The Fault Called Sameness: Jāmi and the Vedic Distrust of Things Too Alike

A Priest Is Told to Change His Voice

Somewhere in the first half of the first millennium BCE, an anonymous ritual theorist sat down to explain why a priest performing two offerings in a row should murmur the first formula and then say the second one out loud. The explanation has nothing to do with audibility, and nothing to do with piety. It is a piece of engineering. Two offerings are going to the same pair of gods, one of melted butter and one of a rice cake, and if they are performed identically the ritualist has a problem on his hands. The text raises the objection in the voice of an imagined critic and then answers it. Julius Eggeling, translating in Oxford in the 1880s, gave the key phrase as “sameness of performance,” and noted the Sanskrit behind it: jāmi. [4]

Now open the Rig Veda, four or five centuries earlier, and look up the same word. It means your sister.

That is the whole puzzle in one line. A word that names the woman in your household who is closest to you by blood becomes, in the hands of the ritual theorists, the technical term for a mistake: two things placed side by side that are too much alike to accomplish anything. Nobody coined a new term for redundancy. They reached for the kinship vocabulary, and everybody understood.

Follow jāmi through the Rig Veda and you are not doing philology for its own sake. You are watching a culture work out, in the only vocabulary it has, a question that turns out to be general: what can sameness produce, and at what point does likeness stop being a resource and start being a dead end? The answer the Vedic poets and their successors arrive at is consistent across contexts that look to us entirely unrelated. Who you may marry, how a chant may be repeated, why two offerings must differ, and what a dawn has to do to count as a new dawn all turn out to be versions of one problem.

What the Word Actually Does

Jāmi (जामि) is both a noun and an adjective, and Vedic poets use it without the slightest self-consciousness in either role. As an adjective it means “of the same stock, belonging together by birth.” As a feminine noun it means a kinswoman, most often a sister. The word sits inside a dense Indo-European kinship cluster whose other members are easier to pin down: jāmātṛ, “son-in-law,” corresponds to Avestan zāmātar and Greek gambrós, “relative by marriage,” and goes back to a root meaning something like “to marry into.” [6] [7] [11] The exact derivation of jāmi itself is less settled than the tidy handbooks suggest, but its semantic centre is not in doubt. It denotes relatedness by origin.

What makes it interesting is the range of things the Rig Veda considers related by origin.

Notice the fourth row. In RV 10.64.13 the poet asks the Maruts whether they remember the common descent the family claims with them, and then adds a request: nābhā yatra prathamaṃ saṃnaśāmahe tatra jāmitvam aditir dadhātu naḥ, “where we first come together at the navel, there may Aditi establish kinship for us.” Kinship is something a goddess has to put in place. It is not a biological given that the ritual acknowledges; it is a state of affairs the ritual is trying to bring about. Anyone who has sat through a wedding knows the feeling, and the wedding hymn at RV 10.85 is precisely a text for manufacturing relatedness where none existed.

The god who gets the epithet most often is Agni. He is the one deity with a physical presence in every household, which is presumably why the poets keep calling him family. RV 1.75 puts the question and answer in two consecutive verses, and the sequence is worth reading as a pair rather than as isolated lines:

kas te jāmir janānām agne ko dāśvadhvaraḥ |

tvaṃ jāmir janānām agne mitro asi priyaḥ

“Who is your kinsman among the peoples, Agni? Who worships you rightly? … You are the kinsman of the peoples, Agni; you are the dear ally.”

Rigveda 1.75, verses 3a-b and 4a-b. My rendering. Griffith (1896) has “Who, Agni, is thy kin, of men?” and “The kinsman, Agni, of mankind.” [2]

The question is asked and then dissolved. Agni has no particular kin among humans because he is kin to all of them. Fire is the one member of the family that every family has. For the broader logic of Agni as the household’s resident god, the epithet is worth more than a dozen theological statements.

Methods note. The counts in this article come from a direct string search of the romanised Saṃhitā text, checking every word containing the sequence jāmi and discarding the ten accidental matches: verbs like sṛjāmi (“I release”) and rujāmi (“I break”), and sandhi collisions like prajām iṣam. What remains is 55 genuine forms in 49 verses across 45 hymns. Grassmann’s 1873 concordance and the Vedic Index remain the authorities for anyone who wants the entries argued rather than counted. [5] [10] Glosses here follow Griffith where quoted and are marked as my own where they are not; Jamison and Brereton’s 2014 translation is the standard modern English rendering and should be consulted for any verse discussed. [1]

Kin and Not-Kin: a Pair That Means Everyone

Six times, the Rig Veda puts jāmi next to its own negation. Every one of these is a totality formula, the figure linguists call a merism: you name two exhaustive opposites in order to mean the whole. English does it with “young and old,” “high and low.” Vedic does it with kin and non-kin.

The formula works because the two categories are both exhaustive and mutually exclusive:

$$J \cup \bar{J} = W, \qquad J \cap \bar{J} = \varnothing$$

where $W$ is the social world as the poets conceive it. There is no third option, no one who is neither. That is a real statement about how Rigvedic society modelled itself: everybody you will ever deal with is either of your stock or not, and the line between is the primary social fact. A world that had a large middle category of people related by contract rather than descent would not generate this figure. Paul Thieme’s monograph on the ari, the stranger who is neither kinsman nor enemy, is the classic attempt to show how uncomfortable the Rigvedic vocabulary is with anything in between. [17]

What is startling is where the formula gets used. Four of the six are battle prayers, and they are chilling in their reach:

ava sthirā tanuhi yātujūnāṃ jāmim ajāmim pra mṛṇīhi śatrūn

“Slacken the taut bows of those driven by sorcery. Crush the enemies, kin and non-kin alike.”

Rigveda 4.4, verse 5c-d. My rendering. Griffith (1896): “destroy our foemen whether kin or stranger.” [2]

The petitioner is asking Agni to kill enemies within the lineage as readily as enemies outside it. Vedic warfare was substantially warfare between related clans, which the battle of the ten kings makes plain, and this line simply declines to make an exception for cousins.

Then there is the sixth passage, which uses the identical figure to mean the exact opposite.

nājāmiṃ na pari vṛṇakti jāmim

“She excludes neither the non-kin nor the kin.”

Rigveda 1.124, verse 6b. My rendering. Griffith (1896): “debarreth from her light nor kin nor stranger.” [2]

The subject is Dawn. The same merism that authorises killing your relatives is used, a few hundred verses away, to make the point that daylight is the one good that arrives without checking your pedigree. The verse goes on to say she does not turn away from the humble any more than from the great. It is one of the cleanest ethical arguments in the collection, and it is made entirely out of kinship grammar: here is a thing that behaves as though everyone were family. The hymns to Uṣas are usually admired for their imagery, which is fair, and the imagery has a logical structure under it, which is less often noticed. Compare the treatment of the guest who arrives from outside the lineage, where the same boundary is crossed by deliberate effort rather than by nature.

The Sisters in the Pressing Shed

Eleven of the fifty-five occurrences are in Maṇḍala 9, the book of soma purification, and they are almost all plural feminine: jāmayaḥ, the sisters. They are not goddesses. They are the fingers.

eṣa vṛṣā kanikradad daśabhir jāmibhir yataḥ |

abhi droṇāni dhāvati

“This bull, bellowing, guided by the ten sisters, runs toward the wooden vats.”

Rigveda 9.28, verse 4. My rendering. Griffith (1896): “Directed by the Sisters ten.” [2]

Ten sisters, two hands, one juice. Elsewhere the sisters are seven and they are the priests, sitting samīcīnāsaḥ, turned toward one another, “filling the station of the One” (RV 9.10.7); or seven and they are the rivers or the poetic metres, urging the pressed drink along with their songs (RV 9.66.8). At RV 1.23.16 the waters themselves come down the roads as jāmayaḥ of the officiating priests, mixing sweetness with milk.

Something specific is being claimed by that vocabulary. A set of things that are alike, of one origin, and acting in concert is exactly what the Vedic poets call a sisterhood. Ten fingers are the paradigm case: identical in kind, indistinguishable in function, useless individually, decisive together. The work of the pressing shed and the division of labour among the priests both get described in this idiom, and the description is doing analytical work rather than decorating the scene.

So sameness in the pressing shed is not a problem at all. It is the productive arrangement. The ten sisters work because they are alike. Hold on to that, because the next section appears to contradict it, and the apparent contradiction is where the interesting content lives.

The One Place Where Kinship Is a Reason Against

Three of the ten Maṇḍala 10 occurrences are in a single fourteen-verse dialogue, and it is the only passage in the Rig Veda where being of one stock is an argument for not doing something. Yamī, the first woman, proposes to her twin brother Yama that they conceive the human race. He declines. She presses. He keeps declining.

The full hymn belongs with the other dialogue hymns and their female speakers, and Yama’s side of it with the funeral hymns where he becomes king of the dead. What concerns us here is narrower: the argument is conducted in the vocabulary of sameness, and neither party treats the outcome as obvious.

Yama’s line in verse 4 is the hinge: “that is our bond, that is our highest kinship.” He is conceding everything Yamī says about how close they are, and turning the concession into the refusal. Being maximally jāmi is precisely the disqualification. Then comes the verse that gave the later tradition its phrase:

ā ghā tā gachān uttarā yugāni yatra jāmayaḥ kṛṇavann ajāmi

“Those later ages will indeed come, when kinswomen will do what is not kin-like.”

Rigveda 10.10, verse 10a-b. My rendering. Griffith (1896): “Sure there will come succeeding times when brothers and sisters will do acts unmeet for kinsfolk.” [2]

Jāmayaḥ kṛṇavann ajāmi: the sisters will do the un-sisterly. The line is a small masterpiece of compression, and it also concedes that the rule is a rule, not a law of nature. Yama does not say it cannot happen. He says that when it happens, it will be a mark of degenerate times. Leonid Kulikov’s study of the hymn and its Iranian relatives sets out how differently the same inherited pair of twins was handled east and west of the Hindu Kush, and how much of the Vedic version’s force comes from its refusal to settle the question by simple prohibition. [8]

Verse 9 contains the third occurrence and is genuinely difficult: yamīr yamasya bibhṛyād ajāmi, which Griffith renders as though the non-kin-like act of Yama should fall on Yamī. Read as Yamī’s own speech, which is the standard assignment of the alternating verses, she appears to be offering to carry the consequence herself. Translators diverge here, and readers should compare two or three before deciding what they think. [1] [2] [20]

graph TD
    A["jāmi: of one stock"] --> B["sister, kinswoman"]
    A --> C["kinsman of all: Agni"]
    A --> D["the sisters: fingers, waters"]
    A --> E["ajāmi: not of one stock"]
    B --> F["RV 10.10: too close"]
    E --> G["merism: kin + non-kin = all"]
    D --> H["likeness that works"]
    F --> I["Brāhmaṇa jāmi: redundancy"]
    I --> J["ajāmitā: required difference"]
    H --> J

Ajāmitā: the Word Becomes a Technical Term

Between the hymns and the ritual manuals something happens that historical semantics does not often let us watch this clearly. Jāmi keeps its shape, keeps its associations, and acquires a second job.

In the Pañcaviṃśa Brāhmaṇa, the manual of the Sāmavedic chanters, the priests face a structural difficulty: a sequence of ājya praise chants that are, in form, the same chant repeated. The text names the problem with the bare word, taj jāmi, “that is sameness,” and prescribes the remedy in the dative of purpose: ajāmitāyai, “for the sake of non-sameness,” the chants are directed to different deities. [3] The reasoning is stated as a fact about causation rather than as a matter of taste. A succession of identical acts produces nothing.

The Śatapatha Brāhmaṇa applies the same principle to the offering fire. Faced with a butter oblation and a cake oblation going to the same divine pair, the text stages the objection in the imagined voice of a critic, kim idaṃ jāmi kriyate, “what is this sameness that is being performed here?” and answers it with a change of voice: the first formula murmured, the second spoken aloud. [4] The offerings stay; the delivery differentiates them.

Three cautions before anyone builds too much on this.

First, the Rig Veda never uses the technical sense. Not once in the fifty-five occurrences does jāmi mean “redundant.” The ritual meaning is a development of the prose period, and reading it back into the hymns produces exactly the kind of anachronism that makes Vedic scholarship difficult to trust. What the Saṃhitā supplies is the raw material: a word that already means “too much the same” in one context, the sexual one, available to be generalised.

Second, the Brāhmaṇas argue by analogy, not by logic. Their characteristic move is to assert a connection, a bandhu, between the ritual and the cosmos, and then reason from it. Brian K. Smith’s study of the resemblance principle is the fullest account of how this system works and how its excesses were policed from inside. [14] Too little resemblance and the ritual connects to nothing; too much and it collapses into sterile repetition. Jāmi names the failure at the second end.

Third, this is not a theory of originality. The ritualists wanted the minimum differentiation required to keep a sequence of acts from cancelling out, and they specified it like a technical standard: change the deity, change the register. Whether the change interests anyone is beside the point.

Aside. The reasoning is explicitly reproductive. A pair that can generate is a pair of unlike things, and the model in front of everyone’s eyes was the fire drill: two pieces of wood, one socket and one shaft, producing a third thing that is neither. The same structure underlies the Vedic fondness for paired deities in the dual, where two gods who are not the same are welded into a single grammatical unit precisely so that they can act as a productive couple.

But the Hymns Repeat Constantly

Here is the objection that ought to be bothering any careful reader by now. The Rig Veda is one of the most repetitive texts ever composed. Whole pādas recur across books. Maṇḍala 9 contains hundreds of verses that describe the same liquid going through the same filter into the same vat with a vocabulary so standardised it reads like industrial process documentation. Refrains repeat verbatim for a dozen verses. If sameness were a fault, the collection would be nothing but faults.

The resolution is in the Dawn hymns, and it is the reason the Uṣas cycle deserves more attention than its reputation as pretty nature poetry allows. Dawn is the Rigveda’s own image of a thing that recurs, and the poets are entirely aware of the problem this poses. RV 1.124, the hymn that gives us the kin-and-non-kin merism, works its way to this:

āsāṃ pūrvāsām ahasu svasṝṇām aparā pūrvām abhyeti paścāt |

tāḥ pratnavan navyasīr nūnam asme revad uchantu sudinā uṣāsaḥ

“Of these sisters who came before, a later one follows the earlier through the days. So may they, new in the manner of the ancient ones, dawn richly upon us now, bright-dayed.”

Rigveda 1.124, verse 9. My rendering. Griffith (1896): “So, like the past, with days of happy fortune, may the new Dawns shine forth on us with riches.” [2]

Pratnavat navyasīḥ. New in the manner of the old. The dawns are sisters, one of a kind, identical in function, and the poet’s word for what saves them from being merely a repetition is that each of them is newer. Each instance is renewed rather than reissued.

The poets apply exactly this standard to themselves. Forms built on navya- and navīyas-, “newer,” appear on the order of eighty times in the Saṃhitā, almost always describing the song currently being sung. The request is stereotyped, which is half the joke: gāyatreṇa navīyasā, “with a newer verse in the gāyatrī metre,” says RV 1.12.11, in the oldest and most conventional metre available. A poet in this tradition is not required to invent a new form. He is required to produce this occasion’s instance of the form, marked as belonging to now and to him. The distinction is the same one the ritualists were making, and it is why the craft vocabulary of Vedic poetry is full of words for fitting, joining and renewing rather than for originating. It is also why a transmission system built on exact memorisation never felt to its practitioners like mere copying.

Repetition of the kind, then, was never the problem. Repetition that fails to differentiate the instance was. The Vedic fault is not “you have done this before.” It is “you have done this twice in a row in a way that makes the second occurrence indistinguishable from the first, and nothing will come of it.”

What the Rule Was Really About

Put the pieces in one line and the shape is hard to miss. Ten fingers, alike and coordinated, press out the soma. Two offerings, alike and undifferentiated, produce nothing until one is murmured and the other spoken. Two twins, maximally alike, cannot begin the human race. Two dawns, alike in kind, work because the second is marked as newer than the first.

The operating principle is that generation requires a pair of terms that are related but not identical. Too little relation and there is no connection at all; the ritual sequence becomes a heap of unrelated gestures, the marriage joins nothing, the verse means nothing. Too much and the pair collapses into a single term and goes sterile. Jāmi names one edge of that band and the Brāhmaṇas built a technical vocabulary around avoiding it.

What I find genuinely arresting about this is not the content of the rule but its scope. We file “whom may I marry” under ethics or law, “must the second chant differ from the first” under aesthetics or professional craft, and “what makes two things productive together” under biology. The Vedic system files all three in the same drawer, under one noun, and nobody in the tradition appears to have thought this odd. That is not muddle. It is the consequence of taking relatedness as the fundamental category and asking, in every domain, how much of it a given arrangement can bear.

The honest coda is that this way of thinking did not stay generous. A principle that begins as a rule about what sameness cannot produce is easy to convert into a rule about who may not mix with whom, and the later legal literature did exactly that, elaborating forbidden degrees of kinship, gotra exogamy and, alongside them, the endogamies that hardened into caste. The energy that had gone into avoiding sameness went into policing it. The Rig Veda is not responsible for that, and the honest reader will neither hold it accountable nor pretend the connection is imaginary. What the Saṃhitā gives us is earlier and stranger: a world in which your sister, your fingers, the dawn and a badly designed sequence of offerings all fall under one word, because they are all answers to the question of what likeness is good for.

Read RV 1.124 straight through in one sitting. It is a Dawn hymn, so it has a reputation for being decorative. In verse 6 it uses kinship grammar to say that daylight discriminates against nobody; in verse 9 it works out how a thing can recur without becoming a repetition. Then read RV 10.10, where the same vocabulary is used to refuse. The two hymns are separated by most of the collection and by a considerable stretch of time, and they are arguing about the same thing. For the larger framework of order that both assume, see the Rigvedic vocabulary of ṛta and dharman.