Samhain: What the Sources Actually Say

Samhain carries more mythological freight than any other point on the modern Wheel of the Year. Walk into any October conversation about witchcraft and you will hear: Celtic festival of the dead. Thinning veil. Celtic New Year. Ancestors returning. Druid bonfires at Tlachtga. These claims have been repeated so often they feel like bedrock. Most of them are not.

The medieval Irish evidence for Samhain is real and specific. The evidence for the familiar modern framing — death-feast, cosmic hinge, year’s end — is much thinner than it’s usually presented. Separating the two is the work of this page.

The name

The word appears in medieval Irish as Samain or Samuin. Its etymology is genuinely contested. The ninth-century glossary attributed to Cormac mac Cuilennáin renders it as “end of summer” — sam, summer, plus a suffix he read as fuin, ending or setting. Later scholars have disputed the derivation; the compound is not phonologically clean. What is clear is that medieval writers understood the feast as marking the close of the warm season, regardless of what the root morphemes technically mean. The tenth-century tale Tochmarc Emire describes it plainly: Samhain, “when the summer goes to its rest.”

The name means something about a seasonal boundary. It does not mean “festival of the dead.” That gloss is a much later imposition, which this page will trace to its actual origin.

The four-festival year: what Tochmarc Emire says

The earliest unambiguous textual attestation of Samhain as a named feast comes from the Tochmarc Emire — The Wooing of Emer — a tale of the Ulster Cycle preserved in manuscripts of the tenth century. In The Stations of the Sun, Ronald Hutton notes that Samhain was the major festival that marked the opening of winter in early medieval Ireland, and that in Tochmarc Emire it is the first of the four quarter-days mentioned by the heroine Emer.

The four festivals named in the text — Samhain (1 November), Imbolc (1 February), Beltaine (1 May), and what later texts call Lughnasadh (1 August) — constitute what modern writers often call “the Celtic seasonal calendar.” The Tochmarc Emire is the earliest source to list all four together. They were known to the Gaels as Beltane, Lughnasadh, Samhain, and Imbolc. The tidy four-fold symmetry, however, is largely a function of the text’s literary structure, not a direct window onto ritual practice. No pre-Christian calendar artifact lists all four as a unified system.

The cereal harvest would likewise have long been completed and the time of warfare and of trading was at an end — making Samhain an ideal moment for the convention of the year’s most important tribal assemblies; and indeed the feis of Samhain, at which local kings gathered their people, is a favourite setting for early Irish tales.

The feis of Samhain: assembly, feast, peace

What do the medieval Irish texts actually describe happening at Samhain? The answer is social, not supernatural: a great assembly.

The festival marked the end of warfare and trading, a time when livestock had been ushered in from summer pastures, and local kings gathered their people for “the three days before Samuin and the three days after Samuin and Samuin itself … and during these seven days there would be nothing but meetings and games and amusements and entertainments and eating and feasting.”

The Echtra Cormaic — Cormac’s Adventure, preserved in manuscripts of around the twelfth century — adds the legal dimension: the tale says that the Feast of Tara was held every seventh Samhain, hosted by the High King of Ireland, during which new laws and duties were ordained; anyone who broke the laws established during this time would be banished.

This is the character of Samhain in the legal and annalistic layers of medieval Irish literature: a peace-time assembly, a moment of political negotiation, a feast. Tribute was rendered. The festival was also a time when people had to give a portion of their harvest to their lord, as described in an extreme version in Lebor Gabála Érenn. The gathering served the same function that any off-season communal assembly serves in a pastoral society: it was when the year’s political, legal, and social business got done, because the farm work was finished and the roads were still passable.

In stark contrast with what they portray about Beltane on 1 May, Irish tales about the feis of Samhain do not allude to any religious observances, though it is believed by Ronald Hutton that some did exist.

That hedge is important. The absence of religious description in the literary sources does not prove an absence of religious practice. It proves only that the literary genre — heroic saga, legal text, mythological tale — was not interested in recording it, or that the Christian monks who copied and redacted these texts had reasons to suppress it. Hutton is careful to note the silence without turning it into a claim.

What the texts do not say: druids, bonfires, Tlachtga

The famous image of druids lighting a sacred fire at Tlachtga (the Hill of Ward, County Meath) on Samhain eve, from which all domestic fires in Ireland were relit, is routinely cited as ancient. The first mention of anything religious at Samhain was in the seventeenth century, by the thoroughly unreliable Geoffrey Keating, who stated without citing his source that the “Druids of Ireland used to assemble on the …” Hill of Tlachtga on that night for a great fire ceremony. Keating was writing Foras Feasa ar Éirinn (c. 1634) at a distance of at least a millennium from the events he describes, with a political agenda — constructing a heroic pre-Christian Irish past — and no surviving source for his specific claim. The druid-bonfire tradition at Tlachtga is not found in any medieval Irish text.

This does not mean bonfires were absent from Samhain practice. Communal fires are deeply attested in Irish and Scottish seasonal custom generally. But the specific theological architecture — druids, sacred flame, domestic relighting — is Keating’s construction, not a medieval report. Every modern account that presents the Tlachtga fire-lighting as ancient Irish fact is, knowingly or not, citing a seventeenth-century antiquarian with no primary source.

Samhain in the mythological cycle: the Otherworld threshold

The supernatural dimension of Samhain in medieval Irish literature is genuine — but it requires precise description.

The invasion of Ulster that makes up the main action of the Táin Bó Cúailnge begins on Samhain, with the off-season timing being part of what surprised the Ulstermen. The Second Battle of Magh Tuireadh also begins on Samhain, and the Morrígan and the Dagda meet before the battle against the Fomorians.

The Second Battle of Moytura (Cath Maige Tuired), a pivotal moment in Irish mythology and one of the most important remaining medieval texts on the subject, was fought between the Tuatha Dé Danann and the Fomoire and is said to have taken place during Samhain.

The pattern is consistent: it seems that when the early writers wished to impart a magical quality to the events they were depicting, they chose the Festival of Samhain for the occasion. Samhain functions in the mythology as a literary hinge — the moment when the ordinary calendar is suspended and extraordinary things can happen.

The most explicitly supernatural tale is Echtra Neraí — The Adventures of Nera — also preserved in tenth-century manuscripts. In Echtra Neraí, Samhain is portrayed as a liminal time when the dead speak and walk and are dangerous, when time moves differently, and when humans can enter the Otherworld realms. It is the closest surviving narrative to what modern pagans and witches describe as “the veil thinning.”

The medieval texts also attach a specific topography to the Otherworld opening. The Morrígan is strongly associated with Cruachan, and with the cave of Cruachan specifically, where several important things happen on Samhain. The Metrical Dindshenchas poem Odras refers to Cruachan as one of the sí or fairy hills and identifies the cave — known today as Oweynagat, from the Irish Uaimh na gCat, “cave of cats” — as the Morrígan’s home.

So the medieval Irish case for Samhain as a time when the Otherworld is more accessible than usual is supported by the literature. The mechanism described in those texts — supernatural beings moving through identifiable landscape portals, the dead walking on the night — is native to the medieval material. The modern phrase “thinning veil” is a twentieth-century idiom for something the medieval texts described in their own terms.

What the medieval texts do not say is that Samhain was primarily a festival for honoring the dead, or that the primary purpose of Samhain observance was ancestor veneration. The literary evidence for the supernatural is mostly incidental — Samhain is the setting for these events, not their cause. The dead-feast framing is an interpretation, not a citation.

---

The Celtic New Year: tracing a Victorian claim

The single most persistent modern claim about Samhain is that it was the Celtic New Year. This is stated as fact in thousands of online pages, witchcraft handbooks, and Halloween explainers. Its source is not ancient.

It was Sir John Rhys who first suggested that Samhain was the “Celtic New Year” in his 1901 work Celtic Folklore, based on very flimsy — and also contemporary, i.e., late-19th/early-20th-century — folklore evidence from the Isle of Man, and having previously cited unreliable or “corrected” Irish texts (see Davis, 2009, pp. 29–30; Hutton, 1996, pp. 363).

Rhys visited the Isle of Man and found that Manx people sometimes called 31 October “New Year’s Night.” He combined this with the Tochmarc Emire listing Samhain first among the seasonal festivals and concluded that the pre-Christian Celtic year began at November. Rhys’s theory was popularised by Sir James George Frazer, though at times Frazer acknowledged that the evidence is inconclusive.

Frazer also said that Samhain had been the pagan Celtic festival of the dead and had been Christianized as All Saints and All Souls. Since then, Samhain has been popularly seen as the Celtic New Year and an ancient festival of the dead.

The Coligny Calendar, a bronze tablet from second-century Gaul representing the only surviving pre-Christian “Celtic” calendar artifact, cuts directly against Rhys’s inference. The Coligny Calendar seems to begin the year in the summer, although the calendar was found in fragments and has been reconstructed in different ways, none of them conclusive. A calendar from Gaul is not identical to an Irish one, and the Coligny evidence is fragmentary — but it does mean that the claim “the Celts began their year at Samhain” cannot be presented as something the archaeological record supports.

Ronald Hutton says the evidence that Samhain was the Celtic or Gaelic new year is flimsy. The phrase “Celtic New Year” applied to Samhain is a scholarly construction of the late Victorian period, built on contemporary Manx folklore and a literary text’s narrative ordering — not on pre-modern Irish calendrical evidence.

The Wheel of the Year: a 1950s assembly

The eight-fold Wheel of the Year in which Samhain sits as the “Witch’s New Year” and the beginning of the dark half is not an ancient Irish structure. Its history is short and specific.

Two neopagan streams in Britain popularised seasonal festival calendars in the twentieth century: the Bricket Wood coven, a Wiccan group founded by Gerald Gardner, and the Order of Bards, Ovates and Druids, a neo-Druidic group founded by Ross Nichols. Legend holds that Gardner and Nichols harmonised an eight-fold calendar during a naturist retreat, merging the four solar stations alongside their four midpoints as a unified festival cycle.

Gardner’s first publications refer to the Celtic festivals as “May eve, August eve, November eve (Hallowe’en), and February eve.” He was working with the four fire festivals and calling them by their popular English names, not their Gaelic ones. The Gaelic names — Samhain, Imbolc, Beltane, Lughnasadh — entered popular Wiccan usage as the tradition developed and sought a more explicitly Celtic identity.

The phrase “Wheel of the Year” was in use by the mid-1960s to describe an annual cycle of eight observances.

Prominent Wiccan Aidan Kelly gave names to the Wiccan summer solstice (Litha) and equinox holidays (Ostara and Mabon) in 1974, which were then promoted by Timothy Zell through his Green Egg magazine. Popularisation of these names happened gradually; in her 1978 book Witchcraft For Tomorrow, influential Wiccan author Doreen Valiente did not use Kelly’s holiday names, instead simply identifying the solstices and equinoxes as “Lesser Sabbats”. Valiente identified the four “Greater Sabbats”, or fire festivals, by the names Candlemas, May Eve, Lammas, and Hallowe’en, while also naming their Gaelic counterparts Imbolc, Beltane, Lughnasa, and Samhain.

The narrative of the Wiccan year — in which Samhain marks the God’s descent into the underworld, the Goddess’s shift into her crone aspect, and the beginning of the dark half of the solar cycle — is a coherent mythological structure. The eight-fold wheel as a unified, named, systematized cycle did not exist in any pre-Christian tradition. No Celt observed all eight. This is a point worth sitting with. It does not make the wheel invalid as contemporary practice. It makes it a twentieth-century construction built from genuine older materials — and practitioners who know that are working more honestly than those who don’t.

Hutton surmises this work with the observation that the notion of a distinctive “Celtic” ritual year with four festivals at the quarter days and an opening at Samhain is a scholastic construction of the eighteenth and nineteenth centuries which should now be considerably revised or even abandoned altogether.

Allhallowtide: the Christian layer

No account of Samhain is complete without the Allhallowtide problem: the relationship between the medieval Irish feast and the Christian observances of All Saints (1 November) and All Souls (2 November) that now occupy the same calendar position.

The feast of All Saints was established at different times in different parts of the Christian world. It was not fixed to 1 November universally until the ninth century, under Pope Gregory IV (c. 835). The coincidence with Samhain has long prompted speculation about Christianization — the Church absorbing a pagan death-feast and redirecting it. Frazer made this the cornerstone of his argument.

Hutton’s analysis in Stations of the Sun treats the relationship as genuinely ambivalent. The evidence does not clearly establish that All Saints was moved to 1 November because of Samhain; the date had other precedents in the liturgical calendar. The accumulation of supernatural folklore around the Halloween period — souls returning, the dead abroad, protective rituals — may reflect the Christian observance of All Souls (explicitly a feast for the dead) as much as or more than a direct continuity with pre-Christian Irish practice.

What the evidence does support is this: by the time we have Scottish and Irish folklore records in any density — from the seventeenth century onward — the end of October and beginning of November is a period dense with belief about the dead, supernatural danger, and liminal passage. Whether the medieval Irish feis of Samhain seeded that folklore, or whether the Christian All Souls feast produced it, or whether the two reinforced each other across several centuries, is genuinely uncertain.

The thin veil question

The “thinning veil” between the living and the dead, the most widespread contemporary Samhain belief, is linguistically modern but not conceptually invented. As noted above, Echtra Neraí and related texts do describe Samhain as a time when the Otherworld is accessible. The sídhe mounds open. The dead speak. Nera enters the Otherworld and encounters things that cannot be seen at other times. Once a year, on Samhain, sídh portals would open to the human world, and one could observe the fires lit within, and the sídh people coming and going.

The shift from the medieval framing to the modern one involves one significant change: in the medieval literature, the Otherworld is populated primarily by supernatural beings — the Tuatha Dé Danann, the sídhe, supernatural enemies — and only secondarily by human dead. The modern emphasis on ancestors and the returning dead of one’s own lineage reads back into Samhain a structure that may owe more to the Christian All Souls (explicitly an annual observance for deceased family members) than to the medieval Irish mythology. The two have been so thoroughly blended in popular practice that unpicking them is largely academic — but it’s worth knowing they are blended.

What the literature actually says about death

From Echtra Neraí (10th century) and Togail Bruidne Dá Derga (11th century), it is clear that Samhain was associated with death, not only of individuals but of entire populations. Togail Bruidne Dá Derga — The Destruction of Da Derga’s Hostel — is one of the most violent texts in the medieval Irish corpus; it sets the catastrophic sack of a king’s hostel at Samhain, and the time-stamp signals that what follows will be outside the normal order of things.

The association of Samhain with death in the medieval literature is genuine. But it is an association with violent, supernatural, and uncontrolled death — the death of armies, the death of kings, the death that comes from the Otherworld breaking through. This is categorically different from the modern Samhain frame of tender ancestor veneration, a place set at the table, the beloved dead invited home. Both responses to the season — terror and tenderness — are legitimate ways of relating to mortality. They are not the same tradition.

---

The honest inventory

A practitioner working with Samhain is working with a layered object. Here is what each layer consists of, and where it comes from:

The medieval Irish layer (10th–12th century, primary texts): A winter-opening feast. The calendar’s most important communal assembly. A favored literary setting for supernatural events. An established association in the mythology with Otherworld access, the sídhe emerging from their mounds, and violent or unruly death. The cave at Cruachan as a named portal. No documented druid ceremonies. No documented ancestor-veneration rites.

The 17th-century antiquarian layer (Keating, c. 1634): Unsourced druid fire ceremony at Tlachtga. Widely repeated since; no earlier source exists.

The Victorian scholarly layer (Rhys 1901, Frazer 1890+): Samhain as Celtic New Year. Samhain as Celtic festival of the dead. Both claims derived from contemporary folklore, not medieval sources, and amplified by Frazer’s comparative religion framework into global currency.

The 1950s Wiccan layer (Gardner and Nichols): Samhain as one of four Greater Sabbats. Placement within an eight-fold Wheel of the Year that combined the four Celtic fire festivals with the four solar stations. The death-and-rebirth narrative of the God and Goddess mapped onto the seasonal cycle.

The 1970s naming layer (Kelly, 1974; Valiente, 1978): The Gaelic names for all eight festivals systematized and standardized in English-language Wicca.

The medieval Irish material is worth taking seriously on its own terms. It is not thin — there is a substantial body of literature, and Samhain occupies a particular place in it. What it documents, though, is a feast of communal assembly and a mythological threshold, not a specific theology of death and return. The death-feast framing is real, but it is Victorian and Wiccan, built from genuine materials and a century of creative interpretation.

For a practitioner, this means the modern Samhain is a legitimate contemporary tradition with deep roots in a specific literary and mythological archive. The roots are real. The shape of the tree is recent.

---

Cross-references on this site: (forthcoming — Wheel of the Year history, All Saints and All Souls, Beltane: what the sources say, medieval Irish otherworld)