Here’s my best attempt at the argument.
A couple disclaimers first: if forced to bet, I’d still wager more on Markan priority. However, I used to accept Markan priority as a given, but then when I heard this argument, it really made me start to question that belief.
For the purposes of this argument, I am going to make two broad assumptions:
First, that none of our versions of the synoptics gospels stood as a source for an intermediate gospel; that is to say, that a proto-version of Gospel A was a source for Gospel B, but then later that same Gospel B was used as a source for a revised version of Gospel A. Note: I actually think an intermediate gospel is a quite high probability, but it very much complicates more straight-forward arguments about synoptic dependence and chronology. So, for now, I’m going assume there was no intermediate gospel.
Second, since the argument outlined below relies heavily on the order of materials in the synoptics — in the main, the so-called Triple Tradition material — I use the terms “pericopes” and “sub-pericopes” (shorthand for “pericope subdivisions”) quite loosely. Depending on how these units of materials are counted, the general consensus among scholars appears to be about 90 pericopes in the Triple Tradition ( see Kurt Aland, BH Streeter, Robert Stein, Joseph Tyson) — 90 separate story units. So when I refer to “pericopes” and “sub-pericopes,” what I am attempting to invoke are the individual and separate units or stories that are in all three synoptic gospels — generally counted at about 90.
Finally, before discussing the most compelling argument for Markan posteriority to me, I should quickly address arguments for the consensus view. The three-legged stool of arguments for Markan Priority seems to be: (i) editorial fatigue; (ii) content and length of Mark; and (iii) rough Greek of Mark. As for editorial fatigue, the best rejoinder, IMHO, is the possibility of Markan redaction; that is to say that Mark found some of Matthew’s text to have inconsistencies, including with Lukan accounts, and sought to smooth those over — for example, Mark changed Herod’s title to “Herod the King” from “Herod the Tetrach,” because he didn’t like the way Matthew used both terms. As for content and length, this argument used to really persuade me (why would Mark create a new gospel, which largely copied from earlier texts, but omitted key materials from his sources, such as the infancy narrative, the genealogies, and the resurrection appearances?); but then I started to read about other ancient works that served as abridged versions of earlier works, which might make sense if Mark saw his purpose in composing his text as different from Matthew’s and Luke’s. Finally, there is the matter of Mark’s rough Greek: why in the world would Mark borrow material, but degrade the literary quality of the source? This topic deserves far more than casual dismissal, but for now I would just point out that we actually have examples of non-canonical texts trying to pass themselves off as early Christian texts by using the synoptics as a source, while heavily butchering the Greek; so, it does happen.
Ironically, order used to be perhaps the leading argument for Markan Priority, as developed by BH Streeter, apparently misunderstanding the arguments of Karl Lachmann. The argument from order developed from the observation that, when Matthew and Luke agree in the order of pericopes, both Matthew and Luke are almost always following Mark’s order, and when Matthew and Luke disagree in order, at least one departs from Mark; thus, Mark’s sequence is the backbone from which the others diverged — implying Mark wrote first.
BC Butler attacked Streeter’s conclusions, arguing that the evidence from order cited by Streeter in support of Markan Priority could just as easily support other answers to synoptic chronology and dependence.
However, it was William Farmer who kind of turned the argument from order on its head, in his support for the Griesbach hypothesis, which posited that Mark was the last of the synoptic gospels. In essence, what Farmer observed was when you (i) examine all the pericopes and sub-pericopes common to all three synoptics, (ii) start from the assumption that Mark wrote first, and then (iii) track the order of pericopes and sub-pericopes, there are many pericopes and sub-pericopes where the order of pericopes diverges as between the three synoptics. In those cases, Farmer observed that Matthew and Luke appeared to alternate in following Mark’s order. Indeed, a repeating pattern — particularly in the first half of Mark — is that Matthew will agree with Mark’s order; but then when Matthew stops agreeing with Mark’s order, Luke will begin agreeing with Mark’s order; and then when Luke stops agreeing with Mark’s order, Matthew will then resume agreeing with Mark’s order; and so on and so forth.
Then, Farmer conducted the same analysis, but changed the assumption to Mark being written last. First, Farmer observed that, using these assumptions, and focusing on the rarer circumstances that Matthew and Luke agreed on the order of pericopes and sub-pericopes, in those cases Mark followed the common order of Matthew and Luke 100% of the time. 100%. Note that alternatives of this are not true: when Mark and Matthew agree on order, Luke commonly disagrees; and when Mark and Luke agree on order, Matthew commonly disagrees. But, when Matthew and Luke agree, Mark ALWAYS agrees. Alternatively, when Matthew and Luke disagreed, Mark sometimes agreed with Matthew, sometimes agreed with Luke, with the variations often displaying divergences in thematic flow.
And, indeed, these observations from order were likewise reflected in common word choice, i.e., when Mark and Matthew have the same order, you see the same vocabulary being used in both accounts; the same applied to common order as between Mark and Luke. The similar vocabulary, however, does not extend to the same degree where Mark and Matthew (or Mark and Luke) are telling the same pericope or sub-pericope, but in different order. When common pericopes are told in diverging order, the word choice is more likely to diverge as well.
From these observations, Farmer concluded that it was more likely that Mark was written last, and used both Matthew and Luke as sources. This conclusion, to Farmer, best explained the data — as Mark would have used Matthew and Luke as sources, copying their order and vocabulary when they agreed on order, and alternating between Matthew and Luke when they diverged in order, while aping the vocabulary of each source as he switched back and forth.
Then came Malcolm Lowe and his 1982 paper, “The Demise of Arguments from Order for Markan Priority” (** you do not have permission to see this link **). In his paper, Lowe analyzed the mathematical implications of each of Markan priority and Markan posteriority using the 90 or so pericopes from the Triple Tradition. In sum, Lowe tested how often the Triple Tradition materials aligned with what would be expected if Mark was a source for both Matthew and Luke, versus what would be expected if Matthew and Luke had both served as sources for Mark.
What Lowe found was that the order and vocabulary for 57 of the 90 pericopes align with expectations when applying the assumption Mark was first and a source for both Matthew and Luke. However, when applying the analysis using the assumption that Mark was last and used both Matthew and Luke as sources, then the order and vocabulary for 88 of the 90 pericopes aligned with expectations.
This is a powerful evidence … and greatly surprised me … and is difficult to explain under any hypothesis that Mark did not use Matthew and Luke as sources. Effectively, the hypothesis that Mark wrote last and used Matthew and Luke as sources better explained the data of order and vocabulary by 54% than the hypothesis that Mark wrote first and was a source for Matthew and Luke.
Now … taking a step back … is there a hypothesis that accounts for all the data I have discussed in this post — a hypothesis that Mark wrote last, using Matthew and Luke as sources; that Mark’s gospel is about 40% shorter than Matthew’s and Luke’s; that the material Mark uses unique to his gospel (“Special Mark”) is only one-third the size of the so-called “Double Tradition” material (i.e., material common to both Matthew and Luke which is not common to Mark); that 97% of Mark is comprised of material common to Matthew or Luke or both; that Mark’s Greek is rough and of inferior quality to the Greek of both Matthew and Luke; and that the Special Mark material is heavy on dramatic turns, secret elements, narrative quirks, and sensory detail?
One hypothesis is that Mark was trying to create a more emotionally-gripping gospel, perhaps to be used to win-over new converts to the faith. Mark perceived Matthew and Luke as being too long and containing less-emotionally-gripping materials for people who are just learning about Jesus for the first time, and need to be hooked in. Mark knew that Matthew and Luke were superior writers in Greek, so he copied them the best he could when he wanted to use their material; but literary quality wasn’t his strong suit. Instead, Mark focused on story-telling and keeping his reader engaged. That was Mark’s secret sauce. Mark was seeking to create an abridged version of the Jesus story that was more gripping, more emotionally-connected for people who didn’t know much about Jesus. It was shorter and punchier. His goal was to get people interested in Jesus, not more fully explain the biography of Jesus.
So, there it is …. an argument against Markan priority.
Hi Robert.
As you’ll read, I’m not really attempting to do a comprehensive refutation of the evidence for Markan priority. I simply gave very cursory possible explanations for some of the leading arguments. But that part is almost superfluous to rest of the post. I was merely trying to suggest there might be explanations for the data supporting Markan priority, in order to partially open up minds to the possibility of Markan posteriority.
Now, let me return to the issue of intermediate gospels — because I think there’s something there.
Let’s assume the hypothesis in the first post — that Mark wanted to write a shorter, more dramatic gospel and used Matthew and Luke as sources to do that. Using that assumption, can we explain some of the data used as arguments in favor of Markan priority by appealing to an intermediate gospel?
What about the argument based on Semiticisms — i.e., that Mark preserves Semitic features more often than Matthew or Luke in parallel passages, where Matthew and Luke smooth them out into more idiomatic Greek. An oft-cited example is Mark 2:4 using a Hebraism that effectively translates to “they unroofed the roof” (ἀπεστέγασαν τὴν στέγην). Let’s suppose Mark used Luke 5:19 (“they went up on the roof and lowered him on his mat through the tiles”) as his source, but knew of the Hebraism and wanted to add it to his version; or, perhaps, a proto-version of Matthew included this Hebraism. Could an early Matthean redactor — who was “finalizing” the text have looked at the Lucan version without the Hebraism, and decided to remove the Hebraism as too clunky? This is highly speculative, but could explain the data.
Now compare this treatment with that in Mark 1:32 and Matthew 8:16 — both of which contain elements of Hebrew (“having become evening”), though Mark’s version contains a redundant time expression (“when the sun had set”). Luke’s version (4:40) is the smoothest in Greek (“as the sun was setting”). A possible explanation — rather than Matthew partially copying from Mark here — is that Mark copied the entire redundant Hebraism from a proto-Matthew. Then, when a Matthean redactor came around, he noticed how much smoother Luke’s version read in Greek, and simply lopped off the redundancy, while maintaining the Hebraic temporal formula in first phrase.
A Matthean redactor could also explain some of the data indicating possible editorial fatigue going both directions as between Mark and Matthew. I will address that in a later post.
In his paper, Lowe analyzed the mathematical implications of each of Markan priority and Markan posteriority using the 90 or so pericopes from the Triple Tradition. In sum, Lowe tested how often the Triple Tradition materials aligned with what would be expected if Mark was a source for both Matthew and Luke, versus what would be expected if Matthew and Luke had both served as sources for Mark.
What Lowe found was that the order and vocabulary for 57 of the 90 pericopes align with expectations when applying the assumption Mark was first and a source for both Matthew and Luke. However, when applying the analysis using the assumption that Mark was last and used both Matthew and Luke as sources, then the order and vocabulary for 88 of the 90 pericopes aligned with expectations.
On my very quick read, it appears you have misunderstood Lowe’s argument and evidence.
He is arguing only that the order of pericopes, taken alone, cannot logically establish that Mark wrote before Mt and Lk. The statistics you cited seem to have come from an analysis by Tyson, cited by Fee, and presented by Lowe on pp 34-35 of the linked article, but what you present does not seem to align with what Lowe presents (as one example, I can’t find 88 anywhere in the text; for another he is not, so far as I can see, looking at coincidence of vocabulary but just order of pericopes).
Just when I thought I was out, they pulled me back in!
I spent so much time years ago parsing this issue that I am loath to dive back in. So I will make general observations.
First, I agree with Robert. “Editorial fatigue” is neither the best nor the strongest evidence for Markan priority.
Then, let me take another approach. It’s clear from the writings of the early church that Mark was always the “red-headed stepchild” of the gospels. They didn’t quite know what to do with it. Some supposed that Mark contained the memoirs of Peter. But the dominant view became that Mark was an abridgement of Matthew. Until the 19th century!
But the reasons why Mark would exist as an abridgment simply don’t make any sense. I would even claim that the existence of Mark only makes sense if it was the first. It’s hard to see why it would have been preserved otherwise. It leaves out so much that has become considered central to the tradition. All those wonderful sayings! Mark’s Jesus is primarily a wonderworker, a healer and an exorcist rather than an ethicist. Who would such an abridgement be intended for? Now none of this is evidence but it is worth thinking about.
Another consideration. It is not often pointed out that the relationship between the synoptic gospels is unprecedented in ancient narrative literature. Nowhere else do you find vast swaths of text shared verbatim by disparate authors. There are examples of magical grimoires and books of recipes where it’s important to copy the wording correctly but nothing like the gospels. Why would any of these writers simply copy each other instead of writing their own version like John did? One idea I’ve tossed about, somewhere between whimsy and speculation, is this.
Given that so much of Matthew and Luke consists of verbatim repetition of Mark, perhaps it might be better to consider them not as separate gospels, but as later versions of Mark? Perhaps what we’re dealing with here are not disparate literary traditions but one single literary tradition? One immediate objection is that each of these texts has its own focus and point of view. But look at the rest of the Bible. It’s full of editing and redactions. In the books of Samuel, for example, we have both a pro-monarchical source and an anti-monarchical source edited together in the same text. Consider multiple Isaiahs. Can one seriously claim that the original versions of these other texts resemble the versions that have survived?
Anyway in this overall discussion just consider me plucky comic relief.
the reasons why Mark would exist as an abridgment simply don’t make any sense. I would even claim that the existence of Mark only makes sense if it was the first. It’s hard to see why it would have been preserved otherwise. It leaves out so much that has become considered central to the tradition. All those wonderful sayings! Mark’s Jesus is primarily a wonderworker, a healer and an exorcist rather than an ethicist. Who would such an abridgement be intended for?
To steelman the other side: Early Christianity was not monolithic. It could be that Mark belonged to an otherwise lost tradition of Christianity that had a very low Christology (and also, deemphasized his ethical teaching; or maybe Mark presumed knowledge of Q among his readers and didn’t feel the need to duplicate the material), so that is the Jesus we see in the gospel he produced. I don’t think that makes the most sense of the data, but I also recognize that, in principle, we can’t reliably date texts based on the ideas in them; rather we must date ideas from the texts that contain them.
Given that so much of Matthew and Luke consists of verbatim repetition of Mark, perhaps it might be better to consider them not as separate gospels, but as later versions of Mark? Perhaps what we’re dealing with here are not disparate literary traditions but one single literary tradition?
I think that is absolutely the right idea. I think they defy our neat categorizations (author vs. scribe-who-took-liberties). On the one hand, they act a lot like scribes given the extensive verbatim copying. On the other, they are clearly authors who have an overarching vision for the work: they don’t just tweak a passage here or omit an offensive line there; they fundamentally rework the material in a coherent way. But if we look at the trend: The earlier we go the less slavish were the copyists (which makes sense given that the sacredness of the text grew with time), if we then extrapolate backwards, I think the synoptics fit the trend line.
An interesting data point is that (I’ve been informed) the Kata + (author) formulation (such as we use to identify the gospels) was routinely used in the ancient world to distinguish different versions of the same text.
Stephen, I’d argue that we have a very good idea of why Mark was preserved — Papias tells us. The belief — even if incorrectly held — among early Christians was that the Gospel of Mark was effectively the stenographer for the teachings of Peter.
It doesn’t really matter if that belief was correct or not, the best evidence we have is that it was a widely held belief.
Once you accept that this belief existed, we can move on from the reasons for preserving Mark, and instead examine why Mark have been composed in the first place.
I used to originally believe that Mark as an abridgment of Matthew and Luke made no sense. But, in re-thinking through the issue, the possible reasons become far more clear. Matthew and Luke are relatively long treatments of Jesus, and include materials that might not matter much for possible converts from paganism. How much would a gentile living in Turkey or Greece care about Jewish genealogies? The infancy narratives and nativity frankly don’t make for a great opening for an exciting story.
Mark can be recited in less than an hour. Mark includes more intrigue and suspense than is present in Matthew and Luke. To use a more modern term, Mark is more of a page-turner and a quick read.
I guess my point is that the ending of Mark at 16:8 is a true mystery. We don’t know what is going on there.
If Mark had the opportunity to write about resurrection appearances, but chose not to, which I think is kind of the majority view among scholars — there’s a reason for that — and I don’t think it’s because he didn’t know the stories.
So for whatever reason there was that resurrection appearances were not included, I’m not sure how relevant it is under the hypothesis that Mark was the first text or whether Mark was an abridgment. Is the explanation, assuming Markan Priority, that Mark either (a) didn’t know about the appearances (highly unlikely) or (b) didn’t know to end his gospel with the appearances? If it’s the latter, that’s a bit tough to reconcile with the creed in 1 Corinthians 15:3-8. A very early liturgical framing of Christ for the Christians seems to be: Christ died for for sins; Christ was buried; Christ rose on the third day; and Christ made appearances to Peter, the Twelve, and others. Mark included the first three, but not the fourth … for whatever reason.
The notion that the text ends with a cliffhanger does reconcile with Mark’s theme of secrecy. It could simply be a narrative technique, and it doesn’t really matter if Mark was the original text or an abridgment of other texts, particularly if the intended audience didn’t know the story before hand.
…but I have not yet been thoroughly convinced.
Robert I would be delighted to be corrected but I have searched rather widely and I cannot find a series of ancient Greco-Roman texts that closely mirror the relationship between the Synoptics. In my opinion the simplest explanation for that is a continuity of belief so while I have fallen under the spell of Robyn Faith Walsh I don’t follow her everywhere. (That would be stalking!) I would love to ask her about it.
I guess my point is that the ending of Mark at 16:8 is a true mystery. We don’t know what is going on there.
I find the view that 16:8 was the original ending fairly compelling but I do pull out the “lost ending” idea for a spin now and then. V 16:7 does seem to at least imply a future episode to be described. What makes my Spidey-sense go off is how attractive an ending at 16:8 is to modern sensibilities. Anachronism and projection is almost inevitable with these ancient texts and anything that strikes me as relevant automatically makes me suspicious. The ancients didn’t think the way we think. They had different assumptions about how the world works and different expectations as to how to respond.
Walsh raised a furor with her suggestion that the gospels writers might not have been writing out of personal piety. She has kind of backed off a bit since there was a danger it would obscure her larger point. But it seems to me we have to at least consider the idea that Mark was a creative writer who found himself a crackling good yarn.
The actual provenance of Mark is completely unknown to us. We can glean something from internal evidence but the same thing could be said for War and Peace or Moby Dick. And as Walsh pointed out in her book, Hellenized Jewish literary practice is one of the things we know least about.
If Mark had the opportunity to write about resurrection appearances, but chose not to, which I think is kind of the majority view among scholars — there’s a reason for that — and I don’t think it’s because he didn’t know the stories.
I’m going to voice agreement vergari on the issue of Mark omitting resurrection appearances: Unless Mark is much earlier than anyone thinks, it is all but certain that they were known to some significant number of Christians well before Mark wrote, and (it seem to me) it is almost certain that they would have been known to all Christians including Mark. That he left them out can’t peg down his date, because it doesn’t seem likely that he left them out because the tradition had not yet developed; rather, assuming he did leave them out (and they haven’t just been lost from the text of his gospel) he likely left them out by choice (a choice motivated by theological or literary concerns that could have been made at any time) and not out of ignorance (because the tradition hadn’t developed or spread).
I agree with all of this up until the end. Scribes and authors are doing fundamentally different things. The trend line among the gospel authors is the opposite of that of scribes.
That is a good point. I’d been thinking of the Synoptics as a single group, not thinking of how the trend would cut among them.
I do in fact believe it was a narrative technique, a classic Aristotelian plot device, and yes it fits very well into the secrecy motif, but did the ending serve the needs of Matthew, Luke, and John? Obviously not. Both the ending, with the disciples apparently never to hear of the resurrection, in addition to their repeated inability to understand Jesus and his teachings are issues that we can see Matthew, Luke, and John reacting to in their respective gospels.
I think it’s more likely than not that Matthew and Luke are responding to Mark. But that is by no means clear to me. It’s a perfectly plausible scenario that Mark looked at the Jesus teachings, particularly in Matthew, and thought that introducing the literary device of secrecy would be more captivating to possible new converts. I don’t see why it’s so much, much more likely that Matthew and Luke are making changes in direct response to Mark.
I suppose this is just an eternal issue I may have with redaction criticism. I would love to see home hard data, some numbers, some math, behind the idea that Matthew and Luke are responding to Mark. Don’t get me wrong; it’s intuitive. And I’m largely persuaded. But then I look at this analysis — which now I can’t seem to find — that appeared to include data indicating that Mark following the order of Matthew & Luke was 50% more likely than Matthew & Luke following the order of Mark.
From Porphyry:
On my very quick read, it appears you have misunderstood Lowe’s argument and evidence.
He is arguing only that the order of pericopes, taken alone, cannot logically establish that Mark wrote before Mt and Lk. The statistics you cited seem to have come from an analysis by Tyson, cited by Fee, and presented by Lowe on pp 34-35 of the linked article, but what you present does not seem to align with what Lowe presents (as one example, I can’t find 88 anywhere in the text; for another he is not, so far as I can see, looking at coincidence of vocabulary but just order of pericopes).
Porphyry, you are correct about the page source on that article: end of page 34; beginning of page 35. Note, in particular, the following statement:
There are thus 57 pericopes which in their sequence corroborate both pure [Markan priority] and pure [Griesbach hypothesis] [a/k/a Markan posteriority], 2 [out of 90, i.e., 88] which contradict the latter [the Griesbach hypothesis / Markan posteriority], but 33 [out of 90, i.e., 57] which contradict the former [Markan priority].
So this is the data from Malcolm Lowe in his 1982 article.
I will check again the source on vocabulary. I believe that is also from Lowe, but it may have been from Peabody.
BDEhrman
FreedomBen
evgendob
Robert
1 Guest(s)