Sweetness Before Sugar: Honey, Bees, and the Secret the Aśvins Stole

The Winds Blow Honey

Three verses near the end of the first maṇḍala say something that should not work as poetry and does. A poet of the Aṅgiras line, praising the All-Gods, runs through the entire visible world and declares each piece of it sweet:

madhu vātā ṛtāyate madhu kṣaranti sindhavaḥ | mādhvīr naḥ santv oṣadhīḥ

“Honey the winds, for the one who keeps the truth; honey the rivers run. May the plants be honey for us.”

Rigveda 1.90, verse 6. Translation mine, checked against Griffith (1896). [1][2]

The next two verses keep going. Honey the night and the dawns, honey the dust of the earth, honey our father Sky. Honey the tall tree, honey the sun, honey our milk-cows. It reads at first like an ecstatic list, the kind of thing that sounds better in Sanskrit than in any translation. But notice what is not happening. The poet is not saying the world is like honey. He is not saying it is pleasant, or good, or blessed. He is using one specific substance, over and over, as the measure of everything worth having.

There is a practical reason for that, and it is the sort of fact that reorganises how you read the text. The Rigveda has no word for sugar. Not cane sugar, not crystallised sugar, not even the sugarcane plant. Search the whole Saṃhitā for ikṣu or śarkarā and you find nothing. Sugarcane makes its first Vedic appearance in the Atharvaveda, in a young man’s love charm that wraps a girdle of cane around the beloved to drive away hatred, and even there the cane is a supporting player in a hymn whose real subject is honey. [3]

So when a Rigvedic poet reached for the idea of sweetness, he had exactly one referent: the comb of a wild bee, cut out of a cliff or a high branch by someone who climbed for it. Every metaphor of sweetness in the oldest layer of Sanskrit is, at bottom, a metaphor about robbing bees.

Follow madhu (Sanskrit: मधु) through the text and you get four things nobody expects from a collection of ritual hymns: a small ethnography of honey hunting, a pair of gods who were shut out of the soma cult and given honey instead, one of the strangest myths in Indian literature involving a severed horse’s head, and a word so old that its relatives turn up in a Norse mead hall and, quite possibly, in the modern Chinese for honey.

Methods note. The counts above come from a plain text search over a romanised, unaccented electronic Saṃhitā, matching the string madhu and its inflected and compounded forms. That method is blunt. It will miss forms hidden by sandhi and it will bundle the noun madhu together with adjectives like madhumant (“honeyed”) and madhuścut (“dripping honey”). Treat the numbers as an order of magnitude, not a concordance. Where a precise count matters, Bloomfield’s Vedic Concordance remains the instrument of record. [4]

A world with only one sweetness

Honey is not one thing in the Rigveda, and separating its senses is the first job. In maṇḍala 9, the great book of soma purification, madhu is overwhelmingly a description of soma itself: the pressed juice runs into a vat that “drips honey” (kośaṃ madhuścutam), a phrase so standard it appears at RV 9.12.6, 9.23.4, 9.36.2 and half a dozen other places with only the metre changed. Roughly a quarter of all the honey vocabulary in the text sits in that one book, describing a plant juice. [1]

That is metaphor, and everyone agrees it is metaphor. The interesting cases are the ones where honey is not standing in for something else.

The distribution is telling. Strip out maṇḍala 9 and the honey language clusters hard around two things: the Aśvins, the divine twins, and the vocabulary of good speech. Neither cluster is about soma at all.

And in a handful of verses, madhu is unmistakably the real substance. RV 8.4.8 has milk “mixed with honey of the saragh”, the bee, which is about as literal as a Vedic poet gets about a recipe. RV 4.45 describes honey being carried in a leather bag. You do not carry a metaphor in a leather bag.

Four verses with a bee in them

Here is the thing that started this article. In a hymn to the Aśvins in the first maṇḍala, in the middle of a list of the twins’ rescues and gifts, a bee talks.

uta syā vāṃ madhuman makṣikārapat

“And that bee murmured honey to you two.”

Rigveda 1.119, verse 9. Translation mine; Griffith renders it “to you in praise of sweetness sang the honey-bee.” [2]

The verse does not stop there. It continues: yuvaṃ dadhīco mana ā vivāsatho, athā śiraḥ prati vām aśvyaṃ vadat, “you two sought out the mind of Dadhyañc, and then the horse’s head spoke to you.” A bee whispers something sweet, and immediately afterwards a severed animal head starts talking. Hold that juxtaposition; the second half of this article is about why it is not random.

The bee is rare enough in the Rigveda to be catalogued exhaustively.

Five entries, four creatures, in more than ten thousand verses. Compare that with the horse, which appears in the hundreds, or with cattle, which are effectively the text’s unit of account.

Aside. Makṣā and makṣikā are genuinely ambiguous. In later Sanskrit makṣikā covers both the housefly and the honeybee, and the lexica list both. Translators render the Rigvedic instances as bees on contextual grounds, since in each case honey is in the same sentence. [5] At RV 7.32.2, where singers cluster around Indra’s pressed soma, “flies” is the better rendering and slightly unflattering to the singers. This is one of those places where translation is a decision, not a lookup.

The other word, sāragha, is the more precise one: an adjective formed from saragh, the bee, meaning “of the bee, bee-produced”. Its two appearances are worth the trouble. RV 10.106 is notorious among Vedicists as one of the least intelligible hymns in the collection, a chain of similes with almost no connective tissue, and translators from Geldner onward have flagged several of its verses as effectively untranslatable. [6] In the middle of that fog sits an image of two bees working at a vessel whose opening faces downward, which is a reasonable description of a wild comb hanging under an overhang, entrance side down. The poet expected his audience to know what that looked like.

The honey-skin on the chariot

The Aśvins travel with luggage. In RV 4.45, a hymn that is essentially one extended honey advertisement, the twins’ chariot carries three pairs of provisions and a fourth item:

pṛkṣāso asmin mithunā adhi trayo dṛtis turīyo madhuno vi rapśate

“Three pairs of provisions are on it, and a fourth, a skin-bag, overflows with honey.”

Rigveda 4.45, verse 1. Translation mine, following Geldner and Griffith. [1][6]

Two verses later the poets ask the twins to “bring the honey-filled skin” (dṛtiṃ vahethe madhumantam), and at RV 8.5.19 the same object turns up again, a madhuno dṛtiḥ set down on the chariot’s track, with an invitation to drink from it. A dṛti is a hide sewn into a bag; the same word describes the dried-up skin in the frog hymn at RV 7.103.2 and the water-skin the rain god empties in RV 5.83.7. It is ordinary equipment.

This is what honey collection looks like when nobody keeps bees. The giant honeybee of South Asia, Apis dorsata, builds a single open comb up to a metre across, hanging from a cliff overhang or a high branch, and it cannot be hived. It is harvested by climbing, smoking, cutting, and carrying the dripping comb down in a container, which for a Bronze Age pastoralist means a leather bag. Eva Crane’s survey of the global history of honey hunting places South Asia squarely in this pattern: hive beekeeping arrives late in the subcontinent, and wild-nest harvesting of dorsata and cerana continued alongside and after it, in some regions to the present day. [7]

The Rigveda never describes the climb. It is not that kind of text; it also never describes ploughing in any detail, though the plough is clearly there in the Sītā verses, and it is famously reticent about the buildings its poets lived in. What it gives you instead is the residue: the bag, the bee, the mixture with milk, the assumption that honey is something a travelling god carries.

Aside. Central Indian rock shelters, including sites in the Bhimbetka complex, carry paintings of figures on ladders taking combs from nests, and these are regularly cited as the earliest South Asian evidence for honey hunting. [7] The caution is that rock art is very hard to date directly; the Bhimbetka paintings span an enormous range, from Mesolithic to historic, and individual panels are usually dated by style and superimposition rather than by any absolute method. The paintings show that honey hunting is old in India. They do not show that any particular panel is contemporary with the Rigveda.

The gods who were not allowed soma

The honey concentrates around the Aśvins, and once you notice that, the rest of the pattern falls out.

The twins have a honey-dripping whip. It is one of their permanent attributes: yā vāṃ kaśā madhumaty aśvinā sūnṛtāvatī, “your whip that is full of honey and full of generosity,” at RV 1.22.3, and again at RV 1.157.4, where the poets ask to be sprinkled with it. A whip that drips honey instead of drawing blood is a strange enough object that the Brāhmaṇas later built a whole rite around it. They drink honey with “honey-drinking mouths”. Their swans are honeyed. They arrive at the pressing “like bees to honey”.

Now set that beside a second fact. The Aśvins are the gods with the most trouble getting soma. The Śatapatha Brāhmaṇa tells the story at length: the twins are physicians, they wander among men, and they are for that reason excluded from the soma draught, until the aged Cyavana, restored to youth by them, negotiates their admission. [8] They then have to acquire the knowledge that makes the offering work, which is where Dadhyañc comes in.

The Rigveda is not the Brāhmaṇa, and it would be a mistake to read the later story straight back into it. But the Rigvedic distribution is consistent with the Brāhmaṇa’s premise. The twins are the gods of the substances that are not soma.

Look at the surā row. It appears at RV 7.86.6 in Vasiṣṭha’s confession to Varuṇa, listed among the things that lead a man astray alongside dice and anger; at RV 8.2.12, where drunkenness is mocked; and at RV 1.116.7, where the Aśvins pour out a hundred jars of it from a horse’s hoof for the poet Kakṣīvant. The one time strong drink appears as a gift rather than a hazard, the Aśvins are handing it over. The same twins own the honey.

Two observations follow. First, the Rigveda’s ritual economy has a centre and a periphery, and honey sits on the periphery with the twins, the physicians, the travellers, the gods who are always arriving from somewhere else. Second, that peripheral status is exactly why the honey material is interesting: peripheral things preserve older layers. The soma cult is an Indo-Iranian institution with a whole apparatus of pressing stones, filters and specialists, described in the workshop language of maṇḍala 9. Honey needs none of that. It needs a rope and a bag.

The horse’s head

Now back to the talking bee, and to the myth that sits behind it.

Dadhyañc, son of Atharvan, appears in the Rigveda in fragments, never as a continuous narrative. Assemble the fragments and this is what you get:

dadhyaṅ ha yan madhv ātharvaṇo vām aśvasya śīrṣṇā pra yad īm uvāca

“When Dadhyañc, son of Atharvan, proclaimed the honey to you two by means of a horse’s head.”

Rigveda 1.116, verse 12. Translation mine; compare Griffith, “made known to you the Soma’s sweetness.” [2]

At RV 1.117.22 the twins are the ones who fit the horse’s head onto Dadhyañc in the first place, and he then “proclaims the honey” to them, described as “Tvaṣṭṛ’s secret”. At RV 1.84.13-15, in a hymn to Indra, the sage’s bones become Indra’s weapon against ninety-nine obstacles, and Indra goes looking for the horse’s head, which he finds hidden in the mountains at a lake called Śaryaṇāvant.

The Rigveda assumes you know the story. The Brāhmaṇas tell it. Indra taught Dadhyañc a secret and threatened to cut off his head if he passed it on. The Aśvins wanted it. So they removed the sage’s own head, kept it safe, attached a horse’s head in its place, and had him speak the secret through the substitute; Indra struck off the horse’s head as promised, and the twins restored the original. [8][9] The secret is called madhu, honey, and in the Bṛhadāraṇyaka Upaniṣad it becomes madhu-vidyā, the honey doctrine: a teaching that every element of the world is the honey of every other, that the earth is honey for all beings and all beings are honey for the earth. The Upaniṣad quotes the Rigvedic Dadhyañc verses directly as its warrant. [10]

graph TD
    A[Indra teaches Dadhyanc] -->|threat: no sharing| B[Dadhyanc bound to silence]
    C[Ashvins want the honey] --> D[Swap in a horse's head]
    B --> D
    D --> E[Horse head speaks the madhu]
    E --> F[Indra beheads the horse]
    F --> G[Ashvins restore his own head]
    E --> H[Pravargya rite: heated pot]
    E --> I["Madhu-vidya, BAU 2.5"]
    G --> J[Bones become Indra's weapon]

Two things about this deserve more attention than they usually get.

The first is the pun, and it is load-bearing. Madhu means honey and it means the secret knowledge, and the myth works only because the two senses are the same word. What the bee murmurs at RV 1.119.9 and what the horse’s head announces are, lexically, the same thing. This is the Rigveda doing something it does constantly: treating a substance and an insight as interchangeable, the way it treats concealment and discovery as a single operation. The honey is real honey and it is also the thing you are not supposed to say out loud.

The second is that the myth is attached to a rite that survives. The Pravargya, an optional preliminary to the soma sacrifice, turns on a clay pot called the mahāvīra, heated until it glows red, into which milk is poured; the offering goes primarily to the Aśvins. [11] The pot stands, in the ritual texts’ own reading, for a head. Jamison’s study of this material reads the myth complex and the ritual as versions of one another rather than as story-plus-illustration, and argues that the Vedic imagination keeps returning to heads that are removed, replaced and restored. [9] The Rigveda knows the hot pot: at RV 1.180.4 the Aśvins make the gharma, the scalding vessel, “full of honey” for the sage Atri.

So the sequence is: a real substance harvested from cliffs, a pair of outsider gods who carry it, a myth about stolen knowledge named after it, and a ritual with a glowing pot that reenacts the theft. That is a lot of load for one word to bear.

Was madhu ever mead?

Here is where the article has to be careful, because this is the point at which popular accounts overreach.

Madhu is one of the most securely reconstructed words in the Indo-European lexicon. Proto-Indo-European *médʰu is continued in at least six branches, and the pattern of meanings is not uniform.

Two features stand out. In the west and in Greek the word drifts decisively toward the fermented drink; Greek méthu means wine, and its derivative gives us “amethyst”, the stone that supposedly prevents drunkenness. In the east it stays with the raw substance. And Greek keeps a second honey word, méli, from a different root *mélit-, which also shows up as Latin mel and Hittite milit. Indo-Iranian lost that second word entirely. [12][13]

There is a further twist that is worth knowing even though it lies outside the Vedic world. Old Chinese mì 蜜, “honey”, reconstructed as something like *mit, is widely thought to be a loan from Tocharian mit, picked up along the Tarim routes. If that is right, the same Indo-European syllable that a Punjab poet used to praise the winds ended up as the ordinary Chinese word for the stuff in the jar. The borrowing is generally accepted but not universally, and the Old Chinese reconstruction carries its own uncertainties, so it belongs in the category of “very likely” rather than “established”. [14]

Now the Vedic question. Did Rigvedic people ferment honey?

The honest answer is that the text does not say, and the silence is loud. The Rigveda describes soma preparation in exhaustive, almost industrial detail: the stalks, the stones, the woollen filter, the mixing with milk, hundreds of verses of process. It describes nothing comparable for madhu. There is no pressing, no straining, no waiting, no vessel of fermentation. Honey appears already made, in a bag, on a chariot.

Three readings are on the table, and I do not think the evidence chooses cleanly between them.

1. Madhu in the Rigveda is simply honey and the sweetness of honey, and the fermented sense visible in Greek and Germanic is an innovation of those branches. Conservative, and defensible on the Vedic evidence alone.
2. Indo-Iranian inherited a honey-based intoxicant and displaced it with soma, leaving madhu as a prestige word applied to the newer drink. This would explain why maṇḍala 9 calls soma honey so relentlessly: the new drink took over the old drink’s vocabulary along with its function.
3. Both drinks coexisted, honey in the domestic and Aśvin sphere, soma in the priestly one, and the Rigveda simply does not describe what happens in the domestic sphere, in line with its general silence about ordinary life.

Reading 2 is the one comparative philologists tend to find attractive, because the semantics of *médʰu across the family point to a drink rather than a foodstuff, and because Vedic religion demonstrably replaced things. It is also the one that outruns its evidence most easily. What can be said with confidence is that honey liquor is thoroughly attested in India later: Kauṭilya’s Arthaśāstra, in its chapter on the superintendent of liquor, lists a dozen kinds of drink, including one made from fermented honey-water. [15] The technique was known. Whether it was known in the Punjab in the second millennium BCE, the Rigveda declines to tell us.

[!NOTE] This is a good place to notice how the soma identification debate and the honey question interact. Every argument that soma was a fermented drink has to explain why the text describes pressing and filtering but never fermentation. Honey has the mirror-image problem: a plausible fermentable, a word whose cousins mean “mead” across half of Europe, and not one line about making it.

Honey on the tongue

The last territory madhu occupies is speech, and it is the one where the Rigveda’s poets are most obviously talking about themselves.

Agni, and his double Narāśaṃsa, are madhu-jihva, “honey-tongued”, at RV 1.13.3, RV 1.44.6 and RV 1.60.3, where the god’s tongue is sweet “even at his birth”. Narāśaṃsa is madhu-hastya, “with honey in his hand”, at RV 5.5.2. A poet of the Rahūgaṇa family signs off at RV 1.78.5 by saying he has spoken madhumad vacaḥ, honeyed speech, and Vatsa says the same of his own composition at RV 8.8.11. The hymn is honey; the mouth that produces it is a honey-mouth; the god who receives it licks it up.

This is not a private Vedic conceit. When Homer introduces Nestor in the first book of the Iliad, the old man rises to speak and “from his tongue flowed speech sweeter than honey” (toû kaì apò glṓssēs mélitos glukíōn rhéen audḗ, Iliad 1.249). Greek uses its other honey word there, méli, not méthu, which is exactly what makes the parallel interesting: the formula survived in both traditions while the lexeme was replaced in one of them. West treats honey-sweet speech as part of the inherited Indo-European poetic stock, one of a set of images that poets in widely separated branches use for the same purpose, and Watkins’s work on formulaic inheritance provides the method for judging cases like this. [13][16]

The comparison is worth making precisely because it is unglamorous. Nobody borrowed anything from anybody here. Two poetic traditions, separated by several thousand kilometres and at least a millennium, independently kept the same habit of describing good speech as the sweetest edible thing their audiences knew.

And in a text where the poet is also a carpenter, weaver and craftsman of the hymn, the honey image adds something the others do not. A chariot is built. A cloth is woven. Honey is found: located in an inaccessible place, taken at some risk, brought back in a bag. The Rigvedic poets who called their verses honey were making a claim about where verses come from.

What kept

Honey outlasted almost everything else in this article. Soma became a botanical mystery within a few centuries, its plant unavailable and its substitutes debated ever since. The Pravargya survives as a specialist rite performed by a handful of ritualists. But the madhuparka, the reception of an honoured guest with a mixture of honey, curds and ghee, moved out of the Vedic ritual sphere and into ordinary Indian social life, prescribed by every Gṛhyasūtra and still performed at weddings. [17] The three verses this article opened with, the madhumatī verses of RV 1.90, are among the most widely recited lines in the whole collection, chanted in contexts that have nothing to do with the All-Gods hymn they came from.

There is a reason those verses travel well, and it is not only that they are pleasant. Read them again with the honey hunt in mind. The poet is not saying the world is nice. He is saying that the world is worth the climb: that the winds and the rivers and the plants and the night contain the same concentrated, difficult, dangerous-to-obtain good thing that a person risks a fall to get out of a cliff. That is a considerably tougher claim than it looks in translation, and it is the closest the Rigveda comes to stating a theory of value.

If you want to test this, do not start with the philosophy. Open RV 4.45, the Aśvins’ honey hymn, and read it as an inventory: a chariot, three pairs of provisions, a fourth bag leaking honey, swans, a road that gets sweeter as they travel it. Then read RV 1.119, where a bee tells two gods a secret and a horse’s head repeats it. Somewhere between those two hymns is a Bronze Age world in which the sweetest thing available had to be stolen from something that stings, and in which that fact was interesting enough to become a theory of knowledge.