Adhiyajña: Towards a Performance Grammar of the Vedas

Abstract

Recent scholarship has challenged the anachronistic projection of the modern category of the poem onto premodern texts. This article attempts to theorize how one might construct an alternative to modern conceptualizations of “the poem” that more closely appropriates the conceptualization of textuality in the Rigveda, an anthology of 1028 sūktas “well-spoken (texts)” that represents the oldest religious literature in South Asia. In order to understand what these texts are and what they were expected to do, this article examines the techniques by which the Rigveda refers to itself, to its performer, to its audience, and to the occasion of its performance. In so doing, this article theorizes a “performance grammar” comprising three axes of textual self-reference (spatial, temporal, and personal); these axes of reference constitute a scene of performance populated by rhetorically constructed speakers and listeners. This performance narrative, called here the adhiyajña level, frames the mythological narratives of the text. By examining the relationship between mythological narrative and performance narrative, we can better understand the purpose of performing a text and thus what kind of an entity Rigvedic “texts” really are. While this article proposes a rubric specifically for the Rigvedic context, its principles can be adapted to other premodern texts in order to better understand the performance context they presuppose.

1. Preliminary Remarks

In her monograph, Beyond Orality: Biblical Poetry on Its Own Terms, Jacqueline Vayntrub argues that the modern categories of poetry and prose have anachronistically imposed themselves on scholars’ interpretations of biblical material. Tacitly, contemporaneous notions of the poetic text impose themselves on past texts as though notions of text itself were universal. In her work, Vayntrub interrogates the categories of verbal art native to the texts themselves, in particular the term mashal,1 frequently translated as “proverb”, is perhaps better thought of as “a speech act.” This is not because of any intrinsically oral character, but because that is how the Hebrew Bible frames the text by situating it in a time and place and attributing it to a speaker. This work represents a huge advance in philological theory, one that scholars of ancient verbal art cannot ignore. What is the nature of that advance? In “Future Philology? The Fate of a Soft Science in a Hard World,” Sheldon Pollock argues that the essence of philology is making sense of texts.2 To make sense of the contents of texts, however, we must first understand the container. I think that is the nature of Vayntrub’s challenge, to really make sense of the contents of texts, understanding the idea of what text is to its producers and to its consumers is a necessary precondition.

The situation in the oldest verbal art from India is somewhat different than the Hebrew Bible. The Vedas are not just represented as oral; they have indeed been orally transmitted for three millennia up to this very day. Their survival has relied on recitation and memorization passed down for generations; so many generations, in fact, that they are often believed to be timeless, authorless and eternal3. Are they oral texts? Surely, but what does that really mean? Vayntrub has shown that “orality” and “textuality” are two sides of a false coin. The orality of the Rigveda, for example, is not casual speech, but a special performative register. It is an oral textuality. So, what is this oral textuality? How does one attempt to understand how these texts theorize what they are doing? What did poets think they were making when they composed a metrical sūkta, which is a substantivized adjective that means something like “a true spoken” something. Is that something a poem? The purpose of this study is not to arrive at a final answer, but instead to take up Vayntrub’s challenge by laying the groundwork for how we might analyze this material to see how it depicts texts and how it locates text in a time and a place. How do the texts depict their performers? How do the texts represent their listeners? If we are to discover native ontologies of text, both what they are what they do, then we cannot import modern notions of author or reader, instead we must investigate how texts construct discursive speakers and listeners in anticipation of historical ones.

The Rigveda is a collection of 1028 sūktas, a term mentioned above, but it derives its name from the word $\underset{°}{r}$c-, “verse.” Why is the $\underset{°}{r}$c a more meaningful organizational unit than the sūkta? This has to do with the history of the use of the text. The Rigveda is one of four Vedas. It appears to be a collection of songs sung in praise of the gods at religious ceremonies in the foothills of the Hindu Kush in the late bronze age, near the end of the second millennium BCE. These ceremonies included major social events such as royal consecrations, weddings, and funerals. At some later point in time, for unknown reasons, this orally memorized, performed, and transmitted collection was redacted into the form we know it. The contents of the other three Vedas were probably created around this time as well. When these anthologies were created, priestly traditions organized themselves into sodalities of oral transmission exclusively dedicated to one Veda, which contained one kind of mantra or sacred utterance. The Rigveda and the Atharvaveda contain $\underset{°}{r}$c, “verse,” the Sāmaveda contains sāman, “melody,” and the Yajurveda contains yajus, “ritual formula.” This division of genre was accompanied by a division of ritual labor: recitation for the Rigvedin, chorus singing for the Sāmavedin, and to the Yajurvedin went all the remaining tasks such as muttering dedications, kindling the fire, and pouring offerings. After the creation of these collections, the Vedas were expanded by texts that commented on the ritual called Brāhmaṇas. This genre division is explicit in later texts, for example the Śāṅkhāyanag$\underset{°}{r}$hyasūtra, which is largely concerned with the proper performance of domestic rituals.

śrutaṃ tu sarvān atyeti | na śrutam atīyād | adhidaivam athādhyātmam adhiyajñam iti trayam | mantreṣu brāhmaṇe caiva śrutam ity abhidhīyate ||4

“Heard (revelation)” surpasses everything; heard (revelation) should not be passed over. (It is) pertaining to heaven, yet also to the self (and) to the sacrifice; (it is) this triplet. Only what is in mantras and the Brāhmaṇas, is defined as “heard (revelation).”

In this passage, śruta-5, the “heard (revelation),” is explicitly limited to Vedic mantras and the sacred Brāhmaṇa commentary. One term in the Rigveda that refers to text is brahman, and it is from this word that brāhmaṇa, “concerning the brahman,” is derived. As the word brahman took on greater theological significance in later Vedic texts6, it makes sense that the term brahman would be replaced by mantra as the unmarked term referring to sacred utterance. This brings me to an important point about this study. The study of the notion of textuality cannot be a study of the uses of a word that means “poem” or “text” for the very reason that individual words change meaning over time. In an oral tradition, there are myriad words that refer to texts. Because the performer memorizes generations of material, what specifically these words refer to is often unclear. In this article alone, we will see sūkta-, brahman-, stoma-, āṅgūṣa-, uktha-, $\underset{°}{r}$ta-, and gīr-, all of which can refer to text in the Rigveda. Are they interchangeable? Are they different genres? Were they distinct at one point in history but not at another point? Because different sūkta are composed by different people at different times, a given semantic can be difficult to pin down. Further, what about words like mati, manīṣā, and dhī? These words seem to refer to a poetic ideas or vision but often refer metonymically to the texts which contain a poetic idea or vision. Thus, we must avoid reifying words as though one word can consistently reflect one idea in all its attestations. Instead, we should think about consistent strategies by which texts conceive of textuality. Let us return to the above passage from the Śāṅkhāyanag$\underset{°}{r}$hyasūtra. In it, the heard (revelation), both the mantras and the Brāhmaṇas are described as adhidaivam, adhyātmam, and adhiyajñam. All three are compound words whose first member is the word adhi, “on, on top of,” but in this context really meaning “pertaining to.” The word adhidaiva means pertaining to what is a daiva, which means “heavenly” or “divine,” the word adhyātma means pertaining to the ātman or “self”, and the word adhiyajña means pertaining to the yajña or “sacrifice.” In other words, the late Vedic tradition conceived of its texts as applicable in three discrete theatres: the cosmic, the personal, and the sacrificial. As the yajña is the performative context presupposed by the Vedas7, I think it is appropriate to use this term, adhiyajña, to refer to the bundle of strategies by which Vedic texts refer to their own textuality, because it is at the yajña that that textuality would have been experienced directly, because it is there that it would have been audible. Of course, there are other performance contexts than ritual ceremony. For example, pedagogical ones, where the oral corpus is transmitted from one generation to the next, or svādhyāya “personal revision” are also performative contexts. This study does not aim to fathom all the possible historical performances of the Rigveda but rather its rhetorical performances. Remember, we want to understand how the Rigveda represents performance in order to understand how its notion of textuality shapes its literary contents.

2. Performative Utterances

So how do we start? If we imagine that Vedic performance is intended to bring about some effect, whether it is a ritual transformation or the simple act of persuasion, then a good place to start would be performative utterances. Eystein Dahl explains that “performative sentences represent a pragmatically marked type of context where the speaker utters the sentence and at the same time fulfils an act of the type specified by the verb8.” The Paradebeispiel of this type is “I promise” in which the sentence describes the act of promising as well as enacting a promise. This “enacting” is the illocutionary point of the sentence; speech brings that promise into being. In (Searle 1979)’s taxonomy of illocutionary acts, the promise would be classed as an assertive because in so promising, the performer asserts this promise is true.

Searle’s other illocutionary categories are relevant to this study too. He classes “I ask” as a directive, for example, because the speaker directs the hearer to act. In presenting illocution as conforming to discrete categories, however, his taxonomy can be misleading. By the same logic that categorizes “I promise” as an assertion of truth, one can categorize “I ask” as an assertion rather than a directive, for by saying “I ask”, I assert the sincerity that I do indeed truly ask. By this logic, many illocutionary acts can be folded into the category of assertion. For Searle, a declarative speech act, in principle, changes reality in accordance with the content of that declaration9, while an assertive merely commits the speaker to the truth of the proposition10. Searle notes that declarations derive their illocutionary force from an extra-linguistic institution. In practice, the distinction between the two types is often blurred. Consider the legal verdict. Searle theorizes the verdict as an overlap of the assertive and the declarative11, because the judge declares someone guilty, making them guilty, yet simultaneously commits to the truth of the proposition that this person is guilty. This double illocution holds for all judicial decisions, which declare legal determinations yet also assert that these decisions are the correct or ‘true’ interpretation of legal precedent12.

The important point here is that the taxonomy of illocutionary acts echoes a relationship with an extra-linguistic institution; the words “guilty” or “not guilty” do not alone change a person’s legal status. If we instead think about this event as a ritual performance, the sentence receives its illocutionary force because it is an authorized ritual act performed by the judge as a ritual actor13. I would add that if we use Searle’s terminology to approach the courtroom holistically, we might say the bailiff performs an illocutionary act when he performs the directive that “all rise”. The illocutionary force of the imperative, however, is secondary to the perlocutionary effect of his utterance14, which identifies the person entering the room to be the proper ritual actor, cueing the audience that this person has special powers of speech at this legal occasion. It is the legitimacy and authority invested in the court which elevates the judge’s assertion to the status of declaration.

Mutatis mutandis, it is the legitimacy and authority of the Vedic sacrifice which determines if assertive utterances function as declaratives, but it is exactly that institution which we cannot access because it is external to the texts. In a sense, however, this is irrelevant, as the assertive is performative by default and only depends on a shared notion of truth between speaker and hearer. Searle claims that “making a statement is as much performing an illocutionary act as making a promise, a bet, a warning, or what have you15.” Consider a typical Yajurvedic mantra from the Kaṭhasaṃhitā16:

devásya tvā savitúḥ prasavè ‘śvínor bāhúbhyāṃ pūṣṇó hástābhyām$\acute{\stackrel{\_}{a}}$dade17

You do I take with the hands of Pūṣan, with the arms of the Aśvins, at the pressing of heavenly Savitar.

What does this yajus tell us? The verb (tvā…) $\acute{\stackrel{\_}{a}}$dade, “(You…) do I take”, seems to be performative; like “I promise”, it describes what it enacts, but exactly what ritual action it enacts is ambiguous. To the mere dilettante of Vedic sacrifice, the adhvaryu appears to be a human priest and one might imagine he comes equipped with human hands and human arms. Kaṭhasaṃhitā 1.2, however, asserts a different truth: that the speaker has the hands of Pūṣan and the arms of the Aśvins. That these are the hands of Pūṣan and the arms of the Aśvins is a reality otherwise invisible save for this assertion18. Through this yajus, the adhvaryu asserts this divine body into existence, while verbally disguising his own human hands and arms.

3. Narrative as Disguise

What does it mean to verbally disguise oneself? In 1983, Boris Oguibénine wrote on masks in Vedic ritual, coming to a singular insight about the ontology of disguise regarding śrauta rituals. The Śrautasūtras are ritual manuals composed after the Vedic period and which the native tradition does not consider śruta or śruti, the “heard (revelation),” but manuals of human composition. The rituals as described by these texts are highly aniconic. Oguibénine remarks that Vedic religion: “remain[s] in the domain of discourse that does announce the disguising of representations”19. In other words, during the ritual one thing is referred to in terms of another thing, as though an act of disguise were taking place. Oguibénine offers as an example of this type of masking the dakṣinā cow who acts as a surrogate for any ritual gift. He adds that in Vedic “disguises indicating virtual masks do not lead to the fabrication of corresponding material images, but the relation between real and virtual remains the same”20. Oguibénine borrows this notion of the “virtual mask” from Claude Lévi-Strauss,21 who used the term “virtual mask” to distinguish the origin myth connected to the material mask used in North American Indian ritual from the material mask itself.

Lévi-Strauss’s observation, that the ritual object derives its significance from a spoken narrative cannot be understated. If ritual actions are symbolic, then they are not meaningful in and of themselves, but their importance is linked to what they signify. In performance, that signification is conveyed through speech. It follows that the narrative associated with the mask, rather than a physical description of the mask, would be the topic of speech at a ritual performance. The physical characteristics of the mask are obvious, and the special origin of the mask is obscure. The mask provides that narrative with a physical anchor, materializing it so that it can affect the material world, while the narrative endows the mask with significance. In that light, even when a physical mask is present, the “virtual mask” is the real disguise. Neither a mask composed of wood nor a mask composed of speech would be a functional disguise outside of the context of performance. This means that, phenomenologically, ritual assertions of disguise function identically to disguises which use ritual props. In the Vedic case, the physical component is not a mask but the performer’s voice and body.

I think that Lévi-Strauss’s idea of the “virtual mask” is a useful way of thinking about the Rigvedic texts, because it frees us from dependence on real historical actors that have not been properly theorized yet. When we ask who produced a Rigvedic sūkta, we might say “the author”, but that imports a modern notion of authorship and imposes it on the text. What is “authorship” from the perspective of the text? If we assume that the text has historical human authors, then we engage in circular reasoning because we are looking for historical actors that resemble the authors of modern texts. Instead, we can expand Oguibénine’s application of the “virtual mask” beyond narratives which euphemize one thing as another, a gift to a priest as a cow, to the entire speaker–listener complex of the text. This has already been done in part by George Thompson22, who uses the term “verbal mask” in his study of poetic impersonation in the Rigveda. The representation of the speaker of the text is a “verbal mask”, but we can expand this to the representation of the listeners of the text too. These listeners need not be real historical actors, but representations of them. They are “verbal masks” too. Thinking this way gives us a crucial epistemological advantage. We can think about textual representations of speaking and listening without needing historical “authors” and “readers”. Surely, historical actors who used the text existed, but by conceiving of rhetorical actors rather than historical ones, we suspend our own conceptualizations of text producers and consumers. We instead allow the text to show us speaking and listening characters: the verbal disguises it would have its historical performers and audiences wear. This is more important from the perspective of these unknown historical participants because, as discussed above, the narrative mask is what confers significance onto human voices and bodies.

4. The Double Scene

While narration in lyrical texts like those of the Rigveda is not sustained in the way it is in epics like the Iliad or the Mahābhārata, yet Rigvedic narratives do exist. In hymns which take the form of lists of divine feats, for example, the narrative may be limited to a single verse, while over the body of the song an argument is constructed by the succession of narratives placed in parallel. Following Laurie Patton’s book Myth as Argument, I take these narratives and sequences of narratives as a strategy of argumentation. These arguments sometimes depend on implied similarities between seemingly unrelated phenomena. Consider Patton’s observation regarding the B$\underset{°}{r}$haddevatā of Śaunaka:

“…the juxtaposition of a grammatical rule next to a cosmogonic myth is a way of “placing,” and therefore making an argument about, both kinds of knowledge; such juxtaposition has its own kind of logic beyond the mere compiler’s whim23.”

How do we make argumentation through narration intelligible to us? As Patton says, the juxtaposition of narrative has its own kind of logic, and that logic is only fully accessible through the extra-linguistic social institution for which the material was compiled24. We do not have access to that social institution, which is the historical sacrifice, but we do have access to a level of narration embedded in the text that is about the sacrifice, that is about this institution of performance interwoven with the other narratives of the text. Examining the juxtaposition of depictions of the performance with other narratives does not tell us about the historical sacrifice, but it does tell us how the historical sacrifice was conceptualized.

The Old Norse text the Vǫluspá or “the prophecy of the seeress” has attracted scholarly attention due to elements indicating it was a performed text. The version of the text I will use is from the Codex Regius (R). Its performative dimensions were first scrutinized by Lars Lönnroth25. He coined the term dubbla scenen, or “double scene,” to capture something he observed in the Vǫluspá. Namely, that the setting of the narrative seemed to mirror or re-create the scene of its historical performance. Lönnroth argues that a vǫlva, “seeress,” addresses Óðinn, but the text is a “double scene” which imports the performance context of an historical speaker and audience located at a farm in 13th century Iceland. On this basis, (Thorvaldsen 2013) argues that this double scene may account for the deictic complexities in the Vǫluspá. Deixis is the system of reference which marks position with respect to the speech event. Because they are defined relative to the speech event, pronouns and verbs which mark the speaker (the first person) and hearer (the second person) of the speech event are inherently deictic26.

While Thorvaldsen distances himself from a fixed historical setting, he studies the way speaker perspective is represented in the Vǫluspá27, finding a speaker–listener complex which shifts between the vǫlva and Óðinn, a human performer and human audience, and a blend of the two. When Óðinn is addressed, he is marked by specific epithets, like Valfǫðr, or the singular form of the second person pronoun: þú. The following represents a vǫlva-Óðinn scene:

vilðo at ec ualf þr

uel fyr telia

forn spioll fíra

þ er fremst um man28

You wish, Valfǫdr, that I tell the past tales of men

the earliest that I can remember.

The audience at the poem’s beginning, however, is in the plural:

Hliods bið ec

allar kindir

meiri oc mi_Ni

mavgo heimdallar29

I ask all families to listen, the greater and lesser sons of Heimdall.

Not only is the poem’s hypothetical audience here explicitly human, it is inclusive of different social strata. For Thorvaldsen, hliods bið ec, “I bid you listen,” is spoken by a human performer. He offers that: “to introduce a performance by asking a crowd for attention must be an almost universal phenomenon30.” Many comparanda from the Rigveda31 corroborate his thought. Consider the following verse:

índrajyeṣṭhā márudgaṇā | dévāsaḥ p$\acute{\stackrel{\_}{u}}$ṣarātayaḥ | víśve máma śrutā hávam ||32

(You) whose chief is Indra, whose gang is the Maruts, (you) gods, whose gifts are of Pūṣan, all hear my call!

This is a common use of the imperative in the Rigveda, in which the divine audience is commanded to pay attention to the performance. Is this the same as commanding a human audience for attention?

In Thorvaldsen’s analysis, he argues that in certain parts the speaker seems to be addressing both Óðinn and a human audience simultaneously, as evidenced by:

hvers fregnit mic

hvi freistiþ min

alt ueit ec oði_N

hvar þv ga falt

ienom me̲ra

mimis bru_Ni33

What do you want to know? Why do you try me? I know everything,

Óðinn, where you hid the eye in the famous well of Mímir.

Although Óðinn is directly addressed, the second person plural verbs fregnit, “you ask,” and freistiþ, “you test,” are directed towards an audience of humans who also wish to know.

Is this a feature of the Indo-Iranian poetic tradition too? Does the Rigveda or the Gāθās34 address their respective audiences in the second person plural? Consider the following from the Avesta.

a$\underset{~}{t}$ frauuaxšiiā nū gūšōdūm nū sraotā35

Next, I will proclaim, now hear for yourselves and hear (it) now!

Just like the opening of the Vǫluspá (R 1.1-4), the poet uses second person plural verbs (gūšōdūm and sraotā) to command his audience to pay attention. Can thinking about this listening audience give us insight into verses like:

y$\stackrel{\_}{ǝ}$vå mazdā ahurā pairijasāi vohū manaŋhā36

I who wish to circumambulate you with good thought, Mazdā Ahurā

Here, the accusative plural clitic vå, “you,” does not agree with the vocative singulars mazdā and ahurā. If we propose a performative context to the Yasna like that proposed for the Vǫluspá, we might speculate that these second person plural verbs and pronouns are deictic traces, and that the singular entity to which that epithet mazdā ahurā refers may be, like Óðinn, only one member of a larger audience. Returning to ṚV 1.23.8, I see no reason why víśve, “all,” from pāda c might not resume the previous dévāsaḥ, “gods,” as well as include the humans present at the sacrifice. If so, both gods and humans present at the performance would be commanded to máma śrutā hávam “hear my call!” The Avestan Gāθās are a fertile site of comparison for the Rigveda because of their closely related languages. The human performer of the Yasna often speaks as Zarathuštra. (Skjærvø 2002) argues that when the poet asserts himself to be the “real” Zarathuštra in Y43.8, the adjective haiθiia- has ontological significance:

“the emphatic adjective “real, true” (haiθiia-, OInd. satya-), as we can see from its other occurrences in the Old Avestan texts, seems to be used to identify objects or person as “real, true” as indicated by their names, as opposed to things or persons that are just “called” something but are not “really” so. In the conceptual universe of the Old Avestan poet-sacrificer this is an important distinction, since, here, the saying “appearances deceive” which seems banal to us, takes on a truly ominous meaning37.”

These assertions of truth are the real reality of the sacrifice: invisible to normal sight but manifest through verbalization. Skjærvø describes haiθiia-, “real” or “true,” as an emphatic adjective used to assert something to be true. This reality is not self-evident; it must be asserted. Perhaps the adjective haiθiia- insists that this is no mere human performer but the figure of Zarathuštra in the same way that the Yajurvedin can verbally disguise his human arms and hands by asserting that he has the arms of the Aśvins and the hands of Pūṣan.

What we can take away from Lönnroth’s double scene is that when performed texts construct their narrative scene, they often superimpose that scene upon the actual historical performance event, just like the verbal mask discussed in the previous section. These two scenes cannot be totally quarantined from one another, and traces of the performance trickle into the narrative that lays on top of it. One epistemological challenge here is that we do not have access to that historical performance, only these trace references to the speakers and listeners that populate it. Another challenge, an ontological one, is that these shadowy figures are certainly not real historical speakers or listeners. The only citizens of the text are rhetorical beings.

5. Para-Narration

To theorize what exactly is this level of narration, which seems to exist only to layer with additional narratives, I want to discuss a notion called para-narration employed by Luz Aurora Pimentel. In her book Metaphoric Narration: Paranarrative Dimensions in À la recherche du temps perdu, Pimentel treats the baignoire scene in Le côté de Guermantes38. In this scene, the narrator goes to the opera, but the narrator’s perceptions of the opera-hall are a blend of details reminiscent of a real opera hall as well as a fantastic watery domain replete with nereids and sea monsters. This conceited metaphor is, for Pimentel, a virtual space which is superimposed on the main narrative space39. Pimentel (1990, p. 155) argues that in the baignoire scene “the main diegetic space, the theatre, is almost obliterated as the metaphoric marine world of nereid and tritons gradually takes over.”

Pimentel qualifies what happens to the main narrative as “almost obliterated” and “gradually take[n] over”. That is, the narrative of an opera-hall and an undersea realm really co-exist, they blend together, repairing the breach in coherence introduced by the extended metaphor. For example, those the narrator identifies as nereids are marked by behaviors appropriate to the ladies of the opera. Proust’s choice to homologize an opera hall to an undersea kingdom seems quite arbitrary, but Pimentel notes that the two narratives are anchored by a play on words: the term baignoire itself. The unmarked meaning of the word baignoire is a bathtub, but in Proust’s time, in early 20th century France, it also referred to the lowest tier of the theatre. Thus, the germ of this metaphoric elaboration is wordplay, and the coherence of the individual metaphors is mediated by its double meaning.

Another example given by Pimentel is from La noche boca arriba by Julio Cortázar. The protagonist of the story is in a motorcycle accident and is rushed to the hospital. In his pain, he begins to dip in and out of fevered dreams. He perceives the hospital less and less. In his dream, he is fleeing the Aztecs through swamp and jungle. The perceptions of the protagonist systematically correlate characters, instruments, and actions; this allows the two separate narrative universes to be mutually intelligible. We learn, for example, of the odor of the hospital through his perceptions of the reek of the swamp. There is no doubt that he sees a surgeon before him in this passage:

…cuando abrió los ojos vio la figura ensangrentada del sacrificador que venía hacia él con el cuchillo de piedra en la mano

…when he opened his eyes he saw the bloody figure of the sacrificer that came toward him with the stone knife in his hand40.

Finally, the protagonist realizes that it was the hospital, the motorcycle accident, that entire world which was the dream. He has now returned to true reality. The reader, however, understands the implication; he has died on the operating table.

Pimentel’s notion of a para-narrative interests me because the concept is essentially an attempt to theorize the reader’s awareness of the relationship between two narrative theatres. She also uses the perceiving character as something of an embedded model reader, who, like the actual reader has access to both worlds and understands the relevance of the narrative levels and the patterned, sustained, and repeated uses of metaphor. A future reader can appropriate the understanding of that perceiving character as a guide since it co-exists alongside the text. It is this conceptualization of para-narrative which I think is applicable to performed oral texts. The two narrative worlds which I will examine are the worlds of myth and legend and that world which is nothing other than the text’s patterned, sustained, and repeated references to its own performance.

In the case of Proust’s baignoire scene, what is the main narrative (the opera house) and what is the para-narrative (the undersea realm) is quite clear. Pimentel offers a compelling way to think about levels of narration, particularly when one level is conceived of in terms of another. The term para-narration should not be applied to Rigvedic performed texts, however, because one level cannot be subordinated to another without imposing the modern distinction of fiction and non-fiction onto the text. Nevertheless, thinking of the Rigveda as comprising homologous levels of interpenetrated narration may be the key to understanding how cosmological narratives make tacit claims about the performance and the performers. It certainly provides us a preliminary etic ontology for the “verbal mask” and the “double scene”, pending an emic one, which would arise from the very research this article is proposing.

6. Textual Self-Reference

If the performance can be conceived of as one narrative level among many in the Rigvedic sūkta, what are the formal features that define this level? In the following sections, I will examine a number of ways the text refers to its adhiyajña level. The strategies presented here are meant to be thorough, but not necessarily exhaustive. This thought experiment is open ended. Tentatively, we can divide these references into explicit self-reference and implicit self-reference. An example of explicit self-reference is “may this song be heard.” Self-reference of this type necessarily breaks away from narratives about the primordial past to fix the textual eye on the present at the very moment in which the song is singing about itself. Textual self-reference often takes the form of wishing for the success of the song. The subject often appears in the plural, and a wish is made in the optative mood.

en$\acute{\stackrel{\_}{a}}$ṅgūṣéṇa vayám índravanto | abhí ṣyāma v$\underset{°}{r}$jáne sárvavīrāḥ ||41

Through this hymn (āṅgūṣa), Indra in our company, may we, all-heroes, be élite in our community.

The pronoun ená, “by this one.” suggests that the song that will make the speakers pre-eminent is none other than this very song (ṚV 1.105). So, the first thing we know about the performance context is that this act of singing is located at the performance. It is important not to trivialize that fact, for if the song conceives of itself as being sung at a performance occasion, and if it can talk about that performance occasion by self-reference and expressions of proximity, then there really is a thin story being told about this song being successfully performed at a competitive social event. That story frames the contents of the rest of the song. That performance narration, then, accounts for the text’s expectation that its audience is located at the performance too, and that its audience understands why a particular text is germane to the event going on at that location. In other words, it is very similar to the expectation that the author of a written text has that the readers can grasp patterned, sustained, and repeated metaphors.

7. Narrative Blending

Since the song, from its own perspective, is always being sung by the performer, traces of this event and this location has the potential to spill into myths and legends. An excellent example of this is ṚV 10.10, which is a dialogue between Yama and Yamī, the first human pair. Each verse of the hymn alternates who is speaking. Although they are brother and sister, Yamī insists that Yama impregnate her in order to create the human race. She claims it to be the will of the gods, but Yama is recalcitrant—he believes it is anathema to the gods’ will. Yama says:

ná yát pur$\acute{\stackrel{\_}{a}}$cak$\underset{°}{r}$m$\acute{\stackrel{\_}{a}}$kád dha nūnám | $\underset{°}{r}$t$\acute{\stackrel{\_}{a}}$ vádanto án$\underset{°}{r}$taṃ rapema ||42

(While) uttering truths, we would whisper something false, (something) which we have never before done; so now what?

In other parts of this dialogue, Yama and Yamī use forms of the grammatical dual, yet here the verbs cak$\underset{°}{r}$m$\acute{\stackrel{\_}{a}}$ and rapema and the present participle vádanto are all grammatically plural. Yet the conversation is set before the existence of humanity, so there should be no humans present other than Yama and Yamī. Why, then, the plurals instead of the dual? I believe the answer is found later in this very hymn. In the fifth verse, Yamī responds to Yama’s claim that they whisper unprecedented things by giving a proper mythological precedent. Specifically, she says that they were created to be a domestic pair just like Earth and Heaven. Yama’s response in verse six mocks her reasoning:

kó asyá veda prathamásya áhnaḥ | ká īṃ dadarśa ká ihá prá vocat |

b$\underset{°}{r}$hán mitrásya váruṇasya dh$\acute{\stackrel{\_}{a}}$ma | kád u brava āhano v$\acute{\stackrel{\_}{ı}}$ciyā n$\underset{°}{\acute{\stackrel{\_}{r}}}$n ||43

Who knows of the first day? Who has seen it? Who will proclaim it here?

Since the domain of Mitra and of Varuṇa is high, what, O floozy,

will you perversely tell the men?

By saying “Who knows the first day? Who has seen it?”, Yama critiques the validity of her knowledge of the primordial precedents. Far more interesting is ká ihá prá vocat “who proclaims it here?” Where is this ihá “here”? The colligation pra + √vac is typically used to describe the act of public performance of song, most famously índrasya nú vīríyāṇi prá vocaṃ44 “I proclaim forth the manly deeds of Indra45.” We might tentatively imagine that, especially in preliterate societies, public memory is conditioned by normative claims made in authorized public performance.46 Yama thus extends his criticism by asking who here, at this present performance, will perform the knowledge of the first day. Note that this too is explicit textual self-reference, since Yama is referring to Yamī’s verses as unsuitable speech. Presumably, Yama mentions the height of the domain of Mitra and Varuṇa because it is in heaven and thus so far away that the gods might not hear the untruths Yamī is telling. This confirms that the scene is terrestrial. Everyone iha, here on Earth at the present performance, however, can hear. So, Yama asks Yamī what falsehoods she will tell the men. Yama and Yamī have stepped through the narrative barrier from their past setting, where they are the only two humans, into the present where an audience of listening men is gathered. It is this audience which I believe accounts for the use of plurals cak$\underset{°}{r}$m$\acute{\stackrel{\_}{a}}$, rapema, and vádanto in ṚV 10.10.4.

8. Deixis as a Marker of Implicit Self-Reference

Beyond explicit textual self-reference, what are other formal markers of the present performance present? We might think of deixis in the terms laid out by Karl Bühler:

“Thus, three axes of reference must be placed at the origin, if this schema is to represent reference in human speech: specifically, the referential axes of here, now, and I47.”

Deixis is essentially a coordinate grid of speech. Because words like “here”, “now”, and “I” define the spatial, temporal, and personal location of the speech act, they can be characterized as forms of proximal deixis. It is just these three axes of space, time, and personal perception that will mark our adhiyajña level of narration because they represent objects, events, and experiences as occurring at the same scene as the speech act.

9. Reported Perception and Text-Deixis

Speaker perceptions and experiential states are marked as belonging to the frame narrative of present performance because that information is normally private and inaccessible except through acts of reporting by the speaker in the present. Consider the following verse in which the speaker reports on his perception:

utéva me váruṇaś chantsi arvan |48

…and appear to me, O racehorse, like Varuṇa!

This verse appears in a hymn dedicated to the sacrificial horse and is part of a mythological narrative about the origin of the horse. We would expect the first person to be the locus of experience, but the point here is that the search must be expanded to verbs in which internal experience is the result of reporting external stimuli anywhere in the speaker’s sensorium. These stimuli may be marked by the second person, like chantsi49. Reports of perception may not be marked by a finite verb at all. In such cases, we must evaluate any narrative assertion as a potential reported perception of the speaker on a case by case basis.

Let us consider a form of deixis which is neither explicitly spatial nor temporal but is better termed text-deixis. That is, a reference to something already said, as in “that is terrible!” in which “that” refers to the speech act to which it is responding. Referring to a previously discussed topic depends on both speaker and hearer knowledge of that previous discussion, and that dependence of shared knowledge belies a dependence on a shared experience of the prior speech act. Text-deixis, in the context of performance at least, operates just like a reported perception.

Kupfer argues that the pronoun etád is text-deictic50 and functions either in a contrastive51 or topicalizing52 capacity. Both these functions are types of text-deixis and rely on shared perceptions of the text between speaker and hearer. Consider ṚV 7.19.10a eté stómā nar$\acute{\stackrel{\_}{a}}$ṃ n$\underset{°}{r}$tama túbhyam “these praise-songs of the men, O manliest one, are for you.” Here, the praise songs (stómā) of the men (nar$\acute{\stackrel{\_}{a}}$ṃ) are characterized by eté, “these,” a text-deictic pronoun. This pronoun connects the praise songs of the men to the poems (ukth$\acute{\stackrel{\_}{a}}$) which men (náraḥ) are announcing in the previous verse: ṚV 7.19.9b náraḥ śaṃsanti ukthaś$\acute{\stackrel{\_}{a}}$sa ukth$\acute{\stackrel{\_}{a}}$ “the men, as announcers of poems, announce the poems.” As a text-deictic pronoun, the etád pronominal paradigm is formally neutral in terms of spatial and temporal deixis, yet it acquires deictic value contextually, by being construed with something independently established as having deictic value. In this case, the present-time reference of śaṃsanti “they announce,” is extended through text-deictic eté to the praise songs (stómā) in the following verse marking them as either being sung in the present or the immediate past if the singing has just finished.

10. Temporal Deixis

The next example comes from a dialogue set in the mythological past in which Saramā speaks to the Paṇis, telling them Indra and the Aṅgirases are coming for the cows. She reports that:

n$\acute{\stackrel{\_}{a}}$háṃ veda bhrāt$\underset{°}{r}$tváṃ nó svas$\underset{°}{r}$tvám | índro vidur áṅgirasaś ca ghor$\acute{\stackrel{\_}{a}}$ḥ |

gókāmā me achadayan yád $\acute{\stackrel{\_}{a}}$yam | áp$\acute{\stackrel{\_}{a}}$ta ita paṇayo várīyaḥ ||53

I know about neither brotherhood nor sisterhood. Indra and the dread Aṅgirases know.

When I came (from there), they seemed to me desirous of cattle.

Go away from here, Paṇis, to somewhere wider!

The human performer impersonating the divine Saramā reports her experience of how Indra and the Aṅgirases appeared. Although the verb achadayan, “seemed,” is marked past tense, the performative act of reporting is happening at the present moment. Notice, too, that Saramā, outside of her reported perception, directs the Paṇis to áp$\acute{\stackrel{\_}{a}}$ta ita, “go away from here54,” locating the scene of the narrative in the same place as the singing of the song itself. The imperative ita like chantsi locates the narrative in the present moment, a timeframe which is temporally proximal to the speaker.

I want to explore a few different ways that the verb can indicate temporal proximity to the speaker. Present indicatives, by virtue of announcing what is happening, are located in the present moment. Imperatives, by virtue of commanding someone to do something which is not yet done, locate the command in the present regardless of stem aspect. Dahl argues against a progressive or imperfective aspect for the present stem and instead for a neutral aspect55. This is a reasonable inference, Dahl claims56, as the present stem is used for performative sentences. Consider the following:

p$\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi tvā páram ántam p$\underset{°}{r}$thivy$\acute{\stackrel{\_}{a}}$ḥ | $\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi yátra bhúvanasya n$\acute{\stackrel{\_}{a}}$bhiḥ |

p$\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi tvā v$\underset{°}{\overset{´}{r}}$ṣṇo áśvasya rétaḥ | p$\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi vācáḥ paramáṃ víoma ||57

I ask you about the far end of the earth, I ask where existence’s navel is

I ask you about the seed of the stallion, I ask you the utmost heaven of Speech.

Like “I promise”, p$\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi “I ask” is a performative utterance.58 Here, it allows the speaker to pose a question without using interrogatives at all, by simply declaring that one is asking. This particular verse has been studied by George Thompson as a brahmodya, “(something) to be uttered by a priest,” which is a kind of ritualized riddle59. The performer is not really asking in order to learn the answer. He knows the answer. In fact, he solves the riddle in the next verse.

iyáṃ védiḥ páro ántaḥ p$\underset{°}{r}$thivy$\acute{\stackrel{\_}{a}}$ | ayáṃ yajñó · bhúvanasya n$\acute{\stackrel{\_}{a}}$bhiḥ |

ayáṃ sómo v$\underset{°}{\overset{´}{r}}$ṣṇo áśvasya réto | brahm$\acute{\stackrel{\_}{a}}$yáṃ vācáḥ paramáṃ víoma ||60

This altar is the far end of the Earth, this sacrifice is existence’s navel.

This Soma is the seed of the stallion, this composition is Speech’s utmost heaven.

This is dramatic irony. The asking feigns ignorance, and the ignorance is performative. Lest we limit performatives to verbs describing speech acts, Dahl also cites ṚV 1.171.1ab: práti va en$\acute{\stackrel{\_}{a}}$ námasāhám emi | sūkténa bhikṣe sumatíṃ tur$\acute{\stackrel{\_}{a}}$ṇām “I go to you with this reverence; with this well-spoken (text); I beg the good will of the mighty.” Here, both active voice emi, “I go,” and middle voice bhikṣe, “I beg,” are first person presents that operate as performatives just like p$\underset{°}{r}$ch$\acute{\stackrel{\_}{a}}$mi, each enacting the very event they describe.

In addition to imperatives and present indicatives, Dahl argues that “the Aorist Indicative in some cases seems to be used as the head of performative sentences61. As an example of such a sentence, he cites ṚV 2.35.1a úpem as$\underset{°}{r}$kṣi vājayúr vacasy$\acute{\stackrel{\_}{a}}$ṃ “Desiring the prize, I release it: my verbal skill.” The verb here is an aorist indicative as$\underset{°}{r}$kṣi. Although the augment marks the verb as being in the past, the aspect of the aorist is perfect62. This perfect aspect indicates that the action has just been brought to completion, and perhaps this explains why the aorist can be used as a performative in much the same way as the present stem. This coheres with the observation of Jamison and Brereton that the aorist is:

“…often used to express the immediate past (in English, “has [just] done” vs. “did”) and is therefore frequently encountered in ritual situations, in which the poet announces a sacrificial act as just completed (like the kindling of the fire) or a poem just composed63.”

Hoffmann, in his ground-breaking work on the subject, Der Injunktiv im Veda, studies a number of so-called aorist injunctives, which are aorists not marked by the augment. I say so-called because, these forms often do not enjoin anything. So-called “injunctives” are finite verbs with secondary endings that lack the augment, and thus are like other finite verbs except that they are not inflected for tense or modality64. We use the misnomer “injunctive” because in some cases the mood is provided by the context, and some of those contexts are injunctive. For example, the particle m$\acute{\stackrel{\_}{a}}$ supplies the “injunctive” verb with prohibitive modality. Dahl65 points out that Hoffman’s aorist injunctives vocam and gāsi66 are performative just like an augmented aorist would be. Indeed, given the potential for ritual performativity which seems fertile in the aorist’s perfect aspect, we might expect the augmentless aorist to surface as a performative verb par excellence. The following verse is a case of such a verb:

ádhā nú asya saṃd$\underset{°}{\overset{´}{r}}$śaṃ jaganv$\acute{\stackrel{\_}{a}}$n | agnér ánīkaṃ váruṇasya maṃsi |67

Then having gone to the sight of him, I realize Agni’s face (as the face) of Varuṇa.

The first-person singular s-aorist maṃsi lacks an augment, leaving the time of this event ambiguous. Is this event set in the past, when the seer first caught sight him? Or does this thought happen whenever he takes sight of him? This, I believe, is a controlled use of ambiguity, prohibiting an audience from restricting the verbal event to the past. Dahl (2010, p. 117) notes that:

“Being radically underspecified with regard to tense and modality, the Injunctive may be hypothesized to pick up its temporal and modal interpretation from the immediately surrounding context and to be assigned a default tense and mood value, probably present tense (t0⊆t’) and neutral/indicative mood, unless otherwise specified by the context.”

As the performer of this hymn claims to be the great seer Vasiṣṭha68, the use of the augmentless aorist here may be a way of effecting an impersonation of the legendary figure, making him speak at the present sacrifice, re-performing his moment of realization.

Occasionally, what appears to be past value in an augmentless form isolated in its verse may benefit from considering the larger rhetorical structures of the sūkta. For example, ṚV 2.11.2a s$\underset{°}{r}$jó mah$\acute{\stackrel{\_}{ı}}$r indra y$\acute{\stackrel{\_}{a}}$ ápinvaḥ | could be taken to mean “Indra, you released the mighty (waters) which you swelled,” with both verbs having past tense value, even though only ápinvaḥ is marked with the augment. This seems reasonable because these events occurred in the mythological past, when Indra slew V$\underset{°}{\mathrm{r}}$tra. Let us now consider the previous verse in this sūkta. ṚV 2.11.2ab s$\underset{°}{r}$jó mah$\acute{\stackrel{\_}{ı}}$r indra y$\acute{\stackrel{\_}{a}}$ ápinvaḥ | páriṣṭhitā áhinā śūra pūrv$\acute{\stackrel{\_}{ı}}$ḥ | follows on the heels of ṚV 2.11.1cd im$\acute{\stackrel{\_}{a}}$ hí tv$\acute{\stackrel{\_}{a}}$m $\acute{\stackrel{\_}{u}}$rjo vardháyanti | vasūyávaḥ síndhavo ná kṣárantaḥ | “For these juices are increasing you, seeking wealth like flowing streams.” Thus, the text introduces the primordial waters that Indra released not randomly but following a simile in which the sacrificial oblations that strengthen Indra are likened to rivers. Ergo, it might be better to translate s$\underset{°}{r}$jó mah$\acute{\stackrel{\_}{ı}}$r indra y$\acute{\stackrel{\_}{a}}$ ápinvaḥ | páriṣṭhitā áhinā śūra pūrv$\acute{\stackrel{\_}{ı}}$ḥ | as “you release [present value] the great (streams) Indra which you swelled [past value], hero, the ancient (waters) stopped by the snake.” This would continue the comparison of the oblations to the primordial waters. Indra is releasing them not in the past but now in the present, just as the oblations are being poured in the rhetorical “now” of the present ritual. A reading of only past value removes context from an argument being made over more than one verse. Isolated pāda analyses risk presupposing that Indra’s mythological activities must always occur in the past, which may seem reasonable but is not the only way mythological narratives are presented in the texts. These texts represent themselves as committed to a sacrificial calendar. Indra’s release of the waters is cosmogonic, yet it is also seasonal. The waters’ release re-occurs each summer when the melting of mountain snow put the rivers of the Hindu Kush in spate.

That augmentless forms of the aorist have present value or simply no value is thus very important, for, as we noted earlier, presents are one of the chief sources of performative verbs. In addition to augmentless aorists, augmentless imperfects may have the potential to be performative too. Dahl argues that:

“the Early Vedic Imperfect has a general past time reference, but that it is not found in immediate past contexts. Moreover, it was argued that the Imperfect is compatible with a completive-sequential as well as a progressive-processual reading and that it is mostly used to denote a single, specific past situation but can also, to some extent at least, be used with an iterative-habitual reading69.”

In other words, the imperfect is an aspect-neutral preterit which has the same scope as the present indicative except it is limited to the past. Since it has non-immediate past time reference, it not an ideal candidate for a performative utterance like the aorist indicative, present indicative, or imperative.70 Stripped of its augment, however, the imperfect is no longer restricted to the non-immediate past and gains all the performative possibilities of the present indicative. This may be another strategy which allows narration about the mythological events can be presented as occurring at the present performance. If so, that would indicate a very strategic use of semantic ambiguity by the composers of the Rigvedic sūktas; further, it may account for the abundance of augmentless forms in the Rigveda.71

Of course, there are ways other than finite verbs and adverbs to mark a sentence for present time value. Consider this verse:

gr$\acute{\stackrel{\_}{a}}$vṇo brahm$\acute{\stackrel{\_}{a}}$yuyujānáḥ saparyán | kīríṇā dev$\acute{\stackrel{\_}{a}}$n námasopaśíkṣan |

átriḥ s$\acute{\stackrel{\_}{u}}$ryasya diví cákṣur$\acute{\stackrel{\_}{a}}$dhāt | súvarbhānor ápa māy$\acute{\stackrel{\_}{a}}$ aghukṣat ||72

The composer having yoked the stones (is) worshipping, with mere reverence seeking the gods. Atri set the eye of the Sun in heaven and banished the powers of Svarbhanu.

Here we see two diptychs in juxtaposition: two actions presented in parallel. The second diptych is marked by the aorist indicative and has a past reference to mythological content, while the first diptych, referring to the ritual performance, lacks a finite verb. Following Patton’s premise that juxtaposition itself is a strategy of argumentation, I argue that this juxtaposition may be presenting two actions in parallel in order to indicate they are connected if not analogous.

In the second diptych, we have a self-contained narrative. Atri, an ancient seer, set the eye of the Sun in heaven and banished the magical powers of the malevolent Svarbhanu. In the first diptych, the brahmán,73 the speaker of sacred speech, has yoked the stones, which means he has made Soma, the stimulating drink which is at the center of Rigvedic ritual. He is doing ritual performance too. Read as a nominative absolute, the two seem utterly disconnected. One half of the verse concerns a priestly figure who makes Soma and does a ritual, and the other half concerns a mythological figure who puts the eye of the Sun in the sky. Because first diptych has no finite verb or copula at all, it thus is temporally ambiguous clause, and thereby achieves much the same effect as a verb which lacks the augment. My hypothesis is that the two are being fundamentally equated, that when a brahmán-priest does the ritual this essentially identical to when Atri set the eye of the Sun in the sky. If so, the present ritual actions are being depicted as re-performances of events of cosmic significance. This constitutes another way in which mythological narratives can be drawn into the narrative frame of the present performance.

In the second diptych, $\acute{\stackrel{\_}{a}}$dhāt may be an attractive candidate for a performative verb. The aorist $\acute{\stackrel{\_}{a}}$dhāt may or may not bear an augment; its phonetic realization is erased by the preverb $\acute{\stackrel{\_}{a}}$. In cases like this, we must resort to the notion of audience perception. Specifically, that only an unambiguous augment can mark a verb as having an unambiguous past reference. It is worth mentioning that among the deictic adverbial particles in the Rigveda, $\acute{\stackrel{\_}{a}}$ deserves special attention going forward. As a free adverb, it marks direction towards the speaker, and thus directs the listener to the here and now. Beware: this is not always the case. For example, when $\acute{\stackrel{\_}{a}}$ is soldered onto a verb stem as a prefix it may not have this deictic value. Another example, when $\acute{\stackrel{\_}{a}}$ follows an ablative, the value of the ablative is no longer “from” but “up to.” Thus, $\acute{\stackrel{\_}{a}}$ as marker of deixis must be considered on a case by case basis.

11. Spatial Deixis

We have seen instances already where spatial proximity to the speaker is marked by adverbs (like iha and atra). There are many instances, however, where proximal spatial deixis is achieved through use of a deictic pronoun. Proximal deictic pronouns typically refer to objects located on the sacrificial ground because they are near the speaker, like the fire altar, while distal deictic pronouns refer to heavenly phenomena, like the Sun, because they are conceived of as far away from the speaker. Consider:

úd asaú s$\acute{\stackrel{\_}{u}}$riyo agād | úd ayám māmakó bhágaḥ |74

Up yon Sun went; up (went) this little lot of mine.

The Sun is qualified with distal deictic asaú. The speaker’s good fortune is depicted as near the speaker with the proximal pronoun ayám. Its close connection to the speaker is emphasized with the first person demonstrative adjective māmaka- “my little.” Regarding asaú, Kupfer explains that an “overwhelming number of attestations of the demonstrative pronouns adás are in support of its distal deictic usage75.” Kupfer offers ṚV 8.91.2 as one example of this overwhelming number:

asaú yá éṣi vīrakó | g$\underset{°}{r}$háṃ-g$\underset{°}{r}$haṃ vic$\acute{\stackrel{\_}{a}}$kaśad | imáṃ jámbhasutam piba |

dhān$\acute{\stackrel{\_}{a}}$vantaṃ karambhíṇam | apūpávantam ukthínam ||76

You over there, the little hero who peeks is coming to house after house;

Drink this pressed-by-jaws, served with grain, gruel, cake, and recitation.

For Kupfer, g$\underset{°}{r}$háṃ-g$\underset{°}{r}$haṃ, “house after house,” is sufficient proof of distance from the speaker. Notice that there are other possible markers of distance here. The proximal imám once again sets up an opposition between the local sacrifice and Indra’s other options, represented by g$\underset{°}{r}$háṃ-g$\underset{°}{r}$haṃ. It is not possible to determine absolutely if éṣi bears the directional $\acute{\stackrel{\_}{a}}$ preverb, but that is a reasonable translation in light of the accent on vic$\acute{\stackrel{\_}{a}}$kaśad, which suggests that it is the main verb of the dependent clause set off by yá. The main clause then should be asaú…eṣi: the verb éṣi would not receive an accent from its location in a subjoined clause77, which suggests that its accent is due to something else. The preverb $\acute{\stackrel{\_}{a}}$, “hither,” is an attractive candidate78.

Kupfer notes that pronouns like ayám and idám explicitly mark proximal spatial deixis, but only when they bear the accent79. Without that accent, they are anaphoric pronouns80. Kupfer cites ṚV 7.74.1, which has two such pronouns, as evidence of the proximal value of the accented pronoun:

im$\acute{\stackrel{\_}{a}}$u vāṃ díviṣṭaya | usr$\acute{\stackrel{\_}{a}}$ havante aśvinā | ayáṃ vām ahve ávase śacīvasū |

víśaṃ-viśaṃ hí gáchathaḥ ||81

These day-rites, Aśvins, (are) heifers calling to you two. As this one here, I have called to you two for help, you two whose goods are powers, so that you will go to clan after clan.

The sequence of time suggests that the performer has just completed the day-rites which are now calling to the Aśvins. The poetic conceit is either personification (rituals can “call” like ritualists can) or metonymy (the rituals really indicate the ritualists). Either device implies a connection between the day-rites and the performer, but im$\acute{\stackrel{\_}{a}}$, “these ones,” formally expresses their proximity to him. In the following diptych, ayáṃ “this one” takes this one step further, as it is the only potential subject for finite verb ahve “I called;” the proximal pronoun ayáṃ is functioning as an alternative to the first-person pronoun ahám.

Let us consider another use of repeated proximal deixis which shifts the setting of a narrative out of myth and into the domain of the song’s performance. ṚV 10.135 opens with this verse:

yásmin v$\underset{°}{r}$kṣé supalāśé | devaíḥ sampíbate yamáḥ | átrā no viśpátiḥ pit$\acute{\stackrel{\_}{a}}$ | purāṇ$\acute{\stackrel{\_}{a}}$ ánu venati |82

Under which tree of good leaf Yama drinks together with the gods,

our father, the clan-master, seeks the ancestors there.

The establishing shot is Yama’s world, the heaven of the ancestors, where he holds symposium under a special tree as lord of the dead. The first verse of this hymn introduces a tension: that the final destination of our dearly departed is unknown and his destiny is in peril. The final verse of this hymn resolves that tension.

idáṃ yamásya s$\acute{\stackrel{\_}{a}}$danaṃ | devamānáṃ yád ucyáte |

iyám asya dhamyate nāḷ$\acute{\stackrel{\_}{ı}}$r | ayáṃ gīrbhíḥ páriṣk$\underset{°}{r}$taḥ ||83

This (is) Yama’s seat; “the house of the gods” is what it is called.

This, his (wind)pipe, is being blown; this one here is perfected by songs.

The ambiguous location of the narrative of Yama’s symposium is now returned to the present with this triplet of proximal deictic pronouns: idáṃ… s$\acute{\stackrel{\_}{a}}$danaṃ, “this… seat,” iyám… nāḷ$\acute{\stackrel{\_}{ı}}$r “this… (wind)pipe84,” and ayám “this one (here).” Like ṚV 7.74.1, the speaker is using ayám to refer to himself, revealing that he is Yama. This epiphany of Yama makes Yama present at the performance, allowing him to speak to the audience directly and resolve the anxiety that Yama’s realm is far away.

12. Concluding Remarks

This method of theorizing an adhiyajña level of narration, which is nothing other than the rhetorical anticipation of speaker, audience, and social occasion built into the texts, is one that I hope will give us much greater insight into what these texts are and what they are trying to accomplish. I have highlighted a range of individual strategies to determine how the text constructs its performance context; these strategies include performative speech acts, personal pronouns, explicit and implicit textual self-reference, and especially deixis, which marks spatial, temporal, and experiential proximity to the speaker. When a poet performs a mythological narrative and refers to the present performance, he is making a claim about the significance of that present performance. He is verbalizing that something is true, even if it is otherwise invisible to the audience. This takes its smallest form in KaṭhS 1.2, which asserts “You do I take with the hands of Pūṣan, with the arms of the Aśvins,” but it allows us to think about mythological narrative in new ways, not as something that exists as an independent cultural monolith, but as something that is evoked to add deep religious significance to performance events, such as a singing a song, lighting a fire, and cooking a meal, that look quite mundane to eyes that see only what is silently visible but are blind to the invisible realities made manifest by speech.