife to leave a husband hostile to Śaiva dharma, and supplies a precedent:
*"If she is unable to do so, she should abandon him and undertake the rites of dharma in her mind. Though she was devoted to her husband, Atri's wife abandoned her husband, supreme among sages."*9
Atri's wife is Anasūyā — and ch 316-6(2), in this same volume, has her as the tradition's exemplar of not leaving: "For three hundred years, Anasuya, Atri's wife, fasted and slept on a pestle… Devoted to her husband, she made the Ganga flow in Chitrakuta." Ch 426-8.2(4) raises her further: *"Anasuya is Uma herself."*9
⚑ The stock exemplar of wifely devotion is cited as the precedent for abandoning a husband, inside one volume, and no note addresses it.
⛔ And there is a second problem in the same passage, which is translator-flagged and which the page must not conflate with the first. The sentences immediately following switch subject with nothing in the English to mark it — "She was not engaged to do this by her guru… and obtained the Pandavas as her husbands" — where note 1979 reads "Anasuya" and note 1980 reads "Droupadi." A reader who does not click both notes reads one woman throughout.9
⚑ Which is the page's whole subject, compressed: a rule, an exemplar who violates it, and an apparatus carrying the difference in two bare names.
You want to know who an institution is actually for. It has a written policy and it has a wall of success stories.
Read the wall, not the policy. The policy tells you what the institution will defend if challenged. The stories tell you what it is proud of — and pride is the better evidence, because nobody curates a case study they would rather not have been asked about.
Then check the stories against the policy, one by one, mechanically. Not impressionistically: take each named exemplar and ask whether the stated rules would have admitted them. ⚑ Here that check returns almost nobody — the demons, the courtesans, the caṇḍālī, the hunters, the tiger and the two who changed sex are all excluded on the same volume's own pages, and none of the litanies notices.
Expect both poles to be stated absolutely, and do not average them. The temptation with a body that says "irrespective of varna or ashrama" in one chapter and "brahmanas, kshatriyas and vaishyas have a right" in another is to synthesise a moderate position neither passage holds. Nothing licenses that. Hold the spread; the spread is the finding.
And look for the third scheme, because it is where the real practice lives. The two poles are rhetoric — one for recruitment, one for order. The graded offer in the middle, the one with a lesser instrument at the bottom and a stated ceiling, is what somebody actually had to administer. It is usually the least quoted clause in the document and the only one that had to work on a Tuesday.
Evidence. Scripted census of the litany form across the body, returning 133 instances, with the three principal litanies counted individually;2 the exemplar categories assembled from those litanies read in place;3 all four restrictive statements and all four inclusive statements located and read at their own chapters;16 endnote 1381 read at its own line;5 ch 453-8.2(31) — the volume's longest chapter, 4,962 words — read in full for the exclusion list and the sect material.7
Tension 1 — the litanies are not fully backed. At least five names in the corpus's litanies index nothing at all: they occur once, inside a list, and are never narrated anywhere in three volumes.8 ⚑ So the wall of photographs is partly a wall of names. That weakens any claim about the lives, and it does not weaken the claim about the selection — somebody chose to fill the slots with demons and courtesans, and the choice is evidence whatever sits behind each entry.
Tension 2 — "the rules exclude them" is doing the work of two different rules. Śūdra exclusion is a worship-rights rule; the women's chain is a saṃskāra rule; the dīkṣā ladder is an initiation rule. A demon is not covered by any of the three. The claim survives because the volume's own inclusive statements name the categories the exemplars fall into — antyaja, śvapacas, all those with bodies — but the three restrictive rules are not a single gate and the page does not treat them as one.
Tension 3 — Sudyumna's disqualification is the translator's. Marked in the body and repeated here. The body gives the outcome and the advisor; the reason is supplied in a note.
Open question. Do the two poles correlate with narrative versus prescriptive material — inclusion where the text is telling stories, exclusion where it is issuing rules? ⚑ Three of the four inclusive statements sit in vrata and japa chapters and three of the four restrictive ones sit in initiation and worship-rights chapters, which is suggestive and is not a count. The sweep that would settle it has not been run.
Against The Taxonomy of Transgressive Women: the two pages are inverses of one another and neither subsumes the other. That page reads a sin-list — the text grading kinds of woman by transgression. This one reads the exemplar-list — the text choosing who to celebrate. ⚑ Cañculā is on both, catalogued as a puṃścalī in one register and as a devotee who obtained the supreme destination in the other, and the volume never puts the two descriptions together. The same person is the category and the counter-example.
Against The Devotees Who Exist Only in the Endnotes: the same 133 instances read for a different property. That page asks whether the entries are backed; this one asks who is in them. Both are true of one object, and the second is sharper because the first shows the filling was a choice — you do not accidentally populate a hundred and thirty-three slots with the disqualified.
Against The Guru's Cow: the same volume, the same section, the opposite finding about mercy. There, four men are sentenced incoherently for one offence and the accidents draw the harshest penalties. Here, demons and courtesans and a tiger are admitted without any test at all. ⚑ The volume is at its harshest on people inside the system and its most generous on people outside it, and both readings come from the Umā Saṃhitā.
Against Only What a Character Asks: the rules and the lives never meet because nothing in the book's structure could put them in one room. The most inclusive sentence is at ch 309 and the most restrictive at ch 430 — a hundred and twenty-one chapters and two Saṃhitās apart, with no listener present at both.
When the Crowd Is Wrong Together — that page's claim is that a testimonial montage works by borrowing "the shape of organic consensus", and that the shape is chosen: relatable, diverse, ordinary, cast to look spontaneous.
The litanies are the same casting decision made a thousand years earlier, and they make a feature visible that the modern case obscures.
The FTX montage casts for relatability — a line cook, a surgeon, a plumber, people the viewer could be. This volume casts for the opposite: a demon king, three prostitutes, a caṇḍālī, a hunter who "caused violence to others", an adulterer, a tiger, a man who died as a dog. Nobody's neighbour. The list is not saying people like you did this. It is saying something stronger and stranger: the worst case you can imagine cleared the bar, so your case is not in question.
⚑ That is a distinct persuasive structure and it has a name in the material rather than in the theory: argument from the floor rather than from the median. Its power comes from being unrepresentative — a montage of ordinary people invites the objection not people like me, and a montage of demons and prostitutes forecloses it. And it explains something the social-proof frame otherwise leaves odd: why the disreputable entries are the ones stated most emphatically, and why Durjana gets the volume's flattest sentence — "he was not touched by any karma." The extreme case is not an embarrassment being managed. It is the load-bearing one.
Running back the other way: the modern practitioner's instinct is to select testimonials for likeness, and this suggests the instinct is wrong wherever the audience's real objection is disqualification rather than difference. Where people fear they are not eligible, the median case is worthless and the floor case is everything — which is a testable prediction about when a customer story should feature somebody nothing like the reader.
Sharpest implication. The volume states its access rules twice in two Saṃhitās — "brahmanas, kshatriyas and vaishyas have a right to the worship" — gives women rights only through a chain of husband, then son, then father, and starts the initiation ladder at vaiśya, so the machinery for climbing is available only to those already above the bottom, with a probation that runs one year for a brahmana and three for a vaiśya. Against which a scripted census finds 133 devotee-litany instances, whose named exemplars are *demons — Hiraṇyakaśipu, Virocana, Bali, Bāṇa — three courtesans, a caṇḍālī, two hunters, an adulterer who is "not touched by any karma", a tiger, a man who died in the form of a dog, and two people who changed sex.* Not one qualifies under the same volume's rules, and none of the litanies notices. The two halves touch exactly once, in an endnote: Sudyumna is on the list of the saved and, per note 1381, "Having been a woman earlier, Sudyumna was not entitled to a kingdom." Qualified for liberation, disqualified from a throne. And the volume says the opposite too, absolutely — "irrespective of varna or ashrama, women, children, male servants, female servants… all those with bodies" — so there are four statements at each pole, both in the register of doctrine, never set against each other, plus a third scheme in the middle where śūdras receive the guru's foot-water instead of the rite. ⚑ The doctrine of who may approach and the corpus of who did approach are in flat contradiction, and the contradiction runs the length of the volume.
Generative questions.