The House the Rigveda Never Describes

Sleep, Mother. Sleep, Dog.

Somewhere in the Punjab, three thousand years ago, someone is standing outside a house at night and a dog will not stop barking.

We know this because the barking got written down. The seventh Mandala of the Rigveda, the family book of the Vasiṣṭhas, contains a short hymn in which a speaker addresses the dog directly, with a mixture of flattery and irritation, and then addresses everyone inside.

“When, O bright Son of Saramā, thou showest, tawny-hued! thy teeth, They gleam like lances’ points within thy mouth when thou wouldst bite; go thou to sleep.”

(RV 7.55, verse 2, Griffith 1896)

Three verses later the dog has been handled and the household is next:

“Sleep mother, let the father sleep, sleep dog and master of the house. Let all the kinsmen sleep, sleep all the people who are round about.

The man who sits, the man who walks, and whosoever looks on us, Of these we closely shut the eyes, even as we closely shut this house.”

(RV 7.55, verses 5 to 6, Griffith 1896)

That last line is the one to hold on to. In Sanskrit it runs teṣāṃ saṃ hanmo akṣāṇi yathedaṃ harmyaṃ tathā: we close up their eyes the way this harmya is closed up. It is a simile drawn from the building the speaker is standing beside. The poet expects his audience to know exactly what closing up a harmya (Sanskrit: हर्म्य) involves, the way we would know what latching a door sounds like.

And that is very nearly the last thing the Rigveda tells us about the building. Ask the text what the harmya was made of, how tall it stood, whether it had a roof of thatch or a wall of mud, whether the door swung or was lifted aside, and the answer is silence. The oldest large body of poetry surviving in any Indo-European language runs to 1,028 hymns and roughly 10,500 verses. It has a word for the stones that crush soma, a word for the felly of a chariot wheel, three separate words for things you can sleep on. It has no word for brick. Its poets, who could spend forty verses on a horse, never once stop to describe the place they lived.

This is not a gap in our knowledge waiting to be filled in. It is a fact about the text, and it turns out to be one of the more revealing facts in it.

A god with almost nothing to do

Every pantheon has a household god, and the Rigveda’s is Vāstoṣpati (वास्तोष्पति), the Lord of the Dwelling. He receives exactly one hymn, RV 7.54, and it is three verses long. Here is the whole of the first:

“Acknowledge us, O Guardian of the Homestead: bring no disease, and give us happy entrance. Whate’er we ask of thee, be pleased to grant it, and prosper thou our quadrupeds and bipeds.”

(RV 7.54, verse 1, Griffith 1896)

Read it twice and notice what is not there. No foundation, no threshold, no post, no beam. The god of the house is asked for recognition, for freedom from sickness, for an easy way in, and for the wellbeing of the two-footed and the four-footed who live under his protection. Cattle and children, not carpentry. The second verse asks him to be pleased with us “as in his sons a father”; the third asks for security in kṣema and yoga, the settled season and the season of moving out.

Outside those three verses the god barely exists. His name appears in only four other places in the entire Samhitā, and each is odd in its own way. In RV 5.41 he is swept up in a rapid list of makers and growers alongside Tvaṣṭar, the trees and the plants. In RV 8.17 the title is handed to Indra, addressed as “Strong pillar thou, Lord of the home, armour of Soma-offerers”, the only time the house-lord is imagined as a piece of structure rather than a presence. In RV 10.61 the gods fashion Vāstoṣpati out of seed that the sky-father spills on the earth while forcing himself on his own daughter, a verse so unwelcome to Victorian sensibilities that Griffith translated it into Latin and left it there, a habit examined in our piece on the history of Rigveda translation.

Then there is the noun itself. Vāstu, the word that would eventually name an entire Indian science of building, never appears in the Rigveda on its own. It occurs only inside the frozen genitive phrase vāstoṣ patiḥ, still written as two words in the transmitted text. The lord of the dwelling arrives in Sanskrit before the dwelling does.

Aside. The later career of that word is spectacular and largely unrelated. By the sixth century CE, vāstuvidyā is a technical discipline with rules for orientation, proportion and the size of house appropriate to each rank, set out at length in the Bṛhat Saṃhitā. Nothing in RV 7.54 anticipates any of it. Reading the Rigveda through vāstu is reading a thousand years backwards.

Fifty houses and no walls

If the god of the house is thin, the vocabulary of the house is not. The Rigveda has plenty of words that translators render as “house”, “home”, “dwelling”, “abode”. The problem is what they mean.

Set out like that, the pattern is hard to miss. The commonest Rigvedic words in this field are not construction terms at all. They are words for belonging (okas, dama), for the social unit (gṛha, pastyā), and above all for protection. Śarman, chardis and varūtha are what you ask the gods for when you are exposed, and the poets ask constantly. A Rigvedic house, lexically speaking, is a thing that covers you. What it is made of is nobody’s business.

Methods note. The counts above come from my own search of the Samhitā text in transliteration, and they count verses containing a form, not tokens. Read them as approximations. Sandhi hides word boundaries, compounds inflate some totals, and homographs deflate others: a naive search for gṛha also catches forms of the root grah, “to seize”. For a term with $n_w$ attestations, the rate per thousand verses is

$$r_w = \frac{n_w}{10{,}528} \times 1000$$

which puts harmya near 1.1 and śarman near 11.7. The ratios are soft. The zeroes are not, and the zeroes are what the argument rests on.

One received claim is worth correcting as we pass it. The Vedic Index of Macdonell and Keith, still the standard reference work on Vedic realia, says of gṛha that the house “was composed of several rooms, as the use of the plural indicates” [1]. That inference is sound for later Vedic prose, where the plural gṛhāḥ is normal. In the Rigveda itself the singular dominates heavily: the locative gṛhe and accusative gṛham account for most attestations, and unambiguous plurals are a handful. The multi-roomed Vedic house is a Brāhmaṇa-period house read back into the hymns. It may well be right. It is not what the Rigveda says.

graph TD
    A["Rigvedic words for 'house'"] --> B["Belonging: gṛha, dama, okas"]
    A --> C["Enclosure: harmya, pastyā"]
    A --> D["Shelter asked of gods: śarman, chardis, varūtha"]
    A --> E["Structure: sthūṇā, vaṃśa, chadis"]
    B --> F["No material ever named"]
    C --> G["A fence around a hearth"]
    D --> H["A gift, not a building"]
    E --> I["Almost always similes"]

Harmya, the warm place

Of all these words, harmya carries the most weight, and it behaves the most strangely. Collect its dozen appearances and a shape emerges that no translation prepares you for.

Something dark, closable, animal-holding, fire-containing, visible against the dawn, and shading unmistakably into the grave. Monier-Williams, compiling from the whole tradition, lists both ends of that range under a single entry: “a large house, palace, mansion”, and also “a stronghold, prison”, and also “a fiery pit, place of torment, region of darkness, the nether world” [2].

The etymology points the same way. The dictionary rejects the traditional derivation from √hṛ, “to captivate the mind”, and proposes instead a link with gharma, heat, the word “perhaps originally signifying the domestic fire-hearth” [2]. Gharma belongs to the Indo-European root that also produced Greek thermos and English warm. On that reading the Rigvedic house is not primarily a structure. It is the warm place. Macdonell and Keith reach the same destination by a different road, glossing harmya as “the Vedic house as a unity including the stabling and so forth, and surrounded by a fence or wall of some sort” [1]: an enclosure with cattle in it and a fire at the middle, defined by what it holds rather than by how it is built.

Which makes the dog in RV 7.55 considerably more sensible. If a harmya is a fenced compound holding a family and its animals, an animal with lance-like teeth in its mouth is not local colour. It is the perimeter.

Posts that hold up the sky

There is one genuine piece of Rigvedic carpentry vocabulary: sthūṇā (स्थूणा), the post. It occurs eight times, and those eight are the whole problem in miniature.

Twice the post is a simile for cosmic stability. Agni is the navel of the peoples, sustaining men “like a deep-founded pillar” (RV 1.59); heaven itself is “stablished like a firm-set pillar” (RV 5.45). Three times the posts belong to gods. Mitra and Varuṇa sit in a sahasra-sthūṇa, a thousand-pillared seat (RV 2.41, RV 5.62), and one verse goes further toward describing an actual building than anything else in the Samhitā:

“Adorned with gold, its columns are of iron. in heaven it glitters like a whip for horses; Or stablished on a field deep-soiled and fruitful.”

(RV 5.62, verse 7, Griffith 1896)

Gold covering, metal posts, sited on good ground. The closest thing to architecture in the text is a house for two gods, in the sky, that nobody has ever seen. Griffith’s “iron” is almost certainly wrong, incidentally: ayas in the Rigveda is more plausibly copper or bronze, a question taken up in our piece on Rigvedic metallurgy.

Once the title of post goes to Indra, at RV 8.17. And once, exactly once in ten thousand verses, somebody plants a post in the earth and we watch them do it:

“I stay the earth from thee, while over thee I place this piece of earth. May I be free from injury. Here let the Fathers keep this pillar firm for thee, and there let Yama make thee an abiding-place.”

(RV 10.18, verse 13, Griffith 1896)

That is a burial. The single occasion on which the Rigveda describes erecting a post is at a grave, marking the boundary between the dead and the living, and the dwelling under construction belongs to Yama. The scene is set out in full in our piece on the funeral hymns.

The roof does no better. Vaṃśa, the bamboo rafter that later Vedic texts treat as the defining member of a house, occurs once in the Rigveda, at RV 1.10, where singers raise Indra with their praises “like a pole”. Chadis, “covering”, appears once as a roof, in the wedding hymn, and the thing it roofs is a cart:

“Her spirit was the bridal car; the covering thereof was heaven.”

(RV 10.85, verse 10, Griffith 1896)

The one roof in the Rigveda belongs to a wagon, and it is the sky. There is more on that remarkable hymn in our reading of the Sūryā wedding.

What the Atharvaveda knew

Here the argument from silence gets its test. If the Rigvedic poets lacked architectural vocabulary because their society had no architecture worth naming, the next stratum of Vedic Sanskrit should be similarly bare. It is not. It is crowded.

The Atharvaveda contains an entire hymn devoted to a house, AV 9.3, the śālā-sūkta. It is a charm for taking one apart, and it names the parts as it goes:

“Of the props (upamít), of the supports (pratimít), and also of the connectors (parimít) of the dwelling (śā́lā) … we unfasten the tied [parts].”

(Atharvaveda 9.3, verse 1, Whitney’s translation [3])

Then, verse by verse, the vaṃśa or bamboo beams, the nahana or ties, the prāṇāha or binding grass, the pakṣa or sides, the saṃdaṃśa or clamps, and the iṭa, the rush-work. The building behind this vocabulary is unmistakable: a bamboo frame, lashed rather than jointed, walled and roofed with reed matting and thatch, and capable of being untied and taken down without being destroyed. That last property is why the hymn exists at all. Nobody composes a charm for dismantling a house unless dismantling houses is a thing people do.

The vocabulary existed, or came into being very quickly. The Rigveda’s poets were simply not using it, because that is not what their poetry was for.

Then the Atharvaveda does something else, and it closes the circle. AV 4.5 is a sleep charm, and four of its verses are RV 7.55.5 to 8, transmitted almost word for word, including the line about shutting eyes “just like this habitation” [3]. This time we know what the charm is for. Whitney records the ritual manual’s application: the Kauśika Sūtra prescribes AV 4.5 “in a rite for putting to sleep a woman and her attendants, in order to approach her safely” [3]. Griffith, translating the Atharvaveda, gave the hymn the title “A lover’s sleep-charm”.

That last verse, present in the Atharvaveda and missing from the Rigveda, gives the game away. Somebody wants the whole household unconscious and himself alert until sunrise. Whether the Rigvedic version was already a lover’s charm, or a thief’s, or an ordinary blessing on a household’s rest that later acquired a more interesting career, is not settled and probably cannot be. What is clear is that RV 7.55 is not a bedtime prayer. It is somebody outside in the dark, dealing with a dog. The Rigveda is fuller of night work than its reputation suggests, as our piece on bad dreams and the Trita Āptya charms shows from another angle.

Aside. Arguments from absence are the most abused instrument in Vedic studies, and they cut in every direction. The Rigveda does not mention rice, tigers, cotton or bricks; it plainly does mention barley, lions, wool and dwellings. The absence of a word establishes something about the register and subject matter of a body of praise poetry. On its own it establishes very little about the world outside it. The disciplined version of the argument is comparative, and that is the version available here: these particular absences are meaningful precisely because the very next layer of the same tradition fills them all in.

The only building they cared to describe

Now invert the question. The Rigveda is not indifferent to construction. It contains a hymn devoted, in loving procedural detail, to putting a wooden object in the ground.

RV 3.8 is addressed to the yūpa, the sacrificial post, and to Vanaspati, the Lord of the Forest who is the tree it was cut from. The hymn follows the object from the axe to the ground:

“Ye whom religious men have firmly planted; thou Forest Sovran whom the axe hath fashioned … O men who lift the ladles up, these hewn and planted in the ground, Bringing a blessing to the field, shall bear our precious gift to Gods.”

(RV 3.8, verses 6 to 7, Griffith 1896)

Later verses have the posts anointed, robed, ringed with bands that “seem to the eye like horns of horned creatures”, and standing in a row to the east “like swans in lengthened line”. The precision is total: siting, orientation, material, finish, decoration, even the direction they face. Everything the hymns withhold from the house is lavished on a stake in a field.

The reason is not mysterious once you say it out loud. The Rigveda is liturgy. It is the working repertoire of poets employed to praise gods at a sacrifice, and its descriptive energy follows the ritual gaze. The sacrificial ground is described because the sacrificial ground is the subject: the vedi, the strewn grass, the post, the fires. The house is not described because the house is merely where everyone happens to live.

Except that the house does get in, sideways, through the one thing inside it that the poets care about enormously. Agni is gṛhapati, master of the house, in twenty-two verses. Agni is damūnas, the domestic one. Every hearth in every compound is a point of contact with a god, which is exactly why RV 10.46 can say Agni is “born in the harmyas”. The Rigvedic house appears in the Rigveda almost entirely as the container of a fire.

Ground truth

What did these compounds actually look like? The honest answer is that nobody knows, and the reason is worth stating plainly: no excavated settlement has been securely identified as Rigvedic. The hymns describe a world of the Punjab and the upper Indus system in roughly the second half of the second millennium BCE, and the material record for that time and place does not come labelled.

The best evidence is a few centuries late. B. B. Lal’s excavations at Hastinapura between 1950 and 1952 found, in the Painted Grey Ware levels, walls of mud and mud brick and wattle-and-daub, and plaster carrying the impressions of the reeds it had been packed against [4]. That is a house of exactly the kind AV 9.3 describes untying: a light frame, reed and thatch, mud packed over it. Text and trowel agree, and both belong to the period after the Rigveda rather than to the Rigveda itself.

The hymns’ own internal evidence points to something less settled than a village. The word grāma, which in classical Sanskrit simply means “village”, occurs nine times in the Rigveda, and in most of them it is a body of people on the move: the Bharatas’ grāma urged on by Indra across a river (RV 3.33), a grāma that a poet watches advancing from a distance, travelling “not by wheels” (RV 10.27), a grāmaṇī who leads one. The single passage where grāma feels like a settlement comes in the hymn to Aranyānī, the goddess of the wild, who is asked why she never enquires after the village (RV 10.146), and the question is put by somebody standing among the trees looking back. Settlement in the Rigveda is seen from the outside, or in motion, or not at all. That hymn gets a full reading in our piece on Aranyānī.

The other structure the text does mention has generated more heat than everything else combined. Pur, usually translated “fort” or “stronghold”, stands behind Indra’s standing epithet purandara, breaker of forts, and the Dāsas are said to hold purs by the hundred. In 1947 Mortimer Wheeler read those verses beside the ruins of Harappa and concluded that Indra stood accused [20]. Wilhelm Rau’s monograph The Meaning of Pur in Vedic Literature (1976) went through the attestations systematically and found nothing resembling a city: the Vedic pur is a rampart or an enclosure, often temporary, a refuge rather than a metropolis [5]. That conclusion has largely held, and it sits comfortably with everything else here. A people whose poets have no word for brick were not composing their war epithets against brick-built cities. The vocabulary of Rigvedic enmity is examined at length in our piece on the Dāsas and Dasyus.

timeline
    title Layers of evidence for the Vedic house
    Rigveda : three-verse hymn to Vastospati : a sleep charm : no house described
    Atharvaveda : Shala hymn AV 9.3 : named parts, ties, thatch : the same charm, now a lover's
    Grhyasutras : rites for siting and raising a dwelling
    Painted Grey Ware sites : mud and wattle-and-daub : plaster with reed impressions

Aside. One trap, and it opens in both directions. The absence of monumental architecture in the Rigveda is sometimes offered as proof that its composers were newcomers to the subcontinent, and sometimes suppressed because that conclusion is unwelcome. Neither move is scholarship. Pastoral societies with real wealth in cattle and metal routinely build in perishable materials, leave almost nothing behind, and produce poetry of extraordinary sophistication. The Rigveda’s silence about walls tells us about its subject matter and about a choice of building material. It tells us nothing about who the builders were.

Three kinds of bed

Go back to the doorway one last time, and notice what the text does hand over.

The final verse of the sleep charm sweeps the women of the household into unconsciousness, and to do that it has to say where they are lying. Whitney’s rendering of the Atharvaveda parallel keeps the distinctions Griffith blurs: “the women that are lying on a bench, lying on a couch, lying in a litter” [3]. Three Sanskrit words, proṣṭha, talpa and vahya, for three different things to sleep on, inside a hymn that will not tell us what the walls were made of.

That is the shape of the whole problem, and it is also the compensation. The Rigveda does not show you its world from the outside, in elevation, the way an excavation report does. It gives you the inside: the dog at the gate, the fire that has to be kept alive, the cattle brought in at dusk, the mother and the father and the master of the house asleep in a row, the beds they are asleep on, the woman somebody is trying to reach without waking anyone. We know what these people were afraid of at night. We know what they slept on. We do not know what they slept under, and after three thousand years that starts to look like the better half of the bargain.

The one building the Rigveda does put up, carefully and in full view, is a wooden post in a field, oriented east, cut with an axe, anointed, banded with metal, raised at dawn and taken down when the sacrifice was over. It stood for a day. It is the best-documented structure in the earliest literature of India, and the reason we can still see it so clearly is that it was the only one the poets were being paid to look at.

Read RV 7.55 and Atharvaveda 4.5 side by side, in that order, and watch a household blessing turn into a burglary. Then read RV 3.8 and notice how much a Vedic poet could see when he was actually looking at a piece of wood.