Ch 319 gives its catalogue of the worst things a person can do, and it gives it twice.
The first list has six items and a stated consequence: "There are six great sins, and they are said to lead to infinite consequences." Every one is sectarian — criticising Śiva, criticising the guru, reviling knowledge about Śiva, stealing gods' property, destroying brahmanas' property, stealing texts about Śiva.
The second list has five and no such clause: "There are five great sins—killing a brahmana; addiction to liquor; theft; violating the guru's bed; and associating with those who commit such acts."
That second list is the classical mahāpātakas — the standard, tradition-wide catalogue of the gravest offences, headed by homicide. It arrives second and it arrives without the infinite-consequences qualifier.
⚑ Between the two lists sits a long roster of further acts declared "equal in impact to an act of criticizing Shiva." One of them is sleeping in a room that has not been swept.
"There are six great sins, and they are said to lead to infinite consequences — criticizing Shiva; criticizing the guru; reviling jnana about Shiva; stealing objects that belong to devas; destroying objects that belong to brahmanas; and foolishly stealing texts that contain jnana about Shiva."1
⚑ Every item is an offence against the sect or its property. Three are speech offences, three are property offences, and none is an injury to a person.
⛔ And the qualifier is doing the ranking work. "Infinite consequences" is not attached to the five that follow. On the page, in the text's own terms, these six are worse — and they are worse in a way that is unbounded.
"There are five great sins — killing a brahmana; addiction to liquor; theft; violating the guru's bed; and associating with those who commit such acts."2
⚑ This is the classical five, in the classical order, headed by brahmanicide. It is the list any reader of this literature arrives already knowing.
⛔ And it is presented as an ordinary enumeration. No infinite consequences, no superlative, no framing clause. The chapter goes straight on to elaborate the first item — what counts as killing a brahmana, including "hurts him with words that pierce the inner organs".2
⚑⚑ The text does not argue for the reordering. It does not acknowledge that there is one. It simply prints the sectarian list first with a stronger consequence attached, then prints the received list second without one, and continues.
Between the two lists is a long series of acts declared equivalent to the first list's head item. The declaration is explicit: *"It is said that all these deeds are equal in impact to an act of criticizing Shiva."*3
The roster includes:
⚑ Most of the roster is omission — things you failed to do, or did inattentively. Two are about where you slept and what you talked about.
⛔ And the equivalence is stated, not implied. Put the chain together in the text's own terms and it runs: an unswept sleeping-place = criticising Śiva = infinite consequences > killing a brahmana.
⚑ The temptation is to read this as hyperbole — sectarian literature inflating its own offences, as sectarian literature does. That reading is available and it is too easy, because it does not account for the structure.
Three features distinguish this from ordinary emphasis.
One: the received list is not deleted. ⛔ A text simply asserting its own priorities would omit the mahāpātakas or subordinate them explicitly. This one prints them in full, correctly, immediately afterwards — and elaborates the first item at length. The competing standard is preserved intact and simply placed second.
Two: the ranking is carried by a technical clause, not by rhetoric. ⚑ "Infinite consequences" is a claim about the exhaustibility of karma, and this Saṃhitā makes exactly that claim elsewhere in load-bearing form — "if it is not enjoyed, it will not be exhausted, even in hundreds of crores of kalpas."4 So the qualifier attaches to a real mechanism in this text's own system.
Three: the equivalence roster is enormous and mostly passive. ⚑ If the sectarian offence were merely being emphasised, the roster would name deliberate acts of hostility. Instead it is dominated by omissions and inattentions — not bowing, not praising, sleeping somewhere unswept, changing the subject.
⚑⚑ Which is what makes the structure worth a page: the offence at the top of the ranking is one almost nobody has to commit deliberately. The classical five require you to do something. The six, and the roster attached to them, largely require you to fail to.
⛔ [n=1 OBSERVED] on the reordering as a deliberate move — one chapter, and the text never states a rationale. What is not in doubt is the arrangement on the page.
⚑ The ranking does not survive the volume unchallenged, and that is worth recording rather than smoothing.
Killing a brahmana is cancelled outright, twice, in the same volume. Ch 335 makes it retroactively void: "Even if a person kills a brahmana, but subsequently fights for his master and gives up his life in a battle, he goes to heaven and does not return from there." And the same chapter licenses it prospectively: *"If a brahmana who knows Vedanta acts like an assassin and attacks, desiring to kill, one can kill him. No sin from killing a brahmana results."*5
⛔ So the item at the head of the classical list is simultaneously the supreme sin, situationally lawful, and retroactively cancellable by a soldier's death — while the item at the head of the sectarian list carries infinite consequences.
⚑⚑ And that sharpens the finding rather than undermining it. The volume has an amnesty architecture that reaches homicide and does not obviously reach criticism. ⛔ Whether the amnesty reaches the six is not stated anywhere this build has read, and no absence claim is made about ranges it has not read.
⚑ One clause inside the second list is easy to skip and it changes the shape of both.
The elaboration of killing a brahmana extends the offence past killing:
*"Out of anger, greed, fear or hatred, if a person kills a brahmana, or hurts him with words that pierce the inner organs, he is said to kill a brahmana and is a great sinner."*2
⛔ Wounding speech is brahmanicide. Not analogous to it, not a lesser form of it — "he is said to kill a brahmana."
⚑⚑ Which means both lists are, in substantial part, about what people say. The six are criticism, reviling, and the theft of texts. The five, at their head, include speech that wounds. The volume's two catalogues of the gravest offences are largely a speech code.
⚑ And the ranking then reads more precisely than sectarian offences beat homicide. It reads: speech that injures a brahmana is homicide, and speech that injures the god's standing carries an unbounded consequence that homicide does not. ⛔ Both offences are verbal; what differs is the target.
⚑ Which is worth putting beside what the same volume does to speech elsewhere. This is a corpus in which spoken testimony persuades nobody — the Vedas are disbelieved, the mantras ignored, the teaching forgotten.7 Speech that cannot convince anyone is nevertheless the medium of its two worst crimes.
You are reading a code of offences and want to know whose code it is.
Look for the received list, and where it sits. ⚑ Most traditions have a canonical list of the worst things. A local code that reproduces it accurately but places it second has told you its priorities without ever stating them — and it is a much stronger move than omitting it, because the omission would be noticed and the reordering is not.
Read the consequence clauses, not the headings. ⛔ Both lists here are called "great sins." The ranking is entirely in a qualifier attached to one and not the other, and the qualifier is technical rather than rhetorical.
Check whether the top offence requires an act. ⚑ The classical five all require you to do something. The sectarian six, and the roster equated to them, are largely failures to do things — not bowing, not praising, not sweeping. ⛔ A code whose gravest offences are omissions is a code that can find almost anyone in breach, and that is a structural property rather than a moral one.
And check the amnesty coverage separately. ⚑ Here the offence at the top of the received list has two stated cancellations, and the offence at the top of the local list has infinite consequences attached. The ranking that matters in practice is not which sin is worse but which sin can be discharged.
Evidence. Ch 319-6(5) (Major Sins), lines 297–306, read in full in place on the source file: the six with their "infinite consequences" clause; the roster of further acts and the equivalence statement; and the five with their elaboration.123 Ch 320-6(6) for the inexhaustibility rule;4 ch 335-6(21) for the two cancellations of brahmanicide.5
⛔ Cited in the Saṃhitā-internal form — the running chapter numbers in this edition are not unique, and the fact is stated nowhere in it.6
Tension 1 — the two lists may not be intended as one ranking. ⛔ They are consecutive but not compared, and "great sins" may simply be a category label used twice for two different taxonomies. ⚑ What resists that reading is the qualifier: one list is given an unbounded consequence and the other is not, in the same passage, by the same speaker.
Tension 2 — this may be a genre feature rather than a claim. ⚑ Sectarian Purāṇas routinely elevate offences against their own deity. ⛔ The page does not claim originality for the elevation — it claims that the mechanism is precise, that the received list is preserved rather than displaced, and that the roster is dominated by omissions. Those three together are what is on the page.
Tension 3 — "almost nobody has to commit these deliberately" is a reading. ⛔ It follows from the roster's content and is not stated by the text. ⚑ The counter-reading is that an unswept sleeping-place stands for negligence toward the sacred generally, and the list is illustrative rather than enumerative. Both fit.
Open question. The volume's amnesty fires on brahmanicide through five separate currencies. Does anything anywhere in the corpus discharge one of the six — and if not, is "infinite consequences" the only genuinely unbounded liability the text carries?
Against The Machine and the Amnesty: the two pages sit on the same chapters and pull in opposite directions. ⚑ That page holds an override firing across nineteen chapters on five currencies; this one holds a class of offence with an unbounded consequence attached and no recorded discharge. ⛔ The volume builds an exhaustive pardon system and a category the pardon system is never shown reaching, and neither passage acknowledges the other.
Against The Hells That Fit the Crime: the same Saṃhitā's penal apparatus, two chapters later. ⚑ That machinery is organ-specific and precisely calibrated to acts. ⛔ It has no obvious instrument for an offence that consists of not bowing — and the volume never asks what the hell for an unswept sleeping-place would look like.
Against Who May Be Initiated: the two codes describe the same population differently. ⚑ The access rules gate entry by varṇa and sex; this catalogue gates conduct by attention. ⛔ A person can satisfy every access rule in the volume and be in breach of the top offence by having slept in an unswept room — which makes the two systems independent rather than nested.
Against The Corpus of the Disqualified: the exemplary lives run straight through this. ⚑ *The litanies celebrate an adulterer "not touched by any karma", three courtesans, and a declared drinker* — three of the classical five — and no litany celebrates anyone who criticised Śiva. The reordering holds in the stories as well as in the code, which is the strongest indirect evidence that it is deliberate.
Against The Supreme Dharmas, and Who Says Them: the same architecture at the other end of the moral scale. ⚑ That page holds a superlative slot filled by a dozen incompatible candidates. This one holds the negative version — and unlike the superlative slot, this ranking is stated once and not contradicted elsewhere in the chapters read.
Selective Enforcement of Legitimate Laws — that page's territory is the code broad enough that everybody is in breach, and what that does to the relationship between the enforcer and the enforced.
The Purāṇic material supplies the structural precondition, and it is a specific one worth isolating: the gravest offence in the code is an omission.
⚑ The classical five all require an act. You have to kill somebody, drink, steal, or go to a particular bed. A code made of those is a code most people are not in breach of, most of the time, and enforcement against them is an event.
The six, and the roster equated with them, invert that. Not honouring worship when you see it being planned. Not bowing at a liṅgam. *"Wilfully sleeping in a spot that has not been cleaned."* *"Ignoring accounts that concern jnana about Shiva and speaking about other things."* ⛔ These are conditions almost everyone is in almost always, and they are placed above homicide by an explicit consequence clause.
⚑⚑ Which is the transferable point. Selective enforcement is usually described as a discretion problem — the same rule applied unevenly. This shows the step before it: the rule can be drafted so that breach is the default state, and once it is, discretion is not an abuse of the code but the only way to operate it. Nobody has to decide to enforce selectively; there is no other option.
⛔ And the corpus supplies the accompanying half, which the enforcement framing usually treats separately. The offence at the top of the received list — homicide — has two stated cancellations in the same volume, one retroactive and one prospective. ⚑ So the code simultaneously makes universal breach the baseline on one axis and makes the gravest classical offence dischargeable on another. What that combination distributes is not punishment. It is standing — and the reason it is worth naming is that neither half looks like a power arrangement on its own.
Sharpest implication. "There are six great sins, and they are said to lead to infinite consequences — criticizing Shiva; criticizing the guru; reviling jnana about Shiva; stealing objects that belong to devas; destroying objects that belong to brahmanas; and foolishly stealing texts that contain jnana about Shiva." Then a long roster of acts "equal in impact to an act of criticizing Shiva", which includes "wilfully sleeping in a spot that has not been cleaned" and "ignoring accounts that concern jnana about Shiva and speaking about other things." And then, second, without the qualifier: "There are five great sins — killing a brahmana; addiction to liquor; theft; violating the guru's bed." ⚑⚑ The classical mahāpātakas, printed accurately, elaborated at length, and placed below an unswept room. The text does not argue for the reordering and does not notice it. ⛔ And what makes the arrangement more than sectarian hyperbole is the shape of the two lists: the classical five all require you to do something; the six and their roster are largely failures to do things — not bowing, not praising, not sweeping, changing the subject. A code whose gravest offence is an omission is a code almost everyone is in breach of, always. Meanwhile the top of the received list, brahmanicide, is cancelled twice in the same volume — "he goes to heaven and does not return from there" — while infinite consequences stays attached to criticism, with no discharge recorded anywhere this build has read.
Generative questions.