Somebody hands you a filing system with twelve boxes in it. Five are labelled primary. Seven are labelled secondary.
You would assume the line between them is about importance, or depth, or how much of the thing each box holds.
Now check which is which. And here is the thing you find, once you lay the twelve out against the tradition this system inherited its vocabulary from: the seven in the secondary pile are exactly the ones somebody else invented.
Not roughly. Exactly.
Viśvanātha's root opens the Northern Division by naming them:
Hāsya(laughter),adbhuta(astonishment),vīra(heroism),karuṇa(compassion),raudra(anger),bhayānaka(fear) andbībhatsa(disgust) — these seven are known asgauṇa-bhakti-rasa.1
Each gets a sthāyibhāva: hāsa, vismaya, utsāha, śoka, krodha, bhaya, jugupsā.1
The shelter is the same for all seven — the five kinds of devotee. These gauṇa-rasas are seen to arise in the five types of devotees previously mentioned. Therefore, the five types of devotees are the āśrayālambana of these secondary rasas.1
And the object is nearly the same for all seven. Kṛṣṇa, his devotees, and persons related to his devotees are the viṣayālambana of six of them.
Then the exception, stated flatly in the root:
Detestable objects, impure meat, blood and so on are the
viṣayaofbībhatsa-rasa. The enemies of Śrī Kṛṣṇa may also be theviṣayaofraudra-rasaandbhayānaka-rasa.1
Bharata's Nāṭyaśāstra gives eight rasas: śṛṅgāra, hāsya, karuṇa, raudra, vīra, bhayānaka, bībhatsa, adbhuta. Śānta is a ninth, added later and contested — the vault's Pollock-derived material has the history.
Now put Rūpa's twelve beside that list.
The seven secondary rasas are Bharata's eight minus śṛṅgāra. All seven. No additions, no omissions, no substitutions.
And the five primary: śānta — the ninth rasa, not Bharata's. Mādhurya — śṛṅgāra transposed and renamed. Dāsya, sakhya, vātsalya — the three the Gauḍīya tradition brought.
So the line running through the twelve boxes is not a line between deep emotions and shallow ones. It is the line between what this tradition inherited and what it added, and exactly one rasa crosses it: śṛṅgāra, which crosses by being renamed.
This is a complete enumeration of twelve items against a complete enumeration of eight, not a sample, so it needs no hedging as arithmetic. The reading — that the primary/secondary distinction is a provenance split wearing a structural costume — is a reading, and it is stated as one. [n=1 OBSERVED] on the interpretation. What is not interpretation is the correspondence, which anyone can check in two lists.
And it makes the retraction legible in advance. When the close of this same Text reclassifies the seven gauṇa-rasas as transitory states of the five, the set it removes is precisely the inherited set.
The other thing you notice reading Text 24 straight through is that it is very short for what it does, and the reason is that most of it is one sentence repeated.
When laughter arises from speech, dress and activities related to Kṛṣṇa and receives nourishment from any of the primary emotions (
mukhya-parārtha-rati) which then recede into the background, it is transformed intohāsa-rati.2
When astonishment arises from seeing the uncommon pastimes of Śrī Kṛṣṇa and receives nourishment from one of the primary emotions, which then recedes into the background, it is transformed into
vismaya-rati.3
When enthusiasm arises in relation to Kṛṣṇa and receives nourishment from
mukhya-parārtha-ratiwhich recedes into the background, it is transformed intoutsāha-rati.4
And again for śoka, again for krodha, again for bhaya, again for jugupsā. Seven times, one word changed.
The mechanism does not distinguish them from each other at all. What distinguishes them is the symptom list — spitting for disgust, flaring nostrils for laughter, reddened eyes for anger — and symptom lists are downstream.
Which is worth saying plainly, because it is the structural reason the retraction costs so little. Nothing individuates a secondary rasa at the level of the machinery. If the machinery is the same for all seven, and the machinery says each of them exists only by borrowing from something else, then removing all seven at once is one move rather than seven.
Six of the seven take Kṛṣṇa or his people as object. The seventh takes impure meat, blood, detestable objects — and Text 24's later section adds the material body.5
So one of the twelve bhakti-rasas is not directed at Kṛṣṇa, at his associates, or at anyone related to them.
This vault has already recorded the same shape one division earlier, where śānta-rasa's viṣayālambana turns out to be four-handed Nārāyaṇa and Paramātmā rather than two-handed Kṛṣṇa — a finding that sits in a 1993 footnote.
Here it is again, in a different rasa, and this time in Viśvanātha's own Sanskrit root.
The two are not the same defect. Śānta's object is a different form of God; bībhatsa's is not God at all. What they share is the structure: a member of an enumerated series of Kṛṣṇa-rasas whose object is off the series. Two instances, two divisions apart, one in the 1993 layer and one in the root. [n=2 SUGGESTIVE]. The decisive test is Rūpa's Sindhu itself, where the full treatments of both would either supply a reconciliation or show there is none.
Text 17 closes with a table assigning each of the twelve rasas a colour and a presiding deity. Bhayānaka is black and Varāha's. Bībhatsa is blue and Matsya's. Vīra is fair and Kalki's. Karuṇa is smoke-coloured and Rāghava's.6
In the range read, the book supplies no reasoning for any of it. Not one line on why fear should be Varāha's, or why the boar-incarnation is the one that presides over being afraid.
That is why this page carries the table as a datum and does not build sections on it. Twelve assignments with no argument attached is not material for analysis; it is material for a footnote, and inventing the reasoning would be the failure mode.
Worth noting only for what it says about the genre. This is a book that argues for almost everything — its definition of bhakti is a repair to two named logical faults, its ālambana correction is a definitional dispute with Bharata Muni. The colour table is the one place in the rasa material where it simply asserts a list.
You are reading a devotional narrative and something in it makes you laugh, or frightens you, or disgusts you.
The apparatus's first question is not which rasa is this? It is whether a primary emotion is underneath it. The formula is the same in all seven cases: the laughter has to be nourished by a receding mukhya-parārtha-rati before it counts as devotional at all.
So the diagnostic runs downward, not sideways. Not is this laughter or astonishment but is there a relationship with Kṛṣṇa doing the feeding. Without that, the same laughter is — by Text 27's own grading — anurasa, a middle grade of counterfeit.
Second scene. You are working with bībhatsa, disgust, which in practice means the disgust at the body that arises in a devotee whose rati has developed. Text 24 calls that vivekajā jugupsā-rati — disgust arising from discrimination — and distinguishes it from prāyikī, the ordinary revulsion at foul-smelling things that anyone has.5
The distinction is worth holding, because only the first is devotional, and the two are indistinguishable from outside.
Third, and this is the one that saves error. If you are using the twelve-rasa scheme to organise anything — content, practice, a reading of a text — the primary/secondary labels are not telling you what you would assume. They are not a ranking of intensity or of value. Sorted by what the tradition actually did, they are a ledger of what came in from outside and what was made at home.
One line in Text 24's root does something the rest of the apparatus is usually careful about, and then Text 27 comes back and fences it.
The root: the enemies of Śrī Kṛṣṇa may also be the viṣaya of raudra-rasa and bhayānaka-rasa.1 So Kaṁsa can be what a devotee is angry at, and Pūtanā can be what a devotee is frightened of.
Text 27's grading: if Śrī Kṛṣṇa is the viṣaya and the enemies of Kṛṣṇa are the āśraya of hāsya or any of the gauṇa-rasas, it is known as aparasa — counted as an extreme case of rasābhāsa.7
And Text 24 says it a third way, from inside: the natural inborn anger that enemies such as Śiśupāla have toward Kṛṣṇa is not born of kṛṣṇa-rati and therefore can never be counted as bhakti-rasa.8 Then again for fear: demons like Kaṁsa, who are always afraid of the Lord, cannot be the āśraya of fear in bhayānaka-rasa because they are devoid of rati.9
So the rule is clean and the book states it three times: an enemy can be what you feel about; an enemy cannot be who does the feeling.
That is the viṣaya/āśraya fork applied with unusual discipline, and it is the same instrument that produces the pūrva-rāga asymmetry one division earlier. It also sets up the dispute at bhayānaka, where Viśvanātha will overrule Bharata on precisely this point — because Bharata's version lets the frightening thing be the ālambana, and Viśvanātha's does not.
Which is worth naming: the fence is not incidental. It is the same distinction three times, in three registers — a root clause, an offence grade, and a definitional dispute — and it is the most consistently enforced rule in the rasa half.
Strongest support. The seven, their sthāyibhāvas, the shared āśraya, the bībhatsa exception and the enemies clause are all in Viśvanātha's Sanskrit root of Text 24. The repeated formula is in the vṛtti and repeats verbatim across all seven sub-sections. The correspondence with Bharata's eight is a matter of comparing two closed lists.
Tension one. One of the twelve bhakti-rasas has an object that is not Kṛṣṇa, not a devotee, and not anyone related to one. In the root, unremarked. Second instance of the shape after śānta.
Tension two. The primary/secondary split tracks provenance exactly and is presented as structural. The book never says so, and the retraction at the close of this Text removes exactly the inherited half.
Tension three. Twelve colours and twelve presiding deities, asserted without argument, in a book that argues for nearly everything else.
Open question. If nothing individuates the seven at the level of mechanism, what makes seven the right number rather than any other partition of the emotions? Bharata's answer was dramaturgical. This book inherits the number and supplies no answer of its own in the range read.
The vault's Pollock-derived page on bhakti-rasa as the twelfth rasa presents the Gauḍīya move as expansive: bhakti-rasa deployed as a meta-category that reorganises the whole taxonomy, the canonical nine not displaced but reinterpreted as modalities of devotion.
Read this Text and the same move looks more like an annexation with a receipt. The inherited eight are indeed reinterpreted — and they are also, all seven of the ones that survive under their own names, filed as secondary, given a shared borrowed mechanism, and then reclassified out of rasa-status altogether three Texts later.
The two accounts are describing the same operation from opposite ends. The historian sees a framework generously extended. The manual shows what the extension does to the extended material: it keeps the names and takes the standing.
And the convergence with this build's own earlier pages is exact. The census found that a book named for Rūpa's ocean contains none of his verses in its rasa half. The ālambana dispute finds Bharata credited and overruled. Here the whole of Bharata's list is credited by inclusion and demoted by classification. Same move, three scales.
Categories that look structural and turn out to be a record of where things came from. Set this against Hindu and Muslim as Contested Categories, whose finding is that a pair of labels treated as describing natural kinds is doing other work entirely — the boundary tracks the history of who was classifying whom, not a real seam in the material.
The parallel is precise and the difference is what makes it useful. In the historical case the classification is imposed from outside and the people classified did not choose it. Here the classification is imposed from inside by the tradition on its own inheritance, and the tradition had every reason to present it as structural rather than genealogical, because a hierarchy of depth is a much better argument than a hierarchy of authorship.
What the pair produces that neither yields alone: a test you can run on any taxonomy that ranks its own contents. Sort the categories by when the classifier acquired them, and see whether the ranking falls out. If it does, the stated criterion is at best redundant and at worst a cover. Here it falls out perfectly — twelve items, one clean cut, one crossing case that crosses by renaming. That is not evidence of bad faith and this page does not claim it is; Viśvanātha may well have believed the depth-ranking on its merits. It is evidence that the depth-ranking and the provenance-ranking are indistinguishable on this data, which is exactly the situation in which a tradition should be asked for a criterion that separates them. It supplies two — relationship-against-reaction, and persistence-under-subordination at Text 25 — and both partition the twelve identically to provenance, so neither can be tested against it.
And a taxonomy generated by filling one sentence seven times. Set this against Fifteen Emotional Rasa Journeys — Newsletter Templates, which the vault holds as a working method: rasa categories used as slots, a template instantiated per emotion to produce a family of content shapes.
The reflexive point is the one worth having. That page treats rasa-as-template as a modern content technique borrowed from a classical source. This Text shows the classical source doing the same thing to itself — one formula (X arises in relation to Kṛṣṇa, is nourished by a receding mukhya-parārtha-rati, becomes X-rati) instantiated seven times, with the emotion name as the only variable.
Which yields a caution that applies to both. A template guarantees coverage and it guarantees that the items it produces will not be individuated by anything the template does not vary. The seven gauṇa-rasas differ in their symptom lists and not in their mechanism, and that is why all seven could be withdrawn in a single sentence. The same is true of any set of content shapes generated by slotting an emotion into a fixed frame: they will be distinguishable on surface and identical underneath, and the moment someone asks what actually differs, the honest answer is the label and the trimmings. The classical case is the demonstration, because it is the one where somebody eventually asked.
Sharpest implication. The seven secondary rasas are Bharata's eight minus śṛṅgāra, exactly — so the line between primary and secondary is the line between inherited and added, with one rasa crossing it by being renamed. They share a single mechanism instantiated seven times, nothing individuates them beneath their symptom lists, and the close of this same Text withdraws precisely the inherited set.
Generative questions.