At 11(24), opening a sāṅkhya exposition, Kṛṣṇa places the beginning somewhere odd:
"In the beginning, at the time of krita yuga and when there were no yugas, people were accomplished in discrimination and because of that knowledge, saw everything as one and without any differentiation. It was one and undifferentiated, beyond the reach of mind and speech."1
⚑ And when there were no yugas. The original state is placed outside the age-system entirely — before there was a first age to be first in.
And what characterises it is not virtue or proximity to the divine. It is undivided perception. People saw one thing. Differentiation had not started, in the world or in the seeing of it.
Then: "that great truth differentiated itself into two" — Prakṛti and Puruṣa — and everything else follows from that split.
Seven chapters earlier, answering a question about varṇāśrama dharma specifically, the same speaker gives the same shape in social terms:
"In the beginning, in krita yuga, there was only one varna among men and it was known as hamsa. As soon as they were born, subjects accomplished their objectives. That is the reason the learned called it krita yuga."2
⚑ One varṇa. Not four with the distinctions unenforced — one, with a name.
And the detail that follows is worth keeping: "In that first age, all the Vedas were there in 'Oum'." The entire textual apparatus, undifferentiated, in one syllable — the same structure as the social one, and the same as the perceptual one at 11(24). ⚑ Three domains, one shape: originally undivided, then split.
The four appear in the next age, and the volume dates that too:
"At the beginning of treta yuga, the three Vedas were manifested from my heart as my breath of life… From the cosmic Purusha, four categories were born—brahmanas from the mouth, kshatriyas from the arms, vaishyas from the thighs and shudras from the feet. They were characterized by their own conduct."2
⚑ Second age. Second-order. The four-fold division is not part of the original arrangement — it arrives when the Vedas divide into three and the cosmic body is parted out.
And the āśramas get the same treatment in the same sentence-run: "Garhasthya appeared from my loins, brahmacharya from my heart, vanaprastha from my chest and sannyasa from the top of my head."
⚠ And here is the clause that must be stated alongside the dating, because leaving it out would make this page a piece of advocacy rather than a record.
"The varnas and the ashramas followed the order in which they were created. The higher the position, the higher the order of men. The lower the position, the lower the order."2
⚑ The ranking is derived from anatomy. Mouth, arms, thighs, feet — and the volume says outright that the ordering follows the elevation of the body part.
So the same passage does both things. It dates the four-fold division to a second age, and it grades the four by where on a body they emerged. ⚠ The dating does not cancel the ranking. It dates it.
You find out that the arrangement everybody treats as permanent has a start date. That is real information and it is not, by itself, an argument that it should end.
The same document that dates the thing also ranks it. Quote both, or you are not reporting the document.
An origin story that says there was one, and then there were four is doing something different from one that says there were always four. Notice which one you are being told, and notice that people cite the first as though it were an abolition.
The ranking is derived from where on a body something came from. That is not a reason and it is not pretending to be one; it is a picture, and pictures are harder to argue with than reasons.
The temptation with this material is severe and has to be named.
It would be easy to build this page as an abolition. Varṇa arrives in the second age; before it there was one class, named; and at 11(12) the roster of those who reached him by association alone runs "vaishyas, shudras, women, outcastes." Assemble those and you have a case.
⚑ The volume does not make it, and the same chapters make the other case. The height-ranking is in the same passage as the dating. The guṇa parts-list at 11(13) has birth among its inputs. And 11(5) has Cāmasa place "women, shudras and others like that" outside the twice-born, as objects of pity.
⚠ What the text supports is narrower and more interesting than either side. ⚑ The division is contingent: it has a date, it did not exist before that date, and the volume says so twice in the mouth of its own protagonist. That is a substantial claim about the status of a hierarchy — and it is compatible with the hierarchy being real, ranked, and operative now, which is exactly what the same passage asserts.
Preserve both. The dating is the finding; the abolition is a reader's addition.
One structural point that makes the dating harder to dismiss as a stray.
It is not one passage. 11(17) dates the four-fold division to tretā, in a chapter explicitly about varṇāśrama dharma, in answer to a direct question about it. 11(24) places the undifferentiated state outside the yuga-system altogether, in a chapter about sāṅkhya, where nothing about caste is at stake.
⚑ Two chapters, two subjects, one speaker, one listener — and the same architecture both times. The second is the more telling, because it has no polemical reason to say it. 11(24) is doing cosmology and mentions the undivided origin because that is where its account starts.
And the shape recurs outside the social material, which is why it reads as a habit rather than a position.
The Vedas: all in Oṃ, then three. Perception: everything seen as one, then differentiated. The metaphysics: one great truth, then two — Prakṛti and Puruṣa. And the practice: at 11(12), the instruction is to "give up the primary injunctions and the secondary injunctions, pravritti and nivritti" — to stop operating the distinctions.
⚑ [n≥3 PATTERN] — this volume's standard account of anything is that it was one, then divided, and its standard instruction is to work back across the division.
⚠ Which is what makes the varṇa dating unremarkable in its own terms and remarkable in ours. The volume is not making a special concession about caste. It is applying its general cosmological grammar to a social fact — and the general grammar happens to entail that the social fact has a start date.
The strongest support is the second instance. 11(24) has no stake in the caste question and supplies the same undifferentiated origin, which makes 11(17) look like an application rather than an apologetic.
A second support. Only one varna among men and it was known as hamsa is specific. It names the single class, which a vaguer text would not have needed to do.
The tension, preserved. ⚑ The four-fold division is dated to the second age and ranked by bodily height in the same passage. Held as printed. Neither half is the volume's real position with the other as decoration.
A second tension. ⚠ The volume's other statements on who may approach do not line up with either. 11(12) has outcastes reaching him by association alone; 11(5) has Cāmasa placing women and śūdras outside the twice-born; 11(14) admits the śvapaca and warns against women six lines later. Five positions, one Skandha — and this page's material is only two of them. The full disagreement belongs with the who-may-approach material and is not resolved here.
Open question. "They were characterized by their own conduct" sits immediately after the bodily derivation. ⚠ Does conduct define the category or follow from it? The clause is ambiguous as printed and the volume does not disambiguate it in the range read.
Set against Your Quality Is Manufactured From a Parts List, the two chapters take opposite views of how much birth does. ⚑ There, birth is one input of nine into a guṇa the chapter explains how to engineer. Here, the four categories emerge from a body and are ranked by height. Both are Kṛṣṇa's, eleven chapters apart.
Against He Lists Fourteen Disciplines by Name, the roster is the hardest evidence on the other side — vaiśyas, śūdras, women, outcastes, and "those who were dominated by rajas and tamas in their nature" — and it is three chapters after the ranking.
Against The Wives Who Walked Out, the Tenth Skandha's narrative runs the same direction as 11(12) and against 11(17). Ritual standing buys nothing there; here it is derived from a cosmic body and ordered by height.
Against The Volume Ranks Its Own Instruments Four Different Ways, the method is the same and the subject differs. ⚑ This text ranks things repeatedly, in incompatible ways, and never stages the comparison — instruments in one case, classes of person in another.
The Founder Problem in Historical Perspective — institutions need an origin account, and the account they get is written later, by people with present interests. The founding is not a fact the institution inherits; it is a claim the institution maintains.
The tretā dating is that structure, and it is unusually explicit. The four-fold division is given a date, a mechanism and a body — and the whole thing is delivered as a direct answer to a question about how varṇāśrama dharma should be practised now.
⚑ What neither field gives alone: founder-accounts normally work by claiming the arrangement is original, and this one gains its authority by admitting that it is not. There was one varṇa, and it was called haṃsa. The account concedes a prior state and still licenses the present one — because the division is attributed to the cosmic body rather than to anybody's decision.
⚠ Which is a stronger move than an origin-claim, not a weaker one. An arrangement said to be original can be attacked by finding evidence of a time before it. An arrangement that includes the time before it, and dates its own arrival, has already absorbed that attack. ⚑ The volume can say there was one class and the higher the position, the higher the order in one passage precisely because the first sentence has been made part of the second's story.
Status Display and Hierarchy Signalling — rank is communicated through legible physical markers, and the markers work because they are read instantly and without argument. Height, position and bodily orientation are the oldest of them.
Mouth, arms, thighs, feet is that vocabulary used as a derivation rather than as a display. ⚑ The ranking is not signalled by height; it is derived from height, and the volume states the rule aloud: "The higher the position, the higher the order of men."
⚑ The insight the pairing produces: a status marker normally sits on top of a hierarchy that was established some other way, and here the marker is doing the establishing. There is no separate argument for why brahmanas rank above śūdras — no claim about function, capacity, contribution or conduct precedes it. The elevation of a body part is the whole of it.
⚠ Which is why the āśramas get the identical treatment in the same breath — loins, heart, chest, top of the head — and why the two orderings come out parallel. ⚑ One picture, applied twice, generating two hierarchies, and the volume adds no reason to either because the picture is not offered as an argument. It is offered as an origin, and origins do not have to argue.
Sharpest implication. *The volume dates its own four-fold division twice — at 11(17) there was "only one varna among men and it was known as hamsa" in the first age, with the four born from the cosmic body at the beginning of tretā; at 11(24) the original state is placed "when there were no yugas", undivided in perception and beyond mind and speech.* And in the same passage as the dating: "The higher the position, the higher the order of men. The lower the position, the lower the order." The dating does not cancel the ranking. It dates it.
Generative questions.