Students of ancient religions read their texts as “sources” in almost the same way as anthropologists approach their material. This is a relationship different from that of philosophers, theologians or literary critics to their texts. Literary critics would not call the novel or poem they read and interpret a source or evidence. Whether philosophers embrace the “truth” of the text they comment on, they would consider it as part of their task to articulate it in terms they consider more adequate. Anthropologists, on the other hand, always remain external to their object, even when they learn from it. The positivism of “source” is by and large out of place in philosophy or literary criticism, whereas it constitutes the basis of the academic study of religion. This is the reason why it is normal to talk of hypothesis or comparative method in anthropology or mythology but not in philosophy. Such an external relation of the mind to its object is susceptible to becoming manipulative, almost unconsciously. One can see why. The theory must be able to accommodate the relevant sources; that is all. Naturally, one's allegiance lies with one's own theory. Thus, the sources are made relevant and affirmative of the theory. The more lacunary and abstruse the sources, the more readily they may be “interpreted” in the light of the theory. We do not want to give up asking comprehensive questions like that of the meaning of a ritual. This increases even further our reliance on theoretical schemas. It is thus imperative to be cautious in our treatment of the sources. Even more importantly, we must ensure the pertinence and probity of our inferences, which perforce bear the weight of our demonstration.
Yasna: a complete, coherent text
The theories I examine here are about the coherence of Yasna, the basic sacerdotal ceremony of Zoroastrianism. They attempt to answer the question whether the segments of Yasna follow each other according to a discernible logic. Does Yasna have a describable structure? From the start, however, this question gets tangled up in their accounts with that of the antiquity of Yasna. Coherence and antiquity of a text are separate matters, even if one must allow that the coherence of an oral composition is different from literary coherence, and that the question of coherence cannot be settled without restoring the text to its authentic historical setting. Then the recovery of the context of its formation, its Sitz im Leben, would be a necessary preliminary step for dealing with the question of its coherence; it may even point the theorist in the right direction, but it cannot answer that question. The relation between these two issues remains unclarified by the theorists we are considering, and causes confusion in their arguments. The “recitative” of the “long liturgy” is not a patchwork of fragments from the Sasanid Avesta but a “coherent text” that was formed in the context of ritual practice and assumed its definitive shape when Avestan was still a productive language – at the Achaemenid time at the latest. In what manner does the “antiquity” of Yasna prove its “coherence”? Comparison with Vedic soma ritual is meant to link the two issues. If it can be shown there is a “structural correspondence” between Yasna and Vedic soma ritual, both questions can be settled at once. I question the validity of this approach below.
The priestly Zoroastrian ceremony that Cantera, following Kellens (Reference Kellens1998: passim), calls the “long liturgy” consists of a “complete, coherent” text recited in the frame of a ritual whose basic design continues the Indo-Iranian ritual tradition. Cantera maintains that these two aspects mutually reinforce each other in the context of an ongoing reflective ritual practice, thanks to which changes in the ritual constituents have been fully integrated. He has formulated this thesis in his recent publications. The more detailed formulations are as follows:
The composition and arrangement of the Yasna is not a late, artificial process, but rather the result of a long process of continuous exegesis of older texts used for similar liturgies and for reflecting upon the nature of the sacrifice (Cantera Reference Cantera2016a: 144).
Sans nier l’évidence que le récitatif actuel est une compilation de textes différents et composés à des époques diverses, on doit signaler que le récitatif de la liturgie longue de même que la performance rituelle de ce récitatif (comment pourrait-on les séparer?) ont déjà acquis leur forme définitive à une époque où on était encore capable de rédiger des textes avestiques … je crois qu'on doit considérer le récitatif de la liturgie longue comme une unité textuelle (avec, bien sûr, une longue histoire derrière elle) créée avec des éléments différents dans le cadre d'un processus de production orale de littérature rituelle en langue avestique. La liturgie longue est une production de littérature rituelle orale et non une composition savante de textes hétérogènes (Cantera Reference Cantera2014: 211).
L'importance de cette découverte n'a été notée que peu à peu, mais elle a des conséquences énormes pour la compréhension des textes avestiques récités dans ces rituels: l'arrangement des textes dans la liturgie n'est pas tardif et secondaire, mais il est bien antérieur à l’époque sassanide. Le texte récité dans la liturgie longue est amalgame des textes rédigés à des époques différentes et peut-être aussi dans des endroits différents, mais ils sont arrangés consciemment pour construire un “texte” (Cantera Reference Cantera2014: 13).
The background to these affirmations is the conception of Zoroastrian liturgy they aim to replace, that is to say, the conception that the texts of the priestly ceremonies are the “residues” or “fragments” of the Sasanid Avesta (Cantera Reference Cantera2014: 313–29).Footnote 1 The textual corpus that Cantera calls the “long liturgy” was constituted in ritual practice when Avestan was still productive.Footnote 2
In fact, during the productive time of the Avestan language there existed not only the conceptual spine of the ceremony, but also its recitative had already taken the form we know today. This recitative is not a late patchwork of fragments of the Great Avesta compiled in Sasanian or post-Sasanian times, but in fact goes back to the productive time of the Avestan language, probably still in the Achaemenid period (Cantera Reference Cantera, Williams, Stewart and Hintze2016b: 62).Footnote 3
In 1998 Kellens reformulated Spiegel's idea that Yasna and its expanded version (Visperad) existed independently of, and much earlier than, the Late Antique collection of Avestan texts (probably under Khosrow I) that is called the Sasanid or Great Avesta in the literature (Kellens Reference Kellens1998). By way of synchronization with the introduction of the Zoroastrian calendar, whose tutelary gods of the days are listed in the right order in Y 16 and seem to constitute the Young Avestan pantheon, Kellens argued that the Yasna corpus goes back to the Achaemenid period. Cantera finds further textual evidence for the antiquity of the long liturgy, namely Avestan references in the Nērangestān to passages in Yasna, and those that parallel ritual instructions found in the Sāde manuscripts (see Cantera Reference Cantera, Williams, Stewart and Hintze2016b: 63–6). If the current arrangement of Yasna texts is from a time when priests still produced texts in Avestan, it must be the work of practising priests, and hence one can and indeed must assume an intention behind it. It is a product of design and not chance. In his compte rendu of Kellens's five-volume study of the long liturgy (Études avestiques et mazdéennes), Cantera approvingly cites Kellens's programmatic statement from the back cover of the fourth volume:
En rédigeant le deuxième volume (2007), j'ai pris conscience de la conséquence imparable de notre nouvelle conception de l'Avesta. Puisque le Yasna est une œuvre ancienne et minutieusement concertée, notre devoir est de chercher à établir le sens que son arrangeur entendait lui accorder (Cantera Reference Cantera2016a: 149, my italics).
Strategies of integration
Cantera never expresses himself in this way, but it is quite evident from his interpretation of Yasna or the Vīdēvdād that he shares this assumption. He maintains that the long liturgy should be treated as “a textual unity”. Although he acknowledges that the text of Yasna is an “amalgam” of texts from different times and milieus, he nonetheless maintains that “they are purposely arranged in order to construct a ‘text’”. How can the acknowledgement of the composite nature of Yasna be reconciled with the assumption that it is a “coherent text”?Footnote 4 In other words, in what way do the texts of the long liturgy carry a unifying meaning? How did they come to embody this assumed meaning? Given the gradual development of the corpus in the context of an oral ritual tradition, which Cantera himself emphasizes, there are two ways one can imagine that the texts carry the putative meaning. 1) This meaning may have been ascribed to it ex post facto by (an) especially authoritative priest(s) or school(s) by way of (a) reinterpretation(s) of the incorporated texts; 2) It could also have been ascribed to the incorporated texts, irrespective of their actual content, based on the circumstance and (presumed) function of the ritual in which they are recited. Or, perhaps more realistically, the ascription of the unifying meaning may have occurred through a combination of these two mechanisms.
Skjærvø’s cosmological schematization of Yasna as the recounting of world history from creation to final renovation is an example of the second strategy (ritual-functional unification).Footnote 5 The aim of the daily ritual is to ensure the rise of the sun and to battle against the forces of darkness and chaos. It is a world-preserving action that imitates, and thus partakes of the power of, Mazdā’s world-creating and renovating actions. There is nothing in the sources Skjærvø uses that suggests the existence of such a scheme. A typical statement in his account is the following: “Y.65 contains a long hymn to the waters, which is directly followed by a request for fravashis to come, apparently in their function as conveyors of the birth waters (Y.65.6)” (Skjærvø Reference Skjærvø, Vahman and Pedersen2007: 81). The collocation of the waters and fravašis in the rite must somehow invoke the latter's cosmogonic role in setting the waters in motion (according to Yašt 13) as a constituent of the world-forming processes and bring to bear the power thereby acquired on the world here and now. The notion of the “birth waters” is Skjærvø’s gloss, coined, perhaps, to drive the point home.
We can see the first strategy at work in Kellens's reflections on the relation between the Staota Yesniia and the intercalated Vīdēvdād texts, which are supposed (by him and Cantera) to be an exegesis of the former. The Vīdēvdād tells the history of the world from the beginning to the end as a story of (universal) purification, enacted in the respective ceremony, which “vise clairement à composer un vaste rite de purification du monde”:
Si le Yašt 19 et le Vidêvdâd font le récit explicit du cours de l'histoire cosmique selon la doctrine des millénaires, ce n'est pas le cas des Staotas Yesniias, qui, comme corpus constitué, n'ont en principe de rapport avec cette doctrine qu'en virtue d'une exégèse ultérieure (Kellens Reference Kellens2015: 4–5; see also Kellens Reference Kellens2011: 78, 137).
The question would then be: how did the Vīdēvdād exegete come up with the (supposed) narrative account? Kellens tries to show how the “ritual course” of the first Gāthā may lend itself to a millenarian framing.
La Gâthâ ahunauuaitī, pas plus qu'une autre, n'a pour but de raconter l'histoire du monde. Son discours est essentiellement la composante verbale d'un processus liturgique qui conduit du plus sombre de la nuit aux premières lueurs du jour, qui ouvrent la voie aux deux offrandes indo-iraniennes traditionnelles: la libation de haoma et la crémation de chair. Ce temps a pour rythme le cycle du feu. D'abord trop humble pour être nommé, le feu devient successivement une aide puissante, puis le langage des dieux, enfin un avaleur de chair qui, rassasié, offre sa protection jusqu’à ce que la perfection finale soit instaurée. Ce processus a ceci de commun avec l'histoire qu'il est conçu comme l'intervalle entre un début et une fin: c'est ce que constatent, en encerclement, Y28.1 paouruuīm et, d'une certaine manière, Y34.15 haiθiiə̄m (Kellens Reference Kellens2015: 13–4).
The “innovative” transposition of the ritual course into a universal history of perfection could have been “inspired” by other elements from the first Gāthā, too. Kellens mentions “la phase déprécative du Y32” (Kellens Reference Kellens2015: 14–5). One cannot be sure how much of this “mise en perspective historique” finds conceptual support in the Gāthā.Footnote 6 The first Gāthā recounts the career of the ritual fire, its inception, its rise and protection against the darkness of the night and what this may be assumed to symbolize, and finally its ushering in of the light of the day at the end of the ritual course, where the god is asked to make existence “perfect and permanent”. The narrative form, with a beginning, a middle, and an end, weaving in the overcoming of an antagonism (Y 32), plus a number of terms (such as paouruuiia- or fraša- or haiθiia-) have inspired a (pseudo-)historical transposition of the ritual course. If this is the case, the exegete must have believed that he found this historical meaning in the Ahunauuaitī. In any case, we can see here an attempt to account for the mythic history that, according to Kellens, is found in the Vīdēvdād by the first mechanism I mentioned above, namely ascription of meaning through interpretation (or misinterpretation) of a (supposed) reference text.
The antiquity of Yasna
Now let me give an example of the combination of the two strategies (exegetical reinterpretation and functional framing) based on the assumption of an inherited ritual structure and purpose. The self-same structure provides a unifying frame of interpretation for the organic development of meaning, generation after generation. Following Kellens (Reference Kellens. and Cantera2012), Cantera points out the continuing presence in the long liturgy of ritual elements that are mentioned in three Young Avestan texts, “which show the same structure that is found in the long liturgy” (Cantera Reference Cantera, Williams, Stewart and Hintze2016b: 62). The list that Kellens gives of these elements is: spreading of soft grass, consecration of firewood, investiture of the zaotar (officiating priest), preparation of the haoma drink, declaration of one's “choice” in accordance with the daēnā, and recitation of an Old Avestan text (for example, the ahuna vairiia). According to Kellens, the presence of these elements in both the long liturgy and the supposed sacrifice the composer of the Young Avestan texts (Y 57.2–8; Y 57.19–26; Yt 10.88–94) has in mind proves the continuation of one and the same ritual structure, with the possibility of minor innovations.
Les trois sacrifices divins des Yašts certifient que l’épine dorsale conceptuelle du Yasna était acquise à l’époque même où on rédigeait des textes du type Yašt en avestique récent. Bien sûr, il faut se représenter l’élaboration du corpus que nous connaissons comme un processus qui a pu être long … Tous les réaménagements dans la sélection et la lettre des textes doivent être considérés comme possibles, mais seulement dans le cadre d'un cursus liturgique bien détérminé. Le seul espace virtuellement ouvert à une réelle innovation est celui occupé par le Bagān Yašt et le Hōmāst (Y 19–26) (Kellens Reference Kellens. and Cantera2012: 57, my italics).
The integrity of the “ritual course” over time ensures textual stability and an enduring regime of meaning.Footnote 7 “La liturgie longue avec son récitatif aurait acquis sa forme actuelle avant la composition du Nērangestān avestique” (Cantera Reference Cantera2014: 211). In fact, Cantera extends the antiquity of the “structure” of the long liturgy to the common Indo-Iranian times. Referring to Tremblay's study (Reference Tremblay2006–Reference Tremblay07) of the “parallels” between Vedic Agniṣṭoma and the long litugry, Cantera writes:
The close parallels found between the Long Liturgy and the ceremonies of Agniṣṭoma provide definitive proof of the antiquity of the actual structure of the Long Liturgy, and help us to identify the role of its different elements through a comparison with the Vedic sacrifice (Cantera Reference Cantera2016a: 149).
Thus the (supposed) persistence of ritual structure is used to argue for continuity in the emic conception of the “Long Liturgy”. The practical outcome is the assimilation of Zoroastrian liturgy to Vedic Agniṣṭoma, which, as we will see, is problematic in a number of respects. But before examining this assimilation, I would like to consider briefly Kellens's assertion of the continuity of the “liturgical course” of Yasna on the basis of the ritual elements mentioned in the three Young Avestan texts. Rituals are generally conservative, but not necessarily in their grammar or semantics. A very limited number of elements (such as animal immolation or libation) can produce a relatively large number of ritual schemes thanks to the symbolic nature of the ritual. The specific meaning ascribed to an element in a particular ritual is purely conventional (see Smith Reference Smith1982). Upon entrance into ritual space an ordinary gesture or object is minutely and rigidly stylized, which is the sign of its differentiation from the profane context.Footnote 8 It is generally not possible to derive the meaning of a ritual from its constituent parts (gestures, words or objects), even in the case of elements that are highly indicative.Footnote 9
As for the items on Kellens's list, we note that in Yt 10.88–92, for example, they do not exhibit any particular structure.Footnote 10 If one, nonetheless, were to turn the sequence in which they are mentioned in the text into a particular structure, it would still not have the order that the matching items have in Yasna (I just cite the phrases from Yašt 10 that mention the items in question): Miθra is worshipped by Haoma with barəsman, zaoθra, and words (88); who Ahura Mazdā appointed as zaotar (89); who as mortar-priest first presented (uzdasta) the heaven-made, star-bejewelled haoma stalks (90); fulfilled will be the man who makes offerings to you (Miθra) time and again, firewood in hand, barəsman in hand, milk in hand, mortar in hand, with washed hands, with washed mortar, with laid-out soft grass (barəsman), with prepared (uzdātāt̰) haoma, reciting the ahuna vairiia (91); in accordance with this daēnā Ahura Mazdā chose (93). The order of the items in this passage that appear on Kellens's list is: 1) barəsman; 2) appointment of the zaotar; 3) firewood; 4) barəsman; 5) haoma; 6) recitation of the ahuna vairiia; 7) choosing in accordance with the daēnā. Compare this with the order of the matching items in Yasna (according to Kellens): 1) Barsom Yašt (Y 2); 2) consecration of the firewood in Āvid (Y 4); 3) investiture of the zaotar (Vr 3–4 intercalated between Y 11.9 and 10); 4) preparation of haoma (Y 9 to Y 11.10);Footnote 11 5) the declaration of choice (Y 12); 6) recitation of Old Avestan texts (Kellens Reference Kellens. and Cantera2012: 56). The items mentioned in Yt 10.91 are presumably the constituents of a basic ritual. Zoroastrian rituals manipulated mortar and haoma, milk and firewood, and so on; and they still do today. But one can hardly conclude from this the persistence of a self-same “ritual structure”. This may well be true, but is not ascertainable from the continuing ritual use of the same items. What we can know from the Young Avestan text is that ritual use was made of barəsman, for example, but not that a section of the ceremony was dedicated to it, and that this (supposed) section appeared in the same relative place in that ceremony as Y 2 does in Yasna. It is clearly the latter knowledge that is required by Kellens's position. If Yt 10.91 gaozastō refers to milk and not sacrificial animal, which is almost certain, then animal immolation is missing among the basic ritual items of Yašt 10. As Kellens and Cantera have themselves argued, animal sacrifice constituted an important, if not the central, part of ancient Zoroastrian ritual. How should we explain its absence from the “ritual structure” allegedly reflected in the Young Avestan text? Yt 10.88 mentions three items in Haoma's rite in honour of Miθra: barəsman, zaoθra, and words. What warrants the differential treatment of these items? If the first one is implicitly understood in Kellens's argument to represent a “phase” of the supposed ceremony, why should not the second and the third?
Proferes has argued that śrauta rites are the result of a “fundamental reorganization” of the earlier clan-based system.
The reformation of liturgical practice that ended the Ṛgvedic period involved a fundamental reorganization of the institutions responsible for the maintenance of orthopraxy, with custodianship of the rites being removed from the individual lineages and invested in priestly offices (in connection to which the various śākhas eventually arose) (Proferes Reference Proferes2014: 200).Footnote 12
According to Proferes, the praiṣas have preserved the trace of an archaic practice from before the standardization of the śrauta system (see Proferes Reference Proferes2014: 211–7). The evidence is in “the sūktavākapraiṣa pronounced in the animal sacrifice contained within a soma rite”. This is the custom of choosing the hotar from among a number of contenders, “which has no equivalent in the standardised śrauta ritual”. The yajamāna is the one who actually chooses the hotar priest. The maitrāvaruṇa addresses the hotar: tvām adya … avṛṇītāyam sunvan yajamāno bahubhya ā saṅgatebhyaḥ “today this … soma-pressing yajamāna chose you … from the many gathered together”. Now, “in neither the pravara nor the ṛtvigvaraṇa, as they are known to us from the post-Ṛgvedic śrauta texts, is there any indication that appointments to the priestly offices followed a selection from among competing candidates” (Proferes Reference Proferes2014: 217; see also Heesterman Reference Heesterman1993: 144–9). The way the officiating priest is “appointed” changed, according to Proferes, following the standardization of the ritual system.
Returning to the passage from Yašt 10, if it is possible and perhaps plausible to think that in saying yim zaotārəm staiiata ahurō mazdā̊ aṣ̌auua, the composer had in mind a concrete rite of appointment of the zaotar, it is impossible to know that this rite was the same as the one we find in Yasna and, even more, that it took the same relative position as the “investiture of the zaotar” does in Yasna. In the Visperad, the verb is used not only for “appointing” the (supposed) sacerdotal college (Vr 3.1) but also for “appointing” the members of the three social classes and the heads of the four social circles (3.2), the young man in the various capacities expected of him, the mistress of the house (3.3), and generally the righteous man and woman (3.4). We must then understand the meaning of āstāiia “I appoint” in such a way that is equally applicable to all these instances. If Vr 3.1 represents the concrete ritual phase of setting in place the priestly college for the following stages of the ceremony, as it is claimed by Kellens and Cantera, among others, why would the identical Vr 3.2–4 not have the same significance? And if they do, in what capacity or sense are the categories enumerated in these sections involved in the rite? The concrete sense of the verb is not clear; and it is even more unclear that it designates a particular ritual action or phase. The continued use of particular ritual terms does not imply survival of the “liturgical course” to which they may have once belonged, even if the items to which they refer remain the same and are manipulated in more or less the same way.Footnote 13 The ritual that the composer of Yašt 10.88–93 had in mind may or may not be the same as Yasna, but one cannot conclude one way or another from the items he mentions.
The process of formation of Yasna
Cantera has taken up the comparison between the course of Yasna and Vedic Agniṣṭoma proposed by Tremblay.Footnote 14 Despite Tremblay's assertion, however, the parallels are insignificant or specious.
Non seulement les actions sont identiques (pressurage d'une liqueur enivrante [Haoma = Soma], immolation, récitations consistant en deux genres majeurs distincts, l'hymne [śástra ou ūkthá (sic.) en Inde, sāsna (sic.) ou uxδa dans les Gāθās] et le chant [stotra ou sāman en Inde, phl. stōm en Iran]), mais leur ordre se correspond (en particulier la succession non triviale d'un pressurage du *Sauma, puis d'un sacrifice sanglant du feu, avec un finale associant eau et feu, l'eau étant donc un troisième élément plus proche du feu que de la liqueur) jusque dans les détails (Tremblay Reference Tremblay2006–Reference Tremblay07: 688).
Animal immolation as such cannot be considered a significant comparative reference since it is an almost universal phenomenon. The process of preparing and ingesting (and offering) of haoma/soma is certainly an important inheritance of Indian and Iranian ritual lore, but its presence can hardly indicate similar ritual structures or a shared conception of sacrifice.Footnote 15 One must give due consideration to the formation of the śrauta system, one of whose aspects was the specialization of the personnel (see Jamison and Brereton Reference Jamison and Brereton2014: 30). The impact of this process on the design of the more elaborate types of offering, especially the soma ritual, was significant. The śrauta systematization threshold separates, perhaps even in basic design, the soma ritual from the pre-śrauta ceremonies (see Heesterman Reference Heesterman1993: 59–60). The formation of Zoroastrian ritual system, too, must have been a complex process that introduced more or less significant discontinuities with past conceptions and practices. Tremblay's claim that Yasna and Agniṣṭoma not only share significant ritual elements but also correspond in order or structure is an impossibly tall order to substantiate.Footnote 16 Yasna is a daily morning ritual whereas Agniṣṭoma is a prestige sacrifice lasting five days involving 16 priests and usually undertaken in spring.Footnote 17 The pressing of the soma takes place on the last day (sutyam ahaḥ) and comprises animal sacrifice. The sacred drink is prepared, ingested and offered in the morning, midday and evening.
The literary “genres” Tremblay invokes for the Gāthās, two in the Annuaire (cited above) and three in the working notes (in Swennen Reference Swennen and Pirart2016: 37–9), are dubious. No context in the Gāthās gives us to understand that uxδa- or sāsnā- has any formal feature. The former (uxδa-) seems to mean something like a discourse or a solemn speech that pursues a point. In Y 35.9 it is used in apposition to vacah-, apparently referring to the discourse underway: imā āt̰ uxδā vacā̊ ahuramazdā … frauuaocāmā “These words now, O Wise Lord, we proclaim as solemn utterances” (Hintze Reference Hintze2007: 95).Footnote 18 If so, the utterances in the Yasna Haptaŋhāiti are understood to be an uxδa-, and this rules out Tremblay's categorization of the term as “hymne”. sāsnā- may have a more specific meaning, but it is hard to be certain from the contexts in which it is used. Its meaning in Y 29.7 and 8 is something like “instruction” or “precept”, which etymologically seems appropriate, too (see Bartholomae Reference Bartholomae1961: col. 1574). But it is not a literary genre. The two substantive nouns derived from √stu “praise”, staota- and stūt-, may designate a particular ritual speech. From Avestan texts we cannot say whether it was characterized at some past stage by formal features (but see further below). The Young Avestan incidences of the terms derived from √stu may incline one to think that they have a special affinity with Haoma, simply because they occur relatively more frequently in the section called Hōm Stōm (Y 9.1–11.10) than elsewhere. Obviously, this valence is not found in the Old Avestan texts, where Ahura Mazdā is the primary object of the staota-.Footnote 19 One cannot ignore this fact. In short, none of these terms evince any ascertainable properties other than the Young Avestan use of derivatives of √stu. On the other hand, Vedic terms such as śástra and sāman denote, precisely, formal genres.
In the published notes Tremblay goes further and suggests that the five Gāthās belong to three (inherited) literary “genres” (uxδa- “hymne”, staota- “chant”, and vaṇta- “charme”): the first Gāthā constitutes a “hymne”, the second and third, a “chant”, and the fourth and fifth (with some hesitation), a “charme”. Based on this nomenclature, he compares the Gāthās with Vedic śastra, stotra, and “récitation de l'Atharvaveda”, respectively. Tremblay does not explain why he classifies, for example, Gāthā Ahunauuaitī as an uxδa, but judging from the synoptic table of the (supposed) correspondences between the two ceremonies, the ground for this categorization seems to be the occurrence of the term in that Gāthā – which is in fact two out of a total of fourteen in the Gāthās and Yasna Haptaŋhāiti. Clearly, this record cannot vouch for the nomenclature. In any case, this mechanical procedure is fundamentally flawed. The connection Tremblay makes between the supposed Avestan “genre” of uxδa and Vedic śástra appears to be based on the incidence of the cognate verb √sąh “declare” in Gāthā Ahunauuaitī (twice: Y 31.1 and 32.7), which is clearly inapt for the purpose it is made to serve.Footnote 20 His statement that each “texte vieil-avestique (i.e. chaque Gāθā et le Yasna Haptaŋhāiti) commence par un verbe-étiquette révélant d'emblée son genre” is thus quite puzzling. A similar “reason” is offered for the classification of the second and third Gāthās as staota- “genre”, while that of the fourth and fifth as vaṇta- is simply stated. Based on the conviction that Yasna is identical in elements and structure with Agniṣṭoma, Tremblay subsumes the Gāthās, by the sheer force of nomenclature, under specific “genres”, pairs the artificial appellations with cognate Vedic terms that designate specific genres used in Agniṣṭoma, and then invokes the spurious accordance of Indian and Iranian “genres” to assert “structural correspondences”.Footnote 21
The question remains whether terms such as yasna- (or yesniia-), staota- (or stūt- or staoma- or staomi-), vahma-, uxδa-, vacah-, vaṇta-, or the like, designate a type of ritual speech which in one way or another is marked? The marking need not be metric or prosodic, but could be based, for example, on the exclusive or particular association of a text with a specific segment of the ceremony, especially where a verbal refrain is present, such as yazamaide “we offer in sacrifice”. There is the further question of whether we actually possess the texts corresponding to the terms which we may plausibly think designate types of speech. In my mind, the only term for which the answer to both these question is positive is yasna-, if in fact it refers to the speech in yazamaide, which in the Young Avestan text becomes a litany. One might also think that Hōm Stōm is an instance of the “praise” speech. However, as I mentioned, the valence with Haoma is absent in the Old Avestan texts. We should also recall that, if Zoroastrian Pahlavi literature refers to Y 9–11 as Hōm Stōm or Stōd, Young Avestan texts (Y 55.6 for example) call the Old Avestan corpus staota yesniia “praise-texts belonging to yasna”.Footnote 22 Further, it is not Haoma that Zarathuštra “praises” in the Young Avestan texts, but Aṣ̌a. There are a number of passages which might suggest that one or the other of our terms referred to particular types of speech. In Y 35.10, uxδa-, staota- and yasna- appear to designate, in one interpretation,Footnote 23 particular ritual segments. But for every such an affirmative evidence there is as good a negative one. In Y 35.9, for instance, uxδa- is coupled with vacah- to characterize the speech being delivered. Is vacah-, too, a type of speech? In Y 41.1, Ahura Mazdā and Aṣ̌a are offered stūtō garō vahmə̄ṇg. Are these “praises, greetings (and) glorifications” or “praise-texts, greeting-texts, (and) glorification-texts”? Perhaps the former, considering Y 34.2cc’ pairigaēθē xšmāuuatō, vahmē mazdā +garōibīš stūtąm “with greeting words of praises at the glorification ceremony … in your honor, O Mazdā”.
As for the “nontrivial” correspondence (“down to details”) between Yasna and Agniṣṭoma in 1) the “succession” of the pressing of haoma/soma and the offering of part of the animal victim into the fire, and 2) the significant association of fire and waters (rather than the latter and haoma/soma) – these, too, are untenable. The pressing process in Agniṣṭoma is complex and forms the axis of the tripartite ceremony. Each of the three ritual periods is framed by an opening pressing and a concluding libation, and each includes an animal offering (the epiploon, for example) into the fire, followed by the ingestion of soma and the sacrificial meal.Footnote 24 Reduced to an abstract pair of markers (pressing and libation), Agniṣṭoma and Yasna have “equivalent structures” – but what is thereby asserted, other than the (already acknowledged) common importance of the inherited element? Would the assertion not be simply a tautology? We could not gain even the most elementary knowledge about the respective conceptions of haoma/soma unless we ask: in what way is it important (to each)?Footnote 25 In any case, what is noteworthy in the tripartite ritual of the final day of Agniṣṭoma is not that the animal offering into the āhavanīya fire follows the pressing – every action does – but that it precedes the ingestion of soma and the sacrificial meal. The precedence of the gods in the ceremony shows that Agniṣṭoma is fundamentally conceived as an offering ritual, involving, of course, demand for reciprocation (see, for instance, Scheid Reference Scheid and Rüpke2007).
During in the concluding (desacralizing) bath, an oblation is made into water – “autant que possible, en une place stagnante d'une eau courante” (Caland and Henry Reference Caland and Henry1906–Reference Caland and Henry07: 397). Just before the oblation, the prastotar intones a song in honour of (ritual) fire. Tremblay sees in this a significant similarity with the contiguity of the sections dedicated to fire (Ātaš Niyāyišn) and water (Āb Zōhr) at the end of Yasna (Tremblay Reference Tremblay2006–Reference Tremblay07: 687–8). The fact that as a part of the desacralizing bath a hymn is sung in honour of Agni just before an oblation into water, however, does not indicate an elective ritual affinity between fire and water, since the oblation is not made for the waters at all. It is made for Agni, or rather for Agni qua Apāṃ Napāt, apparently as the sun descends into the waters.
Cette oblation se fait, non pas au feu, mais dans l'eau, et, autant que possible, en une place stagnante d'une eau courante. On la fait debout, et, selon quelques autorités, orienté vers le point cardinal dans la direction duquel on est allé à l'eau. L'adhvaryu jette à l'eau un brin d'herbe, et y verse avec le sruva ou, suivant d'autres, avec la juhū, une libation qui remplace l’āghāra de l'iṣṭi ordinaire, en récitant T.S. 1.4.45d: “Le visage d'Agni a pénétré dans les eaux / [en sa qualité d’] Apāṃ Napāt qui protège la majesté des Asuras // dans toutes les demeures, honore [ton] combustible, ô Agni, / et que ta langue s’étire vers le beurre. Svāhā!” (Caland and Henry Reference Caland and Henry1906–Reference Caland and Henry07: 397–8).Footnote 26
There can hardly be any doubt that the recipient of the oblation into water is the figure of Apāṃ Napāt. The assimilation of Apāṃ Napāt to Agni is not well understood (see Magoun Reference Magoun1898; Reference Magoun1900). If the basis for it is the speculative identification of the sun and sacrificial fire, as it appears to be (see Proferes Reference Proferes2007: 105–6), then the sun must be the mediating factor, and in particular the setting sun whose reflection in the waters would become the “descendant of the waters”. Tremblay's assertion that water is “plus proche du feu que de la liqueur” is thus malapropos. The close connection between Soma and Agni is displayed throughout the soma ritual, from its name, Agniṣṭoma, to concrete gestures. In the immediate context, for instance, right after the invocation of Agni, two ājyābhāgas are made, accompanied by two Rigvedic stanzas (RV 8.43.9 and 1.23.20) containing the word apsú, “respectivement anuvākyā pour Agni et anuvākyā pour Soma” (Caland and Henry Reference Caland and Henry1906–Reference Caland and Henry07: 398). In the elaborate ceremony of the final day of Agniṣṭoma, soma and fire are omnipresent and constantly interact. The axis of the ritual treatment of fire is its maintenance and its usage for sacrificial oblation.Footnote 27 In other words, in the economy of Agniṣṭoma the proximity of fire and water during the final bath does not stand out, and certainly does not have the meaning Tremblay ascribes to it.
On the Iranian side, the reason for the contiguity of the sections dedicated to fire and water at the end of Yasna is not clear. Kellens has proposed a ritual explanation of it:
Les eaux … abritent un hôte qui est à la fois leur contraire et leur parent le plus intime: *Apām Napāt. Cette association paradoxale est fondée sur la même soumission aux mouvements d'entrée et de sortie du rite. Comme les eaux sont puisées pour servir au sacrifice, le feu est allumé sur l'autel, puis éteint, c'est-à-dire confié aux eaux qui l’éteignent avant d’être rendues au courant … Dans l'Avesta, Apąm Napāt est, avec les eaux, le ratu de l'après-midi. Feu de la cérémonie qui s'achève, il est confié aux eaux qui l'emportent vers la nuit, tandis que le soleil décline et que, paradoxalement, on active les banals feux domestiques. Demain, à la fin de la nuit ou au lever du jour, il sera rallumé pour accomplir un nouveau cycle sacrificiel … L’Ātaš Niyāyišn et l’Āb Zōhr impliquent que le feu était éteint à la fin de la cérémonie et le passage progressif à un feu permanent préservé dans un temple eût fait d'Apąm Napāt un dieu sans emploi si une connivence particulière avec Miθra ne lui avait donné un nouveau rôle (Redard and Kellens Reference Redard and Kellens2013: 8–9).
Apąm Napāt is “le feu éteint du sacrifice”, more precisely, the water-extinguished fire. Further, Kellens suggests that before the establishment of permanent fires at temples the sacrificial fire was extinguished by means of water each day at dusk. This would explain, according to him, the contiguity of the sections dedicated to fire and water at the end of Yasna, and shows, too, why Apąm Napāt and the waters were jointly appointed as the ratu of the uzaiieirina period of the day (i.e. afternoon until sunset). It was imagined that the extinguishing water harboured the fire, which was rekindled on the altar the following morning. As far as I know, there is no evidence for Kellens's schema in the Avesta or later Zoroastrian Pahlavi literature. The distinction between the hearth and cult fires and their (supposed) interactions are not at all clear in the Avesta.Footnote 28 Why does Kellens think that before the establishment of fire temples daily sacrificial fires were put out by water at dusk just before domestic fires were “activated”? The association of Apąm Napāt with the waters is of course implied in the god's appellation, “descendant of the waters”. But to what phenomenon does it in fact refer? The major difficulty with Kellens's theory is that it cannot account for the phenomenology of the god as we know it from Avestan and Vedic sources. In fact, his theory clashes with this evidence. Oettinger explains the god as a deified natural phenomenon. Invoking the description of the xvarənah- in Yašt 19 and comparative Indo-European material, he maintains that it is likely “dass der im Wasser befindliche “Glücksglanz” ursprünglich einmal Teil des unter Wasser feurig glänzenden Wassergottes *Apām Napāt gewesen war” (Oettinger Reference Oettinger, Pirart and Tremblay2009: 193). According to Oettinger, the god originates in the shimmering or glowing rivers and seas. In my mind, this archaeology is basically right. I would add, based on Vedic evidence, that the shimmering is probably the reflection of the afternoon sun. The epithets of Apąm Napāt in the Avesta agree with the description of Apāṃ Napāt in the Rigveda. In the Avesta (Y 65.12, Yt 19.52), the god is characterized as bǝrǝzant-, ahura-, xšaϑriia-, auruuat̰.aspa-, xšaēta-, upāpa- yazata-. Apąm Napāt resides and shines in water; he is a lofty (bǝrǝzant-) lord who drives swift horses. The possession of swift horses seems to connect *Apām Napāt with the sun. In the Rigveda, Apāṃ Napāt is āśuhémā a “horse-driver” (2.35.1), yó anidhmó dīdayad apsv àntár “who, unkindled, shines in the waters” (10.30.4), manojúvo vṛṣaṇo yáṃ váhanti “whom stallions as quick as thought convey” (1.186.5), nādyó “of water” (2.35.1); sá śukrébhiḥ śíkvabhī revád asmé dīdāyānidhmó ghṛtánirṇig apsú “he brilliantly shines in the waters, unkindled, with his shimmering mighty limbs, donned in (golden) ghee” (2.35.4), yó apsv ā śúcinā dāívyena ṛtāvājasra urviyā vibhāti “who shines far-and-wide, with divine flame in the waters, righteous, untiring” (2.35.8); apāṃ nápād ā hy ásthād upásthaṃ jihmānām ūrdhvó vidyútaṃ vásānaḥ “clothed in lightning flash, Apāṃ Napāt has ascended their lap, standing upright, while they lie across” (2.35.9); híraṇyarūpaḥ sá híraṇyasaṃdṛg apāṃ nápāt séd u híraṇyavarnaḥ / hiraṇyáyāt pári yóner niṣádhyā hiraṇyadā dadaty ánnam asmāi “golden aspect, gold-like (is) that Apāṃ Napāt, and he is indeed gold in color, having settled down from a golden lap, the gold-givers give food to him” (2.35.10); tad asyānīkam uta cāru nāmāpīcyaṃ vardhate naptur apām / yam indhate yuvatayaḥ sam itthā hiraṇyavarṇaṃ ghṛtam annam asya “this face of his grows (strong) and (so does his) lovely secret name of Apāṃ Napāt, whom the young women inflame thus: golden ghee is his food” (2.35.11).
In the last two stanzas (RV 2.35.11–12) the poet apparently refers to the establishment (ni + √sad) and maintenance of the ritual fire (Agni), who is identified with Apāṃ Napāt. The “young women” are probably the ten fingers that “inflame” the fire by pouring “golden ghee” in it (it is unlikely that “inflaming” refers to kindling by fire drill, since the fire seems to be already ablaze and because of the conjunctive adverb itthā “in this way”). The golden lap or womb perhaps recalls the celestial home of Apāṃ Napāt/Agni (RV 2.35.6 jánimāsyá ca svàr). Apāṃ Napāt and Agni are certainly distinct gods in the Rigveda, even if speculatively identified in specific situations. Whether one identifies the “descendant of the waters” as chain lightning amidst pouring rain – as Magoun (Reference Magoun1898 and Reference Magoun1900), among others, did over a century ago – or the shimmering reflection of the afternoon sun in seas and rivers (as Oettinger does), the phenomenology of his appearance rules out seeing in the god the water-extinguished ritual fire. Kellens adduces as evidence for his schema the two stanzas I have quoted in full above.
Ce cycle est clairement lisible dans RV 1.35.11–12:Footnote 29 le feu engendre un embryon, que les eaux abritent et nourrissent, puis qui est retiré des eaux pour être de nouveau allumé naturellement, avec des copeaux (Redard and Kellens Reference Redard and Kellens2013: 9).
I do not see how the passage supports Kellens's schema (“cycle”) and his contention that the “embryo” of the fire is “harbored and nourished by the water” that extinguishes it, unless he assumes what is to be demonstrated, namely that *Apām Napāt is the water-extinguished ritual fire. Kellens's conception of the relation between the domestic and sacrificial fires in general is at odds with the śrauta ritual system, as I argued above. One should note that it is the desacralizing role of water that explains its concluding place in the ritual course of the final day of Agniṣṭoma. The ritual bath takes place in the evening, thus providing the occasion and perhaps the inspiration for a final hymn and oblation to Agni qua Apāṃ Napāt, before making the conclusive oblations on the ritual ground and the cremation of the vedi (see Caland and Henry Reference Caland and Henry1906–Reference Caland and Henry07: 405–11). The only datum in Kellens's schema is the placement of the afternoon period (uzaiieirina) under the protection of the waters and Apąm Napāt. Is the schema meant to explain this? If one were to rely on Vedic evidence, the link between the waters and the uzaiieirina period would have to be Apąm Napāt. There may be a trace of the connection between the sacrificial fire and Apąm Napāt in the contest over the xvarənah- in Yt 19.45–52, especially if Oettinger is right in his argument that the xvarənah- originally was an aspect of Apąm Napāt.
Cantera has relied on Tremblay's view about “the close parallels” between Yasna and Agniṣṭoma ceremonies to argue that these “parallels”: 1) “provide definitive proof of the antiquity of the actual structure of the Long Liturgy”; and 2) “help us to identify the role of its different elements through a comparison with the Vedic sacrifice” (Cantera Reference Cantera2016a: 149). I argued above that the first assertion is untenable, and the second generally abounds in risks, and is misleading in the way Tremblay has executed it. But Cantera has gone further. He suggests that Yasna, currently celebrated in the morning, is the result of the coalescence of three daily rites formerly conducted at dawn, noon, late afternoon, similar, in this respect too, to the final day of Agniṣṭoma.
There are certain details that point to an earlier and different timeframe for the celebration of the Long Liturgy that is closer to the three daily pressings of soma in the Vedic rituals … the Yasna is not parallel to a single pressing, but to a complete day's ceremony. This raises the question of whether the Long Liturgy might have been celebrated at a certain time during the whole day (Cantera, Reference Cantera2016a: 168–9).Footnote 30
Cantera proposes to divide the “Long Liturgy” (or rather the Visperad) into three “parts”: Y 1–21; Y 22–59; and from the second Srōš Drōn (before Y 60) to Y 71. His argument, however, is not convincing. He tries to match each of the three parts with one daily ratu (ušahina, rapiθβina, and uzaiieirina) via the respective tutelary god(s). For the connection of the Āb Zōhr with the afternoon “as the most appropriate time”, Cantera relies on Tremblay's interpretation of the sequence of hymn to fire and (supposed) oblation to the waters during the desacralizing bath in Agniṣṭoma and Kellens's view about Apąm Napāt, both of which are doubtful. In the “third part” of the “Long Liturgy” (or rather the Visperad), Cantera also includes the Ātaš Niyāyišn and the second Srōš Drōn. The former must, according to his scheme, be allocated to the ceremony of rapiθβina whose tutelary gods are Aṣ̌a and Ātaš. The latter, dedicated to the tutelary god of the ušahina, is out of place in the uzaiieirina ceremony.
The original duration of the whole ceremony was a whole day: starting at dawn with Srōš Drōn and the Hōm Stōd (with pressing at sunrise?), then Hōmāst and Staōta Yesniia with the animal sacrifice and the meat offering to the fire at noon and, finally, the second Drōn Yašt with the sacrificial banquet and the conclusive libations to the waters in the afternoon. At a later stage, the whole ceremony was compressed and celebrated within a single period of the day (Cantera Reference Cantera2016a: 170).
The texts Cantera wants to assign to the three ratus are, thematically and chronologically, too disparate to allow any arguable division of the kind he proposes. Why should Y 16 or Y 19–21, for instance, be part of the ušahina ceremony as opposed to the other two? If the tripartite structure corresponds to the three pressings of Vedic soma ritual, and thus presumably reflects Indo-Iranian tradition, why are the texts of Cantera's daily ceremonies topo-chronologically so heterogeneous? And why does the pressing of haoma take place in the middle ratu yet its ingestion only in the first?Footnote 31 How to account for the stranded Srōš Yašt (Y 56–57)? One could add further objections, but there is no point. Cantera has also suggested using the presence of three particular sections as the marker of a complete ceremony. The sections are litanies in āiiese yešti, āuuaēδaiiamahī, and yazamaide. Each of the three parts of the long liturgy (erstwhile, the three daily ceremonies) contain all three types of litany. “[T]he Long Liturgy can, indeed, be understood as the threefold repetition of a basic ritual (basically the combination of the āiiese yešti-, āuuaēδaiiamahī- and a yazamaide section) that is extended in three different ways depending on the ritual moment” (Cantera Reference Cantera2016a: 182). This criterion, however, is rather artificial. These three types of litany plus the aṣ̌aiia daδąmi basically constitute the Srōš Drōn, so they would be missing in the “third part” of Yasna, in which there is no second Srōš Drōn.Footnote 32 The yazamaide form is more ancient than the other three and is not limited to litany; as a type of composition, it probably originated in the Yasna Haptaŋhāiti (see Hintze Reference Hintze and Stausberg2004). The litanies were composed to accommodate older texts that had been preserved and chosen to be recited in the Yasna ceremony.Footnote 33 It is not clear at all why the texts that are presently comprised in Yasna were chosen or whether there were significant revisions to the corpus throughout its long history. One must imagine the process of formation of the corpus as one of accommodation of older, disparate texts, rather than as a “threefold repetition of a basic ritual … that is extended in different ways depending on the ritual moment”. In any case, this view of the formation of the corpus does not underwrite but contradicts the claim, implied in the comparison with Agniṣṭoma, that Yasna is derived in structure and purpose more or less directly from Indo-Iranian ritual lore; or, at least, it is not clear how these two perspectives can be reconciled.Footnote 34 It is also difficult to see how the tripartite scheme squares with Cantera's view about the way the Staota Yesniia was understood in the Young Avestan tradition, and with his interpretation of the function and meaning of Vahištō-ištī Gāthā.Footnote 35
Conclusion
In conclusion, the attempts made in recent years to argue for the existence of a coherent text and structure in Yasna have not been successful. In particular, comparison with Vedic data that goes beyond well-defined and justified instances must be abandoned, especially if it is used to establish fundamental theses such as those examined above. It stands to reason that recessive myths and ritual themes have a better chance of having comparative value. This is illustrated by Apąm Napāt/Apāṃ Napāt. Conversely, elaborated elements and elaborate structures are likely to be specific to each tradition. Thus, contrary to Cantera's supposition, it is not true that “comparison with the Vedic sacrifice” helps “us to identify the role of [the] different elements” of Yasna. The sequence of the Ātaš Niyāyišn and Āb Zōhr at the end of the “long liturgy”, if it can be explained at all by the “liturgical course” of Yasna, does not appear to have anything to do with the hymn and oblation to Agni qua Apāṃ Napāt during the ritual bath at the end of the Agniṣṭoma. The ascription of a “ritual course” to Yasna would remain the tautological observation that the sections found in Yasna occur in sequence in Yasna, unless one can show the logic of this sequence.
Appendix: Tremblay's notes on structural correspondences of Yasna and Agniṣṭoma
Philippe Swennen's critical remarks on Tremblay's comparative approach as this appears in his notes are judicious (see Swennen Reference Swennen and Pirart2016: 1–17). I am more sceptical than Swennen about the general value of the undertaking as such and the basic conception which it brings to bear on the data, especially on the Iranian side. According to Swennen, Kellens's two publications in the late 1990s put on the agenda the task of a systematic comparison of the Indian and Iranian ritual traditions (see Kellens Reference Kellens1996 and 1998).
Dès lors que l'unité organique de l'Avesta, appréhendé comme un ensemble de récitatifs liturgiques destinés à être imbriqués les uns dans les autres en fonction de besoins cérémoniels, s'imposait comme l'hypothèse économique … l’évidence de la cohérence et de l'antiquité du matériel, rendue palpable par la présence d'un vocabulaire technique visiblement hérité, appelait une nouvelle confrontation de l'Avesta aux synopsis liturgiques védiques, notamment ceux du Yajurvéda. C'est précisément ce dont Xavier Tremblay prit conscience mieux que tout autre (Swennen Reference Swennen and Pirart2016: 1–2, my italics).
The comparative approach, according to Swennen, should be guided by ritual-technical terms that are common to the two traditions, and it should include “la redoutable question de la permanence du contenu sémantique” of these terms, which implies that first the question of the concrete meaning of the terms must be posed within respective traditions (Swennen Reference Swennen and Pirart2016: 15). This proviso is important for safeguarding, as much as possible, the probity of the comparisons and their results. But it has been ignored time and again by scholars who have advocated the so-called “védisant” approach, which in too many instances has amounted to an automatic assimilation of the Avesta to Vedic data. Contrary to Tremblay's assumption, for example, Swennen shows that the functions of the inherited term nivíd- / niuuaēδaiiemi do not coincide in the two ritual traditions:
En effet, une dérivation verbale équivalente [à niuuaēδaiiemi] est attestée en sanskrit, mais à partir des commentaires en prose. Son contenu sémantique, que je qualifierai de laïcisé, signifie sans surprise “annoncer” et ne va pas de pair avec le début d'une récitation litanique en prose (Swennen Reference Swennen and Pirart2016: 14).
Swennen (Reference Swennen and Pirart2016: 11) notes that Tremblay subordinates the linguistic data to the requirements of his structural approach, creating more problems than it solves. This is Swennen's basic objection to Tremblay's method. But, in my mind, the flaw is deeper and has to do with the “védisant” approach. In practice, it has operated within a set of methodological premises that systematically prejudices the study of the Iranian data. Two such methodical treatments are overinterpretation of Avestan texts and pseudo-systematization of Avestan notions in order to make them suitable for comparison with Vedic data and ultimately reveal their Vedic bearing. The Vedic prejudice is at times present even in Swennen. Referring to Tremblay's notes, Swennen states that he endorses “le caractère hérité de la polarité des racines indo-iraniennes *STU et* ŚAṂS, qui doit suffire à établir que l'habillage de l'action sacrificielle par une alterance de chants et de récitations remonte au moins aux cultes publics indo-iraniens communs, le chant jouissant à l’évidence d'une affinité particulière avec le pressurage de *sauma” (Swennen Reference Swennen and Pirart2016: 11). But where is the evidence for the existence in Avestan tradition of a particular genre comparable to Vedic śástra? Where is the evidence for the assumption that the ritual speech characterized by the verb √stu had any formal features comparable to Vedic stotra? And finally, where is the evidence for the framing of ritual action in Yasna by an “alteration of recitation and singing”? None has so far been shown for any of these postulations, which are rather sourced from Vedic material.Footnote 36 We saw this process in the invention of Avestan “genres”. The distinction between “recitation and singing” is simply immaterial on the Iranian side.
Tremblay's assumption of the inherited nature of Vedic genres and liturgical roles, as Swennen (Reference Swennen and Pirart2016: 16) observes, stumbles on the process of the canonization of Vedic texts and the standardization of the soma ritual. It is not possible to project back into the Indo-Iranian times the characteristics we find in the classical soma ritual, which is what Tremblay does. In effect, he turns the latter into the model to which Avestan texts must conform, ostensibly because the two traditions share a common past. I pointed out in the article the case of the four Avestan “genres” invented to match the four Vedic genres (the four Vedas) and the four classes of priests. I would now like to give a few examples of Tremblay's tendentious interpretation of Avestan data.
According to Tremblay, an oblation of haoma accompanied the recitation of the Yasna Haptaŋhāiti, as it happens in the Agniṣṭoma with prose formulas (in Swennen Reference Swennen and Pirart2016: 63 and 79). No evidence has been adduced so far for libation of haoma into fire at any stage of Yasna. Tremblay cites in evidence Vr 9.3, but this is spurious. He asserts that frā + √yaz “est le verbe technique désignant la récitation du Yasna Haptaŋhāiti” (in Swennen Reference Swennen and Pirart2016: 53), and, based on this artificial semantics, interprets Vr 9.3: “≪nous sommes là≫ pour attribuer, répartir, presser, extraire, filtrer, bien offrir en sacrifice au cours du Yasna Haptaŋhāiti (sens technique de fraoiiaz), bien énumérer en litanies les Haomas présents” (in Swennen Reference Swennen and Pirart2016: 70). The word he translates as “bien offrir en sacrifice au cours du Yasna Haptaŋhāiti” is hufrāiiaštaiiaēca, which simply means “for serving as (or in) sacrifice”. Kellens translates the phrase: “(Ces haomas), les voici pour … le sacrifice-solennel” (Kellens Reference Kellens2010: 80). One wonders what is the value of such a result, produced by a method that manufactures categories on one side in order to assimilate it to the other side. Another example of the same procedure is Tremblay's interpretation of Y 15.
La triade sasti- vanta- rafnah- est un programme: elle indique que l'office qui suit comprendra des hymnes (= uxδa-, véd. ukthá-: texte du genre Ṛgveda), des charmes (vanta- comme au Y 51,22 yazāi… vantā ‘je vais offrir un sacrifice par un charme’: texte du genre Atharvaveda), des salutations (oblations secondaires: rap- est l’équivalent vieil-avestique de xšnauu-…), et le Yasna Haptaŋhāiti (frāoiiaz est le verbe technique désignant la récitation du Yasna Haptaŋhāiti). Ne sont pas mentionnés le quatrième genre, les chants (staomāi-; védique sāman): ce n'est pas un hasard puisque les chants sont plus particulièrement offerts pour l'office de Haoma (in Swennen Reference Swennen and Pirart2016: 53).
The “programme” of the sacrifice mentions rafnah- (the Old Avestan “equivalent” of xšnauu-Footnote 37), which is not one of Tremblay's “genres”, but does not include the staota-. One can see the logic of the interpretation: vaṇta- is a very rare term in the Avesta; and if it is systematically taken to designate a “genre”, it invites the same treatment of other terms with which it happens to be associated where it appears, especially that sasti- is derived from √sąh, which according to Tremblay is the “verbe-etiquette” of “hymne”. Thus, sasti- must denote a “genre”. Ineluctably, rafnah-, too, ends up designating a type of speech (but why the equivalent of xšnauu-?). The reason Tremblay gives for the absence of his fourth “genre”, however, does not follow this logic. The major sacrificial operations on the haoma follow Y 15, namely its pressing, which takes place at the end of Hōmāst, and (in Tremblay's schema) its libation and ingestion during the recitation of the YH.Footnote 38 If “l'office de Haoma” does not mean its pressing and libation, what does it designate?
Pseudo-systematization of Avestan notions and situations thoroughly vitiates the proofs of the “védisant” affirmations. Tremblay's treatment of draonah- in Y 11.7 is an example of this. Since the “major sacrifice” begins with Y 14,Footnote 39 or rather not until the YH,Footnote 40 how can the god Haoma be offered his share of the sacrifice at Y 11.7? The phrase in question in Kellens's translation is: “Coupe vite à Haoma (qui vient de parler) si ferme(ment) sa ration de vache” (Kellens Reference Kellens2007: 89). I cite in full Tremblay's attempt to deal with this quandary.
De fait, Y. 11,7 ne se réfère pas nécessairement à une action accomplie au cours du Hōm Stōm, mais peut annoncer qu'ultérieurement, après que l'hostie sera abattue après le Yasna Haptaŋhāiti, Haoma devra recevoir une part de la viande (la part offerte aux dieux?). Deuxièmement, il n'est pas inévitable que Y. 11,7 affirme que le Haoma doive recevoir de la viande, quoique le zend l'ait compris ainsi: frāoθβarəs- ne signifie pas toujours ‘trancher’, mais souvent en une acception affaiblie ‘créer’, et āofrāoθβarəs- de même peut se référer seulement à l'affectation au Haoma d'une préparation. Dès lors que gauu- désigne autant le lait que la viande de la vache, Y. 11,7 stipule peut-être seulement qu'il faut réserver au Haoma la part de la vache qui lui revient, i.e. son lait (in Swennen Reference Swennen and Pirart2016: 49).
Tremblay, of course, knows well that precisely in this passage the verb means “cut”, and that in relation to Haoma the “share (draonō) of the cow” can only mean the jaws, the tongue and the left eye, as we are told in Y 11.4: “Mon père Ahura Mazdā, pour soutenir l'Agencement, a (prélevé) et m'a attribué comme ration [draonō], à moi Haoma, les mâchoires avec la langue et l’œil gauche” (Kellens Reference Kellens2007: 87). The reason why he cannot acknowledge the obvious meaning of the phrase is also clear. The texts of Yasna belong to different “genres”, each associated with a definite ritual moment and action, which the text accompanies and more or less describes. The “ritual course” determines the content of the text and the meaning of the particular notions occurring therein. Once we decipher the “ritual course”, we can generally know what each text must be about. The structural schema is, of course, taken from the Vedic ritual. If in Agniṣṭoma “le pressurage du matin” precedes “le sacrifice sanglant” (in Swennen Reference Swennen and Pirart2016: 45), Y 11.7 daonō cannot mean parts of the sacrificial cow, but milk, since the “sacrifical phase” is supposed to follow the Hōm Stōm.Footnote 41
The Vedic schematization of Yasna does not help our understanding of the latter. It does not elucidate its structure and content, but obfuscates them. I would like to emphasize the term “schematization”, which is what Tremblay by and large has done. I do not at all mean to discount the value of comparing Avestan data with relevant Vedic material, especially on the linguistic level, as Swennen has urged. Along with mythology, ritual phraseology can yield significant comparative data. Tremblay has given us two thought-provoking examples of this in his derivation of Avestan sraoša- from the postulated Iranian counterpart of Vedic śrauṣaṭ astu: “un dérivé délocutif de *sraošat̰ ‘qu'on écoute’”, and his interpretation of uruuāxš “qu'il avance” (the 3rd sing. inj. sigmatic aorist of √uruuaj “proceed”, i.e. take to the road or start the journey) as the “cri rituel” by which the soul of the sacrificial animal is dispatched to the gods (in Swennen Reference Swennen and Pirart2016: 61 and 69, respectively).
