Quick Answer
The Gathas are seventeen hymns in a language so poorly attested that no two scholars translate them the same way, and the differences are not nuance. Whole verses mean different things in different versions. Before buying a copy, know which translation you are getting and which of two incompatible schools produced it.
Key Takeaways
- Old Avestan is a few thousand words in total. There is no other literature in the language, many words appear once, and the grammar is partly reconstructed.
- Two schools produce two Zarathustras. One is an ethical monotheist reformer, the other an archaic ritual poet composing liturgical riddles. Your translation has already chosen.
- The cheap editions are Mills from 1887. Most inexpensive paperbacks reprint a Victorian translation whose assumptions have long since been set aside.
- Read two translations, not one. A single version gives a false sense that the text says something definite.
- Four things survive every reading: asha against druj, Ahura Mazda as wise lord, a moment of choice, and good thought, word and deed.
🕑 14 min read
What the Gathas Are
Seventeen hymns, traditionally held to be the words of Zarathustra himself, preserved inside the Yasna, the central Zoroastrian liturgy. They sit at Yasna chapters 28 to 34, 43 to 51, and 53, and they are grouped into five sets distinguished by poetic metre: the Ahunavaiti, Ushtavaiti, Spentamainyu, Vohukhshathra and Vahishtoishti Gathas.
Their position matters. They are not a freestanding book. They were embedded in a ritual text and carried inside it, which is why they arrive without narrative, without setting, and without the explanations a reader wants. They are the words of a liturgy whose performance has changed and whose original occasion is gone.
For the beliefs themselves, our guide to Zoroastrian belief covers Ahura Mazda, Angra Mainyu, the Amesha Spentas, the sacred fire and the renovation of the world in full. This article is about the text: what condition it is in, why translations differ, and which one to buy.
A useful comparison: imagine the Psalms surviving only inside a service book, in a language attested nowhere else, with no Hebrew Bible around them, and no other Hebrew literature to check the vocabulary against. That is roughly the situation of the Gathas, and it is the reason this article exists.
The Language Problem
Old Avestan is one of the hardest languages any translator works in, and the reason is quantity.
The entire surviving Old Avestan corpus is a few thousand words. The Gathas are most of it, along with a short liturgy and a handful of formulae. There is no chronicle, no legal text, no letter, no inscription, nothing else in the language at all.
What that means practically is that a great many words appear only once in the whole surviving language. When a word occurs once, there is no way to triangulate its meaning from other uses, and the translator has to reason from three unreliable sources: the shape of the word compared with Sanskrit cognates, the later Zoroastrian tradition's understanding of it, and what the passage seems to need.
Grammar is only partly better off. Old Avestan is closely related to the language of the Rigveda, close enough that a Vedic specialist can make real progress, and it is also its own language with its own developments. Reconstruction from a related language is a powerful method and it produces confident answers that are sometimes wrong.
The consequence is worth stating without softening. Specialists disagree about what the Gathas say, not about how best to render an agreed meaning into elegant English. Two competent scholars can produce versions of the same verse in which the subject, the object and the action are all different. A reader holding one translation is holding one scholar's reconstruction, and the confidence of the printed page conceals that entirely.
The Transmission Gap
The second problem is time, and the numbers are startling.
The Gathas were composed and then carried orally by priests, memorised and recited, for a very long period. Zoroastrian priestly transmission was rigorous and oral memorisation of liturgical texts can be remarkably stable, so this is not a reason to dismiss the text. It is a reason to know what kind of object it is.
Writing came late. The Avestan alphabet was created specifically to record these texts, in the Sasanian period, somewhere around the fifth or sixth century of the common era. That is well over a thousand years after the likely composition, possibly much more.
And the manuscripts are later still. The oldest surviving Avestan manuscripts date to the thirteenth and fourteenth centuries. Everything modern scholarship works from descends from copies made two and a half millennia or more after the hymns were composed.
A further wrinkle: the priests who recorded and copied the text no longer spoke the language. They were transmitting sounds with enormous care and without full comprehension, which preserves some things beautifully and corrupts others in ways that are hard to detect.
Two Schools, Two Zarathustras
This is the part that decides what you will read, and most editions do not tell you which side they are on.
The traditional approach reads the Gathas through the later Zoroastrian tradition, including the Pahlavi commentary known as the Zand, and through the understanding of the living communities that have carried the text. On this reading Zarathustra is a religious reformer: a prophet who preached one supreme god, an ethical choice between truth and the lie, personal responsibility, and a coming renovation of the world. The Gathas are his sermons and his prayers.
The comparative approach reads the Gathas by systematic comparison with the Rigveda, treating them as archaic Indo-Iranian ritual poetry and setting the later tradition aside as a much later interpretation. On this reading Zarathustra is a liturgical poet, the hymns are composed for performance at a sacrifice, and much of what looks ethical is technical ritual vocabulary. The riddling obscurity is deliberate, in the way Vedic hymns are deliberately obscure, because the poet is displaying skill before the gods.
Jean Kellens and Eric Pirart's Les textes vieil-avestiques is the most thoroughgoing version of the second approach, and the Zarathustra it produces is close to unrecognisable next to the traditional one.
The dispute is not a scholarly quibble and it is not settled. It touches whether Zoroastrianism began as an ethical monotheism, which would make it the earliest such religion and would give real weight to the influence arguments, or as a variant of the Indo-Iranian sacrificial religion whose ethical monotheism developed later. The traditional reading has the living community and continuous interpretation behind it. The comparative reading has linguistic rigour and freedom from later assumptions. Both are held by serious people, and a reader should know that the book in their hands has quietly picked one.
The Translations, Compared
Martin Haug (1858 to 1860). The first serious European attempt, and the source of the still-common idea that Zarathustra taught a strict monotheism later corrupted into dualism. Historically important, superseded.
L. H. Mills (1887). The Sacred Books of the East version, translated for Max Müller's series. This is the one you will find reprinted everywhere because it is out of copyright. Careful for its date and a century and a half old.
Christian Bartholomae (1905). In German, and the foundation of modern Avestan lexicography. His dictionary is still consulted. His readings lean on the comparative method.
Stanley Insler (1975). The most useful English scholarly translation for a general reader. Insler produces readable English, explains his choices, and does not pretend to more certainty than he has. If you buy one, buy this.
Helmut Humbach (1959, revised 1991). The revised edition, produced with Josef Elfenbein and Prods Oktor Skjaervo as The Gathas of Zarathushtra and the Other Old Avestan Texts, is the fullest apparatus available in English. The 1991 translations often differ substantially from his own 1958 versions, which tells you something useful about how unstable this field is.
Jean Kellens and Eric Pirart (1988 to 1991). In French, and the most rigorous comparative reading. Read it if you want to see the other Zarathustra.
Community translations. Versions produced by Zoroastrian scholars and organisations for their own communities are valuable for showing how the tradition understands itself, and they are not attempting philological neutrality.
The single most useful thing you can do: put two translations of Yasna 30, the hymn about the two spirits, side by side. It is the most quoted passage in the corpus and the one on which the whole question of Zoroastrian dualism turns. Comparing two versions of it will teach you more about the state of this text in twenty minutes than any amount of reading about it.
What the Cheap Editions Are
Search for the Gathas and you will find inexpensive paperbacks with plain covers, often from print-on-demand imprints, at low prices.
Almost all of them are Mills's 1887 translation, reset. Some are the Sacred Books of the East text with the introduction and notes stripped out, which removes the apparatus that made the original respectable. Some do not name the translator anywhere.
These are not frauds and they are not what you want. Mills was working before a century of Avestan scholarship, in a period when the comparative method was young, and with a Victorian sense of what a scripture ought to sound like. His version reads like the King James Bible because that is what a sacred text was supposed to read like in 1887.
Check before you buy: find the translator's name on the listing. If it is not there, or if it is Mills, and you want a current text, look elsewhere.
What Survives Every Reading
Four things hold across otherwise incompatible translations, and they are the actual content a reader can rely on.
Asha against druj. Asha is order, truth, the way things rightly are, both cosmic and moral. Druj is the lie, deceit, disorder. The opposition runs through the entire corpus and every translator has it. Asha is cognate with Vedic rta, which is one of the strongest pieces of evidence for the Indo-Iranian background.
Ahura Mazda. The wise lord, addressed directly and repeatedly, associated with light, with mind, and with asha. Whether he is the sole god or the chief of a class of beings is exactly what the two schools dispute.
A choice. The hearers are presented with a decision between the two ways, and the decision matters. Whether it is a cosmic ethical choice offered to all humanity or a narrower ritual and communal alignment is disputed, and that something is being chosen is not.
Good thought, good word, good deed. The triad appears in the later tradition as humata, hukhta and hvarshta, and the Gathic material behind it is present in every version.
The One Thing Only the Gathas Have
Set the difficulties aside for a moment, because there is something in this text that exists nowhere else in the ancient Iranian material and that no amount of translation trouble removes.
A voice speaks in the first person, and it is under pressure. The speaker asks Ahura Mazda direct questions and does not receive comfortable answers. He complains about his circumstances. He names his lack of cattle and men, which in a pastoral society is a statement of poverty and powerlessness. He asks how the righteous poor are to be protected from the violent rich. He asks who established the path of the sun and the stars, and the questions come in sequence, unanswered, like a man working something out aloud.
That register is unusual. Most surviving ancient liturgical poetry is confident, formulaic and communal. This has a person in it with a specific situation and specific grievances, addressing a god about them.
Even the comparative school, which reads the ethical content as ritual vocabulary, has to account for that voice. Whether it belongs to a historical prophet named Zarathustra, to a poetic persona conventional in the genre, or to a composite of several composers is disputed. What is not disputed is that it is there, and it is the reason the Gathas hold readers who would put down a liturgy.
Practice: read the questions and leave them open
Find the passage of questions in Yasna 44, where nearly every verse begins by asking Ahura Mazda about the origin of something: the path of the sun, the moon's waxing, who holds the earth below and the sky from falling. Read them in one pass without seeking answers. The text does not supply them. Sitting with a sequence of unanswered questions addressed to whatever you take to be ultimate is the oldest recorded form of a practice most contemplative traditions later formalised, and this may be the earliest example of it we have in any language.
The Dating Question
How old the Gathas are determines a great deal, including whether Zoroastrianism could have influenced anything else.
The linguistic argument places them early. Old Avestan stands very close to the language of the Rigveda, and if the Rigveda is placed somewhere in the second millennium BC, the Gathas belong in a similar range. Many specialists have argued for composition somewhere between roughly 1500 and 1000 BC, and the archaic pastoral society the hymns assume, with cattle, chariots and no cities, fits that.
The traditional argument places them much later. A Zoroastrian and Greek chronological tradition puts Zarathustra a few centuries before Alexander, which would place him around the seventh or sixth century BC and make him roughly contemporary with the Achaemenid rise.
The linguistic case is the stronger one and is now the majority position. The gap between the two datings is close to a thousand years, which is a reminder of how little is fixed here.
That uncertainty bears directly on the influence question. Parallels between Zoroastrian and later Jewish and Christian ideas about angels, cosmic dualism, judgement and resurrection are real, and Persia ruled Judaea for two centuries. Whether anything was borrowed, and in which direction, is difficult to establish precisely because the Avestan texts were written down so long after the period in question. Our piece on Zoroastrian belief takes that argument up in detail.
Who Should Read Them
Read them if you want the oldest layer of one of the oldest surviving religions, and can hold a text loosely.
Read them with a commentary and with a second translation open. This is the rare case where reading one version carefully is worse than reading two versions casually, because one version produces false confidence.
Read a modern account of Zoroastrianism first. The hymns assume everything and explain nothing.
Skip them if you want a wisdom text with usable teaching on the page. These are liturgical poems from a vanished ritual, and much of what a reader takes as profound compression is a passage nobody can construe.
Our position: an extraordinary survival in poor condition, worth reading and impossible to read confidently. Buy Insler for the English, put Humbach's revised edition beside it if you want the apparatus, and avoid the anonymous reprints of Mills. Then read Yasna 30 in two versions and notice that the founding text of the world's first great dualist religion cannot be securely translated. That fact is not a scandal. It is what a three-thousand-year-old oral text in a barely attested language actually looks like when you get close to it.
For the surrounding Persian material, see the Shahnameh for how Iran remembered its own past, and Mithraism for a Roman cult that borrowed an Iranian name and did something entirely its own with it.
The oldest layer, in the least readable condition.
Affiliate disclosure: Thalira earns a small commission on purchases made through the link above, at no additional cost to you. We recommend books we have read and think are worth your time, and we say plainly when we think parts of them are weak.
Frequently Asked Questions
What are the Gathas?
Seventeen hymns in Old Avestan, traditionally attributed to Zarathustra himself and preserved inside the Yasna liturgy at chapters 28 to 34, 43 to 51, and 53. They are grouped into five sets by poetic metre and are the oldest surviving layer of Zoroastrian scripture.
Why do translations of the Gathas differ so much?
Because Old Avestan is barely readable. The entire surviving corpus is a few thousand words, there is no other literature in the language to compare against, many words appear only once, and the grammar is reconstructed. Translators fill those gaps from different assumptions, and the assumptions produce different texts.
What are the two schools of interpretation?
One reads the Gathas through the later Zoroastrian tradition and its Pahlavi commentary, producing an ethical monotheist reformer. The other reads them by comparison with Rigvedic Sanskrit as archaic Indo-Iranian ritual poetry, producing a liturgical poet composing riddling hymns for a sacrifice. The choice determines the meaning.
Which translation should I read?
Insler's 1975 version for a readable scholarly English text. Humbach's revised 1991 edition with Elfenbein and Skjaervo for the fullest apparatus. Kellens and Pirart for the comparative Vedic reading, in French. Read at least two, because reading one gives you a false sense of what the text says.
Are the cheap public domain editions any good?
They are usually reprints of L. H. Mills's translation from the Sacred Books of the East series of 1887, which is more than a century out of date and reflects assumptions later scholarship abandoned. They are legitimate historical documents and they are not what a reader wanting the current understanding should buy.
How old are the Gathas?
Contested. The language is archaic and closely related to Rigvedic Sanskrit, which leads many linguists to place composition somewhere between roughly 1500 and 1000 BC. A later Zoroastrian and Greek tradition dates Zarathustra to a few centuries before Alexander. The linguistic argument is the stronger one.
When were they written down?
Very late. The Gathas were transmitted orally by priests for a very long period, and the Avestan alphabet was invented for the purpose in the Sasanian era, around the fifth or sixth century AD. The oldest surviving manuscripts are from the thirteenth and fourteenth centuries.
What survives every translation?
The opposition of asha and druj, order and the lie. The address to Ahura Mazda as the wise lord. A moment of choice presented to hearers. And the triad of good thought, good word and good deed. Those are stable across otherwise irreconcilable versions.
Did Zoroastrianism influence Judaism and Christianity?
The question is genuinely open. Parallels in angelology, cosmic dualism, judgement and resurrection are real, the Persian empire ruled Judaea for two centuries, and the direction and timing of any borrowing are hard to establish because the Avestan texts were written down so long after the period in question.
Do I need the Gathas to understand Zoroastrianism?
They are the oldest layer and they are not the whole religion. Later Avestan texts, the Pahlavi literature and living Parsi and Iranian Zoroastrian practice all contribute, and a great deal of what people mean by Zoroastrian belief comes from those rather than from these seventeen hymns.
Are the Gathas difficult to read?
Extremely, in any translation. They are compressed, allusive, full of technical ritual vocabulary, and they assume a context no longer available. Readers expecting the plainness of a wisdom text find something closer to fragmentary liturgical poetry with the situation removed.
Where should a newcomer start instead?
With a modern account of Zoroastrian belief and history, then a good translation of the Gathas with commentary alongside it. Starting with the primary text alone is the reliable way to come away confused or, worse, confident about something a different translator would render another way.
Put two versions of one hymn side by side
Take Yasna 30, find two translations of it, and read them in parallel. You will see the same verse assert different things. That experience is not discouraging once you understand it: it is what honest contact with an ancient text looks like, and it will change how you read every translated scripture you pick up afterwards, including the ones that feel settled.
Sources & References
- Insler, S. (1975). The Gathas of Zarathustra. Acta Iranica 8, Brill.
- Humbach, H., with Elfenbein, J. and Skjaervo, P. O. (1991). The Gathas of Zarathushtra and the Other Old Avestan Texts. Winter.
- Kellens, J. and Pirart, E. (1988-1991). Les textes vieil-avestiques. Reichert.
- Bartholomae, C. (1904). Altiranisches Wörterbuch, and (1905) Die Gatha's des Awesta.
- Mills, L. H. (1887). The Zend-Avesta, Part III. Sacred Books of the East vol. 31, Oxford.
- Boyce, M. (1979). Zoroastrians: Their Religious Beliefs and Practices. Routledge.
- Encyclopaedia Iranica. Entries on the Avesta, on the Gathas and their texts, and on Gatha translations. iranicaonline.org.