The sages ask about sadachara — good conduct — and the answer opens with something that should not be possible inside a birth-based caste system.
There is more than one kind of brahmana. There are seven. And they are ranked by behaviour.
A brahmana with sadachara is known as vidvan. A brahmana who possesses sadachara and practices the conduct of the Vedas is known as a vipra. A brahmana who possesses only one of these two is a dvija. A brahmana who possesses a little bit of good conduct and a little bit of the Vedas is known as a kshatriya brahmana. He becomes a servant of the king. With a small bit of good conduct, if a brahmana practices agriculture and trade, he is a vaishya brahmana. One who ploughs the field himself is known as a shudra brahmana. A brahmana who is jealous and causes harms to others is spoken of as a chandala brahmana.1
Seven grades. Same birth at the top and at the bottom.
A man born a brahmana who is jealous and harms people is a chandala brahmana — which uses, for a man of the highest birth, the word the system reserves for its most excluded category.
The ranking runs on exactly two inputs and it is worth separating them, because the text does.
Sadachara — good conduct. Debroy glosses it flatly: "good behaviour."2
The conduct of the Vedas — Vedic practice, the ritual and study obligations.
Both → vidvan, and then vipra for both together. One only → dvija, twice-born, which is the ordinary word for anyone of the top three varnas. A little of each → you begin to fall.
So the top of the scale needs conduct and learning, and possessing only one drops you to the generic term. The system is stricter about its own elite than about anybody else.
And notice which variable survives at the bottom. The chandala brahmana is not defined by having lost his Vedic learning. He is defined by jealousy and causing harm — conduct alone. You can fall all the way to the bottom while retaining every scrap of your education.
The middle of the list does something stranger than a simple ranking, and it is easy to read past.
The grades are not named first-class brahmana, second-class brahmana. They are named after the other varnas.
Kshatriya brahmana — and the text specifies his occupation: "he becomes a servant of the king." Vaishya brahmana — "practices agriculture and trade." Shudra brahmana — "one who ploughs the field himself."
So the descent is measured in the work a man does, and the work is the work of the varna he is being renamed after.
That last one repays attention. The difference between a vaishya brahmana and a shudra brahmana is not the size of the farm. It is whether he ploughs it himself. Supervising agriculture is one grade; putting your own hands on the plough is the grade below.
Which tells you what the taxonomy is actually tracking: not virtue in any general sense, and not piety, but distance from manual labour and proximity to the state. A man near the king outranks a man near the market outranks a man near the soil.
The bottom term is where the chapter's real argument is, and it is a strong one.
A brahmana who is jealous and causes harms to others is spoken of as a chandala brahmana.3
Take the measure of that word choice.
Chandala is the tradition's most heavily loaded exclusion term. Debroy's endnote on the shudra sub-classes defines a chandala by parentage — brahmana mother, shudra father — a person whose category is fixed at conception by a crossing they had no part in.4 The vault holds the historical material on what that meant in practice: a wooden block carried as a warning, an architecture of avoidance built into the fabric of towns.5
And here that word is applied to a man of the highest birth, on the basis of two behaviours: envy, and doing harm.
Three things follow.
First, the criterion is unusually low-bar. Not apostasy, not ritual violation, not eating something forbidden. Jealousy. An interior state anyone can have and most people do.
Second, it is entirely about other people. Both terms — jealous of others, harm to others — are relational. The whole rest of the chapter's conduct material is about your own body, your own timings, your own recitation counts.6 This one clause is the only place in it where the measure of a man is what he does to somebody else.
Third, and this is the part that bites: it is a rhetorical move, not an administrative one. Nobody was reclassifying anyone. There is no procedure here, no authority named, no consequence stated. The text does not say a chandala brahmana loses his thread or may not officiate.
What it does is borrow the moral force of an existing exclusion and point it at the elite. That is a preacher's move rather than a lawgiver's, and it is one of the sharpest things in the volume — because the only way it lands is if the audience already agrees that being a chandala is the worst thing available, which means the sentence relies on the very hierarchy it is criticizing.
You are a brahmana and you have just heard this read out in a tent.
Which one are you?
Not vidvan, because you know what your conduct actually is when nobody is watching, and you know how much of the Veda you have really got.
Not vipra either.
Dvija, probably. Twice-born. Which is the word everyone uses for everyone anyway, and you had never before heard it as a demotion.
You do not plough. You do not trade. You are not in anyone's service. So the middle grades do not touch you.
And then the last one, and you sit with it for a while, because the man three rows in front got the recitation your family used to get, and you have thought about that more than once this month, and the text has just called that jealousy and attached a word to it that you would not use about a stranger.
Nothing happens. No one asks. The reciter has already moved on to the correct hour for defecation.7
But you have been handed a category you can apply to yourself, privately, forever, and no ritual will ever clear it, because the chapter names no remedy for this one.
That is the whole mechanism. A taxonomy with no enforcement is a taxonomy the audience has to run on themselves.
Worth being precise, because the chapter does not stop at brahmanas — and what it does next changes how generous the passage looks.
A kshatriya who rules over the earth is rajan, the others are held to be ordinary kshatriyas. Those who buy and sell grain are vaishyas, the others are spoken of as vanijakas. A person who serves brahmanas, kshatriyas and vaishyas is referred to as shudra. A shudra who tills the land is vrishala, the others are dasyus.8
Every varna gets subdivided. But look at the shape of each subdivision.
The kshatriya split is two grades: ruler, and everyone else. The vaishya split is two grades: grain-dealer, and everyone else. The shudra split is two grades: one who tills, and dasyus — a word meaning bandits, outsiders, the pre-Aryan enemy of the oldest Vedic material.
The brahmana split is seven.
So the resolution of the taxonomy is not uniform. The higher the varna, the finer the grading — and the finest grading by a factor of three and a half is reserved for the class doing the writing.
And the downward version is not conduct-based in the same way. A kshatriya is graded by whether he rules; a vaishya by what he trades; a shudra by whether he tills. Those are occupations, not behaviours. Nobody below a brahmana is offered demotion for jealousy, and nobody below a brahmana is offered promotion for good conduct either.
The regrading mechanism exists in one varna only. Which means the passage's apparent egalitarianism is nothing of the sort. It is an internal disciplinary vocabulary for an elite, dressed in the language of universal conduct.
Be exact about what the chapter claims, because it is easy to over-read in either direction.
It does not say a chandala brahmana becomes a chandala. He remains a brahmana — the noun is still there, with an adjective in front of it. Every one of the seven terms is [X] brahmana, not [X].
It does not say a shudra of good conduct becomes a brahmana. That claim exists elsewhere in this volume — chapter 24 offers it, in the form of two and a half million japa repetitions9 — but it is not here, and the two passages are never connected.
It does not attach a single consequence. No loss of office, no exclusion from rites, no remedy, no procedure.
So the honest summary is: conduct regrades your standing without altering your status. You keep the birth and you lose the name.
Which is a real thing rather than a nothing. In a society where a brahmana's living came substantially from being invited to officiate, a vocabulary that lets a village say he is a shudra brahmana has teeth — social teeth, applied by neighbours, with no institution required.
But it is not what a modern reader is tempted to see here, which is birth being overridden by merit. Birth is untouched. What has been added is a way of talking about people whose birth cannot be argued with.
The structural question, and the reason this page sits in eastern-spirituality with a handshake to history rather than the other way round.10
Why would a birth-based system develop a conduct-based grading of its own top class?
The available answer is that a hereditary elite has a quality-control problem no hereditary system can solve. If standing comes from birth, then birth also produces the incompetent, the lazy and the vicious, and the system has no mechanism to remove them — that is what hereditary means.
So a vocabulary appears alongside it. Not a procedure, because a procedure would contradict the hereditary principle outright. A vocabulary — he is technically a brahmana, and he is a shudra brahmana — which does the sorting work while leaving the principle formally intact.
That reading explains everything odd about the passage. Why the grading is seven-fold for brahmanas and two-fold for everyone else: only the elite has the problem. Why there are no consequences attached: attaching one would create a rival authority to birth. Why the bottom term borrows the most loaded word available: the vocabulary has to carry all the force, because it is not backed by any power to act.
And it explains why the text can hold this passage and chapter 24's japa-based varna escalator without either one disturbing the other. They are not competing theories of caste. One is an internal disciplinary register, and the other is a soteriological promise, and neither was ever an administrative rule.
Strongest support is the asymmetry of resolution — seven grades for brahmanas, two for each of the other three varnas — which is difficult to explain on any reading except that this is an elite's internal vocabulary.
Tension one. The chapter's most severe term borrows its force from an exclusion the chapter is implicitly criticizing. Chandala brahmana only stings if the audience accepts that being a chandala is the worst thing available.
Tension two. Conduct regrades a brahmana downward and never regrades anyone upward. There is no brahmana shudra.
Tension three. Chapter 24 lets a shudra become "a brahmana who possesses the mantra" at two and a half million japas.11 Chapter 20 has no such route. Neither passage acknowledges the other.
Tension four. The volume elsewhere states that "devotion enables a man to rise above his varna and obtain fruits that are disproportionately more."12 That is a fourth position on the same question, in a third chapter.
Open question. Is the seven-fold list this text's own, or a standing dharmashastra formula being quoted? The terms vidvan, vipra and dvija are all common; the compound grades are less so. Debroy glosses sadachara and annotates the shudra sub-classes at length but says nothing about the brahmana list's provenance, which given his habit of flagging sources is mildly suggestive that he did not recognize it as a quotation.
The vault holds a philological reform of exactly this term, and the primary text does not support it.
Brahmana: Not Caste but Knower of Brahman argues that the word properly names a knower of brahman rather than a birth-class, and that the caste reading is a later hardening.
Chapter 20 uses it unambiguously as a birth-class — a man is born a brahmana and then graded within it, and the grading never touches whether he is one. If the reform reading were the text's own, the seven-fold list would be unnecessary: a jealous, harmful man simply would not be a brahmana at all.
That is not a refutation of the philological point, which is about an earlier stratum. But the vault should record that by this text's period the two senses have fully separated, and that this chapter needs them separate — the whole rhetorical force of chandala brahmana depends on the birth-sense being uncontested, so that the conduct-sense can be laid over it as an insult.
The Kali Yuga Varna Audit is the same author making the same move at greater volume — brahmanas selling the Vedas, kshatriyas fleeing, vaishyas rigging balances. Read together, the audit is the diagnosis and this is the vocabulary the diagnosis is written in.
The Chandala's Wood Block: Caste Architecture (history). The historical record shows chandala as an enforced status with physical infrastructure — a required signalling object, routes that could not be taken, an avoidance geometry built into settlements.
Chapter 20 lifts that word out of its enforcement apparatus entirely and uses it as an adjective for a badly-behaved priest.
What the pair produces: a term of real exclusion becomes available as a rhetorical intensifier precisely in proportion to how thoroughly it is enforced elsewhere — the insult works because the reality is unambiguous. And the borrowing runs one way only. Nobody in this literature calls a well-behaved chandala a brahmana chandala. So the traffic between the moral register and the administrative one is asymmetric by design: the elite can borrow the bottom category's stigma to police itself, and the bottom category gets no corresponding loan. That asymmetry is worth stating plainly, because it is the mechanism by which a hierarchy's most humane-sounding internal criticism leaves the hierarchy exactly where it was.
Brand Equity Beyond Purchase (business). The operator material distinguishes a credential — a fact about you that is checkable and permanent — from standing, which is renegotiated continuously by behaviour, and observes that where the credential cannot be revoked, all the real sorting migrates to standing.
Chapter 20 is that migration in a hereditary priesthood. The credential is birth and cannot be revoked. So a seven-grade standing vocabulary grows beside it, with no institution, no procedure and no appeal.
The insight neither gives alone: an unrevocable credential does not prevent quality sorting, it privatizes it — the sorting still happens, but it happens in the language neighbours use rather than in any body that could be held to a standard. That is worse for the person being sorted, not better, because a procedure can be contested and a vocabulary cannot. And it predicts the shape of this passage exactly: maximum descriptive resolution, zero stated consequence, and the most loaded possible term at the bottom. A system that cannot fire anyone will develop very precise words for the people it would have fired.
Sharpest implication. Chapter 20 grades brahmanas into seven conduct-based classes and every other varna into two, which makes it an elite's internal disciplinary vocabulary rather than a theory of caste. The bottom grade — a brahmana who is "jealous and causes harms to others" — is called a chandala brahmana, borrowing the tradition's heaviest exclusion term and pointing it at the highest birth, on a criterion low enough that most people meet it. And it attaches no consequence, names no authority, and offers no remedy. That is the point: an unrevocable credential does not prevent quality sorting, it privatizes it, and a system that cannot remove anyone will develop very exact words for the people it would have removed.
Generative questions.