Something has gone wrong in a household and everyone is angry, and if you watch carefully you can draw the family tree from the anger alone.
The grandmother is angry at the mother for not watching him. The mother is angry at the older boys for taking him somewhere. One of the girls is angry at another girl for sulking so hard the whole afternoon collapsed. Nobody is angry at the child.
You could reconstruct every relationship in that room from the direction of the fury, and you would not need to know what anyone said.
That is precisely how this Text organises devotional anger. Not by heat. By target.
The
viṣayālambanaofraudra-rasais of three types: Śrī Kṛṣṇa,hita(a well-wisher) andahita(one who is hostile).1
Which is already unusual. In most of the twelve, the viṣaya is Kṛṣṇa and the variation is all on the vessel side. Here two of the three objects are other people, and the bulk of the section is about them.
The well-wishers who get shouted at come in three kinds:
(1)
Anavahita(inattentive). Those responsible for protecting Kṛṣṇa who become inattentive to His safety due to carrying out other activities for Him. Śrī Yaśodā became the object of Rohiṇī's anger for tying Kṛṣṇa to the grinding mortar, when the latter saw Him wandering between the fallen trees.(2)
Sāhasī(rash). Those who incite Kṛṣṇa to go to a dangerous place. Kṛṣṇa's friends became the object of Śrī Yaśodā's anger for leading Him to the Tālavana Forest.(3)
Īrṣyu(jealous).Gopīswhose only treasure is indignant anger and whose minds are always besieged by jealousy.1
Note the first one's clause. Yaśodā was inattentive because she was doing something else for him. The negligence is itself service. That is the finest cut in the section and it is made in nine words.
The third type gets a story, and the story folds back on itself:
Śrī Rādhā once displayed such persistent indignant anger toward Śrī Kṛṣṇa that in spite of all attempts, both by Him and the
sakhīs, Hermānacould not be broken. Finally, Kṛṣṇa left in despair. Because of Kṛṣṇa's departure, Śrī Rādhā became distressed and Hermānadesisted. She begged Śrī Lalitā to go and bring Kṛṣṇa back. At that time, She became the object of Lalitā's anger for being so unrelenting in Hermāna.1
So: Rādhā is angry at Kṛṣṇa. The anger drives him off. Rādhā, bereft, asks Lalitā to fix it. Lalitā is angry at Rādhā for having been angry.
Devotional anger, whose object is devotional anger. In the highest rasa, aimed at the tradition's supreme devotee, and counted as legitimate.
The book does not remark on the recursion. It files the case under īrṣyu and moves to the next heading.
The second class of object splits two ways:
(1)
Svasyāhita(hostile to oneself). Someone who impedes one's relationship with Kṛṣṇa. Akrūra became the object of thegopīs' anger because he took Kṛṣṇa away from Vṛndāvana.(2)
Harer ahita(hostile to Kṛṣṇa). The enemies of Kṛṣṇa.2
Akrūra came to Vraja on Kaṁsa's order and Kṛṣṇa's own consent, and drove the chariot to Mathurā because that is what the narrative required of him. He is, in the tradition's own literature, a devotee.
He is filed here under hostile.
Not hostile to Kṛṣṇa — the category is careful about that, and svasyāhita means hostile to oneself, the self being the gopī doing the being-angry. Which is exactly right and also exactly the problem: the classification is entirely indexed to the angry party's interest, and by that index a devotee performing his function becomes an obstacle.
The book supplies no defence of Akrūra here and no note that one might be wanted. [n=1 OBSERVED]. What is not a reading is that the category is named hostile to oneself rather than hostile, so the text has done the work of marking the index — and then used the marked category to house a man who was obeying.
The āśraya line is short and it is a licence list:
When Kṛṣṇa is the object of anger, the
sakhīsandjaratī(elderlygopīslike Jaṭilā) are theāśrayālambana. All kinds of devotees may be theāśrayawhen a well-wisher or hostile person is the object.3
Anger at other people: open to everyone. Anger at God: two classes only. The confidantes of his beloved, and the old women.
And look who is named for the second class. Jaṭilā — Rādhā's mother-in-law, the household's enforcer of respectability, the figure whose entire narrative function is to obstruct the affair.
So the licence to be angry at Kṛṣṇa extends to the person policing the relationship against him. Which is either a very generous reading of Jaṭilā or a very precise observation that the position from which one may be angry at God is a position of standing in his household, and standing does not require approval.
The alternative reading is available and duller: that jaratī names an age-class and Jaṭilā is simply the best-known member of it. The passage supports either.
Now the exclusion, and it is the largest thing in the section.
Krodha-rati divides three ways. Kopa, fury at enemies. Manyu, indignation at relatives, itself subdivided by the relative's seniority — pūjya at worshipful elders (Rādhā's slight indignation at Paurṇamāsī, who had instructed her on chastity toward her husband), sama at equals (Mukharā at Jaṭilā, for accusing Kṛṣṇa of ruining her daughter-in-law's reputation), nyūna at juniors (Jaṭilā at Rādhā, for wearing Kṛṣṇa's necklace). And roṣa, the anger a woman shows her lover.
Then:
In
mādhurya-rasa,roṣais counted as avyabhicāri-bhāva, or transitory emotion, not as a dominant emotion. The anger of a woman toward her lover does not attain to the status ofsthāyibhāva, as is the case with the anger of thesakhīsand elderlygopīstoward enemies (kopa) or relatives (manyu).4
Roṣa is māna — the jealous anger that Text 22 calls famous in mādhurya-rasa, that the whole literature of the gopīs runs on, that Rādhā's unbreakable sulk two paragraphs earlier is an instance of.
And it is not a sthāyibhāva. So the tradition's most celebrated anger at Kṛṣṇa is not raudra-rasa.
The pattern is one this cluster has now seen twice. Mādhurya is excluded from the general grammar of separation and given vipralambha and sambhoga instead. Māna is excluded from the general anatomy of anger and left as a transitory state inside mādhurya. [n=2 SUGGESTIVE] — and consistent with a third observation, that mādhurya receives the thinnest apparatus of the five primary rasas.
The celebrated material is repeatedly handled outside the general machinery, and the decisive test is Ujjvala-nīlamaṇi, which is where all of it is sent.
You are angry inside your practice and you want to know what to do with it.
The passage's first move is not to ask how angry. It is to ask at whom, and the three answers route to three different places. At an enemy of the tradition: kopa. At someone inside it: manyu, and then a second question about their seniority. At Kṛṣṇa himself: a licence the taxonomy restricts to two classes.
That last restriction is worth taking seriously before importing anger-at-God as a devotional posture. The text permits it and it permits it to almost nobody.
Second scene, and it is the common one. Someone in your practice community failed at something while doing something else for the practice. The anavahita category is exactly that case — negligence produced by service — and the taxonomy does not excuse it. Rohiṇī is angry at Yaśodā and the anger is a devotional emotion, not a lapse.
Which is the useful counsel: the tradition does not require you to reclassify a failure as fine because its cause was devotional. It gives the anger a name and a slot.
Third. If your anger is at someone who impeded your access — a teacher who moved, an institution that reorganised, a person who took the thing away — that is svasyāhita, and the category is honest about being indexed to you. Hostile to oneself. Not hostile. The word is doing real work and it is worth keeping when you use the concept, because Akrūra is what happens when it gets dropped.
Lay out what actually varies across this section.
The sthāyibhāva does not vary: it is krodha, the inflammation of the heart that takes place when confronted with hostility, defined once. The anubhāvas do not vary by type — one list of fourteen serves the whole rasa, from pounding the hands to reddishness at the corners of the eyes. The sāttvika-bhāvas do not vary: all eight.
What varies is who the anger is pointed at, and everything the section calls a subdivision is a fact about that person's relation to the angry one.
Kopa/manyu/roṣa = enemy / relative / lover. Pūjya/sama/nyūna = senior / equal / junior. Anavahita/sāhasī/īrṣyu = failed to watch him / led him into danger / drove him away. Svasyāhita/harer ahita = blocks me / opposes him.
Then the vyabhicāri allocation, which confirms it: fierceness is exhibited toward enemies, indignation toward relatives and jealousy toward the beloved.4
This is a kinship-and-standing grid with an emotion word on the front of it. Which is not a criticism — it is a genuinely good instrument, and it is why the section can produce a case as fine as inattentive because she was doing something else for him.
But it does have a consequence the book never draws. If the anger itself does not vary, and only the target does, then raudra-rasa is not measuring an affective state at all. It is measuring a social position. And a devotee's position in Kṛṣṇa's household is exactly what the four servant classes, the four friend classes and the two prītis have already been measuring, by other means.
So the seventh secondary rasa turns out to be reading the same variable as the primary ones, through a different instrument — which is the strongest structural argument available for the retraction four pages later, and the retraction does not use it.
Strongest support. The three viṣaya types, the three well-wisher sub-types with their named cases, the two hostile sub-types, the restricted āśraya, the three-fold krodha-rati with manyu's three grades, and the roṣa demotion are all consecutive in the vṛtti's Text 24 raudra section. The Śiśupāla exclusion closes it.
Tension one. Roṣa — māna — is the tradition's most-written anger at Kṛṣṇa and is explicitly denied sthāyibhāva status here, so raudra-rasa excludes the material it is most associated with.
Tension two. Akrūra is filed under hostile to oneself for performing his narrative function, with no note that the category is being applied to an obedient devotee.
Tension three. Nothing about the anger varies across the section — one sthāyibhāva, one anubhāva list, all eight sāttvika-bhāvas. Only the target varies, so the rasa is sorting relationships rather than emotions.
Open question. Jaṭilā is named as licensed to be angry at Kṛṣṇa. Is that a claim about her standing, or is jaratī simply an age-class with a famous member? The passage supports both readings and settles neither.
The vault's Nāṭyaśāstra-derived material gives raudra as one of the eight: the furious sentiment, krodha as sthāyibhāva, conventionally presented, aimed at an audience. In that setting the object of the anger is whatever the play supplies — a rival, a usurper, an insult — and the theorist has no reason to build a grid of the angry party's relationships, because the audience is not in one.
Transposed into devotion, the audience is in the relationship, and the whole section reorganises around that. Bharata's category is kept and its internal structure is replaced.
Which is now the third distinct thing this build has watched happen to an inherited rasa. Hāsya lost its object, because ridicule could not be housed. Bhayānaka will lose its ālambana definition at the close of this same Text, because Viśvanātha refuses to let the wicked be that in which rati is stimulated. And raudra keeps its name and becomes a relationship map.
Three inherited categories, three different surgeries, none of them announced as a change. [n≥3 PATTERN] on the method: this tradition renovates Bharata's rooms and leaves the nameplates up.
And it converges with the census from the front of the build. Texts 14 through 27 contain no verse of Rūpa's; the rasa half is Viśvanātha summarising in his own words. The renovations are visible precisely because the summariser had to decide, in his own sentences, what each inherited category was going to mean here.
A diagnostic that reads the target of hostile affect rather than the affect. Set this against Insult as Identity Marker, whose claim is that an insult is never a description of the person it lands on — it is a boundary marker, showing the edge of what the speaker cannot survive being seen as. We judge what we cannot afford to be. Nobody reaches for fake unless they fear there is nothing under their own performance.
Both instruments do the same unusual thing: they discard the content and read the vector. Hughes throws away what the insult says and keeps who it is aimed at. This section throws away the anger's intensity — one sthāyibhāva, one symptom list, unvarying — and keeps who it is aimed at.
Then the readings diverge completely, and the divergence is the point. Hughes reads the target as a projection: the vector points back at the speaker's wound, and what it reveals is private, defended, and probably unknown to them. This taxonomy reads the target as a position: the vector points at a place in a household, and what it reveals is public and structural. Rohiṇī's anger at Yaśodā does not tell you what Rohiṇī cannot bear to be. It tells you Rohiṇī has standing to supervise Yaśodā's supervision.
What the pair produces that neither yields alone: anger direction is a high-information signal that two serious systems read for opposite kinds of fact, and the difference between them is whether a shared structure exists. In a household with defined roles, the vector is mostly telling you about the roles. Strip the structure out — strangers, comment sections, people with no standing over each other — and the same vector has nothing structural left to encode, which is exactly the condition under which Hughes' projective reading becomes the better one.
So the two are not rivals; they are the same instrument calibrated for two environments, and the calibration question is answerable in advance. Ask whether the angry party had a position to defend before you ask what they cannot bear to be. Most contemporary anger has no position behind it, which is why the projective reading so often works — and why it fails badly inside families and institutions, where it reads a jurisdictional dispute as a wound.
One handshake rather than two. Akrūra's case — an agent punished for executing the principal's own instruction — has real cross-domain purchase, but the vault's neighbours on obedience and diffusion of responsibility describe an agent excused by orders, and this is an agent blamed by people the order harmed. Forcing the pair would blur exactly the distinction the svasyāhita label is drawing.
Sharpest implication. Nothing about the anger varies across this rasa — one dominant emotion, one symptom list, all eight bodily signs — and everything the section calls a subdivision is a fact about who is being shouted at. Raudra-rasa sorts positions in Kṛṣṇa's household, which is what the primary rasas already sort, by another instrument. And the tradition's most celebrated anger at Kṛṣṇa is expelled from it in one sentence.
Generative questions.
raudra-rasa an emotion category at all — and what would tell us?mādhurya material out of the general machinery. Is Ujjvala-nīlamaṇi a supplement, or a second and incompatible apparatus?māna is called famous in the rasa this one expels it frommādhurya material from a general apparatushāsya, which lost its object