Avesta

The Avesta is the name for the sacred texts of Zoroastrianism, one of the oldest religions still practised in the world. It is not a single book written by one author at one time, like a modern novel. It is a collection of prayers, hymns, ritual texts and religious laws, composed over many centuries and preserved by a community of priests.

The texts are in Avestan, an ancient Iranian language that had long ceased to be anyone’s everyday speech by the time the texts were written down in the Sasanian period. Together with Old Persian, the language of the Achaemenid royal inscriptions such as those carved at Persepolis, Avestan is one of the two known Old Iranian languages, and the Avesta is the only evidence for it. Its oldest parts are thought to be older than the Old Persian inscriptions.

Most of what survives is liturgy: words meant to be recited aloud by priests during ceremonies, or by ordinary believers in their daily prayers. This is an important key to understanding the Avesta. It survived not mainly as a library book but as a living voice, repeated in worship generation after generation.

Today the Avesta is still used by Zoroastrians in Iran, by the Parsi community of India, and by Zoroastrians who have settled in many other countries. For them it is scripture: a holy text whose words are believed to carry spiritual power. For historians and linguists it is also a precious window onto the religion, society and languages of ancient Iran and Central Asia.

Zoroastrians revere Zarathustra (known in Greek and later in European languages as Zoroaster) as the prophet of their religion. According to their tradition, he received a revelation from Ahura Mazda, the “Wise Lord”, and taught people to follow truth, to reject deception, and to care for the good creation. The Gathas themselves name a patron, Kavi Vishtaspa, and later tradition tells how this king accepted Zarathustra’s teaching and helped it spread.

At the heart of the faith is the belief that the world is a place of struggle between good and evil. Ahura Mazda is the creator of everything good. Opposed to him is Angra Mainyu, the destructive spirit. Every person is free to choose which side to support, and good choices are expressed through good thoughts, good words and good deeds — a famous threefold formula in Zoroastrian teaching.

Zoroastrians believe that the Gathas, the oldest hymns in the Avesta, were composed by Zarathustra himself, and many scholars, such as Helmut Humbach, agree. Others, notably Jean Kellens, have questioned whether Zarathustra was their author at all, as a later section of this page explains. Both sides agree that the Gathas are the oldest part of the collection and the foundation on which the rest of the tradition was built.

Because Zoroastrianism is a living religion, it is important to remember that questions about dates and authorship are scholarly questions. They do not decide what the Avesta means to believers, for whom it is the word of their faith recited in their homes and fire temples.

The Avesta that exists today is usually described as having a few main parts. The most important is the Yasna, a text of 72 chapters recited during the central Zoroastrian act of worship, also called the Yasna. In the middle of the Yasna lie its oldest pieces: the five Gathas and a short prose text called the Yasna Haptanghaiti, the “Worship in Seven Chapters”.

The Visperad (from Avestan vīspe ratavō, “all the ratus”, that is, all the spiritual masters) is a text of 24 chapters. It was never meant to stand alone: its chapters are inserted into the Yasna to make a longer, more solemn ceremony. The Vendidad is numbered in 22 chapters called fargards (the original chapter 12 was already lost by the 9th century; some manuscripts contain an 18th-century replacement). It contains an account of the lands created by Ahura Mazda, the legend of the first king Yima, and many rules about purity, illness, death and the care of fire, water, earth and animals.

The Yashts are 21 hymns in praise of particular divine beings, called yazatas (“ones worthy of worship”): for example the waters, the sun, the moon, Mithra (the guardian of contracts and truth), and the fravashis, the guardian spirits of the righteous. Many of them contain vivid stories about heroes and legendary kings of the Iranian past.

Finally there is the Khordeh Avesta, the “Little Avesta”, a prayer book for the faithful in everyday life. It gathers short prayers such as the five Niyayishns (praises of the sun, Mithra, the moon, the waters and fire), the Gahs (prayers for the watches of the day) and blessings used on special occasions; the Sirozeh, invocations of the divinities of the thirty days of the month, is usually transmitted alongside it. Together, these parts make up a book of worship more than a book of history.

The Gathas are five poems, containing 17 hymns in all, arranged in the Yasna according to their metre. The Ahunavaiti Gatha (Yasna 28–34) has seven hymns; the Ushtavaiti Gatha (Yasna 43–46) has four; the Spentamainyu Gatha (Yasna 47–50) has four; the Vohukhshathra Gatha (Yasna 51) and the Vahishtoishti Gatha (Yasna 53) have one each. Together they contain 238 stanzas (100, 66, 41, 22 and 9). A medieval Zoroastrian count, which Helmut Humbach calls approximately correct, gives about 5,570 words for the Gathas together with the Yasna Haptanghaiti, which sits between the first and second Gatha — a small core for such an influential tradition.

The language of the Gathas is highly compressed, allusive poetry, full of rare words and complex grammar. The speaker often addresses Ahura Mazda directly, asking questions and seeking guidance. Central ideas include asha (truth, right order) and its opposite druj (the lie, deception); the free choice every person must make between them; and the “bounteous” powers associated with Ahura Mazda, such as Good Thought (Vohu Manah) and Truth, which later tradition developed into the Amesha Spentas, the “Beneficent Immortals”.

The Gathas also condemn the daevas and those who worship them. In older Indo-Iranian religion the daevas were a class of gods (compare the Vedic devas). The Gathas reproach them for lacking true discernment and for having accepted the bad religion, but do not yet treat them as demons; in the Young Avesta and later Zoroastrianism they became demons. Many scholars read this rejection as a sign of religious reform, although, as Kellens and Herrenschmidt stress, the evidence is patchy and ambiguous.

Just before the Gathas, Yasna 27.13–15 contains three great prayers: the Ahuna Vairya and the Ashem Vohu in Old Avestan, and the Yenghe Hatam in archaized Young Avestan. The greatest is the Ahuna Vairya (Middle Persian Ahunwar), a stanza of 21 words recited in many prayers, including every retying of the sacred cord. Its precise overall sense remains debated. One possible rendering, given by Christopher Brunner, begins: “As a temporal lord (is) chosen, so (is) a spiritual lord, according to Truth …”. For believers its power lies as much in faithful recitation as in a single fixed translation.

Most of the surviving Avesta is in Young Avestan, a later form of the language. Old Avestan and Young Avestan are not simply grandparent and grandchild: they seem to reflect different dialects as well as different periods, but scholars treat them as varieties of one language. Young Avestan texts are more numerous, more varied, and less tightly fixed in wording than the Gathas.

The Young Avesta shows a religion that has grown and absorbed older traditions. Divinities barely mentioned in the Gathas, such as Mithra, the waters (personified as Aredvi Sura Anahita) and the sacred plant Haoma, receive long hymns. Some Yashts preserve heroic legends — dragon-slayers, rival kings, and a royal “glory” (khvarenah) that passes from worthy ruler to worthy ruler. Many of these figures reappear, transformed, over a thousand years later in Persian epic, above all Ferdowsi’s Shahnameh.

Young Avestan texts also mention real places in the eastern Iranian world. The first chapter of the Vendidad lists sixteen lands created by Ahura Mazda, each afflicted by an evil that Angra Mainyu set against it. Gherardo Gnoli identifies most of them with regions of eastern Iran and the Indo-Iranian borderlands — among them Sogdiana, Margiana (Merv), Bactria, Areia (Herat), Gandhara, the Helmand region and the “Seven Rivers” of the Panjab. Some identifications remain debated, notably that of Ragha, which Gnoli distinguishes from Median Raga (modern Rey), and the list mixes geography with mythical features such as the Aryan homeland, Airyana Vaejah. The identifiable geographical background is predominantly eastern Iranian, although some place-names remain disputed. No Achaemenid king is named in the Avesta, while the Gathas give no clear clue to where Zarathustra lived.

The Vendidad deserves special mention. Its name comes from an Avestan phrase meaning “the law repudiating the daevas”. Much of it takes the form of questions from Zarathustra and answers from Ahura Mazda. It sets out rules to protect the good creation from pollution, especially from contact with death, and it gives remarkable attention to the care of dogs, water and fire. It was the nineteenth of the 21 volumes of the Sasanian Avesta, and the only one to survive almost intact, probably because it was recited in a night-time ceremony and dealt with everyday questions of purity.

To understand the Avesta, it helps to picture the Yasna ceremony. It must be performed daily, in the morning watch, by qualified priests only — in present practice usually two priests — in a designated space within a fire temple. They recite all 72 chapters while preparing and offering haoma (a sacred drink made from a plant) and other offerings; a full recitation takes about two and a half hours. Its ultimate goal, in Zoroastrian understanding, is to maintain the integrity of the good creation of Ahura Mazda. Near its end come the āb-zōhr, the offerings to the waters (Yasna 63–69), followed by closing invocations.

Longer ceremonies interweave the Visperad and, at night, the Vendidad with the Yasna. The nocturnal Vendidad ceremony, meant to protect against demonic attack, shows that even the book of laws was used as sacred recitation. This is why current editing projects such as the Corpus Avesticum Berolinense present the surviving Avesta as the texts of the rituals, edited ceremony by ceremony, rather than as the random remnant of a lost library.

The Avesta also shapes daily life. At initiation (navjote among the Parsis, sedra-pushun in Iran), every Zoroastrian boy or girl puts on a white undershirt (sudreh) and ties over it the sacred cord, the kusti, which is wrapped three times around the waist. The cord is untied and retied with prayers in Avestan and Pazand several times a day, for example on waking. It is woven of 72 threads, which tradition links to the 72 chapters of the Yasna. Short prayers such as the Ahuna Vairya and the Ashem Vohu, a praise of asha, are part of every retying.

Recitation must be precise. In Zoroastrian thought, sacred words (manthras) are effective when spoken correctly, with the right sounds and cadence. This concern for precise sound is the reason the texts were preserved orally for so long — and, as we shall see, why a special alphabet was eventually invented to write them.

For many centuries the Avesta existed only in the memories of priests. No Iranian people seems to have used writing in early times, and the old Indo-Iranian tradition gave pride of place to precise oral transmission and to learning sacred texts by heart. The Old Avestan texts appear to have been fixed in wording very early, while Young Avestan material was composed and reshaped over a longer period. Even after the Old Persian cuneiform script was created, probably under Darius I, there is no evidence that the Achaemenids wrote the Avesta down.

Later Zoroastrian tradition, preserved in Middle Persian books such as the Denkard, tells that a complete written Avesta existed in ancient times but was destroyed or scattered when Alexander the Great conquered Iran, and that later kings — an Arsacid king Valakhsh, then the Sasanians Ardashir, Shapur I and Shapur II — gathered and restored it. Modern scholars treat this story with caution: as Jean Kellens puts it, there never was an Avesta set down under the Achaemenids and destroyed by the Greeks. The story may nonetheless preserve a memory of a real break in the religious tradition after the Greek conquest.

The decisive step came under the Sasanian dynasty (224–651 CE), the last great pre-Islamic Iranian empire, which supported Zoroastrianism. Priests created the Avestan alphabet on the basis of the Book Pahlavi script used for Middle Persian, in which several sounds shared one letter and most vowels were not written. The new alphabet has a separate sign for every vowel distinguished in recitation — perhaps on the model of Greek — with 14 (or 16) vowel letters and 37 consonant letters, written from right to left. It was designed to record the liturgical recitation with extreme precision. Karl Hoffmann proposed a 4th-century date for the alphabet, perhaps under Shapur II. The inscription on a sarcophagus lid found in Istanbul, which can be dated no later than about 430 CE, already shows the characteristic forms of the Pahlavi cursive script on which the Avestan alphabet drew.

Later Middle Persian tradition describes the resulting Sasanian Avesta, with its Middle Persian translation and commentary, the Zand, as a collection of 21 nasks (volumes), matched to the 21 words of the Ahuna Vairya prayer, each word naming one nask. How this collection relates to the surviving ritual texts is still debated. Tradition also places a revision of the Middle Persian translation under Khosrow I (531–579). After the Arab conquest and the fall of the Sasanians in 651, the Zoroastrian community gradually became a minority and much of this edition was lost. We know about its contents mainly from the summary in Book 8 of the Denkard, a Middle Persian compendium of Zoroastrian knowledge begun in the early 9th century and completed in the 10th.

What survived the losses were mainly the texts that priests needed for ritual and prayer, copied again and again by hand. Karl Hoffmann reconstructed a common “base manuscript” of about the 9th–10th century behind the surviving tradition; more recent scholarship, including the editors of the Corpus Avesticum Berolinense, doubts that a single Sasanian archetype can or should be reconstructed and stresses how ritual practice shaped each manuscript. The oldest surviving manuscripts in the Avestan script date from the late 13th and 14th centuries: the date of the oldest, K7a in Copenhagen, is read variously as 1268, 1278 or 1288, and important Yasna and Vendidad copies are dated 1323–1324. Older still is a Zoroastrian fragment in Sogdian script from the Dunhuang library cave (British Library, Or.8212/84), whose first lines give the Ashem Vohu prayer in an archaic form. How long the gap is between composition and the surviving written witnesses therefore depends on which text and manuscript are compared.

Manuscripts were copied in two main communities: the Zoroastrians who remained in Iran, especially around Yazd and Kerman, and the Parsis, descendants of Zoroastrians who settled in Gujarat in western India after the Islamic conquest. Scholars classify the manuscripts by the text or liturgy they contain (Yasna, Visperad, Vendidad and so on), by origin (Iranian or Indian), and by whether they add a Middle Persian (Pahlavi) or Sanskrit translation. Manuscripts that contain only the Avestan text are called sāda (“pure”).

Europe learned the Avesta through the French scholar Abraham Hyacinthe Anquetil-Duperron. He sailed in February 1755, reached India in August of that year, and arrived at Surat in March 1758. There the Parsi priests (dasturs) Kaus and Darab gave him a copy of the Vendidad and taught him, through Persian, what they knew of Avestan; he later obtained a better copy from another dastur, Mancherji. In 1762 he deposited about 180 manuscripts — Avestan, Pahlavi, Persian and Sanskrit — in the King’s Library in Paris, and in 1771 he published a three-volume French translation, the Zend-Avesta.

The reception was stormy. The young William Jones, later famous as a pioneer of Indo-European comparative linguistics, immediately accused Anquetil of forgery, and many German scholars agreed. Anquetil was vindicated in the 1790s by Tychsen and Heeren, and in the 19th century Rasmus Rask and Eugène Burnouf showed that Avestan was an ancient Iranian language closely related to Sanskrit. Older books often call the language “Zend”, but this is a misunderstanding: Zand refers to the Middle Persian translation and commentary, not to the Avestan text itself.

Within Zoroastrianism, the Avesta remains the heart of worship. Priests are still trained to recite it from memory; families still pray with the Khordeh Avesta; fire temples in Iran and India continue ceremonies built around Avestan texts. For a community that has endured conquest, minority status and migration, the continuous recitation of these words is a powerful thread of identity.

For linguistics, the Avesta was essential in reconstructing the Indo-Iranian branch of the Indo-European family: Avestan and Old Persian form its Iranian side, closely related to the Sanskrit of India. Old Avestan and the Vedic Sanskrit of the Rigveda are especially close in grammar and poetic language. Shared vocabulary — Avestan haoma and Vedic soma, daeva and deva, ahura and asura — reveals a common inheritance, even as the two traditions came to value some of these terms in opposite ways.

For the history of religion, the Avesta is a key source for ideas that many scholars have compared with later traditions: a single supreme good creator, a cosmic struggle between good and evil, personal moral choice, individual judgment after death, and a final renewal of the world. How far, and through which routes, these ideas influenced Judaism, Christianity and Islam is debated; the Avesta is central evidence in that debate.

For Iranian culture more broadly, the legends of the Young Avesta — of Yima, the dragon-slayers and the Kayanian kings — flowed through Middle Persian literature into the Persian national epic. Modern Iranian nationalism and European thought have also drawn on the figure of Zarathustra, sometimes in ways far removed from the texts themselves. Nietzsche, for example, used Zarathustra as a literary voice for his own philosophy; it is not even certain that he had read Anquetil-Duperron’s translation, so direct dependence on the Avesta should not be assumed.

No event mentioned anywhere in the Avesta can be tied to a historically verifiable chronology, so every date is an inference. Middle Persian sources date Zarathustra by a figure of “258 years before Alexander”, which would place him in the 7th–6th century BCE. In the Bundahishn the 258 years are the sum of the reigns from King Vishtaspa’s acceptance of the religion down to Alexander, so the figure is not a birth date; it comes from a schematic chronology of world history whose context, as William Malandra notes, hardly inspires confidence, although scholars such as W. B. Henning and Ilya Gershevitch defended it. Greek authors, meanwhile, placed Zoroaster absurdly early: Xanthos of Lydia, as cited by Diogenes Laertius, put him 6,000 years before Xerxes’ campaign against Greece, perhaps confusing the 3,000-year ages of the Zoroastrian world calendar.

The main modern argument is linguistic. Old Avestan is very archaic and closely resembles the language of the oldest hymns of the Rigveda, whose own date is uncertain but which mark the earliest plausible limit. The distance between Old and Young Avestan implies several centuries. The Achaemenid empire marks the other limit: not a single Achaemenid king is mentioned in the Avesta, and its geography is eastern Iranian. Serious proposals therefore fall between the mid-2nd millennium and the early 6th century BCE. If anything approaching a consensus exists, it places Zarathustra and the Gathas around 1000 BCE, give or take a century.

The place of Zarathustra’s activity is also debated. The Gathas themselves give no clear clue, but every passage of the Avesta that sheds light on its geography points to eastern Iran and the Indo-Iranian borderlands. Linguistic evidence distinguishes Avestan from Old Persian but does not securely locate its homeland: Avestan shows few of the features that later distinguish the Eastern Iranian languages. Proposals include Chorasmia (Henning), Sogdiana (Gershevitch), Margiana–Bactria (Humbach) and Sistan (Gnoli); Mary Boyce’s attempt to place him on the Inner Asian steppes has been sharply criticised. Later traditions that place Zarathustra in Azerbaijan and Media are generally set aside in favour of eastern Iran.

More radically, Jean Kellens and Éric Pirart (1988) denied that Zarathustra was the author of the Gathas, pointing out that his name appears in the third person or is addressed in the vocative, and reading the Gathas as the expression of a whole religious group; Prods Oktor Skjærvø has supported this view. Others, including Helmut Humbach and William Malandra, continue to regard Zarathustra as a historical figure and the poet of the Gathas. The debate illustrates a general point: the tradition offers a coherent biography, whereas critical scholarship can reconstruct only a probable linguistic and cultural setting.

Critical editing of the Avesta began in the 19th century. Niels Ludvig Westergaard, a follower of Rasmus Rask, published a complete edition in 1852–1854; Friedrich Spiegel, who relied on the native Middle Persian commentary, edited the Avestan texts together with their Pahlavi versions. The standard edition remains that of Karl Friedrich Geldner (1886–1896), based on an analysis of more than 120 manuscripts and covering the Yasna, Visperad, Vendidad, Yashts and Khordeh Avesta. James Darmesteter’s French translation (1892–1893) follows the Pahlavi tradition; Fritz Wolff’s German translation of 1910, based on Christian Bartholomae’s dictionary, is generally more reliable.

In 1902 Friedrich-Carl Andreas proposed that the Avesta had first been written in an ambiguous Pahlavi-type script in the Parthian (Arsacid) period and later clumsily transcribed into the Avestan alphabet, so that editors should reconstruct how each word had been written in this “Arsacid archetype”. For some forty years scholars followed this method, until W. B. Henning, H. W. Bailey and Georg Morgenstierne undermined it around the Second World War. As Jean Kellens summarises, no form in the transmitted text can be explained with certainty as a transcription error. Karl Hoffmann’s work then placed the written canon in the Sasanian period: the Avestan alphabet was invented to record the liturgical recitation of that time, not the speech of Zarathustra.

This has important consequences. Hoffmann’s model aimed to reconstruct the phonetic shape of a Sasanian archetype — itself more than a millennium younger than the Gathas — but Geldner’s edition rests almost exclusively on Indian manuscripts, and according to Alberto Cantera textual criticism can take us back no further than the later “hyparchetypes” behind the surviving manuscripts. Whether a single Sasanian archetype can be reconstructed at all, and whether that should be the goal of an edition, is therefore disputed. Working directly with the manuscripts, Cantera and his team also found Geldner’s edition less reliable than commonly assumed, both in its choice and display of readings and, above all, in its inaccurate presentation of the liturgies as they appear in the manuscripts.

Current projects put this “liturgical turn” into practice. The Avestan Digital Archive, begun by Cantera in 2007 at the University of Salamanca and now based at the Freie Universität Berlin, aims to make all available Avestan manuscripts of the Zoroastrian liturgies accessible online; since about 2010 it has brought many previously little-known manuscripts from Iran to light. The Corpus Avesticum Berolinense regards the reconstruction of a single Sasanian archetype as neither possible nor desirable, and instead edits the Zoroastrian rituals in Avestan in their ritual context, as they were performed and written down in the manuscripts. Open questions include the precise date and place of the alphabet’s creation, the relation between the 21-nask collection described by later tradition and the surviving liturgies, and how to translate Gathic poetry whose meaning often remains contested.