Scholarship on the Avesta

The text. Of course, Avestan research pales in comparison to the study of biblical texts (as it is a small field) and an enormous amount of work remains to be done in putting together a new critical text and resolving the many corruptions. I don’t mean to compare the two. What I mean is that indeed there has been a lot of literary and linguistic research in the past few decades (with important contributions by Jean Kellens, Almut Hintze, Alberto Cantera, Antonio Panaino, P.O. Skjærvø, Michiel de Vaan, among others) that has led to newer understandings of the composition of the Avesta. In contrast to the older theory that it is a mishmash of fragments from the Great Avesta assembled in the Sassanid period, with the majority of the text lost, the current view is that the Dēnkard Avesta was a compilation of Avestan texts with Pahlavi translation and exposition while the extant Avesta is a separate collection of mainly liturgical texts with just a few inclusions from the Dēnkard Avesta (e.g. the Vidēvdād and the Nērangestān). In particular, the Long Liturgy (as found in the Yasna and other texts) is a coherent ritual text that was arranged around older OAv material in pre-Sassanid times (with its structure and arrangement known in other YAv texts), rather than being a unstructured concatenation of fragments redacted during a time when Avestan was no longer spoken. In other words, we have in the Avesta the earlier and more important liturgical texts while later exegetical compositions were more often lost or subject to scribal corruption (lacking the living memory inherent in liturgical performance). My main point in my prior post however is that the vast linguistic differences between OAv, YAv, and Pahlavi permit a far more robust literary stratification of texts in the Avesta than, say, a biblical scholar would attempt using stylistic criteria and the slighter differences between Classical Biblical Hebrew and Late Biblical Hebrew.
And so Kellens attempts a relative chronology of how the Yasna grew as a piece of ritually performed literature.
The chronological horizon of OAv and YAv texts. To get to the heart of the matter, whether you regard the Avesta as a rolling corpus that was largely canonized by YAv speakers, or as a jumble of texts assembled together by later Middle Persian speakers, the composition of OAv and YAv texts themselves could not have been done in the Sassanid era when the writing system was adopted. By that time Avestan had long ceased to be a living language and was understood only via the translations and expository texts that had already been produced (which reflect an imperfect knowledge of OAv and YAv). Rather the alphabet was a sort of phonetic notation for exactly preserving the proper pronunciation of what was only a liturgical language. So the ideas in the OAv and YAv (irrespective of later oral and scribal corruption) certainly belong to an earlier era; the Pahlavi expository texts meanwhile reflect a later stage in theological development. Moreover scholars believe that the individual texts (such as the Long Liturgy) were fossilized and achieved stability (crystallization) before the languages became extinct as native tongues. De Vaan and other specialists believe that 300 BC “would mark the definite end of the period when new YAv. texts could be composed, or old texts adjusted by the redactors” (Avestan Vowels, 2003, p. 13).
Of course, the dating of Avestan texts beyond this is largely conjectural but there are good reasons why nearly all contemporary scholars date OAv to ca. 1200-1000 BC. YAv shows about the same amount of development from Proto-Iranian (usually dated around 1500 BC) as Old Persian, and so they are generally thought to have been roughly contemporaneous languages; the close similarity between OAv and Vedic Sanskrit meanwhile points to a time of composition not far removed from Proto-Iranian. For example, Proto-IE laryngeals were generally lost in the first millennium BC throughout all Indo-European languages (with various effects such as the coloring of vowels), but in OAv they are maintained intervocalically via the laryngeal hiatus which was likely a glottal stop. This occurs also in Vedic Sanskrit but scholars note that it is done much more consistently in OAv. Laryngeal hiatus is an archaic feature of OAv; laryngeals were fully lost in the Old Persian of the Achaemenid era and in YAv texts as well. As Benjamin Fortson explains the OAv text preserves “a stage from a time before the common ancestor of Avestan and Old Persian split” (Indo-European Language and Culture, 2011, p. 231). Also OAv preserves PIE ablaut and inflections lost in both Vedic Sanskrit and YAv; YAv meanwhile developed its own grammatical features. At the same time, the Gathas were ritually performed into YAv times before the text crystallized (after OAv ceased as a spoken language), which de Vaan dates to around 800-600 BC. Pseudo-OAv redactions suggests that OAv was no longer a living language. The very different stages in the language require that a span of several centuries must intervene between the crystallization of the OAv and YAv texts.
Meanwhile Karl Hoffmann showed that YAv retains pronunciations from Old Persian that were lost in Middle Persian. Almut Hintze notes that this means that “the Avesta was already being recited in Persis during the Achaemenid period and there is a continuity of tradition from that time throughout the Parthian and Sasanian periods” (Das Partherreich und Seine Zeugnisse, 1998, p. 154)
Zoroastrianism in the Achaemenid period? Although the Avesta shows signs of having undergone Old Persian transmission, it could not have originated solely in an Achaemenid milieu. Otherwise it would have been composed in Old Persian; Avestan is an eastern Iranian language, so the religious tradition is probably older and originated outside of a southwestern Persian context. Almut Hintze shows that YAv shares isoglosses with other Iranian languages in the southeast and northeast (such as Sogdian) and the YAv itself contains geographical references to Sogdiana, Margiana, Bactria, Herat, Arachosia, and the Helmand river (Yt. 13.143; Yt. 19.67; V. 1.1) – nothing whatsoever in western Iran; nor does the Avesta contain any references to the Median or Achaemenid empires. There is also potential archaeological evidence in the Yaz culture of Margiana, Bactria, and Sogdia (1500-500 BC), which show a near complete lack of burials – a feature consistent with Avestan practices.
The Avestan evidence suggests two main phases in early Zoroastrianism. The OAv period in the late second millennium BC was imo an anti-daēuua movement in the Margiana-Sogdia region (promoting the ahuras and yazatas as gods worthy of worship over other traditional gods), with the Medes and Persians inheriting this rejection of daevas. In all Iranian languages daeva has derogatory meaning and OP has a neologism bag (< “assign, allot”) for “god” which indicates an early avoidance of daevas. Then in the YAv period a more developed form of the religion (which saw the daevas as demons led by Aŋra Mainyu) arose in southeast Iran (Sistan?) which penetrated Media and Persia by the sixth century BC. A western form of this cult then arose in the Achaemenid period. Jean Kellens pointed out that Achaemenid royal names (Darius, Artaxerxes, Vahyazdata, possibly even Achaemenes) were allusive of sayings in the Gathas (e.g. ašā xšaθrəmcā in Y. 29.10 as the model for OP artaxšac̨a; cf. also the YAv version in Y. 20.3). The most interesting example is that of Darius: dārayavau “holding firm the good (thought)” has the model of dāraiiaṯ vahištəm (manō) in Y. 31.7 and the word for “good” is the Avestan term and not naiba which was the OP equivalent. Darius’ father also was named Vištāspa, the name of Zoroaster’s patron in the Avesta, and it is clearly a borrowing from Avestan or Median (as it lacks OP sound changes), and his wife Atossa bears the name of Vištāspa’s wife Hutaosā in the Avesta. This contrasts with that of the Teispids, who were likely Elamite (Cyrus, king of Anshan), and thus it is not surprising to find a lack of Zoroastrian affinities in the inscriptions and onomastics of Cyrus and Cambyses (who in contemporary scholarship are distinguished from the Achaemenids). Skjærvø also draws attention to Xerxes’ Daiva inscription which clusters together different technical terms and ideas found in YAv passages (XPh 46-56).
The “triad” of Artaxerxes II corresponds to the three ahuras in the YAv, with Anahita supplanting Apam Napat (cf. the long Yašt to Anahita and the lack of one for A.N.). The largely post-YAv form of western Zoroastrianism in the late Achaemenid period took on a character of its own, with its promotion of Anahita, deemphasis of the Amesha Spentas, use of a new calendar, introduction of statue temples under Artaxerxes II, and finally the fire temples. Albert de Jong has shown that eschatological ideas from the YAv were known by the late Achaemenid era. Theopompus and Eudemus of Rhodes (fourth century BC) attribute ideas of resurrection, world eras of 3,000 years, and proto-Zurvanite notions to the Magi. Plutarch, citing Theopompus as a main source of information, quotes from Yt. 8.44, mentions the Amesha Spentas, relates the future fall of Areimanius, and describes the coming Renovation. The YAv view of daevas may also appear in Jewish tradition in Tobit (Asmodeus < Aešma daeva). So for these reasons I view OAv and YAv as legitimate sources for the period despite their late date of writing and textual problems.