In 1928 a farmer on the Syrian coast struck a tomb, and excavation at the mound of Ras Shamra the following year uncovered the city of Ugarit, destroyed around 1200 BCE and undisturbed since. Among its buildings were archives written in a previously unknown alphabetic cuneiform, deciphered within a year of discovery. They include mythological narratives, ritual and offering lists, god-lists, letters and administrative records, and they remain the only substantial body of Canaanite religious literature in a Canaanite language.
This was a transformation, and it is also a trap. Before 1929 the religion was known almost entirely through the writings of its opponents. Since then it has been possible to read a West Semitic tradition in its own words. But Ugarit is one city, at the northern edge of the region, in one period, and its texts were produced by and for a royal establishment. Practice at Hazor, Megiddo, Byblos or Jerusalem, three centuries later, need not have resembled it, and the habit of describing Canaanite religion in general from Ugaritic evidence flattens several centuries and several hundred miles into a single picture.
The ritual texts are in some ways more valuable than the famous narratives, because they record what was done: which god received which animal on which day, what part the king played, how purification worked. They also reveal a pantheon far larger than the handful of names that reach general accounts, and a system of offerings closely comparable in vocabulary and procedure to the sacrificial law preserved in the Hebrew Bible, a similarity that has driven a great deal of the scholarship on both.
The other evidence is thinner: Phoenician and Punic inscriptions, personal names built around divine elements, Egyptian and Mesopotamian references, excavated temples and figurines, and a very late Greek account attributed to Philo of Byblos, which claims to translate a Phoenician source and is of contested value.