Three layers of “Celtic magic”
Ideas about Celtic magic come from three quite different layers. The first is ancient evidence: inscribed lead tablets, small objects that may have been amulets, and comments by Greek and Roman authors. The second is medieval and later material, including Irish and Welsh stories and Scottish and Irish folk charms collected in the nineteenth and early twentieth centuries. The third is modern: neopagan and New Age writing that draws on the other two layers and adds a good deal that is new. Mixing them together produces a picture that is more romantic than any of the parts.
The word magic is also a problem. Ancient people did not separate it as sharply from religion or medicine as we do, and Greek and Roman writers often used the word for foreign or disapproved practices. A healing charm, a prayer to a spring goddess and a curse might all have felt like parts of the same effort to influence unseen forces.
Written spells: the strongest evidence
The most direct evidence is writing. Several lead tablets in the Gaulish language, a Celtic language written in Roman times, have been found in France. The Larzac tablet, discovered in 1983 and dated around the end of the first century CE, is one of the longest surviving texts in Gaulish. It is usually understood as a magical text involving women’s names and appealing to powers of the underworld, but the language is only partly understood and translations remain uncertain. A tablet from Chamalières, found in 1971, appears to call on the god Maponos.
In Roman Britain, many lead tablets are written in Latin. At the spring of Sulis at Bath, over a hundred inscribed tablets were thrown into the water, mostly in the second to fourth centuries CE. Many ask the goddess to punish a thief and recover stolen goods such as clothes or money. Others were found at Uley in Gloucestershire, at a temple of Mercury. These show that the people of Roman Britain used curses in a style also found throughout the Roman Empire, with Celtic names and gods alongside Roman ones. They are better evidence for provincial Roman religion than for pure Iron Age Celtic practice, but they show how local people used the tradition. The offering of gifts at healing springs also resembles practice at Asclepian healing temples elsewhere, though there is no sign of direct borrowing.
Roman law, from the time of the Twelve Tables onward, is reported by later authors to have treated harmful incantations as a crime, so curses were not simply a Celtic habit.
Amulets and charms in the ground
Archaeologists find small objects that may have been worn for luck or protection: wheel-shaped pendants, tiny axes, beads and animal teeth. In a few cases, the context suggests an intent, such as items placed in graves or in wells. But the label amulet is usually an inference, because the objects come with no label. A pendant might just be jewellery. The honest answer to what a given trinket was for is often that we do not know.
Coins, brooches and pins were sometimes deposited in springs and rivers, and many scholars interpret this as offerings, a practice that shows belief in powers in water but not a particular spell.
Classical reports of charms and herbs
Pliny the Elder reports several practices that he attributes to the druids. He says they valued mistletoe from oaks as a remedy for many ills, and that a “serpent’s egg,” supposedly formed by a mass of snakes, brought success in legal disputes. He also describes two plants, selago and samolus, to be gathered with special rituals: for selago, the gatherer wore a white tunic, had bare, washed feet and offered bread and wine; for samolus, the gatherer used the left hand and fasted. These accounts may preserve something real, but Pliny wrote at second hand and in a book that mixes sound observation with unreliable folklore. The reliability of the details is unknown. Mistletoe berries are poisonous, and these descriptions are historical curiosities, not advice. Comparable plant lore from other cultures appears under ancient herbal remedies, and Pliny’s own stories are a poor guide to Celtic practice.
A later Latin medical book by Marcellus of Bordeaux, around 400 CE, includes some Gaulish plant names and charms. Because Marcellus came from Gaul, it may preserve older local lore, though by then the region had been Roman for centuries. It belongs to the tradition of Greco-Roman medicine more than to any independent Celtic system.
Medieval stories and folk charms
Irish and Welsh stories are full of spells, shape-shifting and druids who cast mists or raise storms. In the Welsh Mabinogion, the magician Math and his nephew Gwydion create a woman from flowers. These tales were written by Christian authors, often to entertain, and they show what medieval storytellers imagined, not what Iron Age priests did. A hymn traditionally attributed to Saint Patrick, probably later in date, asks for protection against the spells of women, smiths and druids in the usual translations, which shows that the fear of sorcery existed in early medieval Ireland.
Folk charms offer another source, and Celtic charms and folklore of this kind are often presented as ancient. In the late nineteenth century, Alexander Carmichael collected Scottish Gaelic prayers, blessings and charms, published as the Carmina Gadelica. They contain a rich mix of Christian and older themes. Similar charms for cattle, butter-making, the “evil eye” and curing illness were recorded in Ireland. These are valuable evidence for folk culture, but they cannot be projected back two thousand years. Folklore changes by being retold, and most customs recorded in recent centuries show Christian and medieval influence.
Modern ideas
Many popular beliefs about Celtic magic are modern. The idea of a “Celtic tree calendar” linked to the ogham alphabet comes largely from the writer Robert Graves in 1948, and scholars of the early medieval ogham script, which was used for inscriptions on stone, do not accept it. In the 1790s, the Welsh antiquarian Iolo Morganwg invented rituals and documents that he presented as ancient druidic tradition, and some of these entered modern druid ceremonies. The festival names Samhain, Beltane, Imbolc and Lughnasadh, used in the modern Pagan “Wheel of the Year”, come from medieval Irish sources.
None of this makes modern practice illegitimate as a religion or creative tradition. It simply means that it is a modern movement, not a surviving ancient one.
What can be said
Ancient Celtic-speaking peoples certainly used curses, offerings and probably charms, and Roman authors thought them skilled in plant lore and divination. Beyond that, the surviving evidence is thin. The Egyptian magical practice known from long written spell collections has no equivalent in Iron Age Celtic Europe, which is one reason that so much has been imagined to fill the gap.