Sweet, Honeyed, Sharp, Full of Sap: The Rigvedic Word That Became India's Theory of Art

Four words for one mouthful

A man is holding a cup. He has been handed the pressed liquid that the whole morning’s work exists to produce, and before he says anything about gods or cattle or enemies, he reports what is in his mouth.

svāduṣ kilāyaṃ madhumān utāyaṃ, tīvraḥ kilāyaṃ rasavān utāyam

“Sweet, surely, is this one, and honeyed is this one; sharp, surely, is this one, and full of sap is this one.”

Rigveda 6.47, verse 1. Author’s rendering. Griffith’s 1896 version has it as “pleasant to taste and full of sweetness” and “strong and rich in flavour.” [1]

Four adjectives, two of them doubled by kila and uta, the Vedic equivalents of “really” and “and also.” It reads like tasting notes written by someone who has not yet been taught that tasting notes are a genre. The hymn then swings into its real business, which is Indra flattening the fortresses of Śambara, but the first line stays in the mouth.

The fourth adjective is rasavān, “possessing rasa.” That word is the reason this hymn is worth stopping at.

Anyone who has read a programme note for a Bharatanatyam recital, or an essay on Sanskrit poetics, or the back of a book about Indian food, has met rasa as a technical term for aesthetic experience: the flavour a work of art has, and that an audience tastes. It is the single most consequential concept in the Indian theory of art, the pivot of a tradition of criticism that runs from Bharata’s Nāṭyaśāstra through Abhinavagupta to sixteenth-century Vrindavan and into modern film reviewing. And in the Rigveda, which uses the word fifty-two times, it means none of that. It means the liquid you get when you crush something.

What follows is an attempt to watch a word travel without helping it along. The Rigveda gives us rasa at its physical origin, in a workshop, surrounded by stones and strainers and sieves of sheep’s wool. The interesting question is not whether the Vedic poets secretly had an aesthetics (they did not), but what it was about their very concrete use of the word that made it available, a thousand years later, to describe what happens to a spectator in a theatre.

What the word does

Start with grammar rather than meaning. A good way to find out what a Vedic noun denotes is to ask which verbs it keeps company with, because the verbs encode the physics.

Notice what is missing. Nobody in the Rigveda feels rasa, knows rasa, perceives rasa, or is moved by rasa. The word is never the object of a mental verb. It is milked, poured, strained, anointed, drunk, divided and waded into. Across fifty-two occurrences the semantic centre never leaves the physical world of containers and fluids.

That is a stronger claim than it looks, because Vedic Sanskrit is perfectly capable of abstraction. The same poets built an enormous vocabulary for mental acts, for insight and attention and formulation, and they deified several abstractions outright, including Anger and Trust and the Priest’s Fee. If they had wanted rasa to mean an inner state, nothing was stopping them. They used it for sap.

Methods note. The count of fifty-two is my own, made by pattern search over a romanised Saṃhitā text, matching the inflected forms of rasá- and then reading every hit in context. Three categories were excluded by hand: the eight passages where Rasā is the river name, the forms of the root rās “to bestow” that look identical in an unaccented text (as at RV 1.96.8), and the phrase dvāro devīr asaścataḥ, “divine unfailing doors,” where the sequence rasaś is an accident of sandhi. Vedic is transmitted with words fused across boundaries, so a figure like this is a floor rather than a total; a few occurrences inside compounds will have escaped. Stanza totals here follow the Saṃhitā text as printed by Aufrecht, the standard critical edition. [2] For verifiable word totals, Lubotsky’s concordance lists every word of the Rigveda in its pada. [3]

The uneven spread matters more than the total. Taking the Soma book against everything else:

$$\frac{33}{1{,}108} \approx 0.030 \quad\text{against}\quad \frac{19}{9{,}420} \approx 0.0020$$

which is roughly a fifteen-fold concentration per stanza.

Two of the family books, the second and the seventh, never use the word at all. Whole generations of hereditary poets composed without needing it. Then you reach Maṇḍala 9, where every hymn is addressed to a liquid in motion, and the word appears every thirty-odd stanzas.

The pressing floor

Maṇḍala 9 is the Rigveda’s strangest book: a hundred and fourteen hymns to Soma Pavamāna, “Soma purifying himself,” composed to be sung while the plant was actually being crushed and filtered. Read fifty of them in a row and you are reading industrial process poetry. Stalks swell with water, stones come down, ten fingers work, the liquid runs through wool, it is caught in wooden vats, it is mixed with milk, it is carried off to Indra.

Rasa is the name of the thing the process is for. The stalk is not the point; the liquid inside the stalk is the point. The hymns say so with a directness that modern writing about Vedic religion rarely preserves.

saṃ yanti rasino rasāḥ

“The saps of the sappy ones flow together.”

Rigveda 9.113, verse 5. Author’s rendering; Griffith has “The juices of the juicy meet.” [1]

In one hymn the word is handled as transferable stuff with its own history: Sūrya’s daughter brings the great buffalo, the Gandharvas seize it, and then “they placed the rasa in the soma” (RV 9.113.3). The liquid is prior to the plant it lives in. Elsewhere the god mingles with the other gods “with his own rasa” (RV 9.97.12), and the poets anoint his form “with the rasa of the waters” (RV 9.97.57).

The grammar of all this is the grammar of extraction, and the extraction has three steps that will matter a great deal later: crush the body, separate the liquid, distribute the liquid to those entitled to a share. The Rigveda has a whole workshop vocabulary for the first two steps, discussed in detail in the pressing stones and in the filtered god of Maṇḍala 9. What concerns us is that the word for the product of that workshop is the word that Indian aesthetics will eventually use for the product of a play.

A share of the sap

Outside the Soma book, the most revealing occurrences are in a short hymn to the Waters that opens the tenth maṇḍala. It is one of the few Rigvedic passages a practising Hindu is still likely to have recited, since verses from it survive in the formulas said while bathing.

“Give us a portion of the sap, the most auspicious that ye have, like mothers in their longing love.”

(RV 10.9.2, Griffith 1896) [1]

The Sanskrit is yo vaḥ śivatamo rasas tasya bhājayateha naḥ. The verb is from bhaj, “to apportion, to give a share,” the same root that eventually produces bhakti. The request is not for the water but for its rasa, its most auspicious fraction, and the request is phrased as a claim on a distribution. Seven verses later the same hymn, repeating a line from RV 1.23.23, has the speaker report an accomplished fact: rasena sam agasmahi, “we have come together with the rasa,” which Griffith renders “to their moisture have we come.”

Here the word is doing two jobs that a purely mechanical reading misses. It is the active principle in water, the thing that makes water healing rather than merely wet, and it is something a person can be united with. Nothing mystical is being claimed. The verse is about standing in a river. But the idea that a substance has a best part, and that contact with the best part changes you, is already fully formed.

The hymn to Food, the only one of its kind in the collection, pushes the same thought towards nutrition.

“These juices which, O Food, are thine throughout the regions are diffused. Like winds they have their place in heaven. These gifts of thine, O Food, O Food most sweet to taste, these savours of thy juices work like creatures that have mighty necks.”

(RV 1.187.4 and 1.187.5, Griffith 1896) [1]

The rasāḥ of food are spread through the worlds; the svādmānaḥ of the rasas, their sweetnesses, have the strength of thick-necked animals. The Pitu Sūkta is a small masterpiece of enthusiasm about eating, examined on its own terms in what the Rigvedic world actually ate. For our purposes it establishes that rasa was not a soma specialism. Grain has it. Water has it. Plants have it. It is the general name for whatever in a substance can be got out of the substance and taken into a body.

Dew, and a river nobody could reach

Etymology sometimes confirms what the usage already shows. This is one of those times, with a caveat.

Vedic rasá- is normally connected with the Indo-European family of words for dew: Latin rōs, rōris, Lithuanian rasà, Old Church Slavonic rosa, all meaning dew or moisture. [4] The reconstructed root is given variously in the handbooks, which is a polite way of saying it is unsettled; what the cognate set agrees on is a sense of fine moisture that appears, flows and can be gathered. If the comparison holds, the technical term of Indian dramaturgy and the Latin word in rōs marīnus, “sea dew,” now rosemary, are the same word.

Aside. Resist the temptation to read the cognates back into the hymns. Vedic poets did not know that a Baltic peasant would call dew rasà, and a shared ancestor tells you where a word came from, not what its users meant by it. The cognate set is evidence about prehistory. The fifty-two passages are evidence about the Rigveda. They happen to agree, which is pleasant but not guaranteed.

The same stem gives a proper name. Eight times the Rigveda mentions Rasā, a river, and the striking thing about her is how unreachable she is. The Aśvins make her swell with a flood (RV 1.112.12); she is called “great mother Rasā” in a plea for protection (RV 5.41.15); she stands at the far edge of the world that the hymns can imagine, and crossing her counts as an achievement worth reporting.

“I come appointed messenger of Indra, seeking your ample stores of wealth, O Paṇis. This hath preserved me from the fear of crossing: thus have I made my way o’er Rasā’s waters.”

(RV 10.108.2, Griffith 1896) [1]

The speaker is Saramā, the dog sent to negotiate with the misers who are sitting on the stolen cattle, in the dialogue hymn treated in Saramā and the Paṇis. Her boast is geographic. She went past the last river.

Which river, if any, is contested. The Vedic Index catalogues the references and leaves the identification open. [5] Witzel has argued that Rasā corresponds to Avestan Raṅhā and to the stream the Greeks called Rhā, which would make her the Volga and the hymn a fossil memory of a very distant homeland. [6] The Iranian side of the comparison is secure enough; the Avestan Raṅhā is itself a mythological boundary water rather than a surveyed one, and specialists have treated the Volga identification as a secondary application of an old name. [7] For the history of the word, the useful point is smaller and safe: a people whose ordinary noun for extractable liquid was rasa named their furthest river The Sappy One, and then half forgot where she was.

Griffith’s seven words

There is a practical test for whether a word in one language has a counterpart in another: hand it to a translator and count how many words come back. Griffith’s 1896 version is a convenient instrument, since it is complete, public domain, and the work of a single man applying a single sensibility across all ten maṇḍalas. Jamison and Brereton’s 2014 translation, the modern standard, has the same problem and solves it differently verse by verse. [19]

Moisture, juice, savour, flavour, essence, sap, and the adjective he builds for the taste of milk. Seven solutions, none of them wrong, none of them repeatable, which is the signature of a word that cuts the world at a joint English does not have. The Rigveda is full of these, and some of them defeat the specialists entirely; a selection is assembled in the words nobody can translate.

One of Griffith’s choices deserves a pause. At RV 9.67.31 the thing that has rasa is not a plant but a group of hymns: whoever studies the rasa “stored by the seers” in the Pavamānī verses eats purified food. Monier-Williams would gloss that as “essence,” one of the senses his dictionary lists between “sap” and “sentiment,” [20] and in a very late appendix to Maṇḍala 9 it is the earliest hint in the collection that the word could attach to something made of language rather than something made of liquid. One hint, in a passage nobody dates early.

Aside. The first complete dictionary to the Rigveda was compiled by Hermann Grassmann, the Stettin schoolmaster whose mathematics, the extension theory that underlies modern exterior algebra, was ignored in his lifetime. He taught himself Sanskrit from 1849, built the dictionary for his own use, published it only when Indologists asked him to, and brought out 1,776 columns of it through Brockhaus between 1873 and 1875, followed by a complete verse translation. [8] The dictionary is still consulted. The translation is not. There is a lesson in that about which kinds of work survive.

The turn

Somewhere between the late Rigveda and the early Upaniṣads, rasa stops being only a liquid and starts being a relation. The move is visible in a single famous passage, the opening of the Chāndogya Upaniṣad, which runs a chain of essences: of these beings the earth is the rasa, of earth the waters, of the waters the plants, of plants the human being, of the human being speech, of speech the Ṛg, of the Ṛg the Sāman, of the Sāman the Udgītha. [9][10]

Read that chain with the Rigvedic sense in your head and it is still a pressing metaphor. Each term is the liquid you get when you squeeze the term before it. Max Müller, annotating the passage, noted that the commentators explain rasa here by a whole bundle of notions: origin, support, end, cause, effect. [9] The word has become a way of saying “what a thing most essentially yields,” and once it means that, it can be applied to absolutely anything, including a chant.

The Taittirīya Upaniṣad takes the last step and attaches it to the absolute: raso vai saḥ, “He is indeed rasa; for on obtaining rasa one becomes blissful” (author’s rendering; compare Olivelle’s translation). [10] A Vedic poet asking the Waters for a share of their sap, and an Upaniṣadic teacher saying that the ultimate reality is sap and tasting it is joy, are using the same noun in the same metaphorical direction. Only the referent has been promoted.

graph TD
    A["rasa: sap, extracted fluid"] --> B["Rasa: the far river"]
    A --> C["essence, what a thing yields"]
    A --> D["taste as sense quality"]
    C --> E["rasa-dhatu, bodily plasma"]
    C --> F["mercury, rasashastra"]
    D --> G["six tastes of Ayurveda"]
    C --> H["eight rasas of drama"]
    D --> H
    H --> I["bhakti-rasa in Vrindavan"]
    H --> J["rasika, the connoisseur"]

Physicians, cooks, alchemists

Before the critics got hold of the word, the technicians did, and their usage is the bridge nobody mentions.

Classical Āyurveda made rasa a measurable property of every substance. The Caraka Saṃhitā reports a debate about how many tastes there are, with positions ranging from one to infinity, and settles on six: madhura sweet, amla sour, lavaṇa salty, kaṭu pungent, tikta bitter, kaṣāya astringent. [11] Taste became a diagnostic and a pharmacology: you classified a drug by what it did to the tongue, because what it did to the tongue was evidence about what it would do downstream. The same text uses rasa for the first of the bodily tissues, the nutrient fluid that later writers equate with plasma and lymph. [11] A word that meant the liquid pressed out of a plant now means the liquid the body presses out of food.

The alchemical branch is the most literal-minded of all. Medieval Indian alchemy is called rasaśāstra, the science of rasa, because mercury, the one metal that behaves like a liquid, is itself called rasa. David Gordon White’s study of the Siddha traditions traces the fusion of that mercurial chemistry with haṭha yoga, and the logic is continuous with the Vedic workshop: a body is crushed, something quick and bright is released, and the adept drinks it. [12] Whether you find that plausible is a separate question from whether it is a Rigvedic thought. It is.

Bharata borrows a kitchen word

The Nāṭyaśāstra, the treatise on drama that gathers material from roughly the early centuries of the Common Era, needed a term for what an audience gets. It took one from the kitchen, and it said so.

Bharata’s famous formulation, the rasa-sūtra at chapter 6, verse 33, makes rasa arise from the combination of the determinants, the consequents and the transient states, joined to a stable emotion. [13] Dry enough. What is striking is the analogy he reaches for to explain it: as the six tastes are produced when ingredients like raw sugar, spices and vegetables are combined, so the stable emotion combines with the other states and becomes rasa. [13] The theory of aesthetic experience is explained by cooking, on its first appearance, by its own author.

The eight rasas of the original scheme each pair with a stable emotion, the sthāyibhāva that the performance works on:

A ninth, śānta, the peaceful, was argued over for centuries; V. Raghavan devoted a book to the quarrel. [14] The arguing is the sign of a live discipline, and it continued for a thousand years: whether rasa is in the character or in the spectator, whether it is produced or revealed, whether it can be inferred. Abhinavagupta’s answer, that rasa is relished rather than manufactured, is the one that stuck; Masson and Patwardhan’s study of his position collects the thirteen points at which he compares aesthetic experience to mystical experience. [15] The vocabulary of this whole tradition stays alimentary. The audience tastes. The connoisseur is a rasika, a sappy person. Relishing is āsvāda, the word for savouring food. Sheldon Pollock, whose sourcebook assembles thirty of these texts in translation, renders the technical term simply as “taste” and lets the metaphor stand. [16]

The last act of the word’s career is theological. In sixteenth-century Vrindavan, Rūpa Gosvāmin wrote the Bhaktirasāmṛtasindhu, the ocean of the nectar of devotional rasa, and ran the dramatic categories straight into religion: devotion itself has flavours, and a devotee cultivates one as a connoisseur cultivates a taste. [17] By then the word had been doing metaphorical work for two thousand years, and the sap had become an entire theology of feeling. Edwin Gerow’s survey of the Sanskrit critical tradition is a good map of the intervening road. [18]

What the Rigveda did not have

Two cautions, because this is exactly the kind of story that gets told badly.

First, the Rigveda has no aesthetics of rasa, not in embryo, not in code, not in a hidden layer waiting for the right reader. Every one of the fifty-two occurrences is about liquid, or about the potency of liquid, or about a river. The poets had sophisticated ideas about their own craft, including a strong sense of hymns as manufactured objects, which is the subject of the poet as carpenter. They described those ideas with other words. The gap between Rigvedic rasa and Bharata’s rasa is not a short one, and the Upaniṣadic sense of “essence” is doing most of the work in bridging it.

Second, semantic drift is not inheritance. English “nice” once meant ignorant and this tells you nothing about the history of politeness. What makes rasa worth tracking is not that the meanings are connected but that the metaphor was never dropped. Aesthetic theory in Sanskrit could have gone the way Greek theory went, towards vision and imitation, with a vocabulary of looking and copying. It went towards the mouth instead. The tradition kept telling itself that art is something extracted from a made object and then absorbed by a person, and it kept using, for that process, the noun its earliest poets used for what came off the pressing stones.

The Rigveda’s contribution to Indian aesthetics, if we must name one, is not a doctrine. It is a procedure, and a set of verbs: crush, strain, divide, share, taste. The hymns apply them to a plant. Bharata applies them to a play. The grammar is identical.

This is also why the word never became archaic. Rasa is still ordinary Indian vocabulary for juice, and a Tamil household eats rasam, a thin tamarind broth whose name says nothing more profound than “liquid.” Somewhere between the kitchen and the lecture on poetics, the same syllables carry a theory of how art works. Not many technical terms stay edible.

Open RV 9.97 next, a long hymn where the pressing goes on for fifty-eight stanzas, and read it with the sixth chapter of the Nāṭyaśāstra at your elbow. One text is about a plant being destroyed to get at what is inside it, and the other is about an audience. You will not find a doctrine passing from the first to the second. You will find a workshop whose vocabulary was too good to leave behind.

If the semantics of a single word interest you, the collection has a small shelf of them: māyā before it meant illusion, in form-making power; tanū before there was an ātman, in the word for self; tīrtha before pilgrimage, in the ford; and the Rigvedic vocabulary of sweetness in honey before sugar. The pattern they share is worth noticing. The famous abstractions of later Indian thought nearly all begin as something you could pick up.