Monday, 10 March 2014

Biblical scholarship vs. tradition

Periodically on SN the issue comes up of the “traditional” authorship and dating of the gospels, as opposed to the modern scholarly conclusions. The latest advocate of the traditional position is one “Irenaeus of New York” (one wonders what happened to the real-names policy), who endorses not only Matthean priority but also the idea that Matthew was originally written in a Semitic language (Hebrew or Aramaic).

So I thought a post reviewing the evidence and some resources would be useful.

The Synoptic Problem

The diagram shows the extent to which the three Synoptic gospels (Mark, Matthew, Luke) are interdependent; an outright majority of the content of each of the three is shared with one or more of the others in a way that indicates direct literary dependence—that is, the author actually had a text of another gospel, or a common source text, in front of him—rather than simply telling the same stories from a common oral tradition. The copying extends not only to actual sayings of Jesus but also to narrative and structural material, even to asides aimed at the reader (e.g. “let the reader understand”). If the three Synoptic gospels had been handed in by different students for an assignment, the instructor would have no hesitation whatsoever in concluding that blatant plagiarism had taken place.

The fact that we don't find this same level of textual similarity in the gospel of John is also significant; regardless of whether or not John was familiar with the texts of the other gospels, he does not simply copy the wording even when relating the same story with the same details (e.g. John 6, his version of the feeding of the 5000).


Early church writings are indeed unanimous in saying that Matthew was the first gospel. However, there are serious questions about whether those writers had any basis to know who wrote the gospels and in what order. The gospels themselves are anonymous and give no date of composition. Papias, very early in the second century, refers to works by Mark and Matthew, but the descriptions of those works don't square with the extant gospel texts. (Papias also doesn't give the relative order, nor does he suggest any connection between the two.) Furthermore, the early writers were more concerned about “apostolic authority” than details of dates, and they used the orthodoxy or otherwise of the content to determine the authorship, rather than the reverse. Accordingly, we find that tradition is less consistent about the relative order of Mark and Luke (both attributed to successors of disciples), with Clement of Alexandria apparently saying that Mark was written (or at least openly published—there is disagreement over the meaning) later than Luke, while Origen and others giving the order that became canonical (i.e. Matthew, Mark, Luke, John).

Importantly, though, these writers do not seem to suggest a literary dependence even though one clearly exists. Clement, Origen and Papias all claim that Mark was basing his work on what he heard as a companion of Peter, but unless Peter was reciting from a written copy of Matthew (which seems entirely implausible), this could not have produced the text of Mark as we have it. So something about the traditional account must be wrong.


Jerome quotes in several places a gospel he claims to have translated from Hebrew (possibly Aramaic; the early church writers did not distinguish the two). He attributes this to Matthew, saying that "most believe" it is the original version, and variously calls it “according to the Nazarenes”, “according to the Hebrews”, or used by the Ebionites. However, the scholarly conclusion is that he is conflating three separate sources of which he quotes two: a gospel in Aramaic (which modern scholars call the “gospel of the Nazarenes”), another in Greek (now called the “gospel of the Hebrews”), and finally the gospel of the Ebionites (in Greek, and apparently a harmony of other gospels) is not quoted by Jerome at all but by Epiphanius. Other quotations from Clement and Origen are attributed to these gospels too; but the amount of surviving text is tiny, with all three combined taking up only about 5 pages of Ehrman's Lost Scriptures.

Internal evidence

But the real meat of the issue, as with any plagiarism case, has to be in the internal evidence of the documents themselves. Regardless of what the early Christian traditions were, if the text conflicts with them in some way, then it's obviously not reasonable to believe the traditions over the texts.

Given the extent of the copying between the three Synoptics, the only real question is who copied from where. While pretty much every combination has been defended by someone or other, there are two major positions:

  1. Markan priority, in which Mark was written first, and both Matthew and Luke copy from Mark
  2. Markan posteriority, in which Matthew was written first, Luke derives from Matthew, and Mark is a kind of epitome of Matthew and Luke

Note that the issue of Markan priority or posteriority can be addressed without having to speculate about additional sources; the possible existence of “Q” or other hypothetical sources only comes into play when considering the origin of material that is not in Mark.

For an extended account of the textual evidence for Markan priority, Mark Goodacre's book The Synoptic Problem: A Way Through the Maze is available for free online reading or download (in PDF, Kindle or other formats). To summarize extremely briefly, the evidence includes:

  1. Editorial Fatigue.

    Editorial fatigue is a widely-known phenomenon that occurs when an existing chunk of text is adapted into a new context. There's a strong tendency for changes to be made near the start of the copied text (to make it consistent with the new preceding context) but for these changes not to be sustained through the passage, so that by the end, the original text is being copied verbatim despite the inconsistency.

    An example of this is in Mk 1:40-45 / Mt 8:1-4, in which a leper is healed; Matthew starts the story with Jesus being followed by a crowd, who are presumably watching the event, but Matthew still ends the story the same way as Mark, with an instruction for the leper to tell no one. Crowds following Jesus are common in Matthew, and secrecy is a common theme in Mark, but Matthew's version is inexplicable as an original composition (why have crowds if you're going to do a secret miracle?) and only makes sense as a copy of Mark. Goodacre lists a number of other similar examples in both Matthew and Luke.

  2. Harder readings in Mark.

    When a biblical text is copied manually, there's a known tendency for the copyist to “correct” readings that seem to be incomprehensible or theologically difficult. So when Mark says that Jesus could not do any “mighty works” in his hometown, and Matthew says that he did not do any, this looks like a correction of a hard reading (why would Jesus be unable to do anything?) to an easier one (clearly he could choose not to). It seems much less likely that the change would be made the other way.

  3. Scenes present only in Mark.

    There are only a very few scenes present in Mark but not Matthew or Luke. The first question there is, why would there be any such scenes? If Mark was created as a redaction of Matthew and Luke, why add anything at all?

    But the reverse theory, that the scenes in question were intentionally omitted by Matthew and Luke when copying from Mark seems very strong, for similar reasons to the “harder readings” issue above. Specifically, there are two scenes in which Jesus heals people by spitting on them, and a scene where an otherwise inexplicable naked man shows up only to run away. It seems unlikely that anyone would add those, but highly understandable that someone would omit them when copying.

  4. Omissions from Mark.

    If Mark is a condensation of the common material in Matthew and Luke, why omit the Lord's Prayer, and so much of the other material common to both? This makes much more sense if the extra material was added to Mark by Matthew and Luke, rather than omitted by Mark.

  5. Additions and deletions within individual scenes.

    The diagram at the top shows only one way of counting the degree of similarity between the texts. By another measure, much less of Mark is copied into Matthew and Luke, because they often omit apparently redundant explanatory clauses (e.g. “for there were many who followed him”, “for it was not the season for figs”, etc.). These don't seem like the sort of thing that a redactor would add when producing a condensed shorter gospel, but it makes much more sense for the creator of a longer work to omit them.

There's much more, and many examples of all of the above, in Goodacre's book; but the key point is that this isn't some unfounded speculation based on mere similarity of a few verses, but rather a consistent pattern that encompasses the whole triple-tradition text.


Given all of the above, it seems much more probable that the early traditions about the order of the Synoptic gospels are not based on any reliable knowledge of the order of composition, and likely owe more to the attributed authors' relative apostolic status.