The Avesta

A Liturgy, Not a Book

The first European Avesta, published by Anquetil-Duperron in Paris in 1771, was called a Parsi forgery in the same year by William Jones — whose grievance, on Michael Franklin's account, was that Anquetil had falsely accused Oxford's Thomas Hunt of professing Avestan and complained of the Bodleian's unheated rooms — and Anquetil was not vindicated until Tychsen and Heeren, some twenty years later.

Zoroastrian scripture in Avestan, held as Ahura Mazdā's revelation to Zarathustra and recited to this day.


Avesta is the Zoroastrian name for the texts in Avestan, an Old Iranian language attested nowhere else. The ninth- or tenth-century Dēnkard describes a Sasanian collection of twenty-one nasks; Karl Friedrich Geldner reckoned in 1896 that at least three quarters is lost, and only the Vidēvdād, Alberto Cantera notes, survives in the arrangement the Dēnkard gives. The Avestan alphabet — fifty-four letters on Cantera's count, fifteen of them vowels — was invented under the Sasanians to fix a recitation's sound, not its sense. A Vidēvdād nobody in Europe could read reached the Bodleian in 1723; Anquetil-Duperron sailed for India in 1755, learned from Parsi priests at Surat, and published in 1771; William Jones at once called the manuscripts Parsi forgeries. What survives is a working liturgy, not a canon — which is why "the Avesta says" names no single book.

Light

Priests in India still recite the Yasna from memory. Alberto Cantera observes that the scribes who copied the manuscripts were those same priests, performing without them: the books were a safeguard against forgetting, not the thing remembered.

Shadow

Europe took twenty years to credit it — Duchesne-Guillemin dates the vindication to Tychsen and Heeren, 1791-94. Michael Franklin calls Jones's Lettre of 1771 a clash of egotisms: Anquetil had falsely accused Oxford's Thomas Hunt of professing Avestan, and it retarded Zend-Avesta studies for years.

Sources and verification11 sources · 5 links

Primary source

Scholarship

  • Almut Hintze, "Avestan — Introduction, Part 4: Avestan manuscripts" (university teaching materials), undated slide set, hosted at spw.uni-goettingen.de; read directly — a teaching document rather than a publication, and it carries no yearThat Mihrābān Kayhusraw "copied some of the oldest extant Av. mss.", was a descendant of a priestly family of Sīstān and was invited by Čāhil, son of Sangan, a wealthy Parsi merchant of Cambay in Gujarat; that his Yasna manuscripts of 1323 are J2 (Bodleian, Oxford) and K5 (Royal Library, Copenhagen), and his Vidēvdād manuscripts L4 (British Library, London) and K1 (Copenhagen); and that the Sogdian-script Aṣ̌əm Vohū from Dunhuang, c. 800 CE, British Library Or. 8212/84, is the unique witness of a lost Eastern Iranian recension and "pre-dates earliest Avestan mss. by more than 400 years" — the earliest surviving witness to any Avestan text, and the basis of this entry's dateline.
  • Jacques Duchesne-Guillemin, "ANQUETIL-DUPERRON", Encyclopædia Iranica, 1985, vol. II, fasc. 1, pp. 100-101; iranicaonline.orgAnquetil-Duperron (Paris, 7 December 1731 – Paris, 18 January 1805); that he was shown a few copied lines of Avestan in 1754, from a fragment brought to the Bodleian in 1723; that he enlisted as a soldier in the Compagnie des Indes to get a passage, embarked 7 February 1755 and reached Pondicherry on 10 August 1755, arriving at Surat on 1 March 1758; that he obtained a Vendidad from the reformist dastūrs Kaus and Dārāb, who taught him through Persian what they knew of Avestan, "which was not much"; that the three-volume translation appeared in 1771 and "He was immediately accused of forgery by William Jones"; and — the point on which this entry's twenty-year figure rests — that "It was not until 1791-94 that he was finally vindicated by Tychsen and Heeren," with Kleuker's German translation of 1776 and Herder's support preceding that.
  • Karl Hoffmann, "AVESTAN LANGUAGE i. The Avestan script", Encyclopædia Iranica, 1987, vol. III, fasc. 1, pp. 47-62; iranicaonline.orgThat the Avestan script derives from cursive Pahlavi with signs borrowed from the Psalter script and the Greek alphabet as a model for the vowels; that it carries 14 (or 16) vowel letters and 37 consonant letters — a count lower than Cantera's fifty-four; that so large an inventory "suggests that their invention resulted from an attempt to record an orally recited text with all its phonetic nuances" and must therefore be a deliberate creation by a scholar or group of scholars; that the earliest manuscript dates from 1288; the argument from the Istanbul sarcophagus inscription (before 430 CE) for a fourth-century invention, perhaps under Šāpūr II; and that the extant manuscripts ultimately descend from ninth- or tenth-century copies already full of error.
  • Jean Kellens, "AVESTA i. Survey of the history and contents of the book", Encyclopædia Iranica, 1987, vol. III, fasc. 1, pp. 35-44; published 15 December 1987; iranicaonline.orgThat the Avesta is a liturgical collection rather than a single book with fixed contents; its parts (Yasna, Visperad, Khorda Avesta, Sirōza, Yašts, Vidēvdād, fragments); the Pahlavi legend of the twenty-one nasks written under Vištāspa and destroyed by Alexander, and Kellens's flat rejection of it — "there never was an Avesta set down under the Achaemenids and destroyed or dispersed by the Greek invaders"; K7a, dated 1288, as the oldest surviving manuscript; L4 (1323) and K1 (1324) as the two ancient Pahlavi Vidēvdād manuscripts; Duchesne-Guillemin's estimate that only a quarter of the Sasanian collection is extant; and the outline of the Anquetil–Jones–Rask–Burnouf sequence in Avestan studies. Note that Kellens's survey of the manuscripts later in the same article prints "K7a (1278)", against the 1288 of his own opening paragraph.
  • Clarisse Herrenschmidt, "BURNOUF, EUGÈNE", Encyclopædia Iranica, 1990, vol. IV, fasc. 6, pp. 566-67; iranicaonline.orgBurnouf (1801-52) as "virtually the founder of Iranian linguistics", and the Commentaire sur le Yaçna (Paris, 1833) as the work in which "he established the basis for understanding Avestan" — a claim about method and comprehension, not about proving the manuscripts genuine; that he worked from only four Avesta manuscripts, Anquetil-Duperron's 1771 translation and the fifteenth-century Sanskrit version by the Parsi Neriosengh; that his constant reference to Vedic Sanskrit gave him the intuition of the Indo-Iranian branch; and that he relied heavily on Rasmus Rask, whose 1826 tract Om Zendsprogets og Zendavestas ælde og ægthed argued the age and genuineness of the texts.
  • Michael J. Franklin, "JONES, WILLIAM", Encyclopædia Iranica, 2009, vol. XV, fasc. 1, pp. 5-11; iranicaonline.orgJones's dates (London, 28 September 1746 – Calcutta, 27 April 1794); the Lettre à Monsieur A*** du P*** as his third publication of 1771; that its "stinging invective" "wrongly claimed that Anquetil-Duperron had been duped by the Parsis with modern forgeries" — so the charge was that Anquetil was a dupe, not himself the forger; that the provocation was Anquetil's complaint about the lack of heating in the Bodleian and his false charge that Thomas Hunt professed a knowledge of Avestan; and Franklin's judgement that this "clash of egotisms" "seriously retarded Zend-Avesta studies for many years".
  • Alberto Cantera, "VENDĪDĀD ii. Transmission of the Vīdēvdād in India", Encyclopædia Iranica, 2015, published 9 March 2015; iranicaonline.orgThe dated Pahlavi Vidēvdād manuscripts — "L4 in 1323; K1 in 1324" — both copied by Mihrābān Kayhusraw from the copy of Rōstām Mihrābān, itself descending from a manuscript written by Ardašīr ī Wahman ī Rōzweh in Sīstān in 1205 and carried to India about 1231; that every eighteenth- and nineteenth-century Pahlavi Vidēvdād derives from L4 or K1; and that Anquetil-Duperron's teacher Dastur Dārāb belonged to a reformist Surat school of copyists that had been actively emending its manuscripts since the visit of Jāmāsp Īranī from Kermān in the 1720s.
  • Alberto Cantera, "On the Edge between Literacy and Orality: Manuscripts and Performance of the Zoroastrian Long Liturgy", Oral Tradition, 2022, vol. 35, no. 2, pp. 211-50; open access at oraltradition.orgGeldner's figure that at least three quarters of the Great Avesta is lost (Geldner 1896a:20); that only the Vidēvdād survives in the arrangement the Dēnkard describes, with the Bayān Nask, Vištāsp Sāst and Stōd Yasn approximating it; that the Avestan alphabet has fifty-four letters and fifteen vowels and is phonetic rather than phonological, made "for the faithful reproduction of orally transmitted texts", with the footnote that most manuscripts do not use all of them; that in India the Yasna and Vidēvdād are still performed and recited from memory rather than read from a book, against the near-disappearance of the Long Liturgy in Iran; that the reign of Xōsrō I (531-579) is now widely accepted as the likeliest date for the script's invention, with Tremblay proposing 550-630, against Hoffmann and Narten's fourth-century date and Kellens's earlier post-Sasanian proposal; and the central argument that the manuscripts are "not automatic copies from a Sasanian archetype" but scions of a ritual learned by heart, the scribes being the same priests who recited without them.

Where this claim originates

Linked entries

  • Transmitted intoThe GāthāsAttested

    The Gāthās sit inside the Avesta rather than beside it: seventeen Old Avestan hymns placed within the Yasna at chapters 28-34, 43-51 and 53, surrounded by Young Avestan text several centuries younger in language.

  • Set againstThe Greek Magical PapyriAttested

    The magical papyri are the working handbooks of a practice that ended, preserved because Egypt's sand kept the paper. The Avesta is a late copy of a practice that never ended, preserved because priests kept the words.

  • Set againstThe DruidsAttested

    Caesar reported that the Druids held it wrong to commit their learning to writing, and none of it survives. The Zoroastrian priesthood also transmitted by memory for centuries — and then had an alphabet of some fifty letters built to catch the exact sound of what it said.

  • Set againstEnūma ElišAttested

    Both are misread by being read silently: the Avesta's texts exist to be recited in ceremony, and a Babylonian ritual text has this poem read aloud to Bel on the fourth of Nisannu. The difference is that the Avesta is a corpus of many texts and this is one fixed composition in seven tablets.

  • Set againstVelesA modern connection

    Two Indo-European religions, opposite evidential fates. Iran's survives in its own liturgical corpus, transmitted by a living priesthood; the East Slavic gods survive in two oath clauses and a list of idols written down by the church that pulled them down - which is why a forgery could be offered as the missing scripture.