Spoiling Christmas vs. Spoiling Q

Blame Mark Goodacre if you must (who replied, as I hoped he might, to my previous post about Q). But I find that I must provide details that would spoil most fundamentalists’ Christmas in order to make my defense of ‘Q’. Let me be clear once again that I am not presupposing that Q was a single written document. I do suspect that at least some of the material common to Matthew and Luke but not in Mark came from a written source, in Greek. But I don’t think the case for Matthew and Luke’s independent use of common source material rather than each other depends on our ability to determine in every instance whether it was a written source in a particular language, much less our ability to reconstruct a precise ‘original’ form.

The problem with determining relationships is that, even when there is clear use of a written source, none of these authors is likely to have encountered the majority of the stories and sayings for the first time when they read the written work in question. (That Matthew and Luke seem not to have known how Mark’s Gospel would have continued beyond Mark 16:8 is a remarkable exception to this rule that merits a separate treatment on some other occasion. Remind me to return to it someday). These authors had heard the sayings, been told the stories, and prayed the prayers. The versions they encountered would have diverged through conscious reworking as well as simply through the natural processes of oral transmission. Since the authors would have had this separate fund of knowledge to draw on as well, not every difference will be due to redaction of a source. In some cases, the difference will simply show that the author knew a story so well already in one form that the different written version in front of him could not supplant it.

I digress. My point about the differences between Matthew and Luke providing evidence for Q did not have in mind simply their failure to overlap. It is the cases of genuine incompatibility that are the biggest obstacle to making the case that one drew on the other directly.

And so, Christmas comes early this year – at least as a subject on this blog. The infancy narratives provide significant evidence of the point I was seeking to make.

We may quickly dispense with the genealogies. Their apparent incompatibility has been noticed since the days of the early Church Fathers. The proposed solutions – such as successive levirate marriages – are not more persuasive than the view that these authors did not draw directly on one another.

Let me acknowledge outright that I can imagine some possible changes that one of the authors might have made to the other. For instance, I can understand why Matthew would change Luke’s genealogy to make Jesus more directly the heir to the royal line. I can also understand why Luke would have changed Matthew’s, to avoid the problem of Jeremiah’s pronouncement that no descendant of Jehoiachin’s would sit on the throne (Jeremiah 22:28-30). But neither of these changes explains why they did not keep more of the names the same. The more recent lineage – particularly the name of Joseph’s father, which some who were still alive might have actually remembered – could have been left as it was. Indeed, Luke’s list of names would have provided Matthew with enough additional names to make his last group of 14 actually add up to 14 (which, alas, as it now stands, it does not).

The incompatibilities run deeper – not simply in the date (Matthew places Jesus’ birth before the death of Herod in 4 BCE, Luke connects it with the census under Quirinius in 6 CE), but also in the geographical flow. Matthew’s Gospel places the family in Bethlehem from the outset. They flee to Egypt, and then we are told that they wish to return to the place from whence they had left. They only go to Galilee because they fear Archelaus.

In Luke’s account, they begin in Galilee. They go to Bethlehem simply for the census. After Mary gives birth, they are there for a little over a month, because they head to Jerusalem for the purification rite (Leviticus 12:4). Then as soon as they’ve taken care of that (Luke 2:39) they return to Nazareth. Not only is there no time for a detour to Egypt, but instead of being afraid of the king ruling in Jerusalem, they visit public places there and have impressive things said about their infant.

The two accounts simply cannot be reconciled – in spite of the many large groups of children with towels on their heads who have often seemed to do so. I hope this post doesn’t spoil your enjoyment of your local church’s Christmas pageant. I have nothing against telling the story in its traditional form, as long as one doesn’t think it constitutes a plausible historical reconstruction. I doubt most of the children who enjoy participating in it are worried about that.

But presenting these discrepancies is necessary if we are to be clear on why the common material between Matthew and Luke is best accounted for in other terms than direct use of one by the other. That they fail to follow each other precisely is no surprise. But that they fail to be even compatible with one another when there is no obvious motivation for the divergence, that, in my opinion, provides evidence that they drew on a separate source independently of one another.

I look forward to reading Mark’s reply – which I hope he will entitle “Christmas without Q”. If he does, I will do my best to offer in response a version of “The Twelve Days of Christmas” that lists by numbers points of similarity and difference between Matthew and Luke relevant to this subject. I can hear it already. “On the first day of Christmas my Q-love gave to me…”

Stay in touch! Like Religion Prof on Facebook:
Passion According to Mary
Putting Jesus on ICE
Academia Pro and Con
Slavery to the Fear of Death
  • steph fisher

    I hope you don’t mind me butting in. I agree with you that it is not possible to reconstruct a precise ‘original’ form of the sources used by Matthew and Luke. However I do think cases can be made with particular examples in the double tradition, demonstrating that the source, written or oral, was more that likely in a particular language.Sometimes because particular disagreements between Matthew and Luke are what they are, I think it is possible to propose reconstructions which may approximate historical forms of a particular source. Just small bits, demonstrating a better hypothesis in each case than that of Luke having changed Matthew.On the birth narratives, I agree mostly with Mark Goodacre in The Case Against Q. I think that if Luke read Matthew (which I think more than likely he did) he is likely to think that his own birth tradition was more reliable and Matthew’s narrative was just plain ill informed.Happy Christmas.

  • http://www.blogger.com/profile/10284073533968750655 Rick Sumner

    Why did Renan come up with his account when he was using the GJohn? How does Fredriksen use her sources to come up with a sudden shift in chronology?The difference between Fredriksen and Renan on one hand, and Luke on the other, is that they explain how they get there. Whether we agree with it or not, we know the road they took. Luke doesn’t explain that, but that doesn’t mean he missed the drive.I can’t think of a reason we should assume that Luke was doing any different than a modern exegete, at least in principle: Attempting to reconstruct a life of Jesus from his sources.If the modern exegete can use their sources creatively, why couldn’t Luke? If Renan didn’t give an explanation, would you know he was (at least in his mind), primarily relying on John?I must confess, as arguments for Q go, I find this underwhelming.

  • http://www.blogger.com/profile/10326403777027937887 Doug Chaplin

    While you’re waiting for Mark to give the definitve rebuttal of this argument (!) I’ve posted a rejoinder that was a little to long for a comment.

  • http://www.blogger.com/profile/00559055709208918638 Eric Rowe

    Fundamentalism exists as a response to people who disbelieve the historicity of such portions of Scripture as the infancy narratives. I should hardly expect that the sudden realization that James McGrath belongs to that number would be the final straw that spoils a fundy’s Christmas.

  • http://www.blogger.com/profile/02561146722461747647 James F. McGrath

    Well, I certainly didn’t mean to suggest that my holding certain views on this subject or any other will lead to the downfall of fundamentalism or any other movement! I meant simply that I suspect that fundamentalists who read my post might either encounter some new information that challenged their assumptions, or (perhaps more likely) simply regard me as one of these scholarly types who spoils our enjoyment of holidays by writing these sorts of things!

  • http://www.blogger.com/profile/03126711689901268060 Quixie

    Rick Sumner said:”Why did Renan come up with his account when he was using the GJohn? How does Fredriksen use her sources to come up with a sudden shift in chronology?“This is kinda vague. Would you please explain what you mean by these? Perhaps I’m missing some widely-accepted opinion regarding Renan’s and Fredriksen’s methodologies. I have read both authors and can’t figure out exactly what you mean here.Ó

  • http://www.blogger.com/profile/10284073533968750655 Rick Sumner

    I’m not sure what you find confusing: Fredriksen shifts from the Synoptic to John’s chronology at the point of the trial. She explains why she does so, which is what makes her different than Luke–providing the explanation that Luke doesn’t.Likewise Renan; his use of the fourth gospel leaves it almost unrecognizable as his source. Without his explanations, one would wonder how he came to the conclusions that he did.I’m suggesting that Luke did much the same thing–attempted to discern what was the likely chain of events from his sources. He just doesn’t explain the steps. That he did this seems, to me, to be self-evident: We know he did it with Mark.

  • http://www.blogger.com/profile/03126711689901268060 Quixie

    Rick:- – – – – – -Thanks for kind of explaining yourself a bit.What I specifically found vague was the citation of the work of Renan and of Friedriksen in this discussion, without any elaboration on what nuances in their methodologies were relevant to the case at hand.(eg. When you talk about her—Ms. F—changing the chronology . . . . does that mean ‘chronology of the narrative events in GJohn’? or of the books themselves?)An example (off the top of my pointy head)— read the following:”Having modulated to a chord based on the minor third of the original key, Coleman Hawkins would normally choose between using the ninth or the diminished ninth while improvising on a chord built on the dominant degree of the new scale (still unresolved to the tonic), although Wayne Shorter managed to create a harmonic vocabulary that allowed for the alternating use of both choices.”By saying something like this, without further definitions, I’m assuming that the reader is acquainted, not just with the conventions of jazz improvisation, but with the tiny nuances involving “methodological” or stylistic characteristics espoused by particular “specialists” in the field. If I knew that the readers were all jazz musicians, I would not really need to explain those things. But, for a more general audience, I would make sure that I define the terms I’m using and how they relate to the practicioners that I cite. Otherwise, it’s just a vague reference.————————That said, I follow your argument, but I disagree that the author of GLuke is doing the “same thing” as these modern exegetes. I feel that comparing Luke’s methodology to that of modern historiographical work is anachronistic. Luke has not yet been demonstrably shown to be a “historian” in that modern sense. I realize that every conservative is waving the Bauckham banner these days, but I personally found his book (“J & Eyewitnesses”) extremely underwhelming (to use your phrase).for now . . .Ó

  • John C. Poirier

    James,Sharon Mattila makes the same argument that you make, but I don’t think it goes anywhere.I agree with Steph above (citing Mark G.) that it’s hardly a difficult question to ask why Luke preserved none of Matthew’s birth narrative. The fact that there are two different birth narratives is evidence enough, I think, that two distinct narratives were circulating. So if Luke read Matthew, what’s so hard to believe about him disliking the narrative that Matthew chose and substituting the other one for it? After all, Luke tells us that he wrote his gospel because he wants to write a better account than what’s already out there, so we know that he has disagreements with his sources.

  • http://www.blogger.com/profile/02561146722461747647 James F. McGrath

    When I see how Matthew or Luke altered Mark, I can usually figure out why they would have done so. But whereas I can certainly see distinctive Matthean and Lukan emphases in their infancy stories, I can’t figure out why changes in geographical flow and other details were made, if indeed one of them used the other’s Gospel as a source. Many of them don’t seem to help make a particular point, while they do raise questions of how the narratives relate to one another. That’s why I find it hard to believe that one or the other deliberately created this confusion without having an obvious theological or literary aim.

  • Ian

    Mark mostly convinced me of Farrer, I have to say. But I just can’t buy the whole package. The start and end of the gospels are incompatible solutions to a specific problem. They simply don’t read like Luke is aware of Mark but wants to tell a different version.

    It could be that Luke is aware of Matthew, but actively discounts it as a valid source: using it only for bits of content that he knows orally, say. But in that case, I’d expect Luke to be more pointed in his differences, not just to be ignoring. The sense of a story that says “not that, but this”, which I don’t get.

    Perhaps this means there’s a proto-Matthew that doesn’t have the ends in it, circulating. But again, reconstructions of convenience!

    So yes, Mark is very convincing, but I’ve always come stuck on the nativity and passion.

    • http://www.patheos.com/blogs/exploringourmatrix/ James F. McGrath

      Here’s another post on the importance of differences and not just similarities in the case for Q: http://www.patheos.com/blogs/exploringourmatrix/2007/11/the-difference-makes-q.html

      • Ian

        Thanks, don’t know if I should respond there or here.

        I disagree a little. I think the kinds of similarities mean there’s basically no chance that the dependency between Gmatt and Gluke is entirely oral. There is too much verbatim overlap. Telling the same stories can be oral, telling them with the same grammatical structure isn’t (unless one posits the kind of early oral tradition that is only a holdout of advocates of Gmatt primacy). There is a literary dependency there. The only question is, how much of the unique Matt source does Luke know, nothing (Q), or all (Farrer) or in between (a proto-Matthew).

        Exclude the start and end, as you say, and I think Farrer is convincing, the minor agreements, for example, suggest so. But I do agree, it is the top and tail that hurt it. There it is the disagreements that make it awkward, as you say. I’m not convinced that the sequencing differences b/w Gmatt and Gluke are significant, because both of them also resequence Gmark, and the literary dependency there is uncontroversial.

        But Q has major problems too, that are stretched by positing scribal homogenisation. As for mythicism, we know positing scribal changes is simple, and dubious without actual evidence. So Q has to account for a lot still, when Matt and Luke agree against Mark (i.e. in the non-Q material, unless you posit a much expanded Q).

        When I ran the statistical analyses of the language, neither option came out on top (similar results have been found since, with more comprehensive analyses on more grammatical features), hence my tendency to a kind of intermediate solution. Modulo the insignificance of my opinion, and the amateurism of my understanding generally.

        • http://www.patheos.com/blogs/exploringourmatrix/ James F. McGrath

          Now if you can just think of another option that the statistics supports better than those offered by the major scholarly hypotheses, you would be remembered forever. Any ideas?

          • Ian

            I wish! My hubris-filled plan was to solve the problem *with math*(TM).

            The problem is the statistics show what they often do (and what, incidentally, Carrier never even considers with his probabilities): that the statistical error is so large in the data set, that no conclusion is possible.

            By that I mean, no conclusion on purely syntactic features.

            It is simple to come up with a new theory that matches the stats better: that’s just curve fitting to the data. Likewise it is easy to take subsets of the data, and generate more compelling conclusions (hence my ‘if you leave out the start and end’). But it doesn’t tell us anything. My conclusion was (and prepare yourself for the shock, since I know how much you think math is always the answer to historical questions!) the truth will only be iterated towards from careful linguistic research and debating each minutiae. A radical view, I know :)

          • http://www.patheos.com/blogs/exploringourmatrix/ James F. McGrath

            You mean we have to rely on human cognition, research, and at times intuition?! Oh, how very disappointing – my next suggestion was going to be to feed theological systems through computer analysis to determine which if any are at least internally coherent… 😉

  • arcseconds

    Mark Goodacre’s blog NT Blog is now hosted on Blogspot, so the URL no longer directs there.

    The post in question is here.

    • http://www.patheos.com/blogs/exploringourmatrix/ James F. McGrath

      Thank you!