the term "blue moon" for "intercalary month" arose by folk etymology,
the "blue" replacing the no-longer-understood belewe, 'to betray'.
I shouldn't consider myself a scholarly source of course, but it is so glaringly obvious so that it might go as common sense knowledge.
If eleven, twelve contained as is usually thought an element leven, left (en.WT: eleven) and if the intercalary month was precisely those days that were left at the end of a period in calendar (see below), then "belewe" cannot be anything but related to this idiom. be- may be the common prefix. Moon is related to month. This is surface analysis.
It's conceivable that a sense "left" could also come to mean desertion, betrayal, though I do disagree with the -leven theory of eleven, twelve, because it seems literally backwards, counting down – looking forward it requires a sense of "extra" rather than "left".
There is a lot to say about superstition in numerology and religion.
- So Judas, one of the twelve apostle often painted as the twelfth in order, is the prototype of a betrayer. However, the unlucky number usually associated with this is thirteen and the origin of this superstition is not clear. For example:
A year with 13 full moons instead of 12 posed problems for the monks in charge of the calendars. "This was considered a very unfortunate circumstance, especially by the monks who had charge of the calendar of thirteen months for that year, and it upset the regular arrangement of church festivals. For this reason, thirteen came to be considered an unlucky number."[23]
[en.WP: Thirteen (number)]
- It is notable that time keeping was not an exact science. The two-faced Roman god Janus gave it's name to January, technically the eleventh after December (decem- "10") with New Year's occasionally in March. The year "consisted of ten months, beginning in spring with March and leaving winter as an unassigned span of days before the next year" (en.WP: January), "Plutarch's Parallel Lives recounts that Romulus's calendar had been solar but adhered to the general principle that the year should last for 360 days. Months were employed secondarily and haphazardly, with some counted as 20 days and others as 35 or more." "According to Livy, it was Numa Pompilius, the second king of Rome (715–673 BC), who divided the year into twelve lunar months (History of Rome, I.19)." (en.WP: Roman Calender)
It is notable, therefore, that Latin liba is a potential epithet of Janus, "which remains purely hypothetical." (en.WP: Janus), sacrifical cake, cp. λιπάω, libate.
Une des épithètes de Janus est dérivée de l'offrande qu'on lui présentait le plus souvent. Lydus raconte que Yarron, dans le quatorzième livre Rerum divinârum mentionnait cette épithète, et il la traduit par Ποπάνων. Je pense que la forme latine aura été Libo. Il était d'usage de présenter au dieu des liba (πόπανα) les jours des calendes.
[J. S. Speÿer, "Le dieu romain Janus", Revue de l'histoire des religions 26 (1892): 28; emphasis mine, @vectory]
Note also "the Saxon term Wulf-monath (meaning "wolf month")" (en.WT: January) corresponds approximately to Dutch Low Saxon elf, German Low German elf, eleve, ölve, ölven, Old Saxon ellevan (en.WT: PGem. *ainalif "eleven") and English one /w-/. Janus and January were sometimes equated with goddess Juno or IUNO, whose name is possibly cognate to Proto-West-Germanic *aiw-, it may still be difficult to connect the forms of Old Saxon ēo, ēu, ē and Old High German ēo, but there is no good reason either to think of a wolf month. Compare eve, "with a pre-1200 loss of the terminal '-n'" (en.WT) or rather advent, aventura. So, leave can be associated with hollidays, cp. German Urlaub.
- On another note, leap year is "Probably from a much older formation related to the Old Norse hlaup-ár (“leap year”), Old English mōnan hlȳp (“moon's leap”) (compare Latin saltus lunae used in reckoning lunar months on a 19-year cycle)." (en.WT: leap year)
The chief competitor of the system which Dionysius introduced into the West was that constructed in the fifth century by Victorius of Aquitaine, which held its ground in Gaul for nearly three hundred years. Both were based on the lunar cycle of nineteen years, but they differed in four points: the earliest permissible date of the vernal new moon, the earliest day after this on which Easter could be kept, the latest day on which Easter could fall, and the place in each cycle in which the lunar year should be shortened by one day (the saltus lunae).
[Poole, Reginald L. “The Earliest Use of the Easter Cycle of Dionysius.” The English Historical Review 33, no. 129 (1918): 57.]
Note that saltus lunae is Medieval not Classical Latin (etymonline: leap year) so it does not constitute independent evidence for leap year, inasmuch as a folk etymology is always subject to interpretation, because: "There was considerable confusion about dates and the systems upon which bishops, abbots, and secular courts depended in the eighth and ninth centuries, and it was a problem to say which calendar applied to any event, or just how it applied." (Stevens, Wesley M. Speculum 78, no. 1 (2003): 144) With German Schaltjahr ("leap year") in mind, we can safely identify saltus as a corrupted form of kalendae / calendae, cp. yell, see also shall and the irregular change of German sollen.
- L. linquo "leave" compares better to left (hand side) than levis, light, cp. sinister and sine (if you dpn't know what this means you can safely ignore it and head on over to etymonline).
en.WP: Twelve with one reference to OED ed. 1 and one reference to IEW *-lif, *lib