There are two things which “everybody knows” about Samhain: that it was a festival of the dead and that it was the “Celtic New Year”. Ronald Hutton, however, having investigated both of these claims finds little evidence to substantiate either in the earliest documentation. Not only that, he only finds evidence for the celebration of Samhain in Ireland: evidence in the early Brythonic literature a Calan Gaeaf festival is wanting, the most prominent calendar dates being the (Julian) New Year in January and Calan Mai at the beginning of May.
It is not, therefore, an unfair question to ask why on earth I am beginning this Brittonic year in late autumn with a festival attested only in Ireland. In my opinion, Hutton overlooks an extremely important piece of evidence: that of the calendar of Coligny. The seventeenth day of first month of the year is marked with the following notation: trinox samo sindiv, which is normally restored as trinoxtion Samoni sindiu “the three-night festival of Samonis today”. This notation is particularly precious: it is one of only three or so records of a genuine pre-Christian Celtic festival.
While there are always naysayers, the academic consensus is quite firm in that the Gaulish samon- and the Irish samain are cognate. Let us examine the details: the Irish festival lasted three days and three nights, the calendar indicates that Samonis was similarly a three-night festival. The Gaulish festival takes place in the first month of the year, which lends support to the otherwise circumstantial evidence that Samhain was a new year festival. The names of the months in the Coligny calendar following Samon suggest that the year began in the late autumn or early winter, which is the attested season in which Samhain occurs. Taking the Gaulish and Irish evidence together, I do not think it unreasonable to assume that the Celtic year began in the autumn.
That the three nights of Samonis do not occur at the very beginning of the month is no obstacle to it being a new year celebration. In the pre-Julian Roman calendar, the first festival of the new year, that of Anna Perenna, does not occur until the Ides (the fifteenth day) of the first month. Like the atenoux of marked in each month of the Coligny Calendar, the Ides originally marked the full moon. Notably, perhaps, trinox samo sindiv occurs three days after the atenoux of the first month: perhaps it marks not the beginning of the festival but its end. Therefore, in this calendar, we will place the beginning of Samonis at the first full moon of the year: the 15th day of Centugiamos.
Happy then, that there is substantial enough evidence for Samonis being the “Celtic New Year”, let us turn our attention to the second thing which “everybody knows”, that Samhain was a festival dedicated to the dead. Here, I believe, Hutton is on more solid ground. In the earliest Irish texts, Samain has no connection with the dead: nor, in fact, does the 31st October elsewhere in Northern Europe. As it happens, Early Christianity had no common festival dedicated to the dead. A need for such a festival only arose with the crystallisation of the doctrine of purgatory towards the end of the dark ages, and even then there was no common date for the commemoration: the Irish and British churches, in common with those of France and Northern Spain, celebrated All Souls in the spring (which, interestingly, corresponds well with the pagan Roman commemoration of the dead, the Lemuria). The current 2nd November date for All Souls actually originated in 10th century Burgundy. So the early Celtic evidence doesn’t indicate a feast connected with the ancestors for this time of the year at all: the only evidence for ancestral rituals being carried out around this time is firmly Christian in nature, and post-dates the adoption of the 2nd November All Souls date considerably.