The Gathas are 17 Avestan hymns believed to have been composed by Zarathusthra (Zoroaster) himself. They form the core of the Zoroastrian liturgy (the Yasna). They are arranged in five different modes or metres. The Avestan term gāθā ("hymn", but also "mode, metre") is cognate with Sanskrit gatha, both from the Indo-Iranian root *gā "to sing".
The Gathas are in verse, metrical in the nature of ancient Iranian religious poetry, which is extremely terse, and in which grammatical constructs are an exception.
The 17 hymns of the Gathas consist of 238 stanzas, of about 1300 lines or 6000 words in total. They were later incorporated into the 72-chapter Yasna (chapter: ha or had, from the Avestan ha'iti, 'cut'), which in turn is the primary liturgical collection of texts within the greater compendium of the Avesta. The 17 hymns are identified by their chapter numbers in the Yasna, and are divided into five major sections:
With the exception of Ahunavaiti Gatha, which is named after the Ahuna Vairya prayer (Yasna 27, not in the Gathas), the names of the Gathas reflect the first word(s) of the first hymn within them. The meter of the hymns is historically related to the Vedic tristubh-jagati family of meters. Hymns of these meters are recited, not sung.
The sequential order of the Gathas is structurally interrupted by the Yasna Haptanghaiti ("seven-chapter Yasna", chapters 35–41, linguistically as old as the Gathas but in prose) and by two other minor hymns at Yasna 42 and 52.
The language of the Gathas, Gathic or Old Avestan, belongs to the old Iranian language group which is a sub-group of Eastern families of the Indo-European languages. The dependency on Vedic Sanskrit is a significant weakness in the interpretation of the Gathas, as the two languages, though from a common origin, had developed independently. Sassanid era translations and commentaries (the Zend) have been used to interpret the Gathas, but by the 3rd century the Avestan language was virtually extinct, and a dependency on the medieval texts is often discouraged as the commentaries are frequently conjectural. While some scholars argue that an interpretation using younger texts is inadvisable (Geldner, Humbach), others argue that such a view is excessively skeptical (Spiegel, Darmesteter). The risks of mis-interpretation are real, but lacking alternates, such dependencies are perhaps necessary.