It is often truly remarked that our classic Christmas story—the one we celebrate in pageants every December and teach to our children—blends elements from two quite distinct narratives, one found in the first two chapters of Matthew and the other in the first two chapters of Luke. In both accounts, Jesus is born in Bethlehem to Joseph and Mary during the reign of Herod the Great, and in both he ends up growing up in Nazareth. But beyond that, they have little in common. Matthew has an angelic visit to Joseph, the wise men, a warning in a dream, the flight into Egypt, the murder of innocents in Bethlehem, the death of Herod the Great, a journey partway to Judea, worry over a new ruler, and a detour northward to settle in Galilee. None of these things are mentioned in Luke. Luke, on the other hand, has the story of Elizabeth’s pregnancy, the annunciation to Mary, a census, shepherds and angels, and the reactions of Simeon and Anna. None of these things are mentioned in Matthew. For two accounts of the same event, they are about as different as one could imagine.
Could they really both be true?
To start answering this question, we need to look a bit more closely at each narrative. Start with what we find in Matthew. A striking feature of those first two chapters is how much of them is told from the perspective of Joseph. Before an angel ever appears on the scene, he is painfully perplexed by the discovery that Mary is pregnant, and we are told what plan he formed—a plan on which he did not act, and which could therefore be known only if Joseph chose to tell it—for dealing with the problem privately (1:18-19). The angelic message is to Joseph, apparently when he is alone (1:20-23). The birth of the child is told from the standpoint of Joseph’s obedience to the message and his naming of the child (1:24-25). The warning to flee from Herod’s murderous wrath comes to him (2:13), and he immediately obeys (2:14). The word that Herod is dead comes to Joseph in a dream (2:20). Throughout the drama, Mary’s point of view is never seen; Joseph takes her to be his wife, avoids having relations with her until Jesus is born, takes her and Jesus to Egypt, and takes them back again. Anyone who can read should be able to see at least this much: if this is not fiction, then it is an account that goes back to Joseph himself.
Why should anyone think that it is fiction? There is, of course, the supernatural element, a virgin conceiving and angels appearing in dreams to deliver divine messages. That is an issue that deserves its own discussion. (Interested readers who cannot wait might want to start exploring the issue of miracles here.) Setting that aside for a moment, there is really just one major objection: the slaughter of the innocents in Bethlehem, a grisly, public event, is not attested in any other first century source we now have.
But should it be? Start with the other Gospels. The story of the massacre in Bethlehem forms a link in Matthew’s story, explaining why the young family had to flee into Egypt. But since no other Gospel tells of the flight into Egypt, there is no particular reason for them to repeat the event. What of the other books of the New Testament? Here there is, if possible, even less reason; Acts deals entirely with later events, and the various epistles are designed to clarify doctrine, rebuke wrong behavior, and offer practical advice, not to be biographies of Jesus.
How about non-Christian sources? Roman authors wrote late and far away and had little interest in Christianity; often their understanding even of the Jews shows that they had few and poor sources of information. Philo of Alexandria lived hundreds of miles from Jerusalem, which he visited, so far as we can now tell, only once, and his latter years were consumed with the political fate of his people at the hands of the Romans.
That leaves Josephus, the first century Jewish writer who in his Antiquities of the Jews gives us in some detail the doings of Herod the Great. But Josephus’s accounts are for the most part centered on matters of civic or religious importance. Herod the Great had his rival Antigonus killed, his favorite wife executed, her brother murdered, her mother and grandfather executed, his own grown sons Alexander and Aristobulus strangled—those are matters of civic importance, affecting the power struggles within the Herodian family and the line of the succession after his death. He had the chief men of Palestine locked up in the Hippodrome with orders to slay them all at the hour of his death in order that there might be a great mourning on this occasion (Antiquities 17.6.5). (Apparently he was under no illusions as to whether anyone would spontaneously mourn his own passing.) No doubt we would have heard more about this malevolent plan if his relatives had not relented once the old villain was well and truly dead and let all of those citizens go home unharmed.
If prior plausibility means anything in history, the story of the slaughter in Bethlehem rings true. A man who could murder so many members even of his own family because of threats real and imagined to his throne is, undoubtedly, a man who could order a dozen male infants killed in a hamlet seven or eight miles south of his seat of power. But there is no particular reason that the event should have caught Josephus’s attention or merited a place in his history. There were larger matters afoot.
The argument from silence, therefore, has negligible weight here. But to say that is only to undermine the negative case. Is there any positive evidence for the events of Matthew 1-2, beyond the mere fact that the Christians believed they really happened?
As it turns out, there is, in Matthew 2:22:
But when he heard that Archelaus was reigning over Judea in place of his father Herod, he was afraid to go there, and being warned in a dream he withdrew to the district of Galilee.
Three things stand out in this verse. First, the news that Archelaus was in charge was troubling to Joseph. Second, there is a peculiar term used here, a participle: Archelaus is not said to be a king but rather to be kinging, ruling as a king. Third, the indication that Galilee would afford sanctuary is striking.
Take these in reverse order. Herod the Great had been ruler of both Judea and Galilee. Why, then, should shifting northward have removed the holy family from danger? As Josephus explains, when Herod the Great died, his kingdom was divided up among his surviving sons. Archelaus received Judea for his portion, but Galilee went to the younger Herod Antipas. So going to Galilee would indeed have taken them beyond Archelaus’s reach.
Then there is that participle, a delicate way of not quite saying that Archelaus was a king while saying quite plainly that he was behaving like one. This, too, is true to what we know from Josephus. Archelaus took power immediately upon his father’s death, but he never did receive from Rome the title of King; he was allowed, provisionally, to retain power with a view to seeing whether he deserved the title. He didn’t, and a dozen years later he was deposed for gross mismanagement. Thus began the period of procuratorial governance of Judea, culminating in the life of Jesus with the procuratorship of Pontius Pilate.
But most importantly of all, why was the reign of Archelaus so worrisome to Joseph? He was the oldest surviving son of Herod the Great; Judea was the central seat of his power. Knowing that Herod was dead, Joseph, if he knew anything about Herod’s sons at all, could very plausibly have inferred that Archelaus would inherit a sizeable chunk of his father’s kingdom. Why, then, should hearing that Archelaus was, in fact, in charge in Judea cause him to change his plans?
The question takes us back to Josephus once more, and in Antiquities 17.9.3 (Loeb #213-18) we discover a startling answer. One of Archelaus’s first problems when he stepped into his father’s position in Judea was a public uproar over the fact that the now-deceased Herod had, in his final weeks, executed some Jews for “sedition.” Their “crime” was that they had instigated the cutting down of some Roman shields posted over the gate of the Temple, and the Jewish perception was that this was merely their asserting their right to worship according to their own traditions, which forbade graven images (Exodus 20:4).
The execution had been wildly unpopular, and since Archelaus took power just at Passover time when hundreds of thousands of pilgrims flocked to Jerusalem, the story was spreading through the crowds like a wildfire through dry grass. A large group of Jews got into an argument with a much smaller regiment of Roman soldiers; matters escalated, and the Jews stoned the soldiers, most of whom died. The perpetrators then ran into the Temple with their sacrifices, thinking perhaps that they would simply disappear into the throng.
Archelaus panicked, feeling his newfound throne sway beneath him. He sent a troop of horsemen to surround the temple, with strict orders not to let anyone outside go in and not to let anyone inside leave. Then he sent troops directly into the temple. According to Josephus, they slaughtered some three thousand Jews. Passover was canceled. The visiting Jews were sent home.
None of this background information is related in Matthew, but it sheds a flood of light on Joseph’s decision. He had left Judea to escape a homicidal maniac on the throne; now, returning, he was met by a flood of erstwhile Passover pilgrims fleeing the temple massacre. Small wonder he was troubled. Small wonder that he revised his plans, taking Mary and the young Jesus into Galilee where they would be under the authority of a tetrarch who did not yet have blood on his hands.
The combination of indirectness and precision in Matthew’s narrative is evidence that it comes from a source who knew what had really happened. His passing but accurate reference to Archelaus’s status and the unspoken but correct implication that Archelaus’s dominion did not stretch northward into Galilee dovetail exactly with what Josephus tells us. A comparison with the apocryphal literature of the second century suggests that such minute precision in matters incidental to the narrative is not characteristic of the sorts of legends that later generations of Christians invented about the young Jesus. This has the ring of truth.
There is one other point of interest about the reference to Archelaus. He was deposed by the Romans and banished to Gaul in the year AD 6, about ten years after he took the throne and in the twelfth year after Jesus’ birth. Curiously enough, Luke’s narrative, which never mentions Archelaus, tells us that Jesus’ parents went up to Jerusalem for the feast every year, but the first time it mentions his going with them is when he is twelve years old. The chronological coincidence is not a guarantee, but it is at least plausible that they took him with them only when Archelaus was no longer in power and the shadow of their harrowing journey to Egypt had passed.
This is the fifth and final response to your comment from the 25th. This deals with the use of Bayes Thereom for historical arguments.
I should perhaps recast my dislike for Bayesian arguments. My problem is not with the viability of a Bayesian analysis but rather with the futility of reconciling a Bayesian argument between parties. The contentions regarding the appropriate set of inputs and their probabilities seems interminable. If you can get the parties to agree on a set of inputs and their probabilities, then you’ve got something. The prospect of doing this, however, is doubtful. The situation gets even worse for historical topics with theological implications. You eventually have to start accounting for the supernatural and factoring in theological interpretations and motives.
As an example of how this process starts to grow, note that your framing assigned my argument with three conjunctive assumptions and yours with none. Aside from the fact that I don’t even agree with the assumptions, I don’t stand a chance. If we were to be fair, you would have to include the assumptions which roughly mirror the ones you assigned me. Perhaps this might be something like (1) the probability that Joseph would have (could have?) recorded an account, (2) the probability that Joseph encountered Jews returning to Egypt and they told him about Archelaus, and (3) the probability that Joseph recounted the events with 100% accuracy and they were accurately transmitted to Matthew.
I’m not suggesting that we go down this road of actually trying to construct the Bayesian argument. I’m not interested in wading through the unmanageable behemoth that it will inevitably become and then engaging in countless debates over the proper probabilities for the inputs. I would rather just lay out the reasons for my views and hope they make sense. My goal isn’t to prove anybody wrong or even prove that I’m right. I’m primarily interested in showing that there are other factors and other views which should be included in the consideration. The alternative view isn’t just that some guy grabbed a few obscure verses from the Old Testament and used them to build a story. The theory is far more nuanced than that.
This is the fourth response to your comment from the 25th. This deals with the scope of redaction.
You asked for a clarification of the role of the redactor. To answer this, I’ll summarize the sequence that I think makes the most sense out of the data.
It starts with the quote from Papias, which seem to infer that the first version was something like a “sayings gospel”. I wouldn’t be surprised if the hypothetical Q and the original Matthew are one in the same. I also wouldn’t be surprised if this has its origins with Matthew himself. At some point this is merged with Mark to yield a “proto-Matthew”. The evidence for this comes from the synoptic problem. This gospel becomes prominent, particularly among the Jewish Christians. Patristic sources tell us that the Nazoreans and Ebionites used it exclusively. They also tell us that the Ebionites’ version doesn’t include the first two chapters we have now. The Nazoreans, however, were essentially orthodox, so we shouldn’t be surprised if their version eventually wins out. That’s where I think the redactor for the nativity comes in. The redactor is a Nazorean (see Matt 2:23) and is building upon a text that already has the content from Mark and Matthew’s sayings source. In addition to the nativity, the redaction includes allusions to Moses and multiple claims of prophecy fulfillment, just like we see in the nativity.
I arrived as this independently but, if you’re interested in a more rigorous treatment, my understanding is that this is not too distant from the view offered by Dale Allison. I will admit that further study is needed on my part and the ideas are certainly subject to change. That said, I think this does a reasonable job of connecting the dots, including explaining how the traditional assignment of authorship came about.
This is the third response to your comment from the 25th. This deals with the redactor’s use of Archelaus.
I will readily admit that I have no evidence that literate Jews around 70 or 80 AD would have known that Herod and Archelaus had committed multiple atrocities and that Archelaus was not in control of Galilee. Regardless, I think that the most important thing to note is that this isn’t just any first century Jew, this is a literate Jew who had formal training and exposure to information and traditions that most would not have had. It just doesn’t seem at all unreasonable to me. I don’t see how either of us have enough evidence to argue one way or the other.
I also think that you drastically overstate the assumptions regarding the redactor’s use of Archelaus. There are only three basic facts necessary to pen the words we have:
1) Archelaus took over after Herod (and this was near the time of Jesus’ birth)
2) Archelaus had a bad reputation
3) Archelaus didn’t rule over Galilee
I see no reason to suspect that a redactor would have intended anything more than to employ his knowledge of these facts in producing the narrative. For the storyteller, this result is superior to just saying “then they moved to Nazareth” because it offers opportunity for the parallels to Moses, links to OT texts and a dramatic explanation for the Bethlehem/Nazareth problem.
Regarding “kinging”, I question whether you have made too much of this. Isn’t this just the verb one would use? Please explain why this phrasing implies a certain class of ruler. Furthermore, why should we think that Joseph would have been privy to a subtle distinction in Archelaus’ political status?
This is the second response to your comment from the 25th. Here I will address the slaughter of the innocents, though I think we’re really not very distant from each other on this point, so it may be a bit trifle to continue this part of the thread.
Regarding the scale of evidence we should expect, this comes down entirely to Josephus. What bothers me is that you think we can faithfully rely on Josephus to detail the color of the paint on Herod’s wall and recount numerous atrocities committed against the Jewish people by Herod and his family, yet, when it comes to the slaughter of the innocents, the omission is of negligible significance. This just seems inconsistent. While we certainly can’t say that Josephus definitely should have included it in his writings, I disagree that its absence counts for nothing.
Tim,
It’s clear that the whole discussion pretty much boils down to the question of authorship, which is almost certainly unresolvable here. So, for now, I would just like to step back and restate my contention, which is that the evidence presented to argue for the Matthean nativity as a 100% accurate eyewitness account is also consistent with a redactive history.
In an attempt to simplify things, I will split my responses into separate comments so that different topics can operate on separate threads. For this thread, I’ll discuss the style of Matthew’s infancy narrative.
I am in the camp that sees parallels between the literary style of the infancy narrative and literary style of midrash. As I’m sure you’re aware, midrash tells a story as a means of explanation. I am not suggesting that the infancy narrative is a midrash to explain the prophecies which are quoted; rather, it is employing the storytelling style of midrash to fill in the gap regarding Jesus’ birth and to express the theological agenda, as we have already discussed.
A recent post at vridar.org offers a review of the way this use of creative license permeated Jewish literature at that time. As I said, I don’t see that the author of Matthew’s nativity was trying to make it look like the story came from Joseph – he was telling a story about a child and Joseph was the head of that child’s household. The resulting story would naturally contain many events which would be known to Joseph. Not only that, but if you take away the dreams then Joseph’s role is drastically diminished. All this means, from a redactive perspective, is that the author didn’t split the dreams between Mary and Joseph.
My point regarding Josephus was that he often employs a similar literary style, not that he attempted to make it appear as if the story came from a particular individual. If you don’t think that Josephus ever filled in the gaps, embellished details or creatively attributed words, thoughts, motivations and emotions to the characters in his writings, then you have a substantially higher expectation for Josephus’ sources than I do. If you agree, however, that Josephus did employ creative license, or at the very least relayed other persons’ creations as if they were fact, then I don’t understand why we can’t acknowledge that the gospels may be subject to the same practice. Even if you do count Josephus to be completely trustworthy, I think that the prevalence of this literary style in Jewish culture should weigh on our assessment of the gospel texts.
Travis,
Re: “It’s clear that the whole discussion pretty much boils down to the
question of authorship, which is almost certainly unresolvable here.”
Tim resolved the matter with his presentation on Gospel authorship, to which he gave you a link. Did you bother to watch it? If so, what do you think of the data he provided?
I haven’t watched it yet but I will; maybe I’ll have time tonight. I do think that it’s rather presumptuous to assert that the presentation has managed to completely undermine the reasons given by multitudes of scholars over the last 100+ years and only the traditional, orthodox view is left standing. But hey, I haven’t watched it yet, so maybe my mind will be blown and I’ll see something more persuasive than the arguments that I’ve encountered so far from other apologists. That’s why I’m here, to learn and explore all the different arguments for and against the various views that are out there.
Tim,
I’ve been waiting for the opportunity to continue our dialogue from part 2. Let me start by acknowledging a few things:
1) I agree that if the story is true then it would almost have to trace back to Joseph. However, you don’t seem to rely on this for any arguments, so there’s nothing more to say.
2) I agree that absence of evidence is not evidence of absence regarding the story of the slaughter of the innocents. I think that it may inform the motives of the author, but we already discussed that.
3) I agree that the history surrounding Herod and Archelaus fit with the narrative. However, this would be true for both a recounting of true events and for a synthesis of a theological agenda with a known history. Just as Josephus knew of these events, we should expect that the author of the Matthew narrative also knew these things in the preceeding decades. More on this later.
Overall, I don’t see much in this post that isn’t consistent with either view. I was hoping for more of a defensive against the conflicts between Matthew and Luke. Is that a forth-coming post? On the assumption that it’s not, I offer the following questions. I’m interested to see how you fit these into the assertion that both Matthew and Luke offer 100% truthful accounts.
1) Last time you mentioned that Luke 2:39 added support for your view, however, when I review that passage, it indicates to me that Mary and Joseph returned to Nazareth from Bethlehem about 1 month after his birth. This obviously doesn’t mesh with the story in Matthew.
2) Matt 2:19-23 indicates that Joseph intended to go back to Judea and the preceding narrative seems to assume that they were always in Bethlehem in the first place. The most lucid reading of Matthew is that they lived in Bethlehem and only ended up in Nazareth to avoid Archelaus upon their return from Egypt. However, Luke undeniably has them living in Nazareth prior to the birth, and then apparently returning shortly thereafter (#1). The Matthean assumption that Mary and Joseph always lived in Bethlehem until the escape to Egypt is further supported by several additional points:
a) The Magi come to a house in Bethlehem, presumably Mary and Joseph’s house.
b) The slaughter of the innocents was said to be for children under the age of two, according to the timing given by the Magi. It seems reasonable to think then that it had been at least a year since the birth and the family was still in Bethlehem.
c) It took a second dream to redirect Joseph to Galilee. Then the text says that he came to Nazareth and settled there. The structure implies that it is for them a new city and a new home.
d) In the second half of Matt 2:23 the Nazarene prophecy is claimed to have been fulfilled by the redirection to Nazareth, not on the assumption that it was their home town to begin with.
e) Matt 1:18-2:21 offers absolutely zero reason to ever think that Nazareth was ever in the picture.
Lastly, a comment regarding how the harmonization of Matthew’s account with Josephus’ history can be seen as the work of a redactor. A key element in both the Matthean and Lukan nativities is the need to explain how Jesus was known to be from Nazareth despite the prophecy that the messiah would come from Bethlehem. The Matthean author, knowing Archelaus’ reputation among the Jews and the division of power that put Nazareth out of his rule, could have seized upon this to resolve the conundrum. When this is coupled with the agenda of portraying Jesus as messiah, a second Moses and the son of God from birth (see my comments on Part 2) the result is quite conceivably the narrative that we have today.
This last part is admittedly speculative but I contend that it makes the most sense when combined with the other reasons to question the veracity of the narrative, such as: the conflict with Luke, the use of astrology, the overt theological agenda found in Matthew’s differences from Mark, reporting of private information (e.g., Herod and the Magi’s actions, conversations and thoughts) and, as you noted, miracles (including several revelatory dreams). The conflict with Luke is perhaps the most disruptive of these, so I’m particularly interested to see how you resolve that. Looking forward to a response, or the next post, if that’s your plan.
Travis,
Thanks again for the civil, thoughtful engagement.
On your points, using your numbering:
1) Your agreement on this point narrows the explanatory options considerably. Either it’s a genuine history coming from Joseph, or someone else was deliberately trying to make it look as though it were. This consideration will come back in below.
2) I’m glad you agree, up to that point, regarding the weakness of the argument from silence.
3) You write: “I agree that the history surrounding Herod and Archelaus fit with the narrative. However, this would be true for both a recounting of true events and for a synthesis of a theological agenda with a known history.” But how likely is such a synthesis? Not all that likely, not during the lifetimes of people who had been adults when the events occurred.
“Just as Josephus knew of these events, we should expect that the author of the Matthew narrative also knew these things in the preceeding decades.”
I think you’re moving much too quickly here. Christian authors from later eras (authors of the Acts of Paul and Thecla, the Gospel of Thomas, or other apocryphal works from the second century) seem to have very little grip on the history of the first. Josephus was writing with the full force of imperial patronage and had access to formal histories. We find no trace of these in the NT authors.
“I was hoping for more of a defensive against the conflicts between Matthew and Luke. Is that a forth-coming post?”
Yes, a post or two. Luke’s narrative is up next; depending on how long that is, I may have to defer consideration of the consistency of the two accounts until one subsequent to that. Regarding your particular points, so as not to leave you hanging for too long:
1) Luke 2:39 – actually, it’s less specific than that, but it does, I think deliberately, skip over the flight and return. If the document comes from Mary (as I will argue it does), then there is a straightforward explanation for the omission.
2) I disagree that Matt 2:19-23 “seems to assume that they were always in Bethlehem in the first place.” There are plenty of plausible reasons for their returning to the spot they left last, particularly since it is unlikely they knew about the slaughter, having left town there before it occurred and in response to warnings in a dream. Remember that to come from Egypt, they would have to swing north, and Judea is several days’ travel south of Galilee. So Judea (where, according to Luke, Mary had family) would be a very natural stopping point.
Regarding your subsidiary points:
a) The magi come some time after the birth, unlike our pop-art pictures that show them shoulder-to-shoulder with the shepherds.
b) “The age of two” for a Jewish audience would mean anything past the first birthday. “Under the age of two” would be approximately what we would call “not yet 1.” So you’re about a year off here.
c) I disagree with your claim that “[t]he structure implies that [Nazareth] is for them a new city and a new home.” That seems to me to be an overreading in search of a contradiction with Luke, not something driven by the text itself.
d) “In the second half of Matt 2:23 the Nazarene prophecy is claimed to have been fulfilled by the redirection to Nazareth, not on the assumption that it was their home town to begin with.” I don’t see what the assumption or lack thereof has to do with the argument. The text doesn’t say anything either way about whether it was their original home town.:
e) “Matt 1:18-2:21 offers absolutely zero reason to ever think that Nazareth was ever in the picture.” This is another argument from silence, which means that absent some pretty stringent conditions that are missing here, it has pretty close to absolutely zero force. Sorry.
As far as trying to see “how the harmonization of Matthew’s account with Josephus’ history can be seen as the work of a redactor,” I think this is the sort of exercise that gives literary criticism a bad name in the discipline of history. You’re making assumptions that require argument, among them, (1) that it is plausible that there was a redactor at all, or (2) that supposing there was a redactor, it is plausible that he would know so specifically about Archelaus (including all three details I mentioned), and (3) that if he existed, and if he did, he would drop the reference into the story in such an indirect way while leaving so many internal clues to Joseph as the source that he must have been deliberately deceiving his readers to do so.
You need the probability of all three of these assumptions to be quite high in order for your hypothesis to stand on something like an equal footing with Matthean authorship. I think all three are quite low, and their product is spectacularly low. But let’s be really generous and suppose that they each have a probability of 0.5. Then their conjunction has a probability of 0.0625. Suppose that the probability of our having the narrative we do (N), given that it does not come from Joseph (~J) and that these three assumptions are all true, is 0.8: P(N|~J & (A1 & A2 & A3)) = 1. Suppose also that P(N|J) = 0.8. But suppose (and again, I think this is far too generous) that P(N|~J & ~(A1 & A2 & A3)) = 0.1. Then
P(N|~J) = P((A1 & A2 & A3)|~J) P(N|(A1 & A2 & A3) & ~J) + P(~(A1 & A2 & A3)|~J) P(N|~(A1 & A2 & A3) & ~J) = (0.0625 * 0.8) + (0.9375 * .01),
which comes out to just a tad less than 0.06. That means that the likelihood ratio P(N|J)/P(N|~J) is more than 13 to 1, which is a pretty substantial evidential factor in favor of J. And I think that number is arrived at with a huge overestimate of the strength of your case for ~J. The actual force of the argument in favor of J is, I believe, a great deal stronger.
In non-mathematical terms: you acknowledge that your argument is speculative, but I don’t think you realize just how speculative it is or how much the multiple layers of sheer conjecture weaken your case.
Best,
Tim
Travis,
Thanks again for the civil, thoughtful engagement.
On your points, using your numbering:
1) Your agreement on this point narrows the explanatory options considerably. Either it’s a genuine history coming from Joseph, or someone else was deliberately trying to make it look as though it were. This consideration will come back in below.
2) I’m glad you agree, up to that point, regarding the weakness of the argument from silence.
3) You write: “I agree that the history surrounding Herod and Archelaus fit with the narrative. However, this would be true for both a recounting of true events and for a synthesis of a theological agenda with a known history.” But how likely is such a synthesis? Not all that likely, not during the lifetimes of people who had been adults when the events occurred.
“Just as Josephus knew of these events, we should expect that the author of the Matthew narrative also knew these things in the preceeding decades.”
I think you’re moving much too quickly here. Christian authors from later eras (authors of the Acts of Paul and Thecla, the Gospel of Thomas, or other apocryphal works from the second century) seem to have very little grip on the history of the first. Josephus was writing with the full force of imperial patronage and had access to formal histories. We find no trace of these in the NT authors.
“I was hoping for more of a defensive against the conflicts between Matthew and Luke. Is that a forth-coming post?”
Yes, a post or two. Luke’s narrative is up next; depending on how long that is, I may have to defer consideration of the consistency of the two accounts until one subsequent to that. Regarding your particular points, so as not to leave you hanging for too long:
1) Luke 2:39 – actually, it’s less specific than that, but it does, I think deliberately, skip over the flight and return. If the document comes from Mary (as I will argue it does), then there is a straightforward explanation for the omission.
2) I disagree that Matt 2:19-23 “seems to assume that they were always in Bethlehem in the first place.” There are plenty of plausible reasons for their returning to the spot they left last, particularly since it is unlikely they knew about the slaughter, having left town there before it occurred and in response to warnings in a dream. Remember that to come from Egypt, they would have to swing north, and Judea is several days’ travel south of Galilee. So Judea (where, according to Luke, Mary had family) would be a very natural stopping point.
Regarding your subsidiary points:
a) The magi come some time after the birth, unlike our pop-art pictures that show them shoulder-to-shoulder with the shepherds.
b) “The age of two” for a Jewish audience would mean anything past the first birthday. “Under the age of two” would be approximately what we would call “not yet 1.” So you’re about a year off here.
c) I disagree with your claim that “[t]he structure implies that [Nazareth] is for them a new city and a new home.” That seems to me to be an overreading in search of a contradiction with Luke, not something driven by the text itself.
d) “In the second half of Matt 2:23 the Nazarene prophecy is claimed to have been fulfilled by the redirection to Nazareth, not on the assumption that it was their home town to begin with.” I don’t see what the assumption or lack thereof has to do with the argument. The text doesn’t say anything either way about whether it was their original home town.
e) “Matt 1:18-2:21 offers absolutely zero reason to ever think that Nazareth
was ever in the picture.” This is another argument from silence, which means that, absent some pretty stringent conditions that are missing here, it has pretty close to, well, absolutely zero force. Sorry.
As far as trying to see “how the harmonization of Matthew’s account with Josephus’ history can be seen as the work of a redactor,” I think this is the sort of exercise that gives literary criticism a bad name in the discipline of history. You’re making assumptions that require argument, among them, (1) that it is plausible that there was a redactor at all, or (2) that supposing there was a redactor, it is plausible that he would know so specifically about Archelaus (including all three details I mentioned), and (3) that if he existed, and if he did, he would drop the reference into the story in such an indirect way while leaving so many internal clues to Joseph as the source that he must have been deliberately deceiving his readers to do so.
You need the probability of all three of these assumptions to be quite high in order for your hypothesis to stand on something like an equal footing with Matthean authorship. I think all three are quite low, and their product is spectacularly low. But let’s be really generous and suppose that they each have a probability of 0.5. Then their conjunction has a probability of 0.125. Suppose that the probability of our having the narrative we do (N), given that it does not come from Joseph (~J) and that these three assumptions are all true, is 0.8: P(N|~J & (A1 & A2 & A3)) = 1. Suppose also that P(N|J) = 0.8. But suppose (and again, I think this is far too generous) that P(N|~J & ~(A1 & A2 & A3)) = 0.1. Then
P(N|~J) = P((A1 & A2 & A3)|~J) P(N|(A1 & A2 & A3) & ~J) + P(~(A1 & A2 & A3)|~J) P(N|~(A1 & A2 & A3) & ~J) = (0.125 * 0.8) + (0.9375 * .01),
which comes out to just a tad less than 0.11. That means that the likelihood ratio P(N|J)/P(N|~J) is more than 7 to 1, which is a pretty substantial evidential factor in favor of J, enough to move someone who started with P(J) = 0.5 to P(J|N) > 0.875. And I think that number is arrived at with a huge overestimate of the strength of your case for ~J. The actual force of the argument in favor of J is, I believe, a great deal stronger.
In non-mathematical terms: you acknowledge that your argument is speculative, but I don’t think you realize just how speculative it is or how much the multiple layers of sheer conjecture weaken your case.
Best,
Tim
Travis,
Thanks again for the civil, thoughtful engagement.
On your points, using your numbering:
1) Your agreement on this point narrows the explanatory options considerably. Either it’s a genuine history coming from Joseph, or someone else was deliberately trying to make it look as though it were. This consideration will come back in below.
2) I’m glad you agree, up to that point, regarding the weakness of the argument from silence.
3) You write: “I agree that the history surrounding Herod and Archelaus fit with the narrative. However, this would be true for both a recounting of true events and for a synthesis of a theological agenda with a known history.” But how likely is such a synthesis? Not all that likely, not during the lifetimes of people who had been adults when the events occurred.
“Just as Josephus knew of these events, we should expect that the author of the Matthew narrative also knew these things in the preceeding decades.”
I think you’re moving much too quickly here. Christian authors from later eras (authors of the Acts of Paul and Thecla, the Gospel of Thomas, or other apocryphal works from the second century) seem to have very little grip on the history of the first. Josephus was writing with the full force of imperial patronage and had access to formal histories. We find no trace of these in the NT authors.
“I was hoping for more of a defensive against the conflicts between Matthew and Luke. Is that a forth-coming post?”
Yes, a post or two. Luke’s narrative is up next; depending on how long that is, I may have to defer consideration of the consistency of the two accounts until one subsequent to that. Regarding your particular points, so as not to leave you hanging for too long:
1) Luke 2:39 – actually, it’s less specific than that, but it does, I think deliberately, skip over the flight and return. If the document comes from Mary (as I will argue it does), then there is a straightforward explanation for the omission.
2) I disagree that Matt 2:19-23 “seems to assume that they were always in Bethlehem in the first place.” There are plenty of plausible reasons for their returning to the spot they left last, particularly since it is unlikely they knew about the slaughter, having left town there before it occurred and in response to warnings in a dream. Remember that to come from Egypt, they would have to swing north, and Judea is several days’ travel south of Galilee. So Judea (where, according to Luke, Mary had family) would be a very natural stopping point.
Regarding your subsidiary points:
a) The magi come some time after the birth, unlike our pop-art pictures that show them shoulder-to-shoulder with the shepherds.
b) “The age of two” for a Jewish audience would mean anything past the first birthday. “Under the age of two” would be approximately what we would call “not yet 1.” So you’re about a year off here.
c) I disagree with your claim that “[t]he structure implies that [Nazareth] is for them a new city and a new home.” That seems to me to be an overreading in search of a contradiction with Luke, not something driven by the text itself.
d) “In the second half of Matt 2:23 the Nazarene prophecy is claimed to have been fulfilled by the redirection to Nazareth, not on the assumption that it was their home town to begin with.” I don’t see what the assumption or lack thereof has to do with the argument. The text doesn’t say anything either way about whether it was their original home town.:
e) “Matt 1:18-2:21 offers absolutely zero reason to ever think that Nazareth was ever in the picture.” This is another argument from silence, which means that absent some pretty stringent conditions that are missing here, it has pretty close to absolutely zero force. Sorry.
As far as trying to see “how the harmonization of Matthew’s account with Josephus’ history can be seen as the work of a redactor,” I think this is the sort of exercise that gives literary criticism a bad name in the discipline of history. You’re making assumptions that require argument, among them, (1) that it is plausible that there was a redactor at all, or (2) that supposing there was a redactor, it is plausible that he would know so specifically about Archelaus (including all three details I mentioned), and (3) that if he existed, and if he did, he would drop the reference into the story in such an indirect way while leaving so many internal clues to Joseph as the source that he must have been deliberately deceiving his readers to do so.
You need the probability of all three of these assumptions to be quite high in order for your hypothesis to stand on something like an equal footing with Matthean authorship. I think all three are quite low, and their product is spectacularly low. But let’s be really generous and suppose that they each have a probability of 0.5. Then their conjunction has a probability of 0.125. Suppose that the probability of our having the narrative we do (N), given that it does not come from Joseph (~J) and that these three assumptions are all true, is 0.8: P(N|~J & (A1 & A2 & A3)) = 1. Suppose also that P(N|J) = 0.8. But suppose (and again, I think this is far too generous) that P(N|~J & ~(A1 & A2 & A3)) = 0.1. Then
P(N|~J) = P((A1 & A2 & A3)|~J) P(N|(A1 & A2 & A3) & ~J) + P(~(A1 & A2 & A3)|~J) P(N|~(A1 & A2 & A3) & ~J) = (0.125 * 0.8) + (0.9375 * .1),
which comes out to just a tad less than 0.2. That means that the likelihood ratio P(N|J)/P(N|~J) is more than 4 to 1, which is a pretty substantial evidential factor in favor of J, enough to move someone who started with P(J) = 0.5 to P(J|N) > 0.8. And I think that number is arrived at with a huge overestimate of the strength of your case for ~J. The actual force of the argument in favor of J is, I believe, a great deal stronger.
In non-mathematical terms: you acknowledge that your argument is speculative, but I don’t think you realize just how speculative it is or how much the multiple layers of sheer conjecture weaken your case.
Best,
Tim
Hi Tim,
Thanks for the thorough response. Before I offer anything in return, I would like to agree on two things:
A) We’re both doing a lot of speculating. We’re both trying to tell a story to make sense of the data. Neither of us should be under the assumption that we’re necessarily telling the right story. These are just tools for connecting all the dots in hope of painting a reasonable picture that fits all of the data.
B) If you’re planning future posts regarding Luke and resolving the conflicts between Matthew and Luke then I think it makes sense to wait and engage those topics there. Let’s table those until you’ve presented your case and then we can revisit as necessary.
Now, regarding the specific points, I’ll keep the same numbering:
1) I agreed that if the story is 100% true then it was likely to have come from Joseph for all the reasons that you pointed out. However, if the story is not true then I do not agree that the author was necessarily trying to make it look like it came from Joseph. We are also presented with insider information for Herod and the Magi, yet we don’t speculate that they were the sources for those portions. This only tells us about the literary style employed and, as far as I’m concerned, points toward some degree of invention. We probably agree that Josephus didn’t actually have first-hand knowledge of the minutia and the exact words, thoughts, dreams and emotions that are often expressed in his narratives. It was simply a product of his literary style, not a sign of personal testimony by those parties.
2) I did not mean to infer absolutes when I said that the absence of evidence is not evidence of absence. Obviously the level to which missing evidence counts against a claim is commensurate with the level evidence expected for the claim. In the case of the slaughter of the innocents, I agree that the expected evidence would not be substantial. I would not go so far as to say it would be negligible, as you suggest. My intent was to simply note that we’re not very far apart on that point and that it’s not a big factor to be relied upon. I will point out, however, that if the scale of the event truly was historically negligible then it is an odd choice for fulfillment of prophecy. Conversely, this is expected if it was only introduced as part of a larger attempt to link Jesus to Moses and explain why they left Bethlehem.
3) As to whether a literate, 1st century Jew would have knowledge of the murder of 3000 Jews and the perpetrator’s loss of power in the aftermath, we may simply have to disagree. In your argument, you say of Herod’s execution of 40 rebels that “the story was spreading through the crowds like a wildfire through dry grass”. How much larger a blaze the slaughter of 3000 and the canceling of Passover must have been. Then you note that the slaughter was relayed to Joseph by a group pilgrims fleeing back to Egypt. Why should I accept these claims yet reject the notion that the collective memory of this event would not have survived to reach the redactor?
If I skip past the discussion related to Luke, you then assign me with three assumptions.
1) “there was a redactor” – As I’m sure you’re well aware, this idea isn’t pulled out of thin air. It comes from my review of the synoptic problem and patristic claims regarding the early variants of Matthew. That’s a whole separate discussion but the point is that I think there is good reason to believe that the nativity is redactive material.
2) “the redactor knew about Archelaus” – this was addressed in #3 above.
3) “the redactor was attempting to make it look like the story came from Joseph and indirectly referenced the events with Archelaus” – See #1 above regarding my disagreement with the first part. Regarding the second part, I see it more like this: The redactor knew that the messiah had to be from Bethlehem and also knew that Jesus was from Nazareth, born to Joseph and Mary, and the general time period when he was probably born. The redactor also wanted to portray Jesus as the messiah, a second Moses and the son of God from birth. The redactor knew about Herod and Archelaus and that Archelaus lost control of Galilee, where Jesus was known to be from. In the course of constructing the narrative, these all come into play. There wasn’t intent to directly or indirectly incorporate specific details, only a desire to couple the agenda with what was already known.
Finally, I’m not going to get into the game of trying to formulate Bayesian probabilities for historical events which have no statistical backing. This kind of argument rarely means anything to me because (a) there’s almost no basis for thinking that the assigned probabilities are remotely accurate, and (b) it’s very easy to pick and choose the items to include, so as to favor one data set over another. This is likely to happen even if it’s unintentional. I think we would share a common dislike for Richard Carrier’s tactics, which rely heavily on this. That said – nice catch on the math. That was the first thing I noticed when I saw the original comment email.
Travis,
Agreed on points A and B both.
1) You write: “[I]f the story is not true then I do not agree that the author was necessarily trying to make it look like it came from Joseph.”
I disagree. The magi came directly to Bethlehem where they found Mary and Joseph; there is no mystery as to how Joseph would have known these things. But the multiple references to Joseph’s private point of view (1:19ff, 2:13, 19, 22) do make it look like it comes from Joseph. I am not saying that Joseph spoke it to Matthew; there is some reason to believe that Joseph was dead by the time of Jesus’ public ministry. But judging from the text, I should say that the most straightforward explanation would be that Joseph wrote it down or dictated it to someone who wrote it down, and that document was passed through the family to Matthew. If it is not that, or something much like that, it is the work of a forger who has gone out of his way to place himself in the position of Joseph and to write it very carefully that way.
The reference to Josephus’s “first-hand knowledge” does not help much here. Josephus never moved in the circle of the Herods, but he gets details right about them down to the color of the paint on Herod’s bedroom wall (Keener, Matthew (2007), p. 25). Jesus’ disciples were much closer to Jesus’ family than that. Plutarch and Livy are creative enough, but there is no reason to believe that they invented events out of whole cloth and placed them within living memory. If you think the focus on Joseph’s perspective is paralleled in some story from Josephus and is therefore plausibly attributable to Matthew’s use of a similar technique, you’ll need to give a citation and make that case in detail.
2) We will have to disagree on the paucity of evidence regarding the slaughter of the innocents, which I do indeed think is negligible evidence against the occurrence. The analysis of such arguments is, I think, best done in Bayesian terms, despite your dislike of the use of numbers in such arguments. Even where the numbers are merely illustrative, I think they can shed light on structure. The mere attempt to use Bayes’s Theorem in history is not the point at which I would disagree with Carrier; much more important is how one uses it and how one handles the historical data.
You write: “I will point out, however, that if the scale of the event truly was historically negligible then it is an odd choice for fulfillment of prophecy. Conversely, this is expected if it was only introduced as part of a larger attempt to link Jesus to Moses and explain why they left Bethlehem.”
I think this is exactly backwards. It is only if it really happened that we can explain why Matthew, doing a bit of pesher, thought of Jeremiah 31:15 as a point of resonance with the real events. But as a fictional means of getting Mary and Joseph out of Bethlehem, it is cumbersome and complex and risks discovery of the fabrication. Why not simply, “… and when the days of her purification were accomplished, Mary and Joseph made the proper sacrifices in the Temple, and they left Judea and settled in Nazareth”? If the events really took place, the narrative (and the suggestive connections to odd, minor points of prophecy) are natural enough; if they did not, then the invention of so tangled a tale is inexplicable.
3) You write: “As to whether a literate, 1st century Jew would have knowledge of the murder of 3000 Jews and the perpetrator’s loss of power in the aftermath, we may simply have to disagree.”
Who, exactly, do you think wrote the first two chapters of the first gospel, and when? If you think the story took form among agrarian followers of the Jesus movement after AD 70, then I submit that you are too quick to assume among the peasants of Palestine unlimited historical knowledge and dramatic imagination. (The case here, incidentally, extends well beyond the nativity story.)
You write: “In your argument, you say of Herod’s execution of 40 rebels that ‘the story was spreading through the crowds like a wildfire through dry grass’.” That was within a couple of weeks of the execution itself, in Jerusalem, where it had just taken place.
“How much larger a blaze the slaughter of 3000 and the canceling of Passover must have been.”
Much larger – but also removed by over half a century, even on a conservative dating, from the time of the writing of the first gospel. And again, you are not grappling with the triple precision in Matthew 2:22, all dropped in casually, without explanation.
You object: “Then you note that the slaughter was relayed to Joseph by a group pilgrims fleeing back to Egypt. Why should I accept these claims yet reject the notion that the collective memory of this event would not have survived to reach the redactor?”
First, because there is no independent evidence for the existence of “the redactor,” in the sense in which you need him, whatsoever. (See below.) The testimony of the early church (with whom the first Gospel was of chief importance) is overwhelming and unanimous that Matthew (otherwise not a major figure among the apostles) was the author. One must go forward over three centuries, to Faustus the Manichean around the year 400, to find someone trying to raise doubts about Matthean authorship.
Second, the pilgrims in question would have been fleeing Jerusalem within days of the massacre. There is a very significant difference between a gap of six days and one of sixty five years.
Does that make it impossible for the event to be remembered by a purely hypothetical redactor? No. Does it make it unlikely? I think somewhat unlikely, but I am not prepared to say that it would be desperately improbable. What of the probability that it would be remembered by a purely hypothetical redactor who, without access to the resources that Imperial patronage put at Josephus’s disposal, and writing after the fall of Jerusalem, would have recalled accurately the outrage perpetrated by Archelaus, would refrain from mentioning it at all, would nevertheless drop Archelaus’s name into a fictional story he was writing with the intention of recalling to his readers’ minds the events of more than half a century past, would choose with some precision a proper word to describe Archelaus’s position – kinging, but not king – and would send Joseph and Mary and the child to Galilee without bothering to mention that this was outside of Archelaus’s domains, all for the purpose of getting Jesus to Nazareth, when it would have been fully adequate to the same task simply to say, “And they went to Nazareth and settled there”? Yes, I think here we’ve reached desperate improbability.
On the three points:
1) I think you need to clarify what you mean by “there was a redactor.” If you mean merely that the texts of the gospels show here and there some signs of light editing, we probably don’t have a disagreement. But you’re turning this “redactor” into a full-fledged novelist. That is an entirely different matter. And you need that strong claim, because you want to conclude that “the nativity is redactive material.”
2) I have addressed this, above.
3) Ditto. You write: “The redactor knew that the messiah had to be from Bethlehem and also knew that Jesus was from Nazareth, born to Joseph and Mary, and the general time period when he was probably born. The redactor also wanted to portray Jesus as the messiah, a second Moses and the son of God from birth. The redactor knew about Herod and Archelaus and that Archelaus lost control of Galilee, where Jesus was known to be from. In the course of constructing the narrative, these all come into play. There wasn’t intent to directly or indirectly incorporate specific details, only a desire to couple the agenda with what was already known.”
If there was a novelist-redactor, and if he wanted to have Jesus born in Bethlehem but raised in Nazareth, then the entire story of Matthew 1-2 is a bizarre way to do it. Parallels with Moses? Why the flight into Egypt for that? If you’re unconstrained by actual facts, just have the boy Jesus speak to a rock and make water come out of it.
I think you should reconsider your dismissive attitude toward Bayesian reconstructions of arguments. In many cases, even where we are not in a position to give statistical arguments for precise numerical values, we can approximate them well enough for rough-and-ready purposes and can compare them to some simple stochastic probabilities. (What’s the probability that you’ll be rear-ended next Tuesday by someone driving a Yugo? I’ll bet you think it’s lower than the probability that you’ll roll double sixes with a pair of fair six-sided dice but higher than the probability that you’ll win the powerball – and not because you have precise statistical information on either of those propositions at your fingertips.)
The Bayesian apparatus still sheds light on our inferences even without the precise numbers. For the case at hand, it boils down to this: you need to show that the those three claims are each a lot more probable than their denials in order to have even a moderately good hope of making your case that your theory does as good a job explaining the facts as the straightforward realist hypothesis does. And I don’t think you can do it.
As for the initial slips – yeah, that’s the price I get for drafting things fast and hitting “Enter” before going back and checking. Brain says one thing, fingers type another. Sorry.
Best,
Tim
Travis,
As an addendum, and to save you time, you can find a presentation of some of my reasons for maintaining the traditional position on the authorship of the Gospels here: