An Examination of Anderson's Chronological Errors
Regarding Daniel 9's First 69 Weeks
by Bob Pickle
But 476 × 365 =
Add (14 March to 6th April, both inclusive)
Add for leap years
And 69 weeks of prophetic years of 360 days (or 69 × 7 × 360) 173, 880 days. (p. 224)
That is Sir Robert Anderson's theory about the first 69 weeks, and it doesn't hold water.
The first problem that is readily apparent about Anderson's scenario was his confusion about the Julian and Gregorian calendars.
And secondly, the Julian year is 11m. 10 46s., or about the 129th part of a day, longer than the mean solar year. The Julian calendar, therefore, contains three leap years too many in four centuries, an error which had amounted to eleven days in A. D. 1752 when our English calendar was corrected by declaring the 3rd September to be the 14th September, and by introducing the Gregorian reform which reckons three secular years out of four as common years; ex. gr., 1700, 1800 and 1900 are common years, and 2000 is a leap year. (p. 225)
This is true. But then why did Anderson use Gregorian years when calculating the number of days between two Julian dates? If we use Julian dates, we must use Julian years, and if we use Gregorian dates, we must use Gregorian years. We cannot mix the two calendars in the way that he proposes.
Anderson was thus 3 days off in his calculation, for there are really 173,883 days inclusive between Friday, March 14, 445 BC, and Sunday, April 6, 32 AD. Instead of adding 116 days for leap years, Anderson should have added 119, for that is precisely how many leap years there are in 476 years in the Julian calendar.
If Anderson had wanted to use Gregorian years, he should have started off with the Gregorian dates of Saturday, March 9, 445 BC, and Sunday, April 4, 32 AD (Mar. 9, 445 BC Gregorian = Mar. 14, 445 BC Julian; Apr. 4, 32 AD Gregorian = Apr. 6, 32 AD Julian). But when we add 116 days for leap years to the number of days between these two dates, we still end up with 173,883 days. Only by mixing the two calendars does it falsely appear that there are 173,880 days. (See Figures 1 and 2.)
Another freeware program is Online Bible. You can use its Astro-Calendar module to ascertain dates for new moons, to convert Gregorian dates to Julian and vice versa, and to verify the Julian Day number for any date in history. Finding the difference between two Julian Day numbers is the easiest way to compute the number of days between two dates.
Nor was this Anderson's only error. If the new moon occurred at 7:09 AM on March 13, 445 BC, as he supposed, the new crescent would not be old enough and large enough and far enough away from the setting sun by that evening to be seen. As Anderson himself quotes:
"The month began at the phases of the moon . . . and this happens, according to Newton, when the moon is eighteen hours old. . . ." (p. 217)
At sunset on the evening of March 13th it would have been but 11 hours old, too young to be seen. Thus the new month could not have begun until the evening of March 14th, making March 15th the first day of the new month, not March 14th.
Jewish months back then, as Islamic months do today, began when the new crescent moon could barely be seen with the naked eye soon after sunset. MoonCalc is a freeware Islamic calendar program one can use to calculate when the new crescent can first be seen, weather permitting. You can download this program by clicking either here or here. Figure 3 contains a screenshot of its calculations for when the new crescent can be seen in March of 445 BC.
We have two resources that tell us something about the calendar back in the 5th century BC: Jewish scribal papyri from Elephantine, Egypt, and cuneiform tablets from Mesopotamia. Both indicate that March 14 is too early in the year to be considered Nisan, the first month of the Jewish year. At Elephantine, Nisan 1 ranged from March 26 (in 446 and 428 BC) to April 24 (in 465 BC) (Siegfried Horn and Lynn Wood, The Chronology of Ezra 7, p. 157-159). In Babylonia, Nisan 1 ranged from March 26 to April 23 for the years 464 BC to 400 BC (Richard Parker and Waldo Dubberstein, Babylonian Chronology, p. 32, 33). Thus in 445 BC, Nisan would have begun after the new moon of April, not after the new moon of March, making April 13 Nisan 1, not March 14.
Besides, it is hard to conceive how the barley could have been ripe enough for the wave sheaf offering that took place on Nisan 16 if Nisan 1 was as early as March 14. That date is just too early in the year.
Another error concerns the date of 445 BC. One issue involved with this is when Artaxerxes took the throne. According to Anderson:
The murder of Xerxes and the beginning of the usurper Artabanus's seven months' reign was in July B.C. 465; the accession of Artaxerxes was in February B.C. 464; one or other of these dates, therefore, must be the epoch of Artaxerxes' reign. (The Coming Prince, p. 63)
It is not known for sure when Xerxes was murdered.
According to Aramaic Papyri 6 from Elephantine, the date of January 2/3, 464 BC, was in Artaxerxes' accession year (The Chronology of Ezra 7, p. 100-102). What was his "accession year"? That would be the time between the previous king's death and the next New Year's Day. Artaxerxes' year 1 would thus have begun with the following Nisan (March/April) in the Persian calendar and the following Tishri (September/October) in the Jewish calendar.
But as Nehemiah mentions the Chisleu (November) of one year, and the following Nisan (March) as being both in the same year of his master's reign, it is obvious that, as might be expected from an official of the court, he reckons from the time of the king's accession de jure, that is from July B.C. 465. The twentieth year of Artaxerxes therefore began in July B.C. 446, and the commandment to rebuild Jerusalem was given in the Nisan following. The epoch of the prophetic cycle is thus definitely fixed as in the Jewish month Nisan of the year B.C. 445. (p. 63)
Yet since Persian practice was to number the years of their kings from Nisan, not from their anniversary dates, Anderson's explanation must be wrong.
What these verses in Nehemiah really show is that the Jews, in contrast with the Persians, numbered the reigns of foreign kings from their 7th month called Tishri instead of from their 1st month called Nisan. Every 7th month, the king's regnal year increased by one. This is why Nehemiah describes the 9th month Chisleu as coming before the 1st month Nisan.
If Tishri 464 BC began the 1st year of Artaxerxes, then Tishri 445 BC began his 20th year. And that means that Nisan in his 20th year fell in 444 BC, not 445 BC. So Anderson was a year off on his starting date.
You might wonder if Anderson could have started with 444 BC and ended with 33 AD, especially since 33 AD is considered to be one of the possible years of Christ's crucifixion. The answer is no.
First of all, if we count backward from the proposed date of the crucifixion in March of 33 AD, we end up having to celebrate the Passover around a Julian date of March 18 in 444 BC, which corresponds to a Gregorian date of March 13. However, there is absolutely no way that Jews in 445 BC celebrated the Passover eight days before the spring equinox. That is way too early.
We thus must settle for a Nisan 1 that fell no earlier than April 3, 444 BC. It was on the evening of April 2, weather permitting, that the new crescent was first seen, which would make April 3 Nisan 1. 173,880 days later, counting inclusively as Anderson suggested, brings us to Thursday, April 23, 33 AD.
Since the new crescent could not have been seen before the evening of Saturday, April 18, in 33 AD, that would make Nisan 1 fall on Sunday, April 19, or perhaps a day later, and Nisan 5 on Thursday, April 23, or perhaps a day later. Thus Thursday, April 23, could not have been Palm Sunday on Nisan 10.
Worse yet, Nisan 14 would have fallen on Saturday, May 2, or perhaps a day later. But if Nisan 14 fell on a Saturday or Sunday in 33 AD, how could Christ have died on a Friday if He died in that year? Even a hypothetical Wednesday crucifixion does not work when Nisan 14 falls on a Saturday or Sunday.
The only two possible schemes to make dispensational interpretations of the 69 weeks work is to start either in 445 or 444 BC. If neither one fits, as they most certainly do not, Anderson's theory is proven false.
The previous section hints at the problems we face with Anderson's ending date of April 6, 32 AD. His theory called for it to be Nisan 10. He explains it this way:
For example, in A.D. 32, the date of the true new moon, by which the Passover was regulated, was the night (10h 57m) of the 29th March. The ostensible date of the 1st Nisan, therefore, according to the phases, was the 31st March. It may have been delayed, however, till the 1st April; and in that case the 15th Nisan should apparently have fallen on Tuesday the 15th April. (p. 79)
Thus far his explanation proves that he has chosen the wrong date for the 10th of Nisan. If Nisan 15 fell on April 15, then Nisan 10 fell on April 10, not April 6.
But the calendar may have been further disturbed by intercalation. According to the scheme of the eight years' cycle, the embolismal month was inserted in the third, sixth, and eighth years, and an examination of the calendars from A.D. 22 to A D. 45 will show that A.D. 32 was the third year of such a cycle. As, therefore, the difference between the solar year and the lunar is 11 days, it would amount in three years to 33 3/4 days, and the intercalation of a thirteenth month (Ve-adar) of thirty days would leave an epact still remaining of 3 3/4 days; and the "ecclesiastical moon" being that much before the real moon, the feast day would have fallen on the Friday (11th April), exactly as the narrative of the Gospels requires. (pp. 79, 80)
If that didn't make sense, it's because it doesn't make sense. On average, the Jews would add in a 13th month 7 times every 19 years. Since this 13th month was the length of a lunar month, as Anderson admits above, there was no "epact remaining." Thus Nisan 1 would still have begun with observing the new crescent on the evening of March 31st, weather permitting.
Nisan 10 occurred at the earliest on April 10, not April 6 as Anderson supposed.
No one except those who ascribe to Anderson's theory suggest 32 AD as a possible date for the death of Christ. The simple fact is that Nisan 14 in that year would have been on a Monday or Tuesday. It is simply impossible to reconcile this fact with the gospel accounts of the death of Christ.
All agree that Daniel 9's 70 weeks are an allusion to the Jewish practice, commanded in Leviticus 25, to not sow the ground every seventh year. Thus in Daniel 9 we have 70 sabbatical cycles referred to. It would make the most sense to begin these weeks with the beginning of the first year of a seven-year cycle, and end them with the end of the seventh year of a seven-year cycle.
The first sabbatical year after the cross, whenever that was, began in the fall of 33 AD and ended in the fall of 34 AD (see When Were the Sabbatical Years?). Assuming that this marked the end of the 70th week, counting backwards from the fall of 34 AD, we arrive at the fall of 457 BC as the beginning of the first year of a seven-year sabbatical cycle. As already brought out, Artaxerxes' 20th year began in the fall of 445 BC. We find, therefore, that the fall of 457 BC would mark the end of his 7th year. It was during his 7th year that he issued the decree now found in Ezra 7, which went into effect the following fall.
What we want to notice here is that Anderson's theory ignores these sabbatical cycles. By reducing the 483 years to 476, and by stretching what he called 69 weeks over really 68 weeks of years, Anderson totally disregarded the sabbatical cycles. To be more specific, March 14, 445 BC, was in the middle of the 5th year of a sabbatical cycle, and April 6, 32 AD, was in the middle of the 5th year of a sabbatical cycle. Between these two dates we have 68 sabbatical cycles, not 69.
To this may be added:
The issue of the first 69 weeks is an important one. Because of the problems cited above, much of dispensationalism's theology will need to be rewritten.