Dr. Richard Carrier is an ancient historian who is best known for championing the idea that Jesus of Nazareth is a mythical figure — a view which to call it fringe would be to pay it an undue complement. I debated Dr. Carrier a couple of years back on Premier Christian Radio, and you can listen to the debate here. Carrier recently posted an article on his website discussing and critiquing the argument from undesigned coincidences, an argument which I have promoted in my articles, debates, and public lectures. Carrier specifically mentions Dr. Lydia McGrew’s book, Hidden in Plain View — Undesigned Coincidences in the Gospels and Acts, which develops on the argument originally coined by William Paley and by his successor John James Blunt.
Unfortunately, Carrier’s review only demonstrates that he has not read McGrew’s book (since he claims McGrew uses examples which she does not in fact use — in fact, she states explicitly in her book her reasons for not using them). He also never cites any of McGrew’s blog posts where she deals in detail with many of the points he raises in his article. He also gets some of his facts simply wrong. With this blog post, I begin a series of responses to Carrier’s critique of undesigned coincidences
Carrier begins his article by defining the concept of undesigned coincidences. He writes,
The gist is that we can find places in the Gospels where one author seems to “know” things in another Gospel that that Gospel only alludes to, and this proves there was a real story both authors are writing down.
This isn’t a very clear definition. There are different forms of undesigned coincidences, but the most classic form (which Carrier is attempting to define) is when you have multiple (at least two) accounts that report an event where one account answers in passing a natural question raised incidentally by the other. Contrary to what Carrier asserts, the argument has never been that such features prove there was a real story. Rather, such features are evidence that a real event lies behind the reports found in the gospels — that is to say, the presence of an undesigned coincidence is more probable given the hypothesis of historicism than given the annulment of that hypothesis. Some undesigned coincidences contribute greater evidence than others. Carrier’s lack of nuance in his description of the basic argument does not get his article off to the best start.
Carrier illustrates undesigned coincidences by providing an example:
For example, John has Jesus suddenly wash the disciples’ feet, while his source, Luke, doesn’t mention that but mentions an argument among the disciples that could explain the washing incident, so you can make the scene “make more sense” if you combine the accounts. So Luke and John must have been there! The Gospels are therefore direct eyewitness accounts.
Again, this summary lacks nuance. The conclusion of the argument is not that “Luke and John must have been there.” Rather, it is that the account is grounded in an actual event that really took place. No credible historian thinks that Luke was present at the last supper (in fact, Luke indicates in the prologue of his gospel, in Luke 1:1-4, that he was not an eyewitness of Jesus’ ministry but received information handed down by eyewitnesses). However, undesigned coincidences do provide evidence, at the very least, that the four evangelists, if they were not eyewitnesses themselves, had access to reliable sources concerning the events that they narrate in their gospels — especially in cases which confirm very particular details. Considered cumulatively, the argument from numerous undesigned coincidences in the gospels, together with other categories of evidence, provides an inductive basis for supposing the gospels, taken as whole documents, to be substantially reliable reports of the life and times of Jesus of Nazareth.
Carrier summarizes the history of the argument from undesigned coincidences as follows:
This was an old preacher’s argument from over a hundred years ago (e.g. Undesigned Coincidences in the Writings Both of the Old and New Testament by the Reverend J.J. Blunt, published in 1853, building on an argument from William Paley from the previous century). Which never caught on because objective scholars had already started concluding the Gospel authors know things about other Gospels because they read them, and they changed some of those things because they wanted to.
Every scholar who has made use of the argument from undesigned coincidences (including Dr. Lydia McGrew, her husband Dr. Tim McGrew, and myself) is well aware of the so-called synoptic problem (I prefer to call it the synoptic puzzle, since calling it a “problem” suggests that it presents some sort of challenge to the historicity of the gospels). All scholars are agreed that there is a literary dependence between the synoptic gospels (Matthew, Mark and Luke), with the majority view being the so-called two source hypothesis, which is to say that Mark was the earliest gospel written, and Matthew and Luke utilized Mark as a source. According to this view, when Matthew and Luke agree in wording but diverge from Mark, they are utilizing a hypothetical Q document (from the German word Quelle, meaning “source”) which is now lost to us. The Q source document has come under challenge by a significant number of scholars, such as Dr. Richard Goodacre (author of The Case Against Q). Many scholars, including Carrier, are therefore skeptical of the existence of such a hypothetical Q document. I share this skepticism. Scholars debate about the extent to which John utilized the synoptic gospels as a source for his own gospel, contrary to Carrier’s claim in his article that Johanine dependence on the synoptics is “the conclusion of most leading experts on that Gospel today”.
Even where there is literary dependence between gospel accounts, however, the argument from undesigned coincidences suggests that in many instances Matthew and Luke had access to reliable, independent information of their own. McGrew herself makes this point in the introduction to chapter 3 of her book. The point is that the gospel accounts fit together in casual ways, that are best explained if the recounted events are rooted in truth, rather than one gospel author taking his cues from another. Carrier is confusing possibility with probability. We will review examples to illustrate this as this article series progresses.
This doesn’t even get us to a source much less a witness. John invented the washing incident to comment on Luke. In fact, to “fix” Luke into a better story with better messaging. John didn’t like Luke’s story. So he replaced it with a different one. As John did a lot (see Chapter 10.7 in On the Historicity of Jesus). John’s new story doesn’t even require knowing anything about Luke’s. And Luke clearly knew nothing about John’s story either. It hadn’t been invented yet.
Here, Carrier has simply missed the point. In Luke 22:24-27, we are told of a dispute that arose among the disciples at the last supper regarding who among them was to be regarded as the greatest. John’s gospel also reports the same event, in John 13:1-20, except John doesn’t tell us about this dispute that arose among the disciples. Instead, he tells us about an object lesson that Jesus gave the disciples in servant hood, specifically, washing the disciples’ feet. Luke thus explains John, even though Luke doesn’t tell us about the object lesson in servanthood, and even though John doesn’t tell us about the occasion that gave rise to this object lesson, namely, the dispute among the disciples over who was the greatest. Furthermore, John’s account also explains Luke’s account, where Jesus says, “For who is the greater, one who reclines at table or one who serves? Is it not the one who reclines at table? But I am among you as the one who serves” (Luke 22:22). Only by putting the two accounts together do we have a complete picture. This sort of pattern is clearly best explained by the two accounts being grounded in a real event.
Carrier comments that “John’s new story doesn’t even require knowing anything about Luke’s. And Luke clearly knew nothing about John’s story either. It hadn’t been invented yet.” This is a very odd remark, since the whole point is that John and Luke are reporting the event from independent sources, for the reasons detailed above.
But even such a theory as that only gets you to a source. Not that source being true. The source they are all redacting differently could simply be an Ur-Gospel just as late and fictional as its many subsequent redactions. Rather than “an actual event” witnesses remembered differently, it could simply be “a master story” authors are copying from differently.
Carrier later adds,
Reader Chris Sandoval gives examples of the point. First he notes the Ur-Gospel alternative: “an evangelist may have intentionally omitted minor details knowing that his Christian audience already knew the story well enough to fill in the gap,” because they heard it repeated so often. No eyewitness source is needed for this. Just oral lore that could be as fabricated or exaggerated and embellished as any other tale, by decades of creative storytellers. McGrew literally has no argument to offer against this alternative explanation of every single example in her book.
I am disappointed, but not particularly surprised, that Carrier chose not to cite or interact with McGrew’s blog articles (e.g. here and here) dealing with this very objection. I will make a few observations here in response to Carrier’s point. The argument is normally articulated in terms of the oral tradition that many scholars believe lies behind the gospel accounts (although this in principle could be written sources alternatively). The objection is that undesigned coincidences could be just as well explained by the gospel authors incompletely remembering the oral tradition, or incompletely copying from a common ur-source that contained both parts of the coincidence. However, there is no independent evidence for such an ur-source, either in terms of documentary evidence or written testimony to its existence. It is therefore ad hoc — that is, invoked simply for the purpose of avoiding the most obvious explanation that the gospel documents are grounded in true history. The oral tradition alternative has its own problems (which also apply to the ur-source hypothesis) which I will now discuss.
For one thing, there is no evidence for the type of oral tradition of gospel stories that would be needed to adequately explain most undesigned coincidences. Consider a well known example of an undesigned coincidence relating the involvement of Philip in the account of the feeding of the five thousand in John 6:5, which is explained by Philip’s hometown being in Bethsaida (John 12:21) and the feeding of the five thousand taking place in Bethsaida (Luke 9:10) (if you are not familiar with this example, I cover it in this article). It seems very unlikely that Christians, across geographical areas, would have known the hometown of Philip as being in Bethsaida. Indeed, such an assumption would lead us to conclude that virtually any randomly selected adult Christian residing in Pisidian Antioch could have listed the home towns of Bartholomew, Judas, not Iscariot, and every other member of the twelve. Indeed, on this hypothesis, it appears that one would have to suggest that almost any detailed piece of information contained in the gospels was widely known in the Christian community and widely known extremely early, before the gospels were even written. This seems to me very unlikely.
Dr. Michael Bird, in his book The Gospel of the Lord: How the Early Church Wrote the Story of Jesus, puts it like this (p. 185):
I regard the evidence surveyed as constituting moderate grounds for identifying a conserving force in the transmission of the Jesus tradition, since the gaps in our knowledge are too vast to assert otherwise. At the end of the day most of what is said about the formation of the Jesus tradition is based on a priori assumptions, circumstantial evidence, inference, hypothesis, analogy, conjecture, and sheer guesswork.
A “conserving force” hardly means that everybody knew what town Philip was from already, long before John mentioned it.
Moreover, Luke himself seems to have regarded Theophilus as gaining information from his gospel. Based upon Luke’s introduction itself, we would infer that Theophilus apparently didn’t already know all of the things found in Luke’s Gospel. Luke wrote it to give him additional information. This also seems to be implied by John 20:30-31.
There are also many undesigned coincidences involving multiple different incidents. An example of this is Luke’s listing of the women who followed Jesus from Galilee, including Chuza the wife of Herod’s household manager (Luke 8:1-3) explaining Matthew’s report of Herod speaking to his servants about Jesus, presumably in the privacy of his own palace, in Matthew 14:1-2 (again, see my article here for a discussion of this example). On the ur-source hypothesis, the ur-source would presumably have to include both stories, yet oral stories would typically be stories of a given pericope. If there were multiple circulating traditions containing between them both of the relevant stories, then the undesigned coincidence would still exist between the multiple circulating traditions, and the problem the hypothesis sought to address still remains.
Carrier’s hypothesis also fails to explain the coincidence involving John’s mention of Jesus’ approach into Bethany happening six days before Passover, a detail extraneous to John’s gospel (again, for the details, see my earlier article here). Notice that Mark 11 telescopes the account such that it is masked that Jesus in fact arrived in Bethany the evening before the triumphal entry into Jerusalem and only the morning after sent his disciples to fetch the colt. The passages relevant to the coincidence also span a few chapters in Mark (from Mark 11:1 to 14:1). Furthermore, the setting of the olivet discourse in Mark 13 in the evening is not explicitly stated, but may be inferred from the fact that the mount of olives is midway between Jerusalem (where Jesus had been all day) and Bethany (where Jesus’ accommodation was for the night). In view of all of the above considerations, the view that this coincidence is explicable by some sort of ur-source theory seems wildly implausible.
Of course, such theories are even more wildly implausible when it comes to accounting for undesigned coincidences between Acts and the epistles of Paul. For further discussion of the ur-source theory, I refer readers to McGrew’s articles on the topic linked above.
Thus Blunt’s argument was really quite illogical. So, too, McGrew’s resurrection of it. Which is why mainstream scholars ignore her.
This isn’t true. Endorsements for McGrew’s book (which you can find at the beginning of the book) include relevantly-qualified scholars such as:
For example, that same explanation doesn’t even require an Ur-Gospel or oral lore. A later Gospel author could simply assume (reasonably or erroneously) that his audience is familiar with the Gospel he is redacting, and thus be building on knowledge he assumes they already have—or that he forgets they don’t, because he has it: in the Gospel he’s redacting. We see this in the “Long Ending of Mark,” whose author assumes the reader has read all four Gospels and thus knows the full story behind every reference to them in it (see my discussion in Chapter 26 of Hitler Homer Bible Christ).