The CAB project (Corpus Avesticum Berolinense), carried out at the Institute of Iranian Studies of the Freie Universität Berlin, represents the first attempt to edit the entire Avestan Zoroastrian textual corpus within its ritual context, and to produce and offer both printed and digital editions of the texts.

To date, no single edition of Avestan texts has taken into consideration the actual context of the Zoroastrian ceremonies; their successive editors (Spiegel, Westergaard, and Geldner) delivered these texts deprived from their ritual contexts and from the performative information that accompanies them in the manuscripts. By contrast, CAB will present the Avestan texts in their ritual context and include the performative information found in the manuscripts and in the meta-ritual literature, so that the edition also functions as a historical description of the ritual practice of the Zoroastrian communities.

The Corpus Avesticum Berolinense is a natural continuation of the Avestan Digital Archive (ADA), a project on which a part of the CAB team has already worked for years. Since 2006, ADA shifted its focus towards the search for undiscovered manuscripts, while still working, in parallel, to digitize the copies that were already known. Therefore, the editorial principles that will be applied to this new edition of the entire body of Avestan ritual texts are the result of a complete revision and a new understanding of the transmission process and of the newly uncovered materials. The following are the most important conclusions of this revision that affect the edition of the Avestan texts:

  1. All Avestan texts (with the exception of the meta-ritual treatises Hērbadestān and Nērangestān) are essentially ritual texts, whose current form was most likely already arranged by the fourth century BC. These rituals can only be understood in connection to their performative context and by taking into account the performative variation they offer.

  2. The Avestan script was created in order to spread a homogenous recitation of the Avestan texts, as a faithful reproduction of how the rituals were performed in Western Iran in the sixth century AD. This was also designed to be the normative recitation in the entire Sasanian Empire.

  3. The manuscripts have always kept this double function: the exact representation of the recitation and the normative one. Single manuscripts take different positions regarding this tension. Accordingly, some manuscripts adopt innovations in the recitation when they become canonical, so that they are not true copies of primitive rituals, but descriptions and libretti for the actual correct performance of the rituals

  4. Due to the changes in the performance and in the manner of recitation over space and time, and due to the manner in which they were adopted in the manuscripts, a reconstruction of a Sasanian archetype is not only impossible, but also not desirable for its ahistorical nature (since the exact form of the ritual cannot be determined, despite the great help of the Nērangestān).

  5. Taking all of these points into consideration, the edition will therefore represent the rituals as they were performed at certain point in time and in a certain place. This must be decided depending on the transmission of each ritual. In addition to this, numerous apparatuses will provide the reader with necessary information for a reconstruction of the performance of the ritual at other historical times and in other places.


Besides the edition, the project will provide even the non-specialist readers with tools for an assisted reading of the Avestan texts, including translations, grammatical and semantic analysis of words, etc. Amongst these tools, the project will offer a digital version of the standard dictionary of the Avestan language (Bartholomae’s Altiranisches Wörterbuch).

The first steps to be accomplished in the short-term are be the following:

  • A “static” version of the ceremonies: each single ritual will be displayed as it was performed in a concrete context (day, month…), as they appear in the canonical editions.

  • A ceremony generator: a tool that will allow the user to automatically create a ritual within a specific context (part of the day, date, divinity to which the ceremony is addressed, festivity, etc.).

In successive steps, the CAB team will add grammatical, semantic, and ritual information to the analysis of each stanza. Moreover, parallels in the Avestan, Vedic, and Middle Persian corpora will be systematically included. As a final step, we will provide a selective and a complete text critical apparatus and access to the pictures and transliterations of all the available manuscripts. Thus, we hope to offer a reliable edition for the specialist and a useful approach to the Avestan texts for colleagues from neighbouring areas or fields of study (e.g. religious, historical, or linguistic studies). In order to ensure the durability and exchangeability of the tools that will be produced, the digital edition, the dynamic ceremonies, the grammatical information, as well as the additional information will be encoded following the guidelines of the XML-TEI system.

The CAB team is composed not only of experts in Iranian linguistics, working on the editorial and ritual field, but also of IT support, programming the software for each individual tool. Finally, a group of students is engaged with the transliteration and encoding of the Avestan texts, providing a great part of the basic information with which other tasks can be accomplished.


In contrast to the previous editions, CAB’s goal is to represent the Avestan texts as they are recited in the ritual performances. This means, among others, that we present the texts as they appear in the liturgical manuscripts, although previous editions rely mainly on the exegetical manuscripts (especially for the Long Liturgy). Accordingly, the textual setup is for many rituals quite different from the one of the previous editions. Thus, the problem of numbering arises. While indexing manuscripts for the Avestan Digital Archive, we already noticed that the use of the standard numbering by Geldner need to be completed, since many texts included in the liturgical manuscripts were not reflected in Geldner’s edition. As such, we were forced to invent a new numbering for such ceremonies like the Visperad or Vīdēvdād.

Furthermore, the internal analysis of the texts and the way how the texts indicate divisions, sections, beginnings and closings has shown us that Geldner’s (or rather Westergarrd’s) decisions were inconsequent and sometimes opposite to the indications given by the Avestan text itself and by the tradition. Therefore, we felt the necessity to introduce some changes in the system so that it is more consequent and more respectful of the tradition.

However, we are conscious of the difficulties that it supposes to change the system of references to the Avestan texts and the problems it might cause for using old literature, etc., although we always indicate Geldner’s numbering as well. Therefore, we have not acted with complete freedom, but have tried to keep our numbering as close to Geldner’s at least for the text that he has edited completely or almost completely, for example the Yasna. Consequently, we have not changed the number of the hāiti so that the Old Avestan texts maintain basically the same numbering like in Geldner’s edition. This has led us to strange situations like chapter 23 of the Yasna consisting of only one Frauuarāne. Perhaps we should have been more radical, but we did not dare, because we did not want to make the use of older materials impossible or very difficult.

A short explanation about the main parameters we use in our numbering system can be read here.