This is Part 2 of a critical examination of the MMEL hypothesis of the Star of Bethlehem. Go to the index here.
In Part 1 of this critical overview of the Star of Bethlehem film and its version of history (which I have called the MMEL hypothesis), I looked at the reasons scholars can say we know Herod died no later than 4 BCE given the information we have from Josephus as well as what we can connect with other accounts. The information from Josephus seemed to be overwhelmingly in favor of a 5/4 BCE date for Herod’s death, which would then contradict the time frame needed for the conjunctions of Jupiter and Venus as the MMEL hypothesis requires. However, there is another argument that is focused on, though not detailed, in the documentary, and it concerns the text that we have of Josephus.
When it comes to doing historical analysis, the first thing that needs to be established is the text you are trying to interpret and derive historical information from. This means you need to do what is called textual criticism, in which you try to best represent what the text actually was before various copyists and editors had their way with it. Sometimes it’s fixing mistakes, and sometimes it is removing things added in by later sources. Some errors in copying are simply misunderstandings of a marginal note being taken as something to include into the main body of the text. Other times a copyist gets tired and misses a few lines of the original. And because we have almost no original copies of any text from antiquity, we are relying on those copyists, mostly from the medieval period, to relay to us what the ancients had written.
This means that we need to make sure that the data we derive from the works we have are as close as we can to what was intended by the author. We cannot do so perfectly, and without the original it is impossible to really know. However, we can reconstruct the earliest versions possible of a document, knowing what it was like in, say, the 4th century, even if it was written in the 1st. With some literature we have to worry about the copyists and editors wanting to change the text for their own purposes; this has happened with Christian writing plenty, and we can detect at least some of those alterations (see, for example, Bart Ehrman’s The Orthodox Corruption of Scripture).
Now, there have been textual studies of the works of the historian Josephus, but Rick Larson claims that the newest analysis has suggested a different reading of an important part of Josephus’ history that will cause us to adopt the revised chronology.
What is this new evidence?
Before getting deep into this, remember that the conclusion that Josephus has us believe Herod died in 5/4 BCE was based on several passages. It was based on his length of rule from two different starting points, it was based on his age of death and when he was 25 years old, it was based on the reigns of his sons, and it had outside confirmation where possible. The point is, if there were textual corruptions, they had an amazing tendency to be consistent, and that is rather strange. Changes to a text are far more often accidental, such as misreading a number, and they ought to be random. But to have a collection of numbers, scattered among more than one writing by the same person, all consistently giving the same answer, is not expected on random errors. So already, this argument based on a new textual witness seems unlikely. But if there is are manuscripts that have all the numbers different and still consistent, and those numbers instead indicate that Herod died in 1 BCE rather than otherwise, we should pay attention. However, considering that Josephus was consistent with outside historians and numismatic (coin) evidence, we should still probably side with the version we currently have.
So, with that introductory thought, what are these textual witnesses saying? There is actually only one number change that is considered, and it deals with the length of the reign of one of the sons of Herod the Great, that of Philip. The argument comes from David Beyer who first presented this at a meeting of the Society of Biblical Literature (SBL) in 1995, but strangely it wasn’t printed in the volumes that the SBL publishes of the speeches. Instead, it was published in Vardaman’s edited volume, Chronos, Kairos, Christos II. What Beyer finds is that there are a number of manuscripts that say Philip reigned until the 22nd year of Tiberius rather than the 20th year. That would imply Philip became ruler not in 4 BCE but later in 2 BCE, thus Herod died in about that year. Beyer also has to depend on some coreignency as well so that Philip de facto was ruling in 1 CE but it gets confusing at this point.
Now, this is already an immensely weak premise. A textual variant in a single place is supposed to overturn all other things? Beyer does in fact argue that the number of years Philip reigned was also changed, but he provides no evidence that this was the case for Herod’s other sons, or the other pieces of data we have from Josephus about Herod and his reign. So even if in fact Josephus has different numbers for the length of Philip’s rule, that is hardly going to overturn the other, still-consistent evidence. Beyer and others are thus left with a number of ad-hoc suggestions to fix the problem. In fact, Beyer’s thesis doesn’t make things any more probable because his historical reconstruction requires all the ad-hoc additions as well as his own, so even if his textual analysis is right, his hypothesis is still even worse than what it had before he started.
But we can see that the methods used by Beyer (and accepted uncritically by Larson) are highly suspect. For one thing, his analysis shows that the number of years reigned is highly variant, anywhere between 37 (the standard version accepted by scholars) and 22 years. In fact, because of these textual variants, Beyer argues that Philip actually began to properly rule in 4 CE. So the history here is a mess; Herod dies in 1 BCE, but one of his sons isn’t a ruler for another 5 years, though his other sons are…
Alright, before we get much deeper in this this historical free-for-all, let’s look at how he does his textual criticism. When you look, you discover he is actually not using proper, scientific methodology in his work. What a scholar is supposed to do here is not simply find a large number of manuscripts, count them, and declare the majority the winner. You can have another manuscript be, on its own, a closer exemplar to an older tradition. For example, if there are two copies of a book, one is done well and another not so well, but the second, poorer one happens to have more copies made from it, then by Beyer’s method we would be accepting an inferior textual tradition. Instead, one has to talk about the relationship between texts and effectively create a family tree of texts. You then have to talk about what are the better exemplars of the text, and that takes a fair amount of work. Beyer hasn’t done that, so his analysis is already faulty.
Moreover, the collections of texts he went over are all from the British Museum and Library of Congress in Washington. However, these are neither the oldest nor the best manuscripts. Those are found in France and Italy. Yet these manuscripts do not make it into Beyer’s analysis at all. This is basically trying to make conclusions while not including all the evidence, let alone the best evidence. Add that to an unscientific methodology of textual criticism, and you have untrustworthy conclusions.
On the other hand, the published critical editions of Josephus’ works have done this, namely those of S.A. Naber, B. Niese, and H. Thackeray, and all their editions are consistent: Philip died in the 20th year of Tiberius (Niese’s version of the Greek text is used by the Perseus Project; for discussion of Beyer and the other textual scholars, see here). When you look into the manuscript tradition Beyer was working with, you discover that he was going by a single manuscript tradition, and one from a Latin text. Josephus originally wrote in Greek, so the Latin is a translation. The textual critics of Josephus’ work give Latin texts low priority in reconstructing the history books as they are notoriously inferior. And taking a single manuscript tradition from the inferior range and using it to usurp all of the superior Greek traditions is ludicrous. It is also apparent that the Latin version’s numbers are messed-up. As mentioned, the number of years Philip reigned have a 15-year spread, so we have every reason to look elsewhere for better representative of Josephus’ text.
As for the Greek manuscripts (details here), one of the oldest ones available is Codex Ambrosianae F 128 from the 11th century (a century older that Beyer’s oldest exemplar), and the best one is considered to be Codex Vaticanus Graecus 984 from 1354, and these manuscripts have 20th rather than 22nd year as Beyer wants (in fact, all Greek manuscripts worth looking at say 20th). These Greek manuscripts are also significantly older than the printed Greek copy of Josephus’ Antiquities from 1544 that Beyer is concerned with, so it is not a printing error run-on from there as Beyer seems to suggest.
To conclude this, it is apparent that Beyer’s analysis of the textual tradition of Josephus’ works are deeply flawed and goes against the best assessments and the best manuscript traditions. But even if Beyer were right, we still have the numerous, supporting avenues that point to Herod the Great dying in 5/4 BCE, while the 1 BCE date is filled with numerous problems and can only resolve them with many ad-hoc justifications, such as co-rulerships, antedating of reigns, and more blunders by ancient historians that all just happen to agree on the points that establish Herod died when most scholars think he did. You can only change the date of Herod’s death by committing the logical fallacy of possibly, therefore probably–it’s possible that all these things are true, therefore we are justified to believe our new hypothesis as true.
This means that the MMEL hypothesis is already dead in the water; however, a look into how they want to explain the Star is still in order, which takes place in Parts 3 and 4.