Weird Burial Stories · The Archive

Article

Samhain, All Hallows, and the Graves of Early Winter

By Weird Burial Stories Editorial Desk · 2026-09-27

Halloween sits on top of three older calendars: an Irish feast called Samhain, the Roman feast of All Saints, and Cluny's day for praying at the graves of the ordinary dead. The tidy story that one simply became the next does not survive the documents.

By C. V. Wooster

Every October the same sentence gets repeated: Halloween is Samhain, the Celtic feast of the dead, with a Christian name painted over it. The sentence is tidy. The burial record is not. What can be said, from texts rather than from posters, is narrower and more interesting. Early Irish literature uses Samhain as a date when the world of the living and the world of the sídh are dangerously close. The Western church, on a different authority, fixed 1 November as the feast of all the saints and, later, 2 November as a day of prayer for all the dead. English parish custom then put begging, soul cakes, and churchyard prayers into that same week. Those are three histories. They share a week on the calendar. They are not one ritual wearing three masks.

What the Irish texts actually do with Samhain

Samhain, in the medieval Irish year, is one of the turning points of the seasons, the start of winter. Tales set meetings, raids, and otherworldly intrusions there. That is a literary fact, and it is why the feast has a real claim on the history of the dead: the storytellers treated the date as a time when the boundary failed. It is not the same fact as a documented pan-Celtic burial liturgy, practiced the same way from Ireland to Wales, that the church later stole.

Ronald Hutton's study of the British ritual year is the book to read before repeating the theft story. Hutton's conclusion, working from the medieval records rather than from nineteenth-century revivalists, is that those records do not show 1 November as a major pan-Celtic festival, and do not show religious ceremonies attached to it even in the places where the date was observed. In Wales the day was Calan Gaeaf, the first day of winter, and the earliest Welsh literature does not treat it the way it treats May Eve. A feast can matter in Irish narrative and still not be the ancestor of every later grave custom in the first week of November.

His 2024 essay in Folklore names the people who built the stronger claim. In 1847 the Irish scholar John O'Donovan read one entry in Cormac's Glossary as if the year ended in autumn, and he added the caveat "if there be no error of transcribers." Sir John Rhys spread that reading in lectures in 1888, published in 1892, and used Halloween divination and a bonfire story from Tlachtga as support. James Frazer's The Golden Bough then called the festival a feast of the dead, and said the church created All Saints and All Souls to replace it. By 1911 the folklorist John MacCulloch was repeating a Celtic year as fact, and Walter Evans-Wentz called Samhain "the great Celtic feast of the dead." Hutton's finding is that there is "not a single medieval record" of churchmen instituting either feast to suppress or replace a pagan feast of the dead, and that the evidence for that feast "is pretty near non-existent." He also writes that neither idea "has been conclusively disproved, and probably never will be." That is one historian's verdict. It is still debated. Britannica's Halloween entry says the holiday evolved from Samhain. Britannica's Samhain entry says very little can be said with certainty, because the evidence is sparse. The Tlachtga fire, which has only Keating behind it, is set out in Halloween bonfires and the tail-less black sow.

All Saints, then the ordinary dead

The Roman feast is a different stack of documents. Church histories, including the Encyclopaedia Britannica's account of All Saints' Day, place the November date in Rome with Pope Gregory III, who dedicated a chapel in St. Peter's to all the saints in the eighth century. In the following century Pope Gregory IV extended the observance through the Frankish church. This is a papal and imperial calendar decision. It does not require a Celtic burial ground to explain it, and Hutton's point is that historians should not invent the requirement.

The May date is older, and it is also Roman. The Catholic Encyclopedia says that on 13 May 609, or 610, Pope Boniface IV consecrated the Pantheon in Rome to the Virgin and all the martyrs, and ordered an anniversary. The same entry credits Gregory III with fixing 1 November. Encyclopaedia Britannica dates that dedication to 609 and records the name the church then took: Santa Maria Rotonda, or Santa Maria ad Martyres. Hutton's 2024 essay says the source for that chapel, the Liber Pontificalis, does not mention an anniversary, and he treats the November attribution as a later suggestion. The first date he calls secure is a letter of 799 from Alcuin of York, praising Archbishop Arno of Salzburg for keeping the feast. Gregory IV's order across the empire of Louis the Pious is dated 835 by the chronicler Sigebert, in Hutton's reading, and 837 by Britannica. The only reason Hutton finds anyone giving is from the twelfth-century liturgist John Beleth: it was easier to feed pilgrims in Rome just after the harvest. His conclusion is that there is nothing Celtic in the geography of this process. Britannica's Halloween entry still says Gregory III moved the feast "perhaps in part to supplant pagan holidays." Those are two readings of a thin file.

Prayer for the dead is older than any single feast day. What Odilo, abbot of Cluny, added was a fixed day for all of them. The Catholic Encyclopedia's entry on Odilo says he established All Souls' Day, 2 November, at Cluny and its dependent monasteries, and that the custom spread through the Western church. The same entry refuses the neat year that guidebooks like: the decree was probably not in 998, but after 1030. The traditional date is worth repeating only with that caution attached. Odilo's monks were told to pray, fast, and give alms for the dead. That is a burial custom in the strict sense. It sends the living to intercede for people already in the ground. It is not a survival of Samhain unless a document says so, and the careful ones do not.

The story Cluny told about the decision is a ghost story, and Hutton marks it as one. In Jotsald's Life of Odilo, a monk coming home from Jerusalem meets a hermit who lives near a volcano. The hermit has heard demons complain that the prayers of monks are rescuing too many souls, and Cluny's most of all. When Odilo heard the tale, the legend says, he instituted All Souls. Hutton places the Cluniac extension of prayer to all the Christian dead probably between 1024 and 1033. Odilo was abbot from 994 to 1049. That range sits beside the Catholic Encyclopedia's "probably not in 998 but after 1030." Hutton also notes earlier local services. Eigil, abbot of Fulda, who died in 822, was later said to have held one for all the dead on 17 December, and late tenth-century Saxony kept 1 October for praying for the dead. Unlike All Saints, Hutton writes, All Souls was never formally adopted or ordained by a pope. The volcano is the legend. The feast is the document. Candle nights that still use the week, in Poland and in Mexico, are separate histories: Zaduszki and Day of the Dead at the grave.

Soul cakes at the porch

By Shakespeare's lifetime the English season had a street custom. In The Two Gentlemen of Verona, Speed says a lover would "speak puling, like a beggar at Hallowmas." Editors have long glossed the line with souling: poor people going house to house at Hallowmas, singing and asking for soul cakes, on the understanding that they would pray for the giver's dead. The American Shakespeare Center notes the same practice as part of Allhallowtide, the three days from 31 October through 2 November. A cake handed over a threshold in exchange for a prayer is as close as English Halloween gets to a grave rite you can actually document. It is alms, not a reconstructed druid funeral.

Later folklorists thickened the story. Turnip lanterns in Ireland and Scotland, and the pumpkin that replaced the turnip in North America, belong to that later layer of household custom. They are worth a paragraph in a history of the season. They are not evidence that a medieval Irish burial used a carved root as a grave light. The honest Halloween essay keeps those layers in the order the documents arrive: Irish narrative at Samhain, a Roman feast of all saints, a Cluniac day for every soul, and an English porch where someone whined for a cake and promised to pray.

Sources