Eleven Names Before the Knife: The Āprī Hymns and the Fixed Grammar of Vedic Sacrifice

A Goat, a Post, and Eleven Names for Fire

Before the knife touches anything, someone has to talk to the post. In the older stratum of Vedic solemn ritual, an animal, usually a goat, is led to a squared wooden stake driven into the ground at the edge of the sacrificial enclosure. The stake has a name: yūpa. Before the animal is bound to it, before the fire is fed, before the god the sacrifice is nominally for gets anything at all, the presiding priest recites a short, fixed sequence of verses that greets eleven separate things in a set order: the newly kindled fire, a mysterious double of fire called Tanūnapāt, another called Narāśaṃsa, the sacred grass, the ritual doors, Night and Dawn together, two invisible heralds, three goddesses, the divine craftsman Tvaṣṭṛ, and the post itself, addressed by its own name, vanaspati, “lord of the forest.” Then, and only then, does the actual offering begin.

Ten hymns in the Rigveda do this job, and only this job. They are called the Āprī hymns, and almost nobody outside a small circle of Vedic ritualists and philologists reads them for pleasure. That is a shame: they are the closest thing the Rigveda gives us to a script, a genre so rigid that a single swapped name, in only one of eleven slots, identifies which priestly family composed it. This piece reads that script closely, asks what it tells us about the poets who wrote it, and follows it to a much older, non-Indian ritual it turns out to resemble. Readers of this blog’s close reading of the Puruṣa Sūkta will recognize the idea that sacrifice, for the Rigveda’s poets, was itself a kind of language. The Āprī hymns are where that language shows its grammar most plainly.

Quick facts. Genre: liturgical invocation (āprī, “propitiation”), recited before the prayāja fore-offerings of the solemn animal sacrifice (paśubandha). Corpus: ten hymns, RV 1.13, 1.142, 1.188, 2.3, 3.4, 5.5, 7.2, 9.5, 10.70, 10.110. Composed: c. 1500 to 1000 BCE, family-book examples earlier. Setting: the Sapta Sindhu region, Punjab and the upper Indo-Gangetic plain. Why it matters: the genre’s closest surviving relative is Iranian, the Avestan Āfrīnagān liturgy, a rare window onto a shared Indo-Iranian sacrificial inheritance older than either text.

The Ten Hymns Nobody Reads for Pleasure

Open any full translation of the Rigveda and the Āprī hymns read strangely next to their neighbors. There is no myth, no narrative, no argument, only a numbered checklist recited in verse. Griffith’s 1896 translation, still the only complete public-domain English Rigveda, labels several of them simply “Āprīs,” as though the translator gave up looking for a better title. [1] Jamison and Brereton’s 2014 Oxford translation, the standard modern rendering, treats the genre the same way: brief, functional, subordinate to a ritual the hymn itself never fully describes. [2]

The ten hymns are distributed unevenly across the text, and the pattern of that distribution is itself informative.

Table 1. The ten Āprī hymns of the Rigveda, with the traditional composer attributions given by the Anukramaṇī index and the Tanūnapāt/Narāśaṃsa distribution discussed below. [2]

Notice what is missing. Of the seven “family books” (Maṇḍalas 2 through 7), each traditionally tied to a single priestly lineage, only four contribute an Āprī hymn: the Gṛtsamadas (2), Viśvāmitras (3), Atris (5), and Vasiṣṭhas (7). Maṇḍala 4, the Vāmadeva book, and Maṇḍala 6, the Bharadvāja book, have none. The remaining six hymns sit in the non-family books: three in Book 1, one in the Soma book (9), two in Book 10’s miscellany. Whether the absence in Books 4 and 6 reflects a real ritual gap, a lost hymn, or independent transmission is not something the text answers. It is a clean fact worth sitting with before reaching for an explanation.

Aside. The traditional attributions in Table 1 come from the Anukramaṇī, the ancient indexing tradition that assigns a composer, meter, and deity to every hymn in the Saṃhitā. These indices are old and were compiled with real access to information now lost to us, but they are not proof in the modern historical sense. Where this piece calls a hymn “Vasiṣṭha’s” or “Jamadagni’s,” it means “attributed by the Anukramaṇī to,” not “provably composed by.”

There is also a name problem. The word āprī itself never occurs in the Rigveda Saṃhitā. It is a later label, first attested in the Brāhmaṇas, the Śrauta Sūtras, Yāska’s Nirukta, and the Bṛhaddevatā, texts that grew up to explain the Saṃhitā after it was closed. [3] The poets who composed these ten hymns did not call them Āprīs. Later ritualists did, once they needed a name for a genre the Rigveda itself never bothered to define.

Eleven Steps to a Sacrifice

The setting for all ten hymns is the paśubandha, “binding of the animal,” the solemn Śrauta ritual in which a domestic animal is consecrated and offered to Agni and Soma jointly. [4] Within it, the Āprī hymn supplies the prayāja, “fore-offerings,” a set of preliminary oblations described in later ritual texts, including the Aitareya Brāhmaṇa and the Āśvalāyana Śrauta Sūtra. [5] The hotṛ priest recites the verses on the direction (praiṣa) of a second priest, the maitrāvaruṇa. A mirrored set of closing verses, the anuyāja, bookends the ceremony once the central oblation is finished.

Strip away the individual hymn’s phrasing and the underlying sequence is remarkably stable across all ten:

graph TD
    A["Agni newly kindled"] --> B["Tanunapat (Fire's double)"]
    B --> C["Narasamsa (Fire's double)"]
    C --> D["Ila / sacred food"]
    D --> E["Barhis (sacred grass)"]
    E --> F["The divine doors"]
    F --> G["Night and Dawn"]
    G --> H["The two divine heralds"]
    H --> I["Three goddesses: Ila, Bharati, Sarasvati"]
    I --> J["Tvastr, the craftsman"]
    J --> K["Vanaspati, the sacrificial post"]
    K --> L["Svaha: the closing call"]

Figure A. The fixed eleven-step (plus closing) sequence common to all ten Āprī hymns, reconstructed from Oldenberg’s 1897 notes on RV 1.13 and confirmed independently by Sadovski’s 2018 tabulation. [6] Only steps B and C alternate; see the next section.

Two steps need unpacking, since their names are opaque to a modern reader. Vanaspati, “lord of the forest,” is not a stray reference to a garden tree. In the Āprī hymn it names the yūpa, the wooden sacrificial post, addressed as though it were a living participant in what is about to happen to the animal tied to it. [3] Sāyaṇa’s medieval commentary states the equation directly: vanaspatir vikāra yūpaḥ, “Vanaspati is the transformed post.” [3] The other opaque term, Iḷā (also Iḍā), is one of “three goddesses,” alongside Bhāratī and Sarasvatī, who together personify forms of ritual speech, food, and libation. Griffith’s 1896 translation of RV 2.3 renders the moment plainly:

Sarasvatī who perfects our devotion, Iḷā divine, Bhāratī all surpassing,

Three Goddesses, with power inherent, seated, protect this holy Grass, our flawless refuge!

(RV 2.3.8, trans. Ralph T. H. Griffith, 1896)

The same hymn closes its address to the post directly:

Vanaspati shall stand anear and start us, and Agni with his arts prepare oblation.

(RV 2.3.10, trans. Ralph T. H. Griffith, 1896)

A century of Indological reading has not settled what these three goddesses originally were. Martin Haug, writing in the 1860s and 1870s, treated Iḷā, Bhāratī, and Sarasvatī as genuine minor divinities. K. R. Potdar, in the Proceedings of the All-India Oriental Conference, argued the opposite: that the “Iḷās” are grammatically adjectival forms built on the root id-, with Agni as the real referent throughout. [7] Readers of this blog’s piece on Vāc and the Devī Sūkta will recognize the pattern: the Rigveda often personifies ritual speech into something goddess-shaped, and scholars disagree about how literally to take it.

A Name Optional, a Family Revealed

Here is the single most interesting fact about this rigid genre. Of the eleven items in the sequence, ten are fixed. One is not. The second and third slots, Tanūnapāt and Narāśaṃsa, both manifestations of Agni, do not both appear in every hymn. Some hymns invoke only Tanūnapāt, some only Narāśaṃsa, two invoke both. Which choice a hymn makes is not random: it tracks, with only mild untidiness, the priestly family that composed it.

Yāska’s Nirukta, the oldest surviving systematic etymological treatise in Sanskrit (roughly fifth century BCE), already noticed and classified this. The Bṛhaddevatā, the verse index of Rigvedic deities traditionally attributed to Śaunaka, states the underlying rule with unusual directness:

Now as to the deities which, stanza by stanza, are celebrated in this hymn [RV 1.13], they occur in all the Āprīs; the second deity, however, is subject to option.

(Bṛhaddevatā 2.151, trans. Arthur A. Macdonell, 1904)

Modern scholarship confirms the pattern hymn by hymn: RV 1.13 and 1.142 invoke both; RV 1.188, 3.4, 9.5, and 10.110 invoke Tanūnapāt only; RV 2.3, 5.5, 7.2, and 10.70 invoke Narāśaṃsa only. [2] Velizar Sadovski, in a 2018 comparative study for the Max Planck Research Library, states the significance plainly: “the question of which deity is addressed exactly in the second pre-sacrifice is determining for the (self-)identification of the clans and families of the Vedic priests and poets.” [6] A priest reciting an Āprī hymn was, in this one slot, doing something closer to signing his lineage’s name than making a theological choice.

graph LR
    Vas["Vasistha family (RV 7.2)"] --> N1["Narasamsa only"]
    Atri["Atri family (RV 5.5)"] --> N1
    Grt["Grtsamada family (RV 2.3)"] --> N1
    Vad["Vadhryasva, Bk10 (RV 10.70)"] --> N1
    Ag["Agastya, Bk1 (RV 1.188)"] --> T1["Tanunapat only"]
    Vis["Visvamitra family (RV 3.4)"] --> T1
    Kas["Kasyapa, Bk9 (RV 9.5)"] --> T1
    Jam["Jamadagni, Bk10 (RV 10.110)"] --> T1
    Med["Medhatithi, Bk1 (RV 1.13)"] --> B1["Both named"]
    Dir["Dirghatamas, Bk1 (RV 1.142)"] --> B1

Figure B. Which rishi-families invoke Tanūnapāt only, Narāśaṃsa only, or both, in their Āprī hymn. The pattern is not a scattergram: every attested family book example falls cleanly into one camp, and only the older, non-family Book 1 hymns hedge by naming both. [2][6]

Methods note. This kind of pattern only survives because the Rigveda was fixed, syllable by syllable, by an oral memorization tradition long before it was written down; this blog’s piece on that transmission process covers how such fine-grained textual stability was even possible. A scribal tradition copying and re-copying by hand would have been far more likely to “correct” an odd family variant toward whatever the copyist’s own school used.

Not every scholar treats the paśubandha setting as original to the genre. Jan Gonda’s 1975 History of Indian Literature and Lourens van den Bosch’s 1985 two-part study in the Indo-Iranian Journal, the modern discipline’s most sustained treatment, both raise the possibility that the Āprī hymns began as a family ritual centered on Agni, only secondarily folded into the animal sacrifice. [8][9] The evidence is circumstantial (the hymns describe fire, food, and ritual space, never the animal itself), but it is a caution worth keeping: what a later ritual manual says a hymn is for need not match what its composers originally intended.

What Does Āprī Mean?

The word itself carries a small etymological dispute going back to antiquity. Yāska’s Nirukta (8.4) asks directly: āpriyaḥ kasmāt? āpnoteḥ prīṇāter vā, “whence āpriya? From āp, ‘to obtain,’ or prī, ‘to please’?” [3] Modern scholarship favors prī, giving the standard gloss “propitiatory hymn.” [10] Sanskrit ā- + prī (आप्री) is, on this reading, distantly related to the root behind Old Church Slavonic priyati, “to be favorable.” [3]

Table 2. Key Sanskrit terms in the Āprī sequence, with IAST transliteration and literal gloss. The dual naming of the sacrificial post, functionally as yūpa, ritually as vanaspati, is characteristic of how Vedic liturgical language often gives ordinary objects a second, sacral name once they enter the ritual frame.

Did you know? - The word āprī never appears inside the Rigveda itself; it is a label the exegetical tradition supplied afterward. - Two of the seven family books, Vāmadeva’s and Bharadvāja’s, preserve no Āprī hymn at all. - The ancient tradition sometimes counts eleven Āprī compositions, not ten, folding in a related Yajurvedic Praiṣa formula outside the Rigveda proper. - The ten hymns run roughly eleven to fourteen verses, one verse per named ritual entity. - Griffith’s 1896 translation remains fully public domain, unusual among Rigvedic genres most modern anthologies skip over. - The closest surviving parallel to this genre is Iranian, not Indian.

The Iranian Twin

The comparison that makes the Āprī hymns matter beyond Vedic philology is Iranian. The Zoroastrian Āfrīnagān, a liturgy of blessing still performed in the Zoroastrian ritual calendar, shares its name with the Vedic genre at the root: Avestan ā-fri- and Sanskrit ā-prī- descend from the same Proto-Indo-Iranian verb. The Encyclopaedia Iranica states the comparison in terms nearly identical to the Sanskrit dictionaries’ gloss of āprī: “the word is derived from OIr. ā-fri- and invites comparison with Sanskrit āprī-, a class of prayers expressing an invitation to divinities to partake of the sacrifice.” [11] The comparison is not new; Martin Haug drew it as early as 1878. [12]

Sadovski’s 2018 study goes further than the naming coincidence and lines up the actual content of the two liturgies. The Avestan Yasna 71 and Vīsprad 7, part of the “fire list” embedded in the Zoroastrian Long Liturgy, invoke, in comparable order, fire, the ritual straw (barəsman, cognate with Vedic barhis), the waters’ child (Āpąm Napāt), and Narō.saŋha, transparently cognate with Vedic Narāśaṃsa. [6] Sadovski’s own summary is worth quoting directly:

“In the Āprī-litanies in the R̥gveda-Saṃhitā, there appear the same lists of eleven deities… The Āprī-litanies are attested in the whole Rigveda, for every single family of the Family Maṇḍalas, and not only this but also in all four Vedic Saṃhitās, including the Atharvaveda. As a twelfth element at the end of the List of the Eleven, we find the final sacred call svāhā, ‘hail!’”

(Velizar Sadovski, 2018) [6]

Table 3. Vedic Āprī terms and their Avestan counterparts, following Sadovski’s 2018 tabulation. Not a claim of borrowing either direction; both traditions inherited a common Indo-Iranian ritual vocabulary independently. [6] See also this blog’s piece on Mitra, named in a Hittite-Mitanni treaty.

This does not mean the two liturgies are the same ceremony in different languages. The Avesta and the Rigveda are separated by centuries of independent development, different regional histories, and eventually a religious reform, Zoroaster’s, that the Vedic tradition never underwent. What survives is a shared structural skeleton: a fixed list of ritual entities, invoked in something like the same order, at the threshold of a solemn sacrifice, in two closely related tongues. That skeleton likely predates the moment, sometime in the second millennium BCE, when Indo-Iranian speakers who would become “Vedic” and those who would become “Avestan” stopped being the same community.

Reading the Litany Today: Where Scholars Disagree

Three scholarly questions about the Āprī hymns remain genuinely open.

The first is origin. Did the paśubandha framing come first, with the hymns purpose-built for animal sacrifice, or did a pre-existing Agni liturgy get absorbed into the sacrifice once it took its classical Śrauta form? Van den Bosch and Gonda lean toward the second option; Sāyaṇa’s commentary and the Śrauta Sūtra tradition assume the first. [3][8][9] The text alone cannot settle this, since the hymns describe fire, food, and ritual space but never mention the animal.

The second is the theological status of the “three goddesses.” Haug’s reading takes Iḷā, Bhāratī, and Sarasvatī as independent minor divinities; Potdar’s reduces them to grammatical epithets of Agni. [7] Jamison and Brereton’s modern translation, more cautious than either, treats the question as unresolved. [2]

The third is what the Tanūnapāt/Narāśaṃsa variation actually proves. It is tempting to read it as hard genealogical evidence, a signature no family could fake. Michael Witzel’s broader work on dialectal features across the family books supports the idea that they carry real linguistic fingerprints, but his work does not, so far as this piece can verify, address the Āprī variation by name; the caution against overreading a single data point applies here too. [13]

Key Insight: A fixed liturgical genre is not evidence of a static religion. The Āprī hymns show a script so rigid that ten independent hymns, composed across roughly five centuries by poets who may never have read each other’s work, agree on eleven ritual entities and their order almost perfectly. The one place they disagree is the one place that mattered for identifying who was reciting.

A Short Chronology

Table 4. A chronology of the textual tradition and the scholarship built on it. Note how thin the twentieth century is between Oldenberg’s 1897 notes and van den Bosch’s 1985 article, a gap reflecting the hymns’ marginal position in the corpus rather than any lack of interesting content.

Āprī Hymns Per Book

Table 5. Distribution of the ten Āprī hymns across the ten Maṇḍalas. Book 1 alone holds three of the ten, consistent with its character as a later, editorially assembled collection rather than a single family’s inherited repertoire. The family-book examples (2.3, 3.4, 5.5, 7.2) always name exactly one of Tanūnapāt or Narāśaṃsa; only the Book 1 pair hedges by naming both, a further small sign that Book 1 is a later, deliberately comprehensive compilation.

Glossary

Āprī: “propitiation,” the post-Rigvedic name for these ten hymns. Anuyāja: the mirrored “after-offerings” closing the paśubandha. Barhis: sacred grass strewn on the ritual ground as a seat for invited gods. Hotṛ: the priest who chants the Āprī hymn during the sacrifice. Maitrāvaruṇa: the priest who directs (praiṣa) the hotṛ’s recitation. Narāśaṃsa / Tanūnapāt: “praise of men” / “son of himself,” the two alternating names for Agni’s second invocation. Paśubandha: the solemn Vedic animal sacrifice the Āprī hymn accompanies. Prayāja: the “fore-offerings” preceding the central oblation. Svāhā: the exclamation “hail!” closing an oblation. Vanaspati: “lord of the forest,” the ritual name of the yūpa. Yūpa: the wooden, or later stone, post to which the sacrificial animal is bound.

What to Notice While Reading

• Count the ritual entities in order: fire, the second Agni-form, sacred food, grass, doors, Night and Dawn, the two heralds, three goddesses, Tvaṣṭṛ, the post, the closing cry.
• Check which of Tanūnapāt or Narāśaṃsa appears, the closest thing the hymn gives to a signature.
• Notice what is absent: no myth, no narrative tension, no mention of the animal itself.
• Compare a family-book example (RV 7.2) against a Book 1 example (RV 1.13) side by side.

Frequently Asked Questions

What is an Āprī hymn? A short, formulaic hymn recited before the prayāja fore-offerings of the paśubandha animal sacrifice, invoking a fixed sequence of ritual entities: fire, the sacred grass, the sacrificial post, and others.

How many Āprī hymns are there? Ten, within the Rigveda Saṃhitā proper. The Bṛhaddevatā sometimes counts eleven by including a related, non-Rigvedic Yajurvedic formula, the Praiṣa hymn. [15]

Why do only four family books have one? Maṇḍalas 2, 3, 5, and 7 each preserve exactly one. Maṇḍalas 4 and 6 have none; the text does not explain why.

What does vanaspati mean here? “Lord of the forest,” the ritual name for the yūpa, the wooden sacrificial post, addressed directly in the closing verses.

Is the paśubandha still performed today? Some Śrauta communities in India maintain the solemn sacrificial tradition, but animal sacrifice within it is rare in modern practice and remains debated.

How confident are scholars about the family attributions? Reasonably confident for the Tanūnapāt/Narāśaṃsa pattern, attested independently by Yāska’s Nirukta and the Bṛhaddevatā. Less confident about the named individual composers, which rest on the Anukramaṇī tradition rather than internal proof.

What is the Avestan connection? Sanskrit āprī and Avestan ā-fri- (root of Āfrīnagān) descend from the same Proto-Indo-Iranian verb. Sadovski’s 2018 study matches the Vedic sequence’s entities against a comparable Avestan “fire list.”

Scholarly Voices

“The hymn belongs to the class of Āprī hymns, which were classed by the ancient arrangers of the Saṃhitā among the Agni hymns… destined for the Prayāja offerings of the animal sacrifice.” Hermann Oldenberg, Vedic Hymns, Part II, Sacred Books of the East vol. 46 (1897) [16]

“On the Āprī hymns see the detailed examination by Lourens P. van den Bosch.” Stephanie Jamison and Joel Brereton, Rigveda commentary project (UCLA) [2]

“The Āprī-litanies in the Rigveda-Saṃhitā… are attested in the whole Rigveda, for every single family of the Family Maṇḍalas.” Velizar Sadovski, Indo-Iranian Sacred Texts and Sacrificial Practices (2018) [6]

Reading the Grammar Backward

There is a temptation, faced with ten hymns this rigid, to read them as evidence that Vedic religion was a closed system running the same script for five centuries without deviation. The opposite is closer to the truth. A script this fixed only becomes visible because someone, at some point, chose to vary exactly one part of it, and later ritualists preserved that variation rather than harmonizing it away. The Āprī hymns are not proof of Vedic uniformity. They are proof that priestly communities cared enough about their own distinctiveness to leave a fingerprint inside a genre built for the opposite purpose.

That fingerprint is also one of the clearest surviving threads connecting the Rigveda to a sacrificial world older than India itself, one the Avestan priests of Iran inherited too, under a different sky, in a closely related tongue, using recognizably the same words for fire, straw, and praise. The Rigveda rarely lets a reader see that far back. The Āprī hymns, dull as their surface can seem, are one of the few places it does.

Open RV 7.2, the Vasiṣṭha family’s hymn, and read it beside RV 3.4, Viśvāmitra’s. The two lineages were, according to later tradition, rivals; the hymns will not tell you that story, but they will show you the same eleven-item liturgy handled by two different hands, one naming Narāśaṃsa, the other Tanūnapāt, both bound to the same yūpa in the end.