The representation of the Avesta

Listen to audio

The person who embodied the revival of Avestic studies between 1950 and 1989 was Karl Hoffmann (1915-1996). The synthesis of his work was published in 1989 with Der Sasanidische Archetypus, written with Johanna Narten. But how did we get here?

Between 1889 and 1896,Geldner edited the Avesta, which he divided into 5 books: Yasna ("sacrifice"), Yašts ("sacrifices"), Vīdēvdād ("ritual arrangements to keep demons at bay"), the Visperad, which exists only as a complement to the Yasna, and the Xorda Avesta, an anthology of extracts from either the Yasna or the Yašts. The history of the Avesta, according to him, is in three stages: 1. the original Avesta, containing the Gāθās and the great Yašts, had been written and archived, but was burnt by Alexander; 2. the Sassanid Avesta is made up of preserved parts of the original Avesta and new compositions, which was also written and archived; 3. the Avesta "Books of the Parsis", handed down to us by manuscripts, attests to the haphazard remains of the great Sassanid Avesta. With the discoveries made in Central Asia at the beginning of the ^20th century, we realize that it is difficult to imagine a text written before a certain period.

The Andreas theory

Carl Friedrich Andreas also sees three phases: 1. the Urtext is the original text; 2. the Arsakidischer Text is the first written version, notated in a consonant script; 3. the Vulgata is the vocalized text, albeit often incorrectly. Andreas' hypothesis was to enjoy considerable success.

The triple refutation

This was the work of three independent scholars. Georg Morgenstierne and Walter Bruno Henning set out to show that there is no false vocalization in the Avesta text, but simply phonetic developments specific to Avestic. As for Harold Walter Bailey, his focus is not on the linguistic aspect, but rather on the historical. By analyzing the testimony of the Dēnkard, he draws up a new history of the transmission of the Avesta: 1. oral transmission phase; 2. the Avesta was first published during the reign of Xosrō I; 3. the text we have is a second edition, based on the fragments that were saved from this first edition.

Four aspects of Andreas' theory overlooked by Karl Hoffmann

1. Theory not systematized, but shared for a long time

In 1851, in an article by Jules Oppert, the idea of an Avesta originally written in a consonantal alphabet appeared; Friedrich Spiegel also defended this idea; Christian Bartholomae, in §268 of the Grundriss, writes that the Avesta must have been written in a simpler script (so it can only be understood as consonantal).

2. Progressive aspect

Andreas gives greater weight to the analysis of textual data than to testimony such as that of the Dēnkard, which, as always, must be taken with the methods of historical criticism.

3. Not impeccable rebuttal

Morgenstierne's and Henning's interventions are excessive, as there are phonetic features that cannot be explained within the framework of the phonetics of the original language.

4. Long agony

Indeed, Andreas' theory was still being practiced after the three famous articles. Helmut Humbach, in 1959, explains certain aberrant spellings by false vocalization: skt. dyumná-"fate", av. diuuamna- instead of *diiumna- or skt. arámati, av. Ārmaiti and not *arǝmaiti. Jacques Duchesne-Guillemin, in 1962, proposed that oral transmission should coexist with a written standard in a consonant script. F.B.J. Kuiper, from 1960 to 1964, also defended this theory. According to him, there are three causes when a fact is unexpected in relation to the linguistic postulate: 1. the spelling, perhaps a false vocalization; 2. the linguistic divergence between Sanskrit and Avestic; 3. the idiomatic factor and hence the language of the copyists.

Karl Hoffmann's position

In 1958, he wrote a description of Old Iranian, i.e. Avestic and Old Persian. At the time, he had not yet taken a position on Andreas's theory, but one example perfectly illustrates what Andreas's theory was. sraotā is the^2nd pl. of the injunctive or imperative from the root sru-"to listen". Bartholomae makes it a radical present. However, the present radical requires degree zero in plural forms. We would therefore expect *srutā. In consonantal writing, we'd have *srwt', which would have been incorrectly vocalized here. But Karl Hoffmmann will show that sraotā is not a present tense, but an aorist for which the full degree is expected. As a result, there is no reason to speak of false vocalization. In 1969, Karl Hoffmann had a critical word to say about Andreas's theory: "a theory in which vowels count for nothing and consonants for not very much". He then turned his attention to the transmission of the Avesta, which led to three articles between 1969 and 1971 on the graphic system of the Avesta, articles which marked the definitive death of Andreas's representation. In 1975, he elucidated the difficult aṣ̌ā.yecā form of Y 30.1. Kuiper had explained it as a gen. sg. *aṣ̌ahiiā "Arrangement" followed by -ca "and" positing a form °'y' + ... where the h disappears because the sign is interpreted by ā, which changes to -e in the palatal environment. Karl Hoffmann parallels the Vedic r̥tā́ya, dat. sg. with an enlargement -a. Now, if we apply this widening in avestic, we can end up with our form: aṣ̌āi-ā-cā > aṣ̌āyācā > aṣ̌āyecā with evolution into -e- due to the palatal environment.