This page was planned as a comparison of two worship manuals. The read turned up three.
The Vidyeshvara Samhita gives a complete daily protocol in chapter 20. The Rudra Samhita gives one in chapter 43, and another in chapter 45.1
All three cover the same ground in the same order: wake before dawn, go outside, clean with earth, wash, rinse, clean the teeth, bathe, put on a fresh garment, sit down, worship.
And the specifics do not match — not between the Samhitas, and not between the two Rudra chapters either.
That last part is what makes this page worth building. Two Samhitas disagreeing can be explained away as two sources joined. Two chapters disagreeing, two apart, in the same voice, about the same act, cannot.
Set out plainly, because the argument is entirely in the specifics.
| Ch 20 (Vidyeshvara) | Ch 43 (Rudra) | Ch 45 (Rudra) | |
|---|---|---|---|
| Waking | face east; the direction determines the day's fruits | brahma muhurta | brahma muhurta |
| Defecation direction | north (others if obstructed) | south | south |
| Positional taboos | not facing water, fire, a brahmana, or an image; cover with left hand; do not look at the stool | — | — |
| Earth-cleaning | 7, 5 or 3 times — a range | by varna: 5 / 4 / 3 / 2 | earth and water, unspecified |
| Mouth rinse | eight times | — | sixteen times |
| Teeth | leaves or twig; avoid the index finger | twig by varna: 12 / 11 / 10 / 9 angulas | before sunrise |
| Teeth forbidden | — | shashthi, navami, Sunday, vow days, shraddha days | shashthi, navami, Sunday, first tithi of the month |
| Bathing | immerse to the knees; mantras; garment washed after, by pounding | following the norms | facing north or east; the sun in tirthas; hot-water prohibitions |
| Oil | — | — | by weekday; mustard, fragrant, with exceptions |
| Hand/foot washing | hands and feet | left hand ×10, feet ×7 each, hands ×3 | wash |
Read down the columns and the shape comes out. These are not three drafts of one document. They are three different kinds of document — one procedural, one caste-indexed, one calendrical — describing the same forty minutes.
Nothing anywhere flags the overlap.
There is no "as was said before." No "some hold otherwise." No "in the Vidyeshvara Samhita it is stated that..." Chapter 43 opens by having the rishis ask a question the volume has already answered at length, and Suta answers it as if for the first time.2
This matters because the volume is capable of cross-reference. It names five of its own Samhitas as deleted, with their shloka counts. It tells you which sections are full of vijnana. Its narrator states precisely how much of the whole he possesses — a quarter.3
So the silence is not a general inability to refer to itself. It is specific to this case.
The likeliest explanation is the dull one and it is probably right: whoever compiled the Rudra Samhita's worship material was not working from the Vidyeshvara Samhita's. Two bodies of practical instruction, from two sources, bound into one volume without anyone reading across.
Here is the part that will not fit that explanation.
Chapters 43 and 45 are two apart. Both are Brahma speaking. In chapter 44 he is interrupted by Narada asking for more detail, and chapter 45 opens "After this, I will tell you about an excellent method of puja" — so chapter 45 presents itself explicitly as a continuation of chapter 43's subject.4
And they still disagree.
The forbidden-teeth days differ: chapter 43 bars vow days and shraddha days; chapter 45 bars the first tithi of the month. Three days are shared; three are not.
The rinse count appears in chapter 45 at sixteen and is absent from chapter 43 — while chapter 20 has it at eight.
The varna grids are all in chapter 43 and vanish completely in chapter 45, which grades nothing by birth at all.
Two chapters, presented as one continuous teaching, running different rules. Whatever produced that, it was not a compiler joining two Samhitas. It was already happening inside one.
Put yourself in the position the text implies: someone who wants to do this properly and has access to the whole volume.
Try to resolve by authority. All three are spoken by figures the volume treats as maximally authoritative — Suta reporting Vyasa, and Brahma to Narada. There is no rank between them.
Try to resolve by recency. The volume gives no chronology of its own composition, and its status is disputed at the level of whether it is a Mahapurana at all.5 There is nothing to date.
Try to resolve by detail. Chapter 20 is the most detailed on the body, chapter 43 on the social grid, chapter 45 on the calendar. Each is fullest in a different dimension.
Try to resolve by frequency. South wins two to one — except the two are contiguous chapters of one Samhita, which is one tradition speaking twice.
Every route fails, and the failure is total rather than close.
Which points to the actual answer: nobody was ever in this position. You did not own the volume. You heard it recited over many evenings, and you already had a way of getting up in the morning, learned from the household you grew up in. The text ratified a practice you had; it did not issue one.
A manual is only a manual for a reader who has no practice. For everyone this was made for, all three chapters said the same thing — do it properly — and the specifics were confirmation, not instruction.
You have been asked to write down how your team does something. Onboarding, say, or a release.
You interview three people who have all done it for years.
The first tells you a sequence and is confident about it. The second gives you the same sequence and one step in a different order, and when you point at the difference, says "well, either way really." The third gives you the same sequence with two extra checks the others never mentioned, and is surprised the others did not mention them.
Nobody is wrong. Nobody is lying. And there is no way to reconcile them into one document without simply deciding.
Now notice what happened when you tried to write it down. The disagreement did not exist until you asked. These three people had worked alongside each other for years without friction, because the parts they disagreed on were the parts that did not matter, and the parts that mattered they all did identically.
That is the exact structure here. The three protocols agree completely on the sequence — the thing that would break if you got it wrong — and disagree on the counts and directions, which is the layer that only exists once somebody starts writing.
And the document you produce will be more precise than any of your three sources and will describe something none of them actually does.
This is the methodological point, and it is why this page carries more weight than its subject matter suggests.
Most claims that a text is composite rest on style: this passage has a different vocabulary, a different metre, a different theological temperature. Those arguments are real, and they are contestable forever, because an author is allowed to change register and a tradition is allowed to have range.
A procedural contradiction is a different kind of object. North and south are not a matter of tone. Eight and sixteen are not a matter of emphasis. There is no reading of the volume on which both instructions are simultaneously followable, and no amount of literary sensitivity dissolves it.
This build now holds four independent mechanical lines of evidence for composite construction, and none depends on any theory of style:
Four mechanisms, four failure modes, complementary blind spots. A compiler could evade any one of them and would have to be actively defending against all four to evade the set.
Three consequences, stated so they are not lost.
One. The vault's chapter 20 pages are now one recension of three. The Morning Body is not wrong; it is local. Its most distinctive details — the north-facing rule, the index-finger prohibition, the injunction not to look at the stool — appear in neither Rudra version.
Two. The volume's "escape clauses" need re-reading. Eight instances were catalogued across Units 1–2 as a compositional signature: build an exhaustive rule-structure, then place a clause at its base that makes it optional.9 Some of those may be something duller — a compiler who inherited two incompatible specifications and papered the join with "there are no rules about this." Chapter 45 does exactly that three times about domestic worship.4 Not all of them; the ash clause and the place-multiplier collapse are too pointed for that. But the category is no longer clean, and the vault should hold both readings.
Three, and it cuts the other way. The Question Asked Twice argued that some Units 1–2 "structural findings" might be artefacts of reading a performed text as a document. This one is not. A listener four evenings apart genuinely would not notice — but the contradiction is in the content, not in the delivery, and a household trying to obey both would face it on the first morning regardless of how the text reached them.
Strongest support is the defecation direction — north in chapter 20, south in chapters 43 and 45. Binary, explicit, unhedged in all three, and not the kind of instruction that admits a range.
Tension one, and it is the page's core. The two Rudra chapters, presented as one continuous teaching, disagree with each other on the forbidden-teeth days and on whether the rite is graded by varna at all.
Tension two. Chapter 20's range for the earth-count becomes chapter 43's caste grid and then disappears entirely in chapter 45. The same specification appears as latitude, as a birth-table, and as nothing.
Tension three. The volume cross-references itself freely elsewhere — deleted Samhitas, shloka counts, the narrator's own quarter-knowledge — and is silent here.
Tension four. Chapter 45's three releases from its own rules ("there are no rules about worshipping the gatekeepers"; "there are no rules about this"; "this rule is not binding") read either as the volume's escape-clause signature or as a compiler papering a join. Preserve both readings.
Open question, and it is cheap. Chapters 47–150 have not been checked for a fourth protocol. Three variants can only be counted; four or more could be sorted into families, and families would let the recensions be ordered. This is the single highest-value thing to watch for on the Units 4–5 read.
The Second Morning Protocol holds the detailed comparison and the finding that what is invariant across the three is the bare sequence of physical acts, while everything that drifts is a count, a direction or a date. This page takes that and asks what it proves.
The Second Column of Fire is the same phenomenon at narrative scale — the founding myth told twice with the crime deleted from the second telling. Procedure and narrative both doubled, in the same volume, neither doubling acknowledged.
Those two pages together are the strongest pair this build has produced, and they are independent: one could be true without the other.
Three Incompatible Origin Accounts and Brahma's Three Origins add the same failure mode applied to provenance and to a god's birth. Four domains — text-origin, god-origin, founding narrative, daily procedure — and the volume tolerates unreconciled multiplicity in all four.
At that point it stops being a defect to be catalogued and becomes a property of the object. This text does not treat mutual exclusivity as a problem, and any vault page that reads one of its accounts as the account is importing an assumption the source does not hold.
The Ritual-Systematization Gap — WarYoga vs Greek (cross-domain). The observation there is that the Vedic-yogic material hands you procedures — this threshold-moment, this many breaths, cow dung and unbroken rice and copper — while the Greek hero material hands you stories and no steps at all. One tradition transmits a recipe book you could give to a stranger; the other transmits a hall of fame.
This page shows what the recipe book costs.
A story cannot contradict itself procedurally, because it does not specify anything. A procedure can, and this one does — and the more precisely a tradition specifies, the more surface it exposes. Chapter 20 is contradicted by chapter 43 exactly because both were willing to state a number.
The insight neither gives alone: the transmissibility that makes a procedural tradition portable is the same property that makes its seams visible. A tradition that hands you steps can be given to strangers, which is its enormous advantage; it can also be audited, which no story-tradition can. So the Greek material's apparent weakness is a kind of armour — it never made a claim specific enough to be caught disagreeing with itself. And it suggests a rough rule for reading any corpus: you can only detect composition where the text was willing to be precise, which means the parts of a tradition that look most consistent may simply be the parts that never committed to anything.
Bushidō: Code vs. Myth vs. Reality (history). That page separates three objects wearing one word: the actual fragmented clan precepts of the samurai era, the idealised code reconstructed after samurai ceased to exist, and the documented gap between either and how warriors actually behaved.
The three morning protocols are the same problem caught earlier, and that is the value of the pairing.
Bushidō's three layers are separable because they are separated in time — we can date the Meiji reconstruction and see it is not the clan precepts. Here all three protocols sit inside one volume with no dates, no attributions, and identical narratorial authority. There is nothing to pull them apart with.
What the pair produces: the reason a canon is harder to audit than a history is that canonisation deletes the timestamps. Every layer of the Śiva Purana speaks in the same voice with the same warrant, which is precisely what canonisation is for — it makes the material one thing. And that has a consequence worth carrying: where a tradition has been successfully canonised, the only remaining handles are the places where the layers make incompatible factual claims. North versus south is not a minor curiosity in that light. It is one of the few things left that a canon cannot absorb, and it is why a disagreement about which way to face is worth a page.
Sharpest implication. The volume gives the same morning three times, and no two versions agree. You face north in chapter 20 and south in chapters 43 and 45. You rinse eight times in one and sixteen in another. A permitted range of earth-cleanings becomes a caste grid and then vanishes. And crucially, chapters 43 and 45 disagree with each other — two apart, in Brahmā's single continuous voice, on which days the teeth may not be cleaned and on whether the rite is graded by birth at all — which no story about two Saṃhitās being joined can explain. Every route to resolving it fails: no rank between the speakers, no chronology, each fullest in a different dimension. And that is the answer: nobody was ever in the position of having to choose, because the text ratified a practice people already had rather than issuing one. What this contributes to the build is a fourth mechanical line of evidence for composite construction, independent of the other three and immune to the usual defence — north and south are not a matter of tone.
Generative questions.