The Jewish Calendar
a study of its elements
Introduction
The Jewish calendar is a lunar calendar, adjusted so that: the
first month of the year is always in the spring, the days begin
and end at sunset, and the months follow the phases
of the moon. The phases of the moon are a complete cycle of the
apparitions of the moon, from waxing crescent
to first quarter, to full
moon, to last quarter and
to waning crescent, as illustrated by
the animated image on this page.
Between the waning cresent at the end of the month and the
waxing wrescent at the beginning of the next month, there is a
period of time in which the moon can not be seen. During this
period of time a new lunar cycle begins all over again, and is
called the time of the new moon, or the
time when the cycle of the phases of the moon begins anew.
The determination of this time of the new moon is fraught with
difficulty, because the moon is actually invisible during this
period of time. As a result, there have been many different means
employed to determine the time of the new moon. This
determination is important, because it defines the beginning of
the lunar month, the beginning of the
cycle of the visible phases of the moon.
The New Moon
In ancient times, the new moon was determined by several
different methods which became more reliable as knowledge of
astronomy increased. By the time of the prophet Samuel, it was
known that the lunar month was between 29 and 30 days long; so it
became possible to predict within one day the time of the new
moon. Visual observation was used around the predicted time of
the new moon to decide if a month should end at 29 days or at 30
days. Each day was defined to begin and end at sunset.
For example, if at the end of the 29th day a faint waxing
crescent could not be seen in the west just after sunset, then
the end of the month was delayed (i.e., postponed) by one more
day, making the month 30 days in length. Otherwise, if at the end
of 29 days at sunset a faint crescent could be seen, then the
current month was terminated at 29 days and the next day was
declared to belong to the new month. That sighting of the faint
crescent meant that a new moon had occured, because the cycle of
lunar phases had started all over again.
This method of visual determination required fair weather. If
the sky was cloudy, then it would not be possible to sight the
new moon and decide if the passing month should be 29 or 30 days
long. As a result, a special rule was formulated to deal with the
problem of cloudy skies. This rule stipulated that there could be
no more than eight 30-day months in a
year. That way, if bad weather occured during the time of all the
new moons in a year, then at least four of
those months would have to be 29-day months. The reason for this
was so that the average length of a calendar month would be close
to 29.5 days.
If the average length of the calendar months deviated too much
from 29.5 days, then some months might start several days after
the waxing crescent. That would defeat the whole purpose of a
lunar calendar, which is for the calendar months to follow the
phases of the moon.
Because of the necessity to keep the months close the the
cycle of the lunary phases, another rule was established similar
to the first rule. This second rule stipulated that there could
be no more than eight 29-day months in
a year. That way, if unusually good weather and keen eyesight
resulted in terminating eight months at 29 days, then no matter
what happened in the year after that, all the months would have
to be 30 days long, until the beginning of the next year. These
two rules kept the average length of the calendar months close to
29.5 days.
Another way of summarizing the above two rules, is to say that
in ancient times there could be no more than
eight 29-day months, and no less than
four 29-day months. All the remaining
months of the year would be 30 days in length.
The Postponements
By the time of the 9th century C.E., these two rules were
refined by the use of advanced astronomical knowledge. Although
this resulted in a more complex calendar, the result has been a
more accurate one. No other modifications have been made to the
calendar since that time. The ancient rules were replaced with
the new requirements that there can be no more than
seven 29-day months in a year and no less
than five 29-day months in a
year. The increased precision with which these two rules have
been stipulated is the result of advanced astronomical knowledge.
The exact method by which it is determined if a year should have
seven 29-day months, six 29-day months, or five 29-day months is
stipulated by what are known as the postponement
rules, which will be covered in a later section.
The Intercalations
One other aspect of the calendar in ancient times was
harmonizing the calendar with the seasons, so that the first
month always occurs in the spring. The ancient Babylonians
determined that the actual average length of a lunar cycle is 29.53
days, so when it became necessary to adjust the the lunar
calendar to the seasons, the solution was to add a month of 30
days to the end of the year, thus moving the start of the new
year into the spring. The reason for this becomes appearent when
you compare the length of 12 lunar months (354 days = six 30-day
months + six 29-day months) to the length of the seasons (365.25
days). If a calendar had only 12 lunar months, in just three
years the first month of the year would be in the winter, rather
than the spring. By adding one extra 30-day
month approximately every three years,
it was possible to keep the first month in the spring.
A special rule was adopted in ancient times regarding when the
30-day leap month should be inserted into the calendar. It was
decided to insert it just before the last month of the year,
between the 11th month and the 12th month. This same ancient
tradition continues today in the Jewish calendar.
By the time of the 9th century C.E., a set of rules were
established to determine exactly when to insert a leap month; and
it still is exactly as it was in ancient times -- always a 30-day
month, and always inserted between the 11th and 12th months. The
insertion of a leap month is called an intercalation
(pronounced as in-turk-uhl-a-shun).
Thus, there are only two sets of rules used to adjust the
calendar: a set of postponement rules (which determine how many
29-day months there will be in a year) and a set of intercalation
rules (which determine when to insert a leap month of 30 days). I
will take up both these topics in later sections.
The Calendar
Unless one or both of the two sets of rules mentioned above
are invoked, the Jewish calendar simply has six 30-day months,
and six 29-day months, as follows:
Season |
Month |
Days |
Month |
Days |
Season |
Spring |
1. Nisan |
30
|
2. Iyar |
29
|
|
|
3. Sivan |
30
|
4. Tammuz |
29
|
Summer |
|
5. Ab |
30
|
6. Elul |
29
|
|
Fall |
7. Tishri |
30
|
8. Heshvan |
29
|
|
|
9. Kislev |
30
|
10.Tebet |
29
|
Winter |
|
11.Shebat |
30
|
12.Adar |
29
|
|
Note that the average length of all the calendar months is 29.5
days, which is very close to the actaul period of the lunar
phases. Thus, if Nisan starts out in the spring within one day of
the astronomical new moon, then all of the calendar months will
follow the lunar phases to within a day. This is as close as a
calendar based on a 24-hour day can get to the true lunar phases,
because a month may have only 29 or 30 days, not an arbitrary
fraction of days, which would be necessary if one were to track
the lunar phases exactly. If a day were only 12 hours long,
instead of 24 hours, then it would be possible to track the lunar
phases more precisely, since an astronomical phase can occur at
anytime of the day or night; but, a day can only begin and end at
sunset.
Thus, there are three competing criteria
used to determine the Jewish calendar: the day (defined from
sunset to sunset), the month (defined by the cylce of lunar
phases) and the year (defined from spring equinox to spring
equinox). It is not possible to harmonize all of these 3 things
exactly, but the Jewish calendar comes as close as possible to
doing this.
The next two sections explain the adjustments that are
necessary to keep the day, the month and the year harmonized.
Determination of the Leap Year
As mentioned above, 12 lunar months are about 11 days short of
a solar year. A solar year is the cycle of the seasons, from
spring to summer, to fall, to winter, and back to spring again.
Just as the phases of the moon complete a full cycle every lunar
month, so the seasons complete a full cycle every year. So, if we
want to keep the first lunar month in the spring, we will need to
add a leap month approximately every three years. This is exactly
the reason why the Roman calendar has a leap day added every 4
years. It turns out that the cycle of the seasons, from spring to
spring, is 365 and 1/4 days. Therefore, we need to add one day
every four years to the Roman calendar in order to keep the start
of the Roman calendar in the dead of winter. Otherwise, in just
240 years, the Roman calendar would be starting January 1 in the
fall, rather than in the winter.
It turns out that the lunar cycle averages 29.53 days. If we
divide 365.25 days by 29.53 days, we find out that there are
almost exactly 12 and 7/19 months in a year. That means that in
19 years there will be 7 more months than the usual 12. Or, to
put it another way, we need to add seven
leap months in 19 years, in order to keep the first lunar month
in the Spring.
Of course, we would have to spread these seven extra months
evenly over the 19 years, in order to keep the first month of
every year as close to spring as possible. This is a little
different from the Roman calendar, where only one day has to be
added every four years. But, with the lunar calendar, we need to
add seven months every 19 years; so, we have to make a decision
where to put those seven months. We know we need to keep them
spread out evenly, otherwise we would have a wild variation
between the first month and the Spring.
Since there are 4 seasons in a year, and 12 lunar months in an
ordinary year, there are going to be 3 lunar months per season.
But, in a leap year one of those seasons is going to have
four months, since the leap year has 13 months.
This turns out to be the key in determining where to place the
seven extra months. We need to find out what years have four
lunar months in the winter.
Since Passover and Unleavened Bread begin at the time of the
full Moon in the spring, we need to determine which years have four
full moons in the winter. Those years will be the
years when we need to add a leap month, because we require the
Passover (time of the full moon) to be in the spring.
Here is a table of 19 years, showing the number of full Moons
in the winter season. Those years with four
full moons during the winter season are the years where we need
to put a leap month. Note that there are seven
such years.
|
19 years |
Winter of
|
Full Moons
|
Approx. Winter |
Dates of Full Moons
|
Year of
Cycle
|
|
1
|
1981 - 1982
|
3
|
22 Dec - 20 Mar
|
01/09;02/08;03/09
|
4
|
|
2
|
1982 - 1983
|
3
|
22 Dec - 21 Mar
|
12/30;01/28;02/27
|
5
|
*
|
3
|
1983 - 1984
|
4
|
22 Dec - 20 Mar
|
12/20;01/18;02/17;03/17
|
6
|
|
4
|
1984 - 1985
|
3
|
21 Dec - 20 Mar
|
01/07;02/05;03/07
|
7
|
*
|
5
|
1985 - 1986
|
4
|
21 Dec - 20 Mar
|
12/27;01/26;02/24;03/26
|
8
|
|
6
|
1986 - 1987
|
3
|
22 Dec - 21 Mar
|
01/15;02/13;03/15
|
9
|
|
7
|
1987 - 1988
|
3
|
22 Dec - 20 Mar
|
01/04;02/02;03/03
|
10
|
*
|
8
|
1988 - 1989
|
4
|
21 Dec - 20 Mar
|
12/23;01/21;02/20;03/22
|
11
|
|
9
|
1989 - 1990
|
3
|
21 Dec - 20 Mar
|
01/11;02/09;03/11
|
12
|
|
10
|
1990 - 1991
|
3
|
22 Dec - 21 Mar
|
12/31;01/30;02/28
|
13
|
*
|
11
|
1991 - 1992
|
4
|
22 Dec - 20 Mar
|
12/21;01/19;02/18;03/18
|
14
|
|
12
|
1992 - 1993
|
3
|
21 Dec - 20 Mar
|
01/08;02/06;03/08
|
15
|
|
13
|
1993 - 1994
|
3
|
21 Dec - 20 Mar
|
12/28;01/27;02/25
|
16
|
*
|
14
|
1994 - 1995
|
4
|
22 Dec - 21 Mar
|
12/18;01/16;02/15;03/17
|
17
|
|
15
|
1995 - 1996
|
3
|
22 Dec - 20 Mar
|
01/05;02/04;03/05
|
18
|
*
|
16
|
1996 - 1997
|
4
|
21 Dec - 20 Mar
|
12/24;01/23;02/22;03/24
|
19
|
|
17
|
1997 - 1998
|
3
|
21 Dec - 20 Mar
|
01/12;02/11;03/12
|
1
|
|
18
|
1998 - 1999
|
3
|
22 Dec - 21 Mar
|
01/01;01/31;03/02
|
2
|
*
|
19
|
1999 - 2000
|
4
|
22 Dec - 20 Mar
|
12/22;01/20;02/19;03/19
|
3
|
As can be seen from the above table, unless a full moon occurs
within week or less of the winter solstice, there will be only
three full moons during the winter season; but, if the full moon
does occur within a week of the winter solstice, then that year
will be a leap year. The basic pattern is such that there are
four full moons squeezed into the winter season, rather than the
usual three.
The patterns are nearly identical in adjacent 19-year cycles,
so it doesn't matter which year you choose as your starting year,
as long as you are consistent in repeating the same pattern of
leap years in each 19-year cycle. In the above table, I started
with 1981. In the 19-year cycle starting in 1981 there were seven
leap years in years 3,5,8,11,14,16, and 19 of the cycle.
However, the Jewish calendar, for historical reasons, starts
its 19-year cycles in the year 3761 B.C.E. That means that 1981
was the 4th year of a 19-year cycle, which explains the meaning
of the last column above. From the above table, it should be
clear that the years 3,6,8,11,14,17 and 19 in the Jewish 19-year
cycle (which starts in 3761 B.C.E) are leap years.
Thus, to determine if a year is a leap year, one has to find
the location of the year in the 19-year cycle. If that year is
any one of the years 3,6,8,11,14,17 or 19, then that year is a
leap year. In a leap year, one extra month of 30 days is inserted
between the 11th and 12th months.
For example, 1994 was a leap year and here are the lengths of
the months for 1994:
Season |
Month |
Days |
Month |
Days |
Season |
Spring |
1. Nisan |
30
|
2. Iyar |
29
|
|
|
3. Sivan |
30
|
4. Tammuz |
29
|
Summer |
|
5. Ab |
30
|
6. Elul |
29
|
|
Fall |
7. Tishri |
30
|
8. Heshvan |
29
|
|
|
9. Kislev |
30
|
10.Tebet |
29
|
Winter |
|
11.Shebat |
30
|
12.Adar I |
30
|
|
|
|
|
13. Adar II |
29
|
|
Note that the extra 30-day month has been inserted between
Shebat and Adar II (which has 29 days and is the last month of
the year.)
The next section further explains the means by which it is
determined how many 30-day months and how many 29-day months the
year will have.
The Postponement Rules
As mentioned above, in ancient times there could be as few as
four 29-day months, and as many as eight 29-day months in a year.
But, ever since the 9th Century, that variation has been reduced
by one month, so that now there can be as few as five 29-day
months and as many as seven 29-day months -- but, no more and no
less. If it were not for the postponement rules, the calendar
would vary much more in the number of 29-day months allowed.
There has been much debate about the postponement rules for
many centuries because their history is obscure. Some have argued
that they have only a religous basis and have nothing to do with
astronomy. Others have argued that they have everything to do
with astronomy and nothing to do with religion. However, one of
the greatest Jewish scholars, Maimonides, believed that they had
an astronomical basis. From the astronomical facts, it becomes
clear that the postponement rules, whether they were motivated by
astronomy or religion or both, clearly serve an astronomical
purpose.
Before looking at the postponement rules in detail, it is
necessary first to understand what they accomplish. As mentioned
earlier, they determine if a year should have five, six or seven
29-day months. If no variation at all were allowed in the number
of 29-day months, then each year would have six 29-day months and
six 30-day months. A leap year would have six 29-day months and
seven 30-day months.
That means that in every 19-year cycle, there would be 12
years of 354 days, and 7 years of 384 days. That is a total of
6936 days. However, 19 years, counting from spring to spring
equinox, is actually 6939.6 days. That means that some years
simply can't have six 29-day months and six 30-day months;
otherwise, we would be off by nearly four
days every 19 years, and in just 280 years Passover would be
starting in winter, rather than the spring. Cleearly, there has
to be some variation in the number of 29-day months, or the
calendar will not follow the seasons. The calendar would be short
by four days every 19-years, and it
would be cumulative.
The first postponement rule attempts to make this correction
for the missing four days. Since days
begin and end at sunset, and since the astronomical elements of
the Jewish calendar are based on ancient astronomy, which
measured astronomical events from the mean sun at noon, there are
6 mean hours from mean noon to mean sunset. If a new moon is
calculated to fall during that part of the day between noon and
sunset, then the length of the lunar year is increased from 354
days to 355 days by adding 1 day to the 8th month (which
ordinarily would have 29 days). This reduces the
number of 29-day months from six down to five
29-day months. The change is made in the 8th month in order not
to affect the Holy Days (annual Sabbaths in addition to the
weekly Sabbath), which occur only during the first seven months
of the year.
The first postponement rule occurs approximately once every
four years, because a new moon can occur at any time of day or
night, and since 6 hours is one quarter of a day, there is 1
chance in 4 that a new moon will fall in the interval between
noon and sunset.
Since there is only one new moon that needs to be checked each
year to determine the length of the year, there are two logical
choices for the new moon to be selected for determining the
length of the year. One choice is the first new moon of the year,
at the very end of the 12th or 13th month. Another choice would
be the last month of the holy day seasons. Since there are in
fact a fixed number of days, always, in the first seven months of
the year, the choice of which of those seven months to select for
determining the length of the year is somewhat arbitrary. Since
the new moon of the 7th month, the Feast of Trumpets, is the only
new moon of those seven months that is a Holy Day (annual
Sabbath), it is the 7th month that is selected to determine the
length of the year.
Thus, the first postponement rule is: if the new moon of the 7th
month falls between noon and sunset, then the 8th month of the previ
ous
year is increased to 30 days. Note that the extra
day is added in the previous year, not in
the month after Trumpets. This is
important to notice, because the aim is to keep the new moon of
Trumpets in one year from postponing the Trumpets for the
following year. This very fact is often little understood by
those who have argued against the postponement rules. They have
argued that the postponement in one year causes a postponement in
the next year; but they are wrong. The Jewish calendar has very
strict rules to prevent this from ever happening!
However, as can be seen, adding one extra day to the length of
the year approximately once in four years is too much of a
correction. That means that there must also some years in which
either the first rule is cancelled, or there must be some
neighboring year shortened by one day. The remaining postponement
rules accomplish just that.
As mentioned before, there has been much debate as to whether
the remaining postponements needed to be specified for religious
or for astronomical reasons, or if they are the only means for
either cancelling the first postponement or shortening a
neighboring year by 1 day, in order to reduce the over-corrections
caused by the first rule. In any case, the remaining rules either
cancel the first rule, or they cause the year to be shortened by
1 day. This is done by increasing the the number of 29-day months
from six to seven 29-day months, by subtracting one day from the
9th month.
It can also happen that the remaining rules can lengthen a
year, so two additional rules were also adopted to guarantee that
the year can never vary by more than one
strong>
day from the normal length of 354 days in an ordinary year, or
384 days in a leap year.
Again, all of the changes are made right after the 7th month,
in either the 8th month (by adding one day) or in the 9th month(by
subtracting one day), so that the start of the next year is
either accelerated by one day or delayed by one day. This gives
the calendar much more stability than it had in ancient times.
Now, the year can only have five, six or seven 29-day months,
whereas in ancient times it could have had four, five, six, seven
or even eight 29-day months!
The net result of the postponement rules is to fix the number
of 29-day months at five, six or seven; and, to keep the length
of the 19-year cycle at 6939 or 6940 days.
It is important to understand these aspects of the
postponement rules before descrbing them in detail, because it is
possible to lose sight of the fact that they actually have a
valid astronomical affect, even though they are couched in
religous terms. This is not to deny that there may very well be a
religious foundation for the rules, but it is not a purely
religous foundation, for if it were purely religous then
abolishing them would have no affect on the synchrony between the
calendar, the lunar phases and the seasons. Yet, these rules
actually do have more than a purely religious affect on the
calendar.
The Religous Statement of the Postponements
under construction.
Bibliography
Goldstine, Herman H.,"New and full moons
from 1001 BC to AD 1651", Amr. Phil. Soc., Philadelphia,
1973.
Hamburger, M.,"Journal fuer die reine und
angewandte Mathematik", Crelle, Berlin, 1896, vol. 116, pp.
90-96.
Hastings, James, ed., "Encyclopedia of
Religion and Ethics", vol. 3, pp. 117-124, Scribner &
Sons, NY, 1924.
Lindo, E. H.,"A Jewish calendar for sixty-four
years", L. Thompson, London, 1838.
Milikowsky, Chaim Joseph,"Seder Olam: a
rabbinic chronography", vol. 1, Yale University, 1981.
Neugebauer, O.,"A History of Ancient
Mathematical Astronomy", Springer-Verlag, 3 vols.
Ptolemy, C.,"The Almagest", Encylopedia
Britannica Great Books series, vol. 16.
Resnikoff, Louis A., "Scripta Mathematica",
1943, vol. 9, no. 3, pp. 191-195 and vol. 9, no. 4, pp. 274-277.
Schwarz, Adolf,"Der Juedische Kalendar:historisch
und astronomisch untersucht",Breslau, 1872.
Seidelmann, P. K., ed.,"Explanatory
Supplement to the Astronomical Almanac", US Naval Obs., Wash.
DC, 1992.