Samhain as a Festival of the Dead
Both writers accept that there was a pagan festival (Samhain) at this time. Frazer’s view was that it was a festival of the dead that was a fire festival of the pagan Celts. It was the time when the souls of the dead were supposed to visit their old homes. When it instituted the All Saints festival on 1st. November the Church gave a “colour of Christianity to the ancient heathen rite” (Frazer 1914 2:82). Hutton accepts that the pre-Christian festival existed and that it was a time when fairies and witches were especially active. Elsewhere he also accepts that it was the time when humans were “most susceptible to divine and supernatural interference” and that it was then that “heroic and royal figures met fated deaths or enchantments” (Hutton 1991, 177). But he says there is no evidence that Samhain was connected with the dead (Hutton 1996, 365, 369-70).
Hutton states that by the year 800 churches in England and Germany were celebrating a festival dedicated to All Saints on 1st November, having previously kept a feast of martyrs at various stages of the spring. Later Pope Gregory IV (9th. Century) followed this practice. In Ireland the spring observance was still happening. Hutton comments that this makes nonsense of Frazer’s claim that the November date was chosen because of a Celtic influence. Rather both Celtic Europe and Rome followed a Germanic idea. This implies that the whole basis of believing in an early November festival of the dead in Europe before the introduction of the November All Saints in England or Germany in about the 8th century has been removed. Hutton states that evidence of the pagan festival of the dead is “utterly wanting” and that Frazer’s chain of reasoning “completely breaks down” (Hutton 1996, 364-5).
But Hutton also mentions that a festival honouring All Souls on 2nd November was gradually established during the 11th and 12th centuries, having previously been held in February. Both All Saints and All Souls are festivals of the dead (though, illogically, Hutton writes as though only All Souls is). So the Church seems on two occasions to have moved festivals of the dead to early November from earlier times in the year. One might suspect from this fact alone that the Church was following its practice elsewhere of supplanting pagan festivals with Christian ones, and that the pagan festival was a festival of the dead. Frazer suggests that Gregory’s motive for moving All Saints was the suppression of an old pagan custom. He suggests that the Church gave a colour of Christianity to the heathen rite by substituting the saints for the souls of the dead as the true object of worship on November 1st. He also suggests that it was the failure of this attempt to divert the devotion of the faithful from the souls of the dead to the saints that caused the Church to later recognize the festival of All Souls on November 2nd (Frazer 1907 315-8; 1914, 2: 81-83).
Since Hutton states that evidence of the pagan rite of the dead is utterly wanting, I can only refer to Frazer’s argument. As noted, he mentions the positioning of the two Christian festivals of the dead, All Saints and All Souls, at the start of November. He also gives many examples of early November, or October, folk or traditional customs connected with the dead across Europe (Frazer 1907, 309-315; 1914, 2:69-81; see also 1936, 364). He mentions many individual practices, eg sometimes people welcome the arriving ghosts, or make “soul-lights” with lard or butter, or feed the dead for a period. Frazer also observes that celebrations of this sort, with practices of divination and new fires, are often held at the beginning of a New Year. The year of the Celts seems to have begun at this time. For this he relies, inter alia, on the researches of Sir John Rhys (as noted above) and on Hallowe’en customs in the Isle of Man and Ireland (Frazer 1913b, 1:224-5; Rhys 1888, 460, 514 et seq.; 1901, 1: 315 et seq.). He also finds a similarity of the Christian customs with various “heathen rites” across the world, eg among the Mixtecs in Mexico, and among various inhabitants of the East Indies. Sometimes they observed New Years in November which were also festivals of the dead (Frazer 1907, 303-309; 1914, 2: ch. IV; 1913a, 145). He might also have mentioned Divali, which is sometimes celebrated as the new year, and which he refers to elsewhere as a feast of the souls of ancestors returning (Frazer 1913a, 145). He also observes that, when objections to All Souls were raised at the Reformation, the ecclesiastical authorities were prepared to abandon it, and this is consistent with the authorities realising that the festival had only been adopted as a concession to ineradicable paganism (Frazer 1907, 316; 1914, 82).
One could reasonably take different views about the strength of this as a case. It is circumstantial, not direct, evidence. I would view Frazer’s evidence about Samhain, taken together, as persuasive, though less than total proof. But if Hutton had contented himself with saying that he found Frazer’s evidence unimpressive and that he took the view that the November festival of the dead had originated with the 8th century Germanic peoples, that might have been reasonable. Instead, Hutton deals with Frazer’s argument in completely black and white terms. He describes it as “nonsense”. But Frazer might be right. It is in Hutton’s argument, not Frazer’s, that the chain of reasoning breaks down. Once again we are in the territory of Hutton failing to read his sources properly. His argument, and his mention of the “Celtic influence”, would have made sense if Frazer had said that the pagan festival had been limited to the Celts and had never been celebrated by other European peoples. But Frazer does not say this. Indeed he explicitly states the contrary. He states that the pagan festival of Hallowe’en, on 1st. November or its eve, marked the transition from autumn to winter “not only among the Celts but throughout Europe” (Frazer 1913b, 1: 224,5). Note that in this passage Frazer’s references to “Hallowe’en” are to the pagan festival, not the Christian one. It is clear from the first pages of section 6 of chapter IV that the kind of Celts he is referring to there are pre-Christian ones. See for instance the reference to the Celtic bisection of the year that “dates from a time when the Celts were mainly a pastoral people, dependent for their subsistence on their herds (Frazer 1913b,1:223).
Elsewhere, referring to the pagan origins of All Souls, Frazer suggests that the festival, later displaced by one day, spread from the Celts “to the rest of the European peoples” (Frazer 1907, 316 and 1914, 2: 81). Again it is evident that, in the context, where Frazer refers to “the Celts” he means the pagan Celts. The “old festival” of the dead of the pagans was on 2nd. November or, on the basis of the theory he then puts, 1st. November (Frazer 1914, 83). Note that the term “All Souls”, as used by Frazer means, before its “recognition” by the Church, the pagan feast. Incidentally, Frazer knew that All Saints was moved to 1st. November earlier in Germanic lands than in Celtic ones (also page 83). Had there been a contradiction here I have no doubt Frazer would have noticed it.
So, to assert that 8th century Celtic Christians followed Germanic ones in moving All Saints to early November from a different time of year does not make Frazer’s argument incoherent. It does not disprove the theory that the pagan festival of the dead influenced the moving of the All Saints festival to November. It is therefore not appropriate for Hutton to dogmatically describe Frazer’s argument as “nonsense”. Hutton is not in a position to show that Frazer was wrong and that an ancient European festival of the dead at Samhain did not exist.
The arguments in favour of Frazer’s case may go beyond the ones he himself put. He was following Sir Edward Tylor (1832-1917). Tylor wrote that annual feasts of the dead are found in Europe, Asia and Africa and coincided with the end of harvest or end of the year. He also describes All Souls as falling into a broader pattern of feasts on which the living leave viands for the dead to feed on (Tylor 1871, 2:32-9). Also W.Y.Evans-Wentz refers to the Irish collection of tales Silva Gadelica that mentions “due worship and sacrifice on the Feast of Samhain to the Tuatha De Danaan, the gods of the dead, at that time supreme”. He also states that “Samhain….was the great Celtic feast of the dead when offerings or sacrifice of various kinds were made to ancestral spirits” (Evans-Wentz 2002, 299, 439, first published 1911). J.A.MacCulloch states that "The great commemoration of the dead was held on Samhain eve....Whatever be the reason, Celts, Teutons and others have commemorated the dead at the beginning of winter" (MacCulloch 1911, 265 ff.). Barry Cunliffe refers to Irish literature indicating that Samhain was a liminal time between the two years at which "the spirits of the dead could roam free" (Cunliffe 1997, 189).
 Samhain is generally identified with Hallowe’en, but it makes little difference to the arguments whether the precise date was 31st. October or 1st. November. Very anciently festivals may have been tides rather than specific days, as suggested eg by the twelve days of Christmas and the ten or twelve days of the celebrations at the Irish festival of Lughnasa, when people took to the tops of hills to enjoy themselves. It is accepted by all that Samhain was a real festival.The word Samhain continues to exist in Celtic languages to mean the month of November. The Irish for 1st November is Lá Samhna, and the Manx name for that day is Sauin (Frazer 1913b 1:241-4).
 On Samhain being a Celtic fire festival see Frazer 1913b, 1:222, 224 and 2:40 et seq.
 Hutton says that “It remains possible that northern European pagans had honoured their dead at this time”, but his point is that there is no reason whatever to believe that this happened.
 Hutton states that Abbot Odilo ordered the mass for the souls in February, and over the next two centuries it became the norm to hold the feast on 2nd. November. However he gives no authority for this. Frazer gives a different account, which is that Odilo ordered the mass on 2nd. November (Frazer 1914 2:82). This was also the version given by Tylor (Tylor 1871, 2:33-4) and Evans-Wentz (Evans-Wentz 2002, 453). But whoever is right here, the question is – why was 2nd. November chosen? And why positioned next to the other feast of the dead on 1st November? If there was an observance in February, this may have been placed at the time of the Roman Parentalia, a festival period in honour of spirits of the ancestors. But why the transfer and why to November 2nd? The All Saints festival may have been transferred from a feast of martyrs in May or (per Hutton 1996) in April. The reason for the transfer is said to have been that in the 8th century the Pope dedicated a new chapel at St. Peter’s in Rome to All Saints on 1st November. Though this could have fallen at Samhain by chance, it seems fortuitous when, presumably, a different date might have been chosen for the dedication.
 I refer to his comment “The dead arrived later”. Frazer correctly treats both the Saints and the Souls as classes of the dead (Frazer 1913b, 225 footnote: “After all the Saints are only one particular class of the Souls of the Dead”). One cannot get canonised as a saint by the Catholic Church unless one is dead.
 Nb circumstantial evidence is not necessarily weak evidence. People are convicted of offences in the courts on the basis of circumstantial evidence, although the standard of proof required is beyond all reasonable doubt.
 In this paper I have concentrated on the issues of the Celtic New Year and of whether the pagan Samhain was a festival of the dead. There are many other matters traditionally associated with Hallowe’en which may also go back to pre-Christian precedents, eg that it was (1) a time when fairies, witches and spirits were especially active and the Otherworld more prominent, (2) a time of divination and prophecy of events over the next twelve months (because of it being the start of the Celtic New Year), (3) a Celtic fire festival, and (4) “a particularly numinous time”. Hutton seems less opposed to most of these possibilities (see Hutton 1996, 365-9).