A woman is terrified of a snake.
Ask what the fear is about and everyone answers the same way: the snake. Obviously the snake. What else would it be about?
Now ask it in two slightly different English sentences and watch them come apart.
The snake is the thing in which her fear is located. — No. Her fear is located in her. The snake is the thing by which her fear is caused. — Yes.
Two prepositions. In English they blur, and most readers would let either sentence stand. In Sanskrit they are different cases and they cannot blur, and this Text stakes the shape of an entire taxonomy on which one the technical term means.
At the close of the bhayānaka section, Jīva Gosvāmī sets out both positions:
Śrī Jīva Gosvāmī comments that acceptance of the wicked as
āśrayaof this rasa instead of a devotee is in accordance with the opinion of previous authorities onnāṭya-śāstralike Bharata Muni.1
Then Viśvanātha's:
In accordance with the ordinary
rasa-śāstras,hāsaand the other secondary emotions are accepted assthāyibhāvasof the seven secondary rasas only because they take on the characteristics ofrati, being nourished bymukhya-parārtha-rati. According to this opinion,dāruṇa, or the wicked, are accepted asālambanaofbhayānaka-rasa. But in the author's own opinion,ālambanarefers to that IN WHICHratiis stimulated (the devotee) and NOT BY WHICHratiis stimulated (the wicked, in the case of fear).1
That is the dispute. A locative against an instrumental.
And it is not a quibble about a word, because ālambana is the term the whole apparatus hangs from. Vibhāva divides into ālambana and uddīpana. Ālambana divides into viṣaya and āśraya. Everything downstream — who can occupy which slot in which rasa, what counts as a legitimate combination, what counts as a counterfeit — is decided by what ālambana is taken to mean.
Under Bharata's reading, Pūtanā is an ālambana of bhayānaka-rasa. She is what the fear is about; she is in the scene as one of its structural supports.
Under Viśvanātha's, she is not. Ālambana is the locus, and fear is located in Yaśodā. Pūtanā drops out of ālambana altogether.
So the correction evicts the frightening thing from the category it names. A rasa called the terrible turns out, on the author's own reading, to have no terrible thing among its supports.
Which is the same operation performed on hāsya two sub-sections earlier, where Jīva rules that laughter has no viṣayālambana at all because a laughed-at object would be ridicule and ridicule is inadmissible. Two of the seven secondary rasas, in one Text, lose the object their names point at.
Having removed the wicked from ālambana, Viśvanātha has to put them somewhere, and where he puts them is new:
When a friend or relative is anxious for Kṛṣṇa's safety due to the threat of a wicked person, the enemy is the CAUSE of fear, the relative is the REPOSITORY of fear (
āśraya), and Kṛṣṇa is the OBJECT (viṣaya) for whom fear is experienced.1
Three roles. Cause, repository, object.
The apparatus, everywhere else in this book, is two-place: viṣaya and āśraya, object and vessel. Text 14 fixes it that way and the five primary rasas are specified by naming exactly those two.
Here a third slot appears, is used, and is never added to the general machinery. It is not defined in the vibhāva material, it does not appear in the compatibility matrix two Texts later, and no other rasa is described with it.
That is worth recording plainly. The correction that makes Viśvanātha's reading of ālambana coherent requires a structural addition, and the addition is made locally and left there. [n=1 OBSERVED] — one passage, one three-place description, in the range read.
He gives two configurations, and the ordering is a surprise:
When there are two different devotees who are both fit for the Lord's mercy (
anukampya), one may become angry with the other due to some situation concerning Śrī Kṛṣṇa. In this case, the devotee who is angry will be theviṣayaof fear, the other devotee will be theāśrayaof fear, and Kṛṣṇa will be the cause. This is a common scenario. The second circumstance is more specialized — when a friend or relative is anxious for Kṛṣṇa's safety due to the threat of a wicked person.1
So the ordinary case of devotional fear is one devotee frightened of another devotee's anger, with Kṛṣṇa as the cause of the quarrel.
Not fear of demons. Not fear of Kṛṣṇa's power. Fear of a fellow devotee who has become angry about something to do with him — which, read against the raudra section's household of women furious at each other's supervision, is exactly the room this taxonomy has been describing all along.
The demon case, the one every retelling of the tradition leads with, is marked more specialized.
Before the dispute, the section runs conventionally and it is worth having on the record.
Bhaya is defined as the extreme agitation and restlessness of the heart experienced upon committing some offence or upon seeing a dreadful object.2 Two sources: your own offence, or a dreadful presence.
A person is dreadful for three reasons: ākṛti — a dreadful form, like Pūtanā; prakṛti — a dreadful nature, like Śiśupāla; prabhāva — awesome power, like that of Indra and Śaṅkara.2
That third one repays attention. Indra and Śiva are named as objects of fear on grounds of sheer power — and both are adhikṛta-bhaktas, appointed servants of Kṛṣṇa, in Text 19's own list. A devotee may legitimately be frightened of another devotee simply for being powerful, which is the same finding as the common scenario above, arriving by a different route.
The āśraya side has two classes: anukampya, one who has committed an offence but remains a fit object of the Lord's favour — Jāmbavān, who fought Kṛṣṇa not realising who he was, and Kāliya after their fight; and bandhu, a relative apprehensive for his safety.3 Yaśodā supplies all three modes of the second: fear on seeing Keśī, on hearing of Keśī entering Gokula, and on merely remembering Pūtanā when a visitor asked about her.
One sāttvika note, easy to skip and rather good: all eight bodily symptoms manifest in this rasa except tears.4 Fear dries.
You are frightened inside your practice and you want to know whether that is a devotional emotion or a problem.
The definition splits it for you immediately. Fear upon committing some offence — that is one thing, and the offence architecture is where it belongs. Fear upon seeing a dreadful object is the other, and the three grounds tell you which kind you are in: a form, a nature, or a power.
If it is power — and in practice it usually is, because what frightens people in a tradition is rarely a demon and often an institution or a senior figure — the passage's own example is Indra and Śaṅkara. Being frightened of somebody powerful inside the household is a catalogued devotional state, not a failure of faith.
Second scene, and this is the one the taxonomy actually optimises for. Another practitioner is angry, the anger is about something to do with the practice, and you are afraid of them.
Viśvanātha calls that the common scenario and gives it the fullest three-role description in the section: they are the object, you are the vessel, and Kṛṣṇa is the cause. Which is not consolation, but it is precision — and it locates the cause somewhere other than in either of you.
Third. Notice what the correction does for you as a reader. Under Bharata's version you would be looking for the frightening thing in the scene. Under Viśvanātha's you look for where the fear is sitting, and the frightening thing is demoted to a cause. Practically, that is the difference between organising your attention around the threat and organising it around your own state — and the whole apparatus, once corrected, points the second way.
The obvious question about this dispute is why it appears here, in the sixth of seven secondary rasas, rather than in Text 14 where ālambana is defined.
The answer is that bhayānaka is the first place the two readings come apart in practice.
Run them both on the five primary rasas and nothing happens. Śānta, dāsya, sakhya, vātsalya, mādhurya — in every one, the thing the emotion is about is Kṛṣṇa, and Kṛṣṇa is also unproblematically a support of the rasa. In which and by which pick out different things in principle and the same thing in fact, so the ambiguity costs nothing.
Bhayānaka is where a rasa's emotion is about something that is not a legitimate support — a demon — and the two readings finally disagree about who is in the scene.
Which means Bharata's definition had a latent defect that only a devotional application could expose. In a dramaturgy, the frightening thing is obviously one of the supports; it is on stage, the audience can see it, and nothing depends on distinguishing the locus of the emotion from its cause because the locus is the audience and the audience is not part of the scene. Move the vessel into the scene — which is the whole transposition — and the distinction becomes load-bearing.
So this is not Viśvanātha catching an error Bharata made. It is a definition that worked for a thousand years in the setting it was built for, failing on first contact with a setting it was not.
And the cost is proportionate. The repair requires a third role the apparatus does not otherwise have, and that role is introduced in a subordinate clause and never generalised. A definition corrected in one rasa; a machine left two-place everywhere else.
Strongest support. Both positions are stated in the same paragraph, Bharata's attributed through Jīva Gosvāmī and Viśvanātha's marked as the author's own opinion, with the in which / by which contrast spelled out. The three-role redescription follows immediately with both configurations worked.
Tension one. The correction evicts the frightening thing from a rasa named for it, and requires a third structural role — cause — that appears here and nowhere else in the range read.
Tension two. The common scenario of devotional fear is a devotee afraid of another devotee's anger. The demon case is marked more specialized, which inverts the emphasis of every popular presentation of this material.
Tension three. Indra and Śaṅkara are named as objects of fear on grounds of power, and are adhikṛta-bhaktas in the same book. Fear of a fellow servant, licensed and unremarked.
Open question. Does the third role have a Sanskrit term in Viśvanātha's own text? The English gives cause; the root of Text 24 does not supply a technical word for it in the range read, and whether Rūpa's Sindhu at 4.6 has one is not answerable from here.
The vault's Pollock-derived material carries the Nāṭyaśāstra apparatus as literary theory, and ālambana in that setting is untroubled: the persons and objects the emotion rests on, present in the depicted situation, available to a spectator's inspection.
That account has no reason to distinguish locus from cause, and this page's argument is that it had no occasion to. The distinction is invisible until the spectator is moved inside the scene.
Which is the same structural point the pūrva-rāga page reaches from the other side. There, the moment the vessel becomes a participant, the other participant has to be reclassified as material — so Kṛṣṇa's own longing becomes uddīpana. Here, the moment the vessel becomes a participant, the frightening thing has to be reclassified as cause. Two reclassifications, one relocation, and neither passage cites the other. [n=2 SUGGESTIVE] on the transposition producing a family of forced demotions rather than isolated ones.
And it converges with the build's standing observation about the method. Bharata is named, credited with the position, and overruled — in a subordinate clause, without argument beyond a definitional preference, in the middle of the sixth secondary rasa. The honouring and the working-around are the same sentence.
A technical dispute conducted entirely at the level of a preposition, with the whole framework downstream of it. Set this against Preposition Shift as Metaphysical Distance, where Macfarlane distinguishes writing about a river, writing with a river, and being written by a river — and treats the three not as stylistic alternatives but as three distinct metaphysical postures, each determining what the writing is, what the writer is for, and how the work is done.
The match is close enough to be uncanny. Both cases turn on a case-marking that English flattens; both treat the flattening as consequential rather than pedantic; and in both, the shift relocates the agent. Macfarlane's by makes the river the author and the writer the conduit. Viśvanātha's rejection of by which refuses to let the frightening thing be a support and leaves the devotee as the sole locus.
The instructive difference is direction. Macfarlane's sequence moves agency outward — from the writer, to a collaboration, to the river writing through a human body. It is a progressive dissolution of the subject, and each step is presented as a deepening.
Viśvanātha moves it inward. The scene may contain a demon, a threat, a cause; none of them may hold the rasa. Only the devotee can. The correction is a hardening of the subject, not a dissolution of it — and it is made in a tradition whose reputation is for self-abandonment.
What the pair produces that neither yields alone: a case-marking dispute is a reliable place to find a framework's real commitments, because prepositions are where a system says who is doing what without having to argue for it. Both writers know this and both use it deliberately. And the specific yield here is that the two traditions, both concerned with an experience overwhelming a person, place the locus in opposite places — the nature writer surrenders it to the river, the devotional theorist refuses to surrender it to the demon. Which suggests the question worth asking of any account of being-overwhelmed is not how much of the self survives but what the grammar does with the agent, and that this can be checked in a sentence rather than argued about in a paragraph.
Sharpest implication. An entire taxonomy corrected at the level of a case-ending: ālambana means that in which an emotion is located, not that by which it is caused — which evicts the frightening thing from the rasa named after it, and requires a third structural role that appears once and is never generalised. And the correction only became necessary because a dramaturgical definition was moved into a setting where the spectator stands inside the scene.
Generative questions.