The representation of the Avesta (II)

Listen to audio

Last week, we saw, among other things, Bailey's representation of the Avesta, which we could summarize using the following diagram:

-
The Sassanid Avesta with a first edition under Xosrō I, in the ^6th century CE;
The Avesta that has come down to us, made up of "fragments rescued" from the Sassanid Avesta.

Karl Hoffmann's depiction of the Avesta

Karl Hoffman's painting is inspired by Bailey's, as we shall see. Karl Hoffmann's representation is as follows:

Arachosian Avesta: in fact, Avestic is an eastern Iranian language that transited before arriving in Persia. Karl Hoffmann believes that there was a period when Avesta passed through Arachosia (the Kandahar region of Afghanistan). The Avesta was therefore transmitted at some point in Arachosia, for three reasons:
a. linguistic: the very name of Arachosia provides an example of a particular phonetic treatment: ^Haraxvaitī- with the treatment^-xv- within the group*-asu̯a- instead of^-aŋvha- as expected (*-asu̯a->*-ahu̯a->*-aŋhuua->^-aŋvha-).
b. event-driven: in the Behistun inscription, the usurper Vahyazdāta splits his army in two: on the one hand against Darius and on the other in Arachosia. According to Hoffmann, there was therefore a particular complicity between the ruling Achaemenid clan and Arachosia.
c. Iconographic: in the Persepolis reliefs, we see an alternation between Medes and Arachosians for those close to the king. Arachosia therefore seems to occupy a privileged place in the king's entourage.

The heterogeneity of the arguments leads Kellens to disagree with this hypothesis. The linguistic argument, for example, is strange, because this internal treatment is exceptional. What's more, it's the regular treatment in the initial, which doesn't make it a completely foreign phenomenon in Avestic.
Sassanid Archetype: was written down under Shāhbuhr II, in the ^4th century CE. This Archetype still existed in the ^9th century when the Dēnkard was written.
Between 2. and 3., Karl Hoffmann reconstructs an intermediate stage: the StammHandschrift, the "basic manuscript", which is the common model for all the manuscripts we possess. Hoffmann postulates this stage because of an error common to all the manuscripts we possess. Y12 attests the form ziiånīm (acc.), whose corresponding Vedic term is jyāním "absolute destruction, desolation". Yet, phonetically, we would theoretically expect *ziienīm in avestic with palatal infection. Consulting the critical apparatus, he realized that most manuscripts attest to the ziiåiienīm lesson. The scribe therefore made a mistake in writing ziiå°and instead of crossing out, he continued with the correct lesson, hence the form ziiåiienīm. The incorrect syllable with -å- has been retained by some scribes, who have noted the aberration of a form ziiåiienīm. If all manuscripts reproduce the same error, it's because they have a common model, which Karl Hoffmann calls StammHandschrift.
Avesta-Ausgabe is the Avesta we possess, the one in Geldner's edition. It consists of the debris of the Sassanid Archetype. The earliest surviving manuscript dates from around 1020 AD.

Jean Kellens' new representation of the Avesta

In preparing his lectures at the Collège de France for the 1995-1996 academic year, Jean Kellens realized that Karl Hoffmann's scheme didn't work very well. The general and most important problem lies in the StammHandschrift, for here we have an extreme reduction of the period of loss of the Avestic texts between the Sassanid Archetype (still alive in the ^9th century) and the Avesta-Ausgabe (^11th century). The texts would therefore have been lost in around 150 years. A sudden and rapid loss of texts that cannot be explained..

Geldner's edition gives the impression that the Avesta consists of five books, but this is not the case, as can be seen from the Prolegomena of his edition. In reality, we're dealing with two sets of texts:

a long liturgy (Yasna, Visperad, Vīdēvdād);
a collection of short liturgies (Yašt, Xhorda Avesta).

These two ritual sets are watertight. They are complete, coherent texts, not haphazard scraps as previously thought. What's more, these rituals are ancient, dating back to well before the Sassanid era, perhaps to the Achaemenids, perhaps earlier.

This representation of Avestan transmission, which was quite different from Karl Hoffmann's, was accepted by de Vaan (Michiel de Vaan: The Avestan Vowels. Amsterdam, 2003), then by Cantera (Alberto Cantera: Studien zur Pahlavi-Übersetzung des Avesta. Wiesbaden, 2004).