Abstract: I show here the misguided attempts to interpret 1 Sam. 20:5 as teaching a conjunction method of setting the month.
Then David said unto Jonathan, “Behold, tomorrow is the new moon, and sitting I should sit with the king to eat” (1 Sam. 20:5). This text is an often cited “proof text” for keeping the Scripture feast days according to the timing of the lunar conjunction. It is cited to give a reason to reject keeping the feast days according to the visual sighting of the new moon. The argument is based on several assumptions: 1. A day always begins at sunset, 2. The word tomorrow always refers to a following day, and 3. The first appearance of the visible new moon could not be predicted in advance. 4. The conjunction day could be predicted in advance. From the first two assumptions it is concluded that the moon was not seen. From the third assumption it is concluded that it could not be calculated or predicted in any way. The last assumption is used to win the conjunction argument by default. It is therefore concluded that the conjunction was predicted and that the conjunction was meant by David.
Let us analyze the logic. Supposing that an actual visual sighting is disproved by the text, then does it follow that a conjunction method is proved, and that a visual method is not being assumed in the announcement? Not at all. This is a classic non sequitor. Disproving an actual sighting does not prove a conjunction was meant. It may be that after 29 days from the previous new moon was called the new moon day regardless of any sighting, and any discrepancy in calling it so rectified only after the new moon was seen again. That is, the next day could simply be called the new moon until proved otherwise. If the moon wasn’t seen, then it would not be the new moon any more, and the next day would become the new moon day. It is clear from the passage that there were two days of the new moon feast. One of them clearly would prove not to be the actual new moon day. The reason for the two days was that visual sighting was uncertain between two days. So it is clear that one of those two days is called the new moon before it can be clear whether it is the new moon or not. David could simply be calling the day after the 29th of the month a new moon day because it was always the custom to call it a new moon day, and a new moon feast was always held on day 29 + 1 no matter what.
So the logic is incorrect. Let us now look at the assumptions. Does a day always begin at sunset? No. Does a day even normally begin at sunset? These assumptions are disproved in When Does the Day Begin? So we must dismiss it. Invariably the word “tomorrow” in Hebrew means the daybreak. This is proved in 1 Samuel 19:11. Michal says to David that if he does not save his life “tonight” then “tomorrow” he will be killed. If then it was already night, a new moon could have been seen, and the morning would be called “tomorrow.” So also in 1 Samuel 20:5. It could already be night, and the new moon could already have been seen, and David could have meant “Tomorrow is the New Moon Feast.” This suggests that the word “new moon” (חֹדֶשׁ) besides referring to two new moon days, the new moon, and the month, also refers to a new moon feast. The New International Version certainly thinks this is what the word means. It has “New Moon festival” in 1 Sam. 20:5. Likewise the NLT, GWT, CEV, AMPC, EXB, GNT, ICB, NOG, NCV. Brown-Driver-Briggs defines the Hebrew word as “1. new moon = day, time, of new moon, as religous festival 1 Sam. 20:5 .... 2. month.” Holladay defines, “1. new moon, the day on which the crescent moon is again visible, 2. month.” It is evident here that new moon does not just mean the moment the new moon occurs. It also refers to the day of the new moon and to the feast on the day even when the new moon does not occur on it, e.g. 1 Sam. 20:27, “Then it was in the tomorrow of the new moon, the second one.” Amos 8:5 uses the word הַחֹדֶשׁ to mean the duration of a new moon feast.
Let us now look at the third assumption, which is the claim that when the new moon would be visible could not be predicted in advance. This is easily disproved. The old month may be established by a first visible sighting on a clear day under unambiguous conditions. This happens frequently enough. Then this month grows to 30 days without a new moon been seen at the end of the 29th day. The 30th day is a clear day, and the new moon has not yet been seen. It is easy to predict beforehand that the new moon will become visible just after sunset based on the fact that a month can be no longer than 30 days.
Let us now look at assumption number four that the conjunction day could be predicted in advance. I listened to one conjunction advocate argue that he could do this by watching the old moon wane. His rule was that when the old crescent was no longer visible in the dawn then the next sunset would begin the first day of his month. The problem with this method is that it is not foolproof. The rule frequently results in the first day of the month not being the conjunction day, as this example from 2002 shows. The old moon would be last sighted at dawn on April 10, 2002. The old moon would not appear on April 11 at dawn, LD0. This conjunction advocate makes April 12 (LD1) the first day of his month. However, the conjunction did not happen until after sunset April 12, which according to this person would be the next day since sunset always marked the start and end of his days. Following this rule often results in the month often starting a day before the conjunction!
Since conjunction prediction cannot be done by any visual means, in order to predict a conjunction in advance, it is necessary to have a valid conjunction date for the start of the old month. In this case the lack of a sighting will not confirm that the month is 30 days long allowing one to predict the next conjunction on the 31st day.
Since conjunction prediction cannot be done by counting from a day on which the conjunction is known to occur, then it is at least necessary to know what time of day the conjunction occurred on the first day of the old month. But then a calculation is required. One adds 29.53059 days to the time of the old conjunction to arrive at the mean time of the new conjunction. But there is a problem with a mean value. A mean value is only an average time between conjunctions and not an actual time in any give case. The actual time between conjunctions varies from 29.273 to 29.820 days! Therefore, if one depends on the mean value to predict the time of the new moon, then one will frequently miss the actual day of the conjunction. The variance from the mean lunation -6.17 hours and +6.93 hours (NASA figures).
The Rabbinic Calendar is based on a mean lunation to calculate its new moons. Therefore a conjunction is not an actual conjunction. The the calculated Rabbinic Calendar could not predict the actual conjunction, then what makes people think the actual conjunction could be predicted by any simple calculation before its inception?
The conjunction day cannot be consistently determined by any set of visual observations. The day the old moon vanishes in the east is no reliable guide to which of two days would be the conjunction. The assumptions applied by conjunctionists to 1 Sam 20:5 are faulty as is their logic in using the assumptions.
Pagan astrologers were superstitious. If an event came into proximity with an astronomical event then they turned it into an omen. This type of reasoning has entered in among people who call themselves either Christian, Messianic, or Torah observant. In this case, I give an example of such reasoning. It involves symbols and comparisons and the logical fallacy called predication, i.e. making things equal that are not equal. It takes the form of because X is similar to Y then X = Y means Z. If a man dies on an eclipse then death at X = Y time of eclipse means Z he was a bad man. Anyone can see how stupid, unscientific, and superstitious this reasoning is. But people calling themselves children of the Almighty are doing it.
This reasoning is used in defense of the conjunction. I have seen it often enough that an example deserves to be made of it. Perhaps some people will learn how to think instead of being lead astray and blown all over by every wind of doctrine. The argument goes that since there was darkness before the Almighty created light (darkness at time X = Y before light) that this means a new moon should be defined in the darkness by a conjunction (i.e. X = Y means Z). Yep, that is the sort of reasoning they go by. Such symbolical and comparison games are only suitable for illustrating an objective truth already known. They are worthless for establishing objectivity in the first place.
This sort of argument is even followed to the point of noting that the moon will not give its light when Messiah returns, i.e. X = Y (moon does not give light at time of return) means Z that the month begins at the conjunction. Visual observers of the moon have approached this situation in the past by simply alternating 29 and 30 day months until the moon appears again. In the modern age we can simply calculate when it should be visible. But not being able to see a new moon by no means means that a month begins at a conjunction.
The claim is made that using the conjunction ensures that the full moon will be on the annual Sabbath (Nisan or Aviv 15). This statement is made based on astronomical ignorance and observational selectivity. In the 2002 case given above the full moon occurred on April 27th at 6 hours 25 minutes JMT. I used Stellarium to do these calculations. But this is clearly lunar day 16 counting the way one advocate of a conjunction method counts. On the first day of non-visibility at dawn he waits to sunset on that day to start his month. The result for 2002 is that the full moon will be on LD16. So all of the annual Sabbath will be before the full moon. Why is this? From the time of conjunction to the full moon may be anywhere from 13.9 days to 15.7 days. This is because the moon travels on an elliptical orbit. If the perigee of the moon is at the last quarter, then the moon will travel the slow part of the ellipse between the conjunction and the full moon taking up to 15.7 days. But if the apogee is at the last quarter, then between conjunction and full moon will be the fast part of the ellipse. The moon will take the minimum time to reach full moon. It is therefore a myth that ANY method of fixing a new moon at conjunction or first visibility can guarantee that the full moon falls on a particular day, or even two particular days. If someone thinks that Scripture requires a feast day to fall on the exact 100% astronomical full moon, then I have news for them. Reality has broken and will continue to break their expectation no matter how they calculate the calendar. They will have to end up by fixing months using the full moon instead of the new moon. If anyone thinks any passage requires this, then they had better take the text to mean almost full, or full as most people would lazily say of the moon.
There is more to this story -- more than can be searched out in the foundations of earth below or in heaven above.