A religion known from the outside
Nobody can write a complete account of Celtic religion. The Iron Age peoples of Gaul, Britain and Ireland are not known to have written down their beliefs, so there is no scripture, prayer book or priestly manual. The evidence falls into three groups: archaeology, which is direct but silent about meaning; Greek and Roman writers, who were outsiders with their own purposes; and Irish and Welsh stories written by Christian monks, often many centuries after the pagan period. Each has strengths and limits, and the main challenge is not to treat any single group as a complete picture. Celtic spirituality varied by place and period, and there was no single Celtic church or creed.
Gods and goddesses
Names of gods survive mainly on Roman-period inscriptions and dedications, which are often in Latin and combine a local name with a Roman one. Julius Caesar, writing about Gaul, said the Gauls most honoured Mercury and also worshipped Apollo, Mars, Jupiter and Minerva. This is an interpretation: he matched Gaulish gods to Roman equivalents by function, which tells us what he thought they resembled, not their real nature.
Other names are well attested. Epona, linked to horses, was worshipped widely, even by Roman soldiers. Taranis seems to be a thunder god, and Lugus gave his name to Lugdunum, modern Lyon, and perhaps to later Irish Lugh. Sulis was the goddess of the hot springs at Bath. Groups of mother goddesses appear on many carved reliefs. An antlered figure on a monument in Paris is labelled Cernunnos. A Roman poet, Lucan, mentions three gods named Teutates, Esus and Taranis and a later commentary describes grim offerings to them, but this is second-hand and difficult to verify.
Many gods were local, tied to a river, a mountain or a tribe, rather than part of a shared pantheon in the way the Olympian gods and monsters of Greek myth are. Comparisons with Greece are tempting, but they risk imposing order on a patchwork.
Sacred places
Celtic sanctuaries were less like temples than enclosed or natural spaces. The Gaulish word nemeton, for a sacred place, is preserved in placenames, and it is often linked with groves. Classical writers mention sacred woods, and Roman accounts of the attack on Anglesey say groves were cut down. Archaeology adds other kinds of site.
- Votive deposits in water. At La Tène itself, and at Llyn Cerrig Bach on Anglesey, large numbers of weapons, tools and fittings were thrown into lakes and pools, apparently as offerings.
- Enclosed sanctuaries. At Gournay-sur-Aronde in northern France, a ditched enclosure held the bones of sacrificed animals and deliberately damaged weapons over several centuries.
- Sculpted shrines. The southern French sites of Roquepertuse and Entremont had stone portals and niches for human skulls.
- Ritual monuments. Navan Fort in Northern Ireland, associated in later stories with the kings of Ulster, had a large timber structure that tree-ring dating places at about 95 BCE. It was filled with stones and burned, an act whose meaning is unknown.
Springs and rivers were also sacred, and Roman-period shrines often grew up around them. The healing spring at Bath, dedicated to Sulis, is the best known.
Ritual and sacrifice
Greek and Roman writers report animal sacrifice, feasting, divination and, more shockingly, human sacrifice. Caesar described huge wicker figures filled with people and burned, and Strabo mentions similar practices. Because Romans had political reasons to portray conquered peoples as savage, these accounts have to be read with caution, though some archaeological finds support the idea of ritual killing in some places. The bog body known as Lindow Man, from Cheshire and radiocarbon dated to around the first century CE, died violently, and some scholars connect him to a ritual, but others warn that we cannot be certain of the circumstances. Radiocarbon dating can place the body in time without explaining why he died.
Other practices included the display of enemy heads, noted by Greek writers and supported by skull finds at several sanctuaries, and the burial of rich grave goods. The idea of feasts as sacred events is reasonable given the scale of imported wine vessels, but again it is inferred.
Afterlife
Caesar claimed that the Gauls, and especially the druids, held that souls pass from one body to another after death, and he said this made them courageous in battle. Other Greek writers compared this to the teachings of Pythagoras. Whether this was a real doctrine, a half-understood belief about an afterlife, or a Greek reading of Celtic ideas is debated. Rich graves, with vehicles, weapons and feasting equipment, suggest that the dead needed or deserved goods for some form of continued existence. Compare the very detailed Egyptian afterlife beliefs, which survive in written texts; nothing like that exists for the Celts.
Seasons and calendars
The most important native calendar is the Coligny calendar, a bronze tablet in Gaulish found in eastern France, from the Roman period. It shows a lunisolar system of months with lucky and unlucky days. It is a rare piece of direct evidence for Celtic timekeeping, though it does not explain ritual. The festivals now called Samhain, Imbolc, Beltane and Lughnasadh come from medieval Irish sources, and their origin in the Iron Age is assumed, not documented. Such a scheme differs from the Roman festival calendar and was probably regional.
The later texts
Irish and Welsh stories such as the Táin and the Mabinogion preserve rich material, including gods turned into heroes and kings, otherworld feasts and magical animals. They were written by Christian scribes between about the seventh and the twelfth centuries. The Tuatha Dé Danann, for example, appear in these stories as an earlier race and have often been called Celtic gods, but the manuscripts treat them as a mixture of mythical history and literary invention. Using them to explain the Iron Age is possible only with extreme caution, because the stories reflect their medieval authors and audiences.
Where druids fit
Druids appear in many sources as priests and learned men. What Greek and Roman writers actually said about them is explored in the history of the Roman accounts of druids; the central point here is that druidry was one part of a wider belief world, not the whole of it. Because so much Celtic religion depends on indirect testimony, it is best to speak of what is evidenced, what is probable, and what is only imagined.