Me:
> ?? Let's get this right. The time it takes for the moon to go
> around the earth is said to take 27.321661 days, no?
Richard:
> No! That's the sidereal period, i.e. with reference to the fixed
> stars. Relative to the Earth-Sun axis (which moves with respect to
> the fixed stars), it's a bit over 29 and a half days [...]
Yes, but isn't that really irrelevant? Look again. The criteria here
is that we use whole numbers, not fractions, and that we want a
_multi-cycle_ coordination of the sun, moon and Venus (the three
brightest objects in the sky, coincidentally). It looks to me that using
a common multiple 7 is the most optimal solution by far under these
stricter rules.
lunar cycle: 7 * 4 = 28 (off by one day, 27.3 or 29.5 days)
Venus cycle: 7 * 32 = 224 (off by one day; 225 days)
solar year: 7 * 52 = 364 (off by one day; 365.25 days)
Everything is off by ONLY one day using seven as a common multiple!
Since 52 in the solar year is divisible by 4, we can say alternatively
that 7*(4*13) = 364, but 7*4 is the length of the lunar cycle as
established in the first line! This is absolutely perfect... almost
sacred in fact. Holy Mother of Joseph! Can you do better with whole
numbers to get all three cycles with more accuracy than that? Can you
use "8" like that? Or "6"? Good luck. If you can, I will retreat from my
mathematical amusement but damn if this ain't looking good.
Another thing I notice is this. Reconfigured from above, we have:
lunar cycle: 7 * 4 = 28
Venus cycle: 7 * 4 * 8 = 224
solar year: 7 * 4 * 13 = 364
Seems that "four" is also a holy multiple here too. It's like all the
cycles can be multiples of the approximate lunar cycle (7*4). Hmm.
So to cleverly tie it back with IE, maybe _that's_ why *septm would have
been borrowed like in many other languages from agriculture-savvy Semites.
Not because of a taboo, but because it's divinely mathematical and so
handy for off-the-cuff calculations of cycles.
= gLeN