The Hebrew calendar or Jewish calendar is the annual calendar used in Judaism. It determines the Jewish holidays, which Torah portions to read, Yahrzeits, and which set of Psalms should be read each day. Originally observational, it is now a rule-based lunisolar calendar, using both lunar months and years defined via a solar cycle. This is in contrast to the Gregorian calendar, which is based solely upon a solar cycle, or the Islamic calendar, which is purely lunar. Although the Chinese calendar is also a lunisolar calendar, its rules are totally different. All seasons mentioned herein are northern hemisphere seasons because the Hebrew calendar developed in the region east of the Mediterranean Sea.
Two major forms of the calendar have been used: an observational form used before the destruction of the Second Temple in 70 and a rule-based form first fully described by Maimonides in 1178. The period between 70 and 1178 is a transition period between the two forms, with the gradual adoption of more and more of the rules characteristic of the modern form. Except for the modern year number, the modern rules reached their final form before 820 or 921, with some uncertainty regarding when (see below). The modern Hebrew calendar cannot be used for Biblical dates because new moon dates may be in error by up to four days and months may be in error by up to four months. The latter accounts for irregular intercalation such as the three successive years which were given extra months during the early second century according to the Talmud.
Jews have been using a lunisolar calendar since Biblical times, but originally referred to the months by number rather than name. Although the Bible never mentions an embolismic (extra) month, it must have existed in order to keep the first month in spring. Only four pre-exilic month names appear in the Tanakh (the Old Testament), Abib (first), Ziv (second), Ethanim (seventh), and Bul (eighth). All are Canaanite and at least two are Phoenician names. Indeed, all of the months may have been identifiable via either native Jewish numbers or foreign Canaanite/Phoenician names, but the other names were simply not recorded in the Bible. During the Babylonian exile, immediately after 586 BCE, Jews adopted Babylonian names for the months. Some sects, such as the Essenes, used a solar calendar during the last two centuries BCE.
In Second Temple times, the beginning of each lunar month was decided by two eyewitnesses testifying to having seen the new crescent moon. Patriarch Gamaliel II (c. 100) compared these accounts to drawings of the lunar phases. According to arguments offered a millennia later, these observations were compared against calculations made by the main Jewish court, the Sanhedrin. Whether an embolismic month (a second Adar) was needed depended on the condition of roads used by families to come to Jerusalem for Passover, on an adequate number of lambs which were to be sacrificed at the temple, and on the earing of barley needed for first fruits. The beginning of each Hebrew month, once decided, was first announced to other communities by signal fires lit on mountaintops, but the Samaritans and Boethusaeans lit false fires, leading to the use of special messengers. But they could not reach the communities outside Palestine within one day, so the outlying communities opted to celebrate scriptural festivals a second day as well, the "second feast-day of the Diaspora". During the time of the Amoraim (third to fifth centuries) calculations were increasingly used, e.g. by Samuel the astronomer during the first half of the third century who stated that the year contained 365 ¼ days, and by "calculators of the calendar" about 300. Jose, an Amora who lived during the second half of the fourth century, stated that the feast of Purim, 14 Adar, could not fall on a Sabbath or a Monday, lest 10 Tishri fall on a Friday or a Sunday. This indicates a fixed number of days in all months from Adar to Elul, also implying that the extra month was already a second Adar added before the regular Adar.
The Roman-Jewish wars of 66–74, 115–117, and 132–135 caused major disruptions in Jewish life, also disrupting the calendar. During the third and fourth centuries, Christian sources describe the use of eight, nineteen, and 84 year lunisolar cycles by Jews, all linked to the civil calendars used by various communities of Diaspora Jews, which were effectively isolated from Levant Jews and their calendar. Some assigned major Jewish festivals to fixed solar calendar dates, whereas others used epacts to specify how many days before major civil solar dates Jewish lunar months were to begin.
The Ethiopic Christian computus (used to calculate Easter) describes in detail a Jewish calendar which must have been used by Alexandrian Jews near the end of the third century. These Jews formed a relatively new community in the aftermath of the annihilation (by murder or enslavement) of all Alexandrian Jews by Emperor Trajan at the end of the 115–117 war. Their calendar used the same epacts in nineteen year cycles that were to become canonical in the Easter computus used by almost all medieval Christians, both those in the Latin West and the Greek East. Only those churches beyond the eastern border of the Byzantine Empire differed, changing one epact every nineteen years, causing four Easters every 532 years to differ.
A popular tradition holds that Patriarch Hillel II revealed the continuous calendar in 359 due to Christian persecution, formerly a secret known only to the 'calendar committee', a council of sages. This tradition was first mentioned quite late by Hai Gaon (died 1038). But the Talmud, which did not reach its final form until c. 500, does not mention it or even anything as mundane as the nineteen year cycle or the length of any month, despite discussing the characteristics of earlier calendars. Furthermore, Jewish dates during post-Talmudic times (specifically in 506 and 776) are impossible using modern rules. Instead, all evidence points to the arithmetic rules of the modern calendar being developed in Babylonia during the times of the Geonim (seventh to eighth centuries). Most of the modern rules were in place by about 820 as described by the Muslim astronomer al-Khwarizmi. A notable exception was the epoch (the beginning of year 1), which was placed one year later than that of the modern calendar.
In 921, Aaron ben Meir , a person otherwise unknown, sought to return the authority for the calendar to Palestine by asserting that the first day of Tishri should be the day of the new moon unless the new moon occurred more than 642 parts after noon, when it should be delayed by one or two days. The Babylonian rules delayed the first day of Tishri when the new moon occurred after noon. One possible explanation is that local time on the Babylonian meridian is 642 parts later than on the meridian of Jerusalem, so ben Meir was asserting that the calendar should be run from Jerusalem, not Babylon. An alternative explanation for the 642 parts is as follows: The calculated time of New Moon during the six days of creation was on Friday at 14 hours exactly (day starting at 6pm the previous evening), assuming that creation occurred in the Autumn to coincide with Rosh Hashana. However, it was at 9 hours and 642 parts on Wednesday if creation actually took place six months earlier, in Spring. Ben Meir may thus have believed, along with many earlier Jewish scholars, that creation occurred in Spring and the calendar rules had been adjusted by 642 parts to fit in with an Autumn date. In any event he was opposed by Sa'adiah Gaon and in the end all Jewish communities ignored his opinion, but accounts of the controversy show that all of the rules of the modern calendar (except for the year) were in place before 921. In 1000, the Muslim chronologist al-Biruni also described all of the modern rules except that he specified three different epochs used by various Jewish communities being one, two, or three years later than the modern epoch. Finally, in 1178 Maimonides described all of the modern rules, including the modern epochal year.
The epoch of the modern Hebrew calendar is 1 Tishri AM 1 (AM = anno mundi = in the year of the world), which in the proleptic Julian calendar is Monday, October 7, 3761 BCE, the equivalent tabular date (same daylight period). This date is about one year before the traditional Jewish date of Creation on 25 Elul AM 1. A minority place Creation on 25 Adar AM 1, six months earlier, or six months after the modern epoch. Thus adding 3760 to any Julian/Gregorian year number after 1178 will yield the Hebrew year number beginning in autumn (add 3759 for that ending in autumn). Due to the slow drift of the Jewish calendar relative to the Gregorian calendar, this will be true for about another 20,000 years.
The Hebrew month is tied to an excellent measurement of the average time taken by the Moon to cycle from lunar conjunction to lunar conjunction. Twelve lunar months are about 354 days while the solar year is about 365 days so an extra lunar month is added every two or three years in accordance with a 19-year cycle of 235 lunar months (12 regular months every year plus 7 extra or embolismic months every 19 years). The average Hebrew year length is about 365.2468 days, about 7 minutes longer than the average tropical solar year which is about 365.2422 days. Approximately every 216 years, those minutes add up so that the Hebrew year is "slower" than the average solar year by a full day. Because the average Gregorian year is 365.2425 days, the average Hebrew year is slower by a day every 231 Gregorian years. During the last century a number of Jewish scholars suggested that the chief rabbinate in Jerusalem consider modifying this rule to avoid this effect.
There are exactly 14 different patterns that Hebrew calendar years may take. Each of these patterns is called a "keviyah" (Hebrew for "species"), and is distinguished by the day of the week for Rosh Hashanah of that particular year and by that particular year's length.
- A chaserah year (Hebrew for "deficient" or "incomplete") is 353 or 383 days long because a day is taken away from the month of Kislev. The Hebrew letter ח "het", and the letter for the weekday denotes this pattern.
- A kesidrah year ("regular" or "in-order") is 354 or 384 days long. The Hebrew letter כ "kaf", and the letter for the week-day denotes this pattern.
- A shlemah year ("abundant" or "complete") is 355 or 385 days long because a day is added to the month of Heshvan. The Hebrew letter ש "shin", and the letter for the week-day denotes this pattern.
A variant of this pattern of naming includes another letter which specifies the day of the week for the first day of Pesach (Passover) in the year.
Every hour is divided into 1080 halakim or parts. A part is 31/3 seconds or 1/18 minute. The ultimate ancestor of the helek was a small Babylonian time period called a barleycorn, itself equal to 1/72 of a Babylonian time degree (1° of celestial rotation). Actually, the barleycorn or she was the name applied to the smallest units of all Babylonian measurements, whether of length, area, volume, weight, angle, or time. But by the twelfth century that source had been forgotten, causing Maimonidies to speculate that there were 1080 parts in an hour because that number was evenly divisible by all numbers from 1 to 10 except 7. But the same statement can be made regarding 360. The weekdays start with Sunday (day 1) and proceed to Saturday (day 7). Since some calculations use division, a remainder of 0 signifies Saturday.
The calendar is based on mean lunar conjunctions called "molads" spaced precisely 29 days, 12 hours, and 793 parts apart. Actual conjunctions vary from the molads by up to 7 hours in each direction due to the nonuniform velocity of the moon. This value for the interval between molads (the mean synodic month) was measured by Babylonians before 300 BCE and was adopted by the Greek astronomer Hipparchus and the Alexandrian astronomer Ptolemy. Its remarkable accuracy was achieved using records of lunar eclipses from the eighth to fifth centuries BCE. Measured on a strictly uniform time scale, such as that provided by an atomic clock, the mean synodic month is becoming gradually longer, but since the rotation of the earth is slowing even more the mean synodic month is becoming gradually shorter in terms of the day-night cycle. The value 29-12-793 was almost exactly correct in 1 CE and is now about 0.6 s per month too great. However it is still the most correct value possible as long as only whole numbers of parts are used. Especially, it is far more accurate than the average solar year due to the 19-years-235-months equality described above — the total accumulated error of 29-12-793 from its Babylonian measurement until the present amounts to only about five hours.
The 19 year cycle has 12 common and 7 leap years. There are 235 lunar months in each cycle. This gives a total of 6939 days, 16 hours and 595 parts for each cycle. Due to the vagaries of the Hebrew calendar, 19 Hebrew years can be either 6939, 6940, 6941, or 6942 days each. To start on the same day of the week, the days in the cycle must be divisible by 7, but none of these values can be so divided. This keeps the Hebrew calendar from repeating itself too often. The calendar almost repeats every 247 years, except for an excess of 50 minutes (905 parts). So the calendar actually repeats every 36,288 cycles (every 689,472 Hebrew years).
Leap years of 13 months are the 3rd, 6th, 8th, 11th, 14th, 17th, and the 19th years beginning at the epoch of the modern calendar. Dividing the Hebrew year number by 19, and looking at the remainder will tell you if the year is a leap year (for the 19th year, the remainder is zero). A Hebrew leap year is one that has 13 months in it, a common year has 12 months. A mnemonic word in Hebrew is GUCHADZaT (the Hebrew letters gimel-vav-het aleph-dalet-zayin-tet, i.e. 3, 6, 8, 1, 4, 7, 9. See Hebrew numerals). Another mnemonic is that the intervals of the major scale follow the same pattern as do Hebrew leap years: a whole step in the scale corresponds to two common years between consecutive leap years, and a half step to one common between two leap years.
A Hebrew common year will only have 353, 354, or 355 days. A leap year will have 383, 384, or 385 days.
Although simple math would calculate 21 patterns for calendar years, there are other limitations which mean that Rosh Hashanah may only occur on Mondays, Tuesdays, Thursdays, and Saturdays (the "four gates"), according to the following table:
| Day of Week
|| Number of Days
| Monday || 353 || 355 || 383 || 385
| Tuesday || 354 || || || 384
| Thursday || 354 || 355 || 383 || 385
| Saturday || 353 || 355 || 383 || 385
Basically, the Hebrew months alternate between a long month and a short month, that is:
- Tishrei (30 days)
- Cheshvan (also spelled Heshvan or Marchesvan) (29 or 30 days)
- Kislev (30 or 29 days)
- Tevet (29 days)
- Shevat (30 days)
- Adar (29 days)
- Nisan (30 days)
- Iyar (29 days)
- Sivan (30 days)
- Tammuz (29 days)
- Av (30 days)
- Elul (29 days)
For leap years, a 30 day month of Adar I is added immediately after the month of Shevat, and the 29 day Adar is called Adar II. This is to ensure that the months remain at the same season rather than continuing to drift earlier by about 11 days per year.
The 265 days from the first day of the 29 day month of Adar (the last of the religious year) and ending with the 29th day of Heshvan forms a fixed length period that has all of the festivals specified in the Bible, such as Pesach (Nisan 15), Shavuot (Sivan 6), Rosh Hashana (Tishri 1), Yom Kippur (Tishri 10), Sukkot (Tishri 15), and Shemini Atzeret (Tishri 22).
The festival period from Pesach up to and including Shemini Atzeret is exactly 185 days long. The time from the traditional day of the vernal equinox up to and including the traditional day of the autumnal equinox is also exactly 185 days long. This has caused some unfounded speculation that Pesach should be March 21, and Shemini Atzeret should be September 21, which are the traditional days for the equinoxes. Just as the Hebrew day starts at sunset, the Hebrew year starts in the Autumn (Rosh Hashanah), although the mismatch of solar and lunar years will eventually move it to another season if the calendar isn't reformed (this will not happen for thousands of years).
Karaites use the lunar month and the solar year, but determine when to add a leap month by observing the ripening of barley in Israel, rather than a fixed calendar. This occasionally puts them a month out of sync with the rest of the Jews. (For several centuries, most Karaites, especially outside Israel, have kept in step with other Jews for the sake of simplicity. However, in recent years many Karaites have reverted to their traditional practice.)
Accuracy of the Jewish Calendar
The length of the month assumed by the calendar is correct within a fraction of a second. There will thus be no significant errors from this source for a very long time. However, the assumption that 19 years exactly equal 235 months is wrong, so the average length of a 19 year cycle is too long (compared with 19 tropical years) by about 0.088 days or just over 2 hours. Thus on average the calendar gets further out of step with the tropical year by roughly one day in 216 years. If the intention of the calendar is that Pesach should fall on the first full moon after the vernal equinox, this is still the case in most years. However, at present three times in 19 years Pesach is a month late (as in 2005). Clearly, this problem will get worse over time and if the calendar is not amended, Pesach and the other festivals will progress through a complete cycle of seasons in about 79,000 years.
- The Code of Maimonidies (Mishne Torah), book three, treatise eight: Sanctification of the New Moon. Translated by Solomon Gandz. Yale Judaica Series XI, Yale University Press, New Haven, Conn., 1956.
- Ernest Wiesenberg. "Appendix: Addenda and Corrigenda to Treatise VIII". The Code of Maimonidies (Mishne Torah), book three: The Book of Seasons. Yale Judaica Series XIV, Yale University Press, New Haven, Conn., 1961. pp.557-602.
- Samuel Poznanski. "Calendar (Jewish)". Encylopædia of Religion and Ethics, 1911.
- F.H. Woods. "Calendar (Hebrew)", Encylopædia of Religion and Ethics, 1911.
- Otto Neugebauer. Ethiopic astronomy and computus. Österreichische Akademie der Wissenschaften, philosophisch-historische klasse, sitzungsberichte 347. Vienna, 1979.
- Ari Belenkiy. "A Unique Feature of the Jewish Calendar — Dehiyot". Culture and Cosmos 6 (2002) 3-22.
- Arthur Spier. The Comprehensive Hebrew Calendar. Feldheim, 1986.
- L.A. Resnikoff. "Jewish calendar calculations", Scripta Mathematica 9 (1943) 191-195, 274-277.
- W.H. Feldman. Rabbinical Mathematics and Astronomy.