Maitreya is narrating creation, and when he reaches the social order the sentences have a consistent and easily missed shape:
The brahman emerged from Purusha's mouth. Brahmanas, preceptors and foremost among the varnas, also emerged from the mouth.1
The power of protection emerged from the arms and so did the kshatriyas, who follow that vow.1
Production, trading and the means of subsistence were created from the lord's thighs and so were vaishyas.2
Servitude that is necessary for attaining dharma was born from the illustrious one's feet. Earlier, the shudras were also born from here and they please Hari through service.2
Read the word order. In every case the function is named first, and the people arrive attached to it — also, and so did, and so were.
The mouth does not produce brahmanas who then take up the Veda. It produces the Veda, and brahmanas come with it. The arms produce the power of protection, and kṣatriyas come with that. The thighs produce trade, and vaiśyas come with trade.
What is being created is a set of offices. The populations are described as arriving in their train.
Three of the four run identically. The fourth does not:
Servitude that is necessary for attaining dharma was born from the illustrious one's feet. Earlier, the shudras were also born from here.2
Earlier.
For the first three, function and people emerge together. For the fourth, the people are placed before the function — the śūdras were already there, and servitude comes after.
That single word inverts the causal order for one varṇa alone. In the other three cases you could say the office called forth its holders. Here the holders precede the office, which means the office was fitted to a population that already existed rather than the population arriving with its work.
Register: this rests on one adverb in a translation. Whether it renders a temporal claim in the Sanskrit or is an artefact of English sequencing is not settled here, and this page does not build a doctrine on it. What can be shown is that the parallel construction runs three times and breaks on the fourth, and that it breaks in the direction of putting the people first.
The passage closes with a claim that does something the assignments alone would not:
Following their own dharma, these varnas worship their own preceptor, Hari, from whom they were born. They follow their means of subsistence and faithfully purify themselves.2
All four worship the same figure, and the ground given for it is the same for all four: he is where they came from.
That is a flattening move inside a ranking scheme. The hierarchy of body parts stands — mouth above feet, and nobody pretends otherwise — but the relation to the source is identical across all of them. A śūdra's claim on Hari is not weaker than a brahmana's; it rests on precisely the same fact.
Which is why this passage cannot be read simply as a legitimation of rank. It is that, and it also installs a channel that bypasses rank, in the same sentence.
The volume assigns the varṇas to body parts three times, and the assignments do not match.
At 2(1), inside the meditation on the cosmic body: "The great-souled one has the brahmanas as his mouth, the kshatriyas as his arms, the vaishyas as his thighs and the dark-complexioned ones as his feet."3
Here at 3(6): the same four sites, but as emergence rather than constitution — the varṇas come out of the parts rather than being them.
At 4(21), brahmanas addressing Pṛthu: "from his mouth... It is therefore said that brahmanas are his heart and kshatriyas are his limbs. Hence, brahmanas and kshatriyas protect each other."4
Three statements. The first says the varṇas are the body parts. The second says they came out of them. The third says brahmanas are the heart and kṣatriyas the limbs — different sites entirely — and derives that from the mouth-origin with a therefore that does not carry.
And the three do different work. Constitution makes the social order a permanent feature of a divine anatomy. Emergence makes it an event with a history. Heart-and-limbs makes it a mutual dependency, which is exactly the conclusion the speakers at 4(21) need from a king whose protection they are seeking.
The version deployed tracks the argument being made. Nobody reconciles them, and the apparatus glosses each locally without noting the others — the volume's established behaviour, landing here on the origin of the caste order.
The choice of image is not neutral and it is worth asking what it buys over the alternatives.
A decree would make the order legislated, and legislation can be repealed. A contract would make it agreed, and agreements have parties who might renegotiate. Emergence from a body makes it anatomical — and body parts do not vote on their arrangement, cannot swap positions, and are not in a relationship that admits of terms.
It also supplies an argument against complaint that no other image supplies as cheaply. Feet are not wronged by being feet. There is no one to appeal to, because there is no decision to appeal.
That is the work the figure does, and naming it is not the same as endorsing it. What the passage gains in stability it pays for in a specific way: an anatomy cannot be argued with, which means it also cannot be argued for. Nothing in these lines gives a reason why trade belongs at the thighs. The image is doing the persuading, and the persuading consists of removing the question.
There is a further cost, and it is the one that shows up at 4(21). An anatomical justification is fixed, and fixed things are useless when the argument changes. The brahmanas addressing Pṛthu need mutual dependency, not rank — they want a king who understands that he needs them — and the mouth-and-arms assignment gives them exactly the wrong shape, since a mouth does not depend on arms. So they reach for heart-and-limbs instead, an assignment with no standing anywhere else in the volume, and derive it from the mouth-origin with a connective that does not hold. The anatomy that was supposed to settle the matter permanently turns out to be reissued whenever it fails to say the needed thing, which is the ordinary fate of a justification that cannot be argued and therefore cannot be amended.
Three absences are worth marking, because readings of this passage routinely supply them.
It does not rank the varṇas explicitly. The ranking is carried entirely by the anatomy — mouth, arms, thighs, feet — and never stated as a claim.
It does not say the arrangement is unchangeable. That too is carried by the image rather than asserted, and an emergence account is in principle compatible with subsequent movement in a way a constitution account is not.
And it does not address birth. Nothing in these lines says a person's varṇa is fixed by their parentage. The passage is about where the four came from at the origin, not about how individuals are sorted afterwards — which is the question the tradition's later disputes actually turn on, and which is not in this text at this point.
None of that makes the passage benign. It makes it narrower than its reputation, and the narrowness is where the argument about it should happen.
What follows the varṇa account, within a few lines, is the volume's most direct statement that the narrator does not know what he is talking about:
Who can dare to measure or describe the illustrious one's divine deeds, atman and forms? ... Despite that, I have recounted according to my intelligence and according to what I have heard.2
And then, of Brahmā himself: "The first wise one used his intelligence to perfect his yoga and meditated on him for one thousand years. Even after this, did he comprehend the greatness of his atman?"2
So the account of where the social order came from is bracketed, by its own narrator, as a best effort from a limited position — immediately after being delivered.
That disclaimer applies to the whole cosmogony and not specially to the varṇas. But it applies to the varṇas too, and it is the kind of qualification that gets dropped when the four lines are quoted on their own.
Someone justifies an arrangement — a reporting structure, a pay band, who gets to speak in the meeting — by describing it as a body, an ecosystem, an organism, a machine with parts.
The move is doing something specific and it is worth naming before responding to it.
An anatomical argument converts a decision into a fact. Decisions have authors, reasons, and alternatives that were rejected. Anatomy has none of those. Once the arrangement is a body, asking why trade sits at the thighs sounds like asking why knees bend forward.
So the useful response is not to argue the assignment. It is to ask for the decision the anatomy replaced. Who decided, when, on what grounds, and what was the runner-up? Sometimes there is a good answer, and it will be better than the metaphor. Often there is no answer, and the absence is the finding — the anatomy was reached for precisely because the decision could not be defended in the form of a decision.
And check whether the metaphor is doing double duty, as it does here. This passage ranks and it also grants everyone the same claim on the source. Organisational versions do the same: we're all one body is a levelling claim and you're the hands is a ranking one, and they usually arrive in the same sentence from the same person.
The strongest evidence that offices are being created rather than populations is the word order, repeated three times. Function first, people appended with also or and so did. A passage creating peoples and assigning them work would run the other way round.
Tension — the fourth case inverts that order. Earlier, the shudras were also born from here puts the population before the function, uniquely, and nothing marks the change. Preserved.
Tension — the volume assigns the varṇas to body parts three times, at 2(1), 3(6) and 4(21), with three different relations and, in the third case, different body parts. The version used matches the argument being made. Preserved unresolved.
Open question. Earlier is one word in a translation. Does the Sanskrit mark a temporal priority for the śūdras, or is this an artefact of English clause ordering — and is there any commentary tradition that reads it as significant?
This vault's material on Puruṣa-sūkta descendants generally treats the body-origin of the varṇas as a single stable trope, varying in wording but not in structure across its Vedic and Purāṇic appearances.
This volume does not support that stability. Within four Skandhas it gives three versions with three different relations between varṇa and body — constitution, emergence, and a heart-and-limbs assignment that abandons the standard sites altogether. The trope is not being transmitted intact; it is being re-deployed, and the deployment tracks what each speaker needs.
Two Incompatible Maps of the Same Body documents the collision between the varṇa assignment and the loka assignment inside a single chapter, at exactly the four rank-bearing sites. Read together, the two pages suggest the body is not a stable map that different schemes are placed onto. It is a surface that whatever needs authorising gets written on.
With Manufactured Legitimacy. That page works the construction of authority that presents itself as discovered — arrangements given a pedigree, an origin, or a natural basis so that they arrive already justified and the question of who decided never surfaces.
The varṇa passage is that operation performed with the minimum possible material: four body parts, four functions, no argument at all.
What the pairing produces: the manufactured-legitimacy frame is built for cases where a specific interest is being served and the manufacture can be traced to a beneficiary. This case resists that in a way worth taking seriously. The same passage that ranks also grants every varṇa an identical claim on the source, and the narrator disclaims his own competence a few lines later — neither of which a straightforwardly self-serving construction would include. Which isolates something the frame tends to collapse: legitimation and stabilisation are separable functions, and an anatomical origin story does both at once. It legitimates a ranking, and it also makes the whole arrangement narratable as a single thing with a history, which is what lets people locate themselves in it at all. The second function is why such stories survive in traditions whose interests have long since changed, and why exposing the beneficiary — the usual critical move — so often fails to dislodge them. You have removed the justification and left the map.
Sharpest implication. The passage creates jobs, not peoples — three times over, in its own word order — and then breaks its pattern exactly once, for śūdras, by putting the people first. Whether that is a translator's clause order or a claim in the source, it marks the one place where the office had to be fitted to a population instead of arriving with one.
Generative questions.