The Nāṭyaśāstra VI.31 declares: vibhāvānubhāvavyabhicārisaṃyogād rasaniṣpattiḥ — "rasa arises from the conjunction of factors, reactions, and transitory emotions." The aphorism omits the stable emotion (sthāyībhāva) from the list of conjuncts. The omission seemed deliberate to some commentators and accidental to others. Either way, it forced every theorist who took the sūtra seriously to ask: WHAT is the relationship between the stable emotion and the conjunction-event that produces rasa? Is the stable emotion a separate thing that the conjunction produces, manifests, infers, or commonizes? Or is the stable emotion ALREADY one of the conjuncts and the sūtra is just listing it under a different name?1
The Sanskrit aesthetic tradition produced four canonical answers across four centuries. Each is named by a different verb of cognitive operation:
Pollock's apparatus makes one critical point that the four-position tradition obscures: the four readings are not arbitrary alternatives. Each reading commits to a specific ontology of rasa AND a specific epistemology of how rasa is cognized, and the two commitments determine each other. Once you settle the ontological question (what kind of thing rasa IS), the epistemological question (how rasa is known) is largely determined. The four positions are four different ontology-epistemology packages, not four different verbs for the same thing.1
Ontology: Rasa is the stable emotion enhanced by the aesthetic elements. The same desire, anger, or grief that exists ordinarily becomes more intense, more saturated, more developed under the influence of the foundational factors, the reactions, and the transitory emotions. Rasa is fully real and located in the character (Rāma, Sītā, etc.) and figuratively attributable to the actor by force of identification.
Epistemology: Rasa is produced (utpatti) — brought into being by the conjunction of aesthetic elements operating on the underlying stable emotion. The conjunction operates causally. When the right factors are deployed in the right configuration, the stable emotion is enhanced into rasa-state. The actor's identification with the character lets the rasa "exist in the actor too" but only in a derived sense.
The strength: The framework is metaphysically conservative. Rasa is an enhanced version of ordinary emotion. The aesthetic elements operate on the stable emotion the way that nutrition operates on a seedling — the seedling is already there; the aesthetic-conditions cause it to develop.
The weakness: The framework cannot explain why the audience experiences rasa. If rasa is in the character, and the character is fictional, what is the audience experiencing? Lollaṭa's framework requires either (a) the audience's rasa is a kind of sympathy-with-character (which the framework does not theorize) or (b) the audience experiences only the represented-stable-emotion-of-the-character (which leaves rasa-in-the-audience unaccounted for).
Ontology: Rasa is the stable emotion of the character, but the character's emotion is itself an imitation of a real emotion. The actor imitates Rāma's desire; the imitation is structurally similar to a painted horse — clearly not a real horse, yet recognized AS a horse. Rasa is the audience's apprehension of this imitated-stable-emotion.
Epistemology: Rasa is inferred (anumiti). The audience sees the actor's behavior (which represents Rāma's behavior), sees the aesthetic elements (which represent the situation Rāma is in), and INFERS the underlying stable emotion. The inference is performed by the audience using the same cognitive operation as any other inference — drawing a conclusion (Rāma feels desire) from premises (these aesthetic elements + the actor's behavior).1
The strength: The framework provides a precise epistemological account. The audience's cognitive operation is named (inference) and the procedure is intelligible (premises → conclusion). The actor's role is also clarified — the actor produces the premises by skilled imitation.
The weakness: The framework treats aesthetic experience as essentially the same as any other inferential cognition. But aesthetic experience does not phenomenologically feel like inference. You do not REASON your way to feeling moved by a play. The framework also fails to account for the quality of aesthetic experience — why inference about Rāma's emotions produces rapture while inference about other people's emotions in ordinary life does not. Bhaṭṭa Toṭa would later devastate this position by pointing out these gaps.
Ontology: Rasa is NOT located in the character. Rasa is something that happens in the audience — specifically, it is the audience's direct experiential apprehension of a universalized emotional condition. The character Sītā is NOT the foundational factor for the audience's desire-rasa; rather, "Sītā" gets commonized into "woman as such," and the audience apprehends this commonized factor and experiences the rasa through it.
Epistemology: Rasa is experientialized (bhukti / bhoga) through a three-stage process: (1) abhidhā (denotation — the words and stage actions communicate what is happening), (2) bhāvanā (commonization / actualization — the specific characters and situations are universalized), (3) bhoga (experientialization — the audience now SAVORS the universalized emotional condition as a direct experience, freed from spatiotemporal particularity).1
The strength: The framework relocates rasa from the impossible-to-verify character to the testifiable-by-the-audience-member's-own-experience. It explains why aesthetic experience differs from real-world emotion (the commonization step strips away particularity). It accounts for the audience's pleasure even in tragic literature (the experientialization is supermundane, not bound by the worldly pleasure-pain calculus).
The weakness: The framework requires a specific theory of how commonization happens — the bhāvanā operation. Critics (including later Abhinava) argued this remained an under-theorized intermediate step. The framework also implicitly relies on Mīmāṃsā ritual-theory (where bhāvanā is the technical term for what scriptural injunctions accomplish in the ritualist's psyche), which made it doctrinally heavy.
Ontology: Rasa is a manifestation event. The aesthetic elements are not causes that produce rasa nor premises from which rasa is inferred; they are signs through which rasa is manifested. The relationship between the elements and rasa is the same as the relationship between language and meaning under Ānanda's dhvani (implicature) theory.
The crucial recoding that Pollock makes explicit: Ānandavardhana c. 875 introduced abhivyakti as a LINGUISTIC term. Dhvani is the technical name for the implicit meanings language can convey beyond its denoted senses. Rasa is the soul of a special kind of dhvani — manifestation where the temporal succession from expressed-meaning to implicit-meaning is not registered. Ānanda's framework operates entirely at the level of LANGUAGE — how language produces meaning-effects in the receiver.1
Abhinavagupta c. 1000 took the term abhivyakti and recoded it from a linguistic operation into a psychological one. After Abhinava, abhivyakti means: a mental state in the audience-member that manifests through reception of the aesthetic elements. The locus shifts from language-processing to consciousness-event. The framework now matches Bhaṭṭa Nāyaka's reception-locus while preserving Ānanda's terminology.1
Epistemology (in the recoded Abhinavan version): Rasa is manifested (abhivyakti) — neither produced nor inferred, but revealed via the dhvani-function operating on consciousness. The aesthetic elements function as the linguistic-and-staging machinery through which the audience's stable-emotion-predisposition is brought into the manifestation-state. The audience does not produce rasa, does not infer rasa, does not experience-rasa-as-savoring-of-something-given; the audience UNDERGOES rasa-revelation.
The strength: The framework preserves Ānanda's linguistic precision while accommodating the reception-theory commitment that the discipline had embraced via Bhaṭṭa Nāyaka. It produces the most elaborated technical apparatus the discipline ever generated.
The weakness: The framework's terminological precision is somewhat false — it uses the linguistic term abhivyakti for a psychological event. Later theorists (Mahima Bhaṭṭa, Ruyyaka) noticed and complained. Ruyyaka explicitly named the recoding as a recoding; Mahima Bhaṭṭa tried to refute it by showing that manifestation reduces to inference.
Pollock's central methodological point: each position's ontology and epistemology are co-determined.1 You cannot separate "what rasa IS" from "how rasa is cognized" because each commitment determines the other:
| Position | Ontology of Rasa | Epistemology of Apprehension |
|---|---|---|
| Utpatti | Stable emotion enhanced; in the character | Caused/produced; recognition by sympathy |
| Anumiti | Imitation of stable emotion; in the apprehended-character | Inferred; via premises-to-conclusion |
| Bhukti | Universalized emotional condition; in the audience-experience | Experientialized via commonization |
| Abhivyakti (linguistic) | Linguistic implicature; rasa-as-soul-of-dhvani | Manifested via dhvani-function in language |
| Abhivyakti (psychological) | Mental state of stable-emotion-revelation; in the audience-consciousness | Manifested via dhvani-function operating on psyche |
A theorist could not, for example, adopt Lollaṭa's ontology (rasa in the character) with Bhaṭṭa Nāyaka's epistemology (apprehended via commonization in the audience), because the ontology presumes the character locus and the epistemology presumes the audience locus. The packages are tightly bound.
The framework's largest tension is the historical-philological one: the "four classical positions" narrative is partly a retroactive construction. Pollock notes that some of the positions (especially the New Scholars' cognitive-error theory, Jagannātha's eleven positions) are almost certainly never-actually-defended-by-anybody extrapolations generated to fill out conceptual possibility-space.1 The "four classical positions" are real in the sense that the four theorists existed and held the views attributed to them — but the cleanness of the four-position taxonomy is itself a later disciplinary construction, useful for pedagogy but partly artificial.
A second tension: the four positions are usually presented as a chronological progression where each fixes the deficiencies of the prior. But the progression is not as clean as the narrative implies. Lollaṭa's utpatti and Bhaṭṭa Nāyaka's bhukti are not the same kind of object — utpatti is a metaphysical claim; bhukti is a phenomenological-mechanism claim. Comparing them as "alternative positions on the same question" obscures the fact that they answer different questions. The four-position narrative levels what were probably more like four-different-projects.
A third tension: the abhivyakti category houses two structurally different positions (Ānanda's linguistic, Abhinava's psychological). The "four positions" narrative groups them, but the philosophical work each is doing is different. The grouping is a historiographical convenience that obscures a major doctrinal recoding.
Bhaṭṭa Lollaṭa and Bhaṭṭa Nāyaka, separated by 75 years, define the production/experientialization opposition that frames the rest of the four-position debate. Lollaṭa holds rasa is real and enhanced from a stable emotion in the character. Bhaṭṭa Nāyaka holds rasa is experiential and constructed via commonization in the audience. The convergence: both think rasa is a real, definite, locatable thing — not just an effect or illusion. The divergence: WHAT KIND of real thing it is. Lollaṭa's framework is realistic-emotional (rasa = real enhanced emotion). Bhaṭṭa Nāyaka's is realistic-phenomenological (rasa = real direct-experience). The split would shape every subsequent debate. Even Abhinava's abhivyakti framework would be measured against the Lollaṭa-Bhaṭṭa-Nāyaka opposition — does it preserve Lollaṭa's emotional realism (no — manifestation is not enhancement); does it preserve Bhaṭṭa Nāyaka's audience-locus (yes — manifestation occurs in the audience consciousness). (See Bhaṭṭa Lollaṭa and the Utpatti-Vāda and Bhaṭṭa Nāyaka and the Copernican Revolution.)
Śrī Śaṅkuka and Abhinavagupta, separated by 150 years, both treat rasa as a cognition-event — but their accounts diverge radically. Śaṅkuka treats it as ordinary cognition (inference) of a special object (imitated-emotion). Abhinava treats it as a supermundane cognition (manifestation) of a transcendent revelation (bliss-of-consciousness with stable-emotion-coloring). The convergence: both refuse the production-account (Lollaṭa) in favor of an apprehension-account. The divergence: whether the apprehension is ordinary-cognition-of-special-object (Śaṅkuka) or supermundane-cognition-of-transcendent-event (Abhinava). The split reveals what the discipline had to learn: cognitive accounts of rasa work only if the kind of cognition is itself special. Bhaṭṭa Toṭa, Abhinava's teacher, did most of the work showing why Śaṅkuka's ordinary-inference framework fails. (See Śrī Śaṅkuka and the Anumiti-Vāda and Abhinavagupta's Teaching.)
The four-position taxonomy operates as a model of how disciplines develop alternative ontology-epistemology packages for a single contested phenomenon. The pattern is generalizable beyond rasa-theory. The framework also produces specific implications for how we think about audience-experience in any contemporary context.
Psychology: Personality Typology and Individual Differences Hub — The four positions on rasa map cleanly onto four contemporary positions on what happens during emotionally-intense aesthetic engagement. Lollaṭa's utpatti corresponds to the appraisal-emotion tradition (we feel real emotions in response to fictional stimuli; the emotions are produced by our appraisals; they exist as real psychological events in us). Śaṅkuka's anumiti corresponds to theory-of-mind and mentalizing accounts (we infer characters' emotional states from their behavior and respond accordingly; the operation is inferential and cognitive). Bhaṭṭa Nāyaka's bhukti corresponds to embodied-simulation and empathic-immersion accounts (we mentally simulate the universalized version of what the character feels; the operation is experiential, not inferential). Abhinava's abhivyakti corresponds to flow-state / aesthetic-revelation accounts (the aesthetic event manifests something that was already structurally present in our psychology; the operation is revelatory, not constructive). The handshake reveals: each of the four positions has a contemporary scientific-psychology analogue, and contemporary debates about audience response to fiction are recapitulating the four-position debate from 9th-11th-century Kashmir. The fourfold typology is structurally robust — it captures four genuinely different things that could be going on in aesthetic engagement, and the same disagreements arise in both the Sanskrit and contemporary traditions because the underlying question (which of these four things is the right account?) remains open. The handshake produces an audit-question: which of the four does my aesthetic-engagement-theory actually presuppose? Most contemporary aesthetic theory operates with all four mixed, without recognizing they are mutually inconsistent.
Creative-Practice: Non-Fiction Writing Craft Hub — The four-position framework offers a precise craft-pedagogy diagnostic. Different writers operate as if different positions are true, and this shapes their craft decisions. Writers who implicitly hold the utpatti view focus on production — they try to produce the right emotional charge in their material, treating the reader as a kind of receptor of the charge. Writers who hold the anumiti view focus on signposting — they make sure their material gives the reader enough premises to draw the right inferences. Writers who hold the bhukti view focus on universalization — they strip particulars to allow the reader to enter into a commonized experience. Writers who hold the abhivyakti view focus on revelation — they design their work to manifest what already exists in the reader's psyche under the right cognitive conditions. The handshake produces a craft-pedagogy practice: writers should identify which of the four positions their craft implicitly endorses, and either commit to that position more deliberately or rotate consciously between them. Most contemporary writing pedagogy mixes the four without acknowledging the mixing; the result is muddled craft advice. Pollock's precise rendering of the four positions offers writers a sharper vocabulary for what their craft is actually attempting.
AI-Collaboration: AI-Assisted Creator Hub — The four-position framework applies directly to questions about AI-generated affective content. An AI system that operates as if utpatti is the right account will try to produce emotional charge in its outputs by direct emotional language and conventional emotional triggers. An AI operating as if anumiti is right will try to provide rich behavior-and-situation cues from which the reader can infer emotional state. An AI operating as if bhukti is right will try to universalize particulars to allow reader-immersion. An AI operating as if abhivyakti is right will try to design outputs that resonate with the reader's pre-existing psychology. The handshake: contemporary AI evaluation tends to assess outputs against a single implicit framework. Pollock's four-position taxonomy provides a more articulated evaluation vocabulary. An AI output can be evaluated as "produces emotional charge well but lacks signposting" (good at utpatti, weak at anumiti); "has good behavior cues but fails to universalize" (good at anumiti, weak at bhukti). The four-position framework lets us evaluate outputs along four dimensions rather than collapsing them into a single quality-judgment.
The Sharpest Implication
The four-position taxonomy reveals that what looks like a single phenomenon (aesthetic engagement with emotionally-charged content) is actually four structurally different cognitive operations that the same word "rasa" has been used to describe. The discomfort: this means that most contemporary discussions of aesthetic experience are talking past each other, because participants assume they are discussing the same phenomenon but are actually defending different positions in the four-fold space. The argument-pattern is recognizable: one participant insists aesthetic experience is "really real" (defending utpatti); another insists it's "an inferential operation" (defending anumiti); another insists it's "experiential immersion" (defending bhukti); another insists it's "revelation of what was already there" (defending abhivyakti). Without the four-position framework, these participants cannot see they are defending mutually inconsistent positions on different questions. With the framework, the debate gets clearer: which question are we asking? What ontology of aesthetic experience are we committed to? The four-position taxonomy is a clarification tool, not a refutation tool — but the clarification it provides is substantial.
Generative Questions