emily,

Starting days with sunset might (this is an unexamined guess) be a feature of lunar calendars (such as the Islamic, the neo-Babylonian, and the neo-Babylonian calendar's modern progeny: the Jewish Calendar and the Christian Easter cycle) which start the month with the new waxing crescent. Even if this generalization is right, I'd be cautious before extending it to lunar calendars (like the ancient Egyptian luncar calendars--not to be confused with the 365-day Egyptian sothic calendar) which start the lunar month by noting the last appearance of the waning crescent. I'd be even more cautious of extending to solar calendars. That doesn't mean it's not true that the Irish started the day with sunset (one of the classical writers asserted this for the Gauls, I vaguely recall), only that we can't jump to the conclusion.

This doesn't mean Samhain wasn't an ancient Irish new year. In the Tochmarc Emire, which mentions the four cross-quarter days, Samhain is the first mentioned. Other sources (I'm not certain of their geographical origins within Ireland, so they may only apply to parts of the Island) suggest that it was the time of important assemblies. So it might have been "a" new year day, even if not "the" new year day. Some societies, including our own (the School year begins in August/September, the federal fiscal year begins on 1 Oct., for some Christians the lectionary year begins on the Sunday in the week of November 27, for Jewish folk there is the 1st of Tishri and 3 other new-year days, etc.) have multiple new years. But even taking the more cautious approach and calling Samhain "a" new year presupposes that the Irish, prior to Christianity, used years at all. Maybe they used half-years, and spoke of "next summer" and "next winter" rather than "next year".

On its own merits, of course, November 1 is just as good as January 1 for us moderns to use for starting a year.

T.