The volume uses two words for social position and treats them as though they were one.
Debroy stops once, in a single endnote, to say they are not.
A jati is the class one has been born into and is different from a varna.1
Twenty-one words, and they are the only place in the entire volume where the distinction is made.
Everywhere else the body text runs both terms through the same sentences — the four varnas, those of mixed varnas, those born from a pratiloma marriage, "irrespective of varna," "depending on one's own varna, qualities and conduct."2 A reader without the note has one category doing two jobs and no way to see the seam.
The standard force of the distinction is worth setting out, because Debroy states it compactly and does not elaborate.
Varna is the four-fold classification — brahmana, kshatriya, vaishya, shudra. Four categories, textually defined, pan-Indian, and the thing every prescriptive passage in this volume keys on.
Jati is the actual birth-group: the endogamous community a person is born into, of which there are thousands, locally variable, and which is what anyone in the eleventh century would have been asked about if a stranger wanted to know who they were.
So one is a scheme and the other is a membership. The four varnas are a map; jatis are the territory.
And the two do not map cleanly onto each other. A jati's varna-assignment is frequently contested, locally variable, and in many cases simply unassigned — which is precisely why a text that legislates by varna and a society that operates by jati produce constant friction.
Debroy's note flags that this text is legislating in the first register and its readers were living in the second.
The distinction turns out to matter in at least three places in this volume, and the body text is unreadable in all three without it.
Chapter 20's shudra sub-classes. The endnote lists eight — vyadha (hunter), vyalagrahi (snake-catcher), vagatita, chandala, and others.3 Those are not four varnas. They are jatis, listed under a varna heading, and the confusion between the two registers is exactly what makes the passage hard to parse.
Chapter 20's seven-fold brahmana taxonomy. Vidvan, vipra, dvija, kshatriya-brahmana, vaishya-brahmana, shudra-brahmana, chandala-brahmana.4 Those are conduct-grades within a varna, and they use varna-names for internal distinctions — which only works if everyone understands varna as a scheme flexible enough to be borrowed metaphorically.
And chapter 28's "those born from a pratiloma marriage." Endnote 450: "where the mother's varna is superior to that of the father's."5 A whole class of person produced by a crossing, who has a jati but whose varna is precisely the thing in dispute.
The clearest case is chapter 24's varna escalator, and it does not work without the two senses being separable.
To counter the status of being a kshatriya, a person must perform japa five hundred thousand times. After having again performed japa five hundred thousand times, he will become a brahmana.6
A man performs a recitation and becomes a brahmana.
Now ask which sense of the word that is.
If it is jati — if the claim is that his birth-community changes — the claim is straightforwardly false in a way everyone around him could verify. His family is his family. He eats where he ate. Nobody's marriage prospects change.
If it is varna, the claim is about his position in a textual scheme, and there is no fact on the ground that contradicts it, because varna-assignment was already contested and abstract.
The escalator only makes sense in the second register. And the text's own phrasing points there: what he becomes is not a brahmana but "a brahmana who possesses the mantra" — a qualified status, ritual rather than social.7
So the ladder is real and it operates in the register where nobody has to move house.
That is not a debunking. It is what makes the passage intelligible: a ritual reclassification inside a scheme, not a claim about anyone's neighbours — and it explains why such passages could circulate for a millennium without producing any observable social movement or any recorded objection.
You are asked who you are and you give your jati.
That is the answer. It is the name of the group, and the person asking now knows where you can eat, who you can marry, and what your family does, which is what they wanted.
Nobody has ever asked you your varna. It is not a question that comes up. If pressed you would say the obvious thing, and if pressed further you would say that the people two villages over dispute it, and that the dispute is old and has a history and nobody wins it.
And then the recitation happens and the text spends the whole evening legislating by varna.
Which mantra by varna, which clay by varna, which rudraksha colour by varna, which lingam material by varna.8
You listen to all of it in a category you do not use.
The text's whole apparatus of assignment runs on a variable nobody in the room actually navigates by — and every escape clause in it, every there is no harm in worshipping a lingam meant for another, is doing the work of making a scheme fit a society it does not describe.
The reading cost is substantial and worth itemizing.
Without the distinction, the shudra sub-class list reads as a subdivision of a varna, which makes it look like the four-fold scheme is finer-grained than it is. With it, the list is visible as jatis being filed under a varna heading — an act of classification rather than a description.
Without it, the brahmana taxonomy's use of kshatriya-brahmana and shudra-brahmana reads as incoherent. With it, the varna-names are legible as borrowed labels, available for metaphorical use precisely because varna is a scheme rather than a group.
And without it, the escalator reads as a claim about birth that anyone could falsify. With it, it reads as a ritual reclassification, which is what it is.
Three passages, all significantly clarified, by twenty-one words in an endnote.
Which is this volume's now-familiar pattern: every taxonomy it invokes as common knowledge is defined only in the annotation layer, and this is the extreme case — the distinction is not merely undefined in the body, it is actively collapsed by the body's usage.9
The general principle, stated as precisely as the evidence allows.
This volume operates two registers of social classification and never labels either.
The birth register is where jati lives. It is local, thousands-fold, endogamous, and it determines the facts of a life. The volume touches it almost never — the chandala's parentage, the pratiloma marriage, the eight shudra sub-classes, all in endnotes.
The ritual register is where varna lives. It is four-fold, textual, and it is what every prescriptive passage keys on. Which clay, which mantra, which colour, which count.
And the two registers behave completely differently under the volume's own rules.
The ritual register is permeable — assignments are announced and waived, materials may be substituted with no harm, japa moves you up it, and "everyone, women and others, have a right to worship the lingam."10
The birth register is fixed — the chandala's status is set by his parents before he exists, and nothing in the volume touches it.
The permissiveness everyone quotes belongs entirely to the ritual register. The fixity belongs to the birth register. And because the text uses one word for both, the permissiveness reads as though it applied to everything.
That is the finding, and it required an endnote to see.
The last observation, and it is about the source rather than the doctrine.
Debroy makes this distinction once, in twenty-one words, annotating a passage where the body happened to use jati.
He is a literal translator with a stated method — word for word, interpretation relegated to footnotes.11 So the note appears because a word appeared, not because he set out to clarify the system.
Which means the clarification is accidental, and its scope is accordingly narrow. He does not go back and apply it to the shudra sub-classes, or to the brahmana taxonomy, or to the escalator. The reader has to carry it forward themselves.
And it raises a question the volume cannot answer: how many other distinctions were collapsed by the body text and never happened to trigger a note?
The two endnote taxonomies this build has already found — transgressive women, shudra sub-classes — both survived only because a word appeared that needed glossing.12 Neither is in the body. There is no reason to think they are the last.
The annotation layer is not a supplement to this text. It is the part of the text that still has its distinctions in it.
Strongest support is that three separate passages — the shudra sub-classes, the seven-fold brahmana taxonomy, and the japa escalator — become substantially more coherent under the distinction and are confused without it.
Tension one. The body text uses both words through the same sentences and collapses them. The distinction exists in the volume only as an endnote.
Tension two. The volume's celebrated permissiveness (waived assignments, substitutable materials, the japa escalator, "everyone has a right") operates entirely in the ritual register, while its fixed exclusions operate in the birth register — and one word covers both.
Tension three. The escalator's "brahmana who possesses the mantra" is a ritual-class term, and its unqualified "he will become a brahmana" one sentence earlier is not. Both in the same passage.
Tension four. Debroy's note is triggered by a word rather than by a plan, so the distinction is never applied to the passages that most need it.
Open question. Does the Sanskrit distinguish consistently where the translation does not? Debroy's method is word-for-word, which suggests the collapse is the text's rather than the translation's — but he also notes elsewhere that he supplies chapter headings and makes interpretive calls where the colophon is silent.13 Volume 1 cannot settle whether the body ever uses jati precisely.
The vault holds a philological reform of the neighbouring term and this distinction sharpens what it can and cannot claim.
Brahmana: Not Caste but Knower of Brahman argues the word names a knower rather than a birth-class.
The jati/varna distinction lets that be partly right in a specific way. Brahmana as a varna term is a position in a scheme, and schemes can carry qualitative senses. Brahmana as a jati term names an actual endogamous group and cannot.
So the reform reading is defensible in the ritual register and not in the birth register — and the volume's collapse of the two is exactly what lets the argument seem either obviously right or obviously wrong depending on which passage you pick up.
Sadachara: The Brahmana Taxonomy and Japa That Changes Varna and Sex should both be read with this page in hand. The first borrows varna-names as conduct-grades, which only works if varna is a scheme. The second moves a man up the scheme without moving him between groups.
The Chandala's Wood Block: Caste Architecture (history). The historical record documents caste as lived and enforced at the level of specific communities — required markers, prohibited routes, avoidance geometry — which is jati-level operation, not varna-level.
The Purana legislates almost exclusively at varna level and touches jati only in endnotes.
What the pair produces: the textual tradition and the lived system operate in different registers, and a text's silence about jati is not evidence that jati was unimportant — it is evidence about what a text is for. Prescriptive literature legislates in the abstract scheme because that is the only level at which pan-regional instruction is possible; a rule keyed to specific communities cannot travel. Which means the entire apparatus of varna-assignment in this volume was portable by design, and its famous escape clauses were the mechanism that let a four-fold scheme land in a thousand-fold society without breaking.
Brand Equity Beyond Purchase (business). The operator material distinguishes a category — an abstract classification used for planning and messaging — from a segment, the actual identifiable group that behaves as a unit, and finds that confusing the two produces strategies that read as coherent and cannot be executed.
Varna is a category. Jati is a segment. This volume plans in the first and its readers lived in the second.
The insight neither gives alone: a prescriptive system built on categories rather than segments will generate rules nobody can quite apply, and its escape clauses are not softness but the interface layer — "there is no harm in worshipping a lingam meant for another" is what a four-fold rule needs in order to survive contact with a society organized differently. That reframes every waiver in this volume: they are not concessions to human weakness. They are translation devices between a scheme and a world, and their frequency is a measure of how badly the two fit.
Sharpest implication. Twenty-one words in a single endnote — "a jati is the class one has been born into and is different from a varna" — are the only place this volume distinguishes its two registers of social classification, and three major passages become coherent under it. Every prescriptive rule in the book keys on varna: which clay, which mantra, which rudrākṣa colour, which liṅgam material. Every fixed exclusion operates at jati level and appears only in endnotes: the chandala's parentage, the eight shudra sub-classes, the pratiloma crossing. The two registers behave completely differently — the ritual one is permeable (assignments waived, materials substitutable, japa moves you up it) and the birth one is fixed — and because one word covers both, the permissiveness everyone quotes reads as though it applied to everything. It does not. And the distinction survives only because a word happened to appear that needed glossing.
Generative questions.