Re: Why not 13 months? (Was La Systeme Metrique)

Michael J. Tuvin (
15 Jul 1995 22:33:27 GMT

The following is a follow-up to "jewish calendar". - I don't know how to
include the posting (Rick Goldstein's) to which I am replying.

Jewish calendar has several unique features. It is neither solar, nor lunar,
but both. In other words it is based on solar year but subdivided by lunar
months. In order to avoid fractionating (1 solar year is not equal to 12
lunar months; 1 lunar month contains a bit more than 28 solar days) an
additional month of Adar is added once in a while. It made Jewish
calendar cyclical but aperiodical. There are no two years which are
identical in terms of distribution of months, weeks and days; each year has a unique pattern of
In ancient times (before the destruction of the Temple and Jerusalem by
Romans) the beginning of every month was actually established and
announced in the Temple court on the basis of witness' testimony that he
saw a new (young) moon. It worked just fine for a couple of millenia, but
then the neccessity arose to have it "tabulated" for the Diaspora
communities (news of a new month arrival travelled too slow before the
dawn on the Internet) and also because of the impossibility to continue
the witness'practice.
The problem is that the length of the lunar month varies rather broadly
because of the complicacy of Moon's movemements. If I am not mistaken it
varies in the range of several minutes (4-6?). Thus it is very
difficult to get a reasonably correct average length. Astronomy of
that period is not known to be able to make such measurements. Ancient Jews
never were notorious astronomers, built any astronomical devices and
weren't all that interested in the subject.
If calendar is based on an incorrect estimate of an average month's
length, error could rapidly accumulate and this solar-lunar balance will
be all screwed up in a matter of a century or so: spring holidays will be
in autumn etc.
Amazingly enough Jewish calendar is based on the lunar month's length of
an impeccable precision - only ~0.3 sec different from that known today
(calculated in 1920-1930-ies [?]).
Go figure how they managed it...
Michael Tuvim