Why Do We Still Believe in Q?

Q is a hypothetical Gospel. The letter Q stands for Quelle, German for “Source.” It has long been a popular hypothesis, and it serves a lot of agendas well. But there is no evidence whatever for it. Q never existed. And there is no rational reason to believe it did. Instead, it is defended with a panoply of lies and blatant violations of logic. A lot like the historicity of Jesus.

Q is one of the things that shocked me when I started my postdoc research on the historicity of Jesus (in pursuance to a fan-generated research grant). I had always assumed everything scholars told me about Q was true. It was just a sayings collection. Luke never followed Matthew instead of Mark. It didn’t contain a crucifixion narrative. All lies. And it was the shock at having been lied to by so many experts and so much of the reference literature that led me to realize you can’t trust experts or the consensus in this field. You always have to double check. Because there are a shit ton of lies in this field.

Even many experts don’t know how awash with lies this field is. Too often, they just naively trust what they are taught in schools. An example is Mark Goodacre having been duped, by someone at some point in his career, into believing Paul wrote that he learned the gospel “from those in Christ before” him. On radio I told him no, no such verse exists (in fact, Paul says exactly the opposite, and so assertively that he even swears by it). Goodacre refused to believe me. Until, on a commercial break, we had a chance to try and find the verse he was sure was there. But it wasn’t.

That’s the lesson I learned from Q. Never trust anyone in this field. Never trust the consensus in this field. Until you have double-checked its validity by some dependable means. Not because they are all liars. But because there are so many lies no one has checked the truth of, so many fallacies no one has vetted the logic of. They instead just believe what they were taught. And keep repeating it. Or just keep citing “the consensus,” without ever actually checking if that consensus rests on sound facts or valid logic.

Q is one more product of that hall of fallacies and lies.

What Is Q?

We all know Matthew copied Mark verbatim (without telling anyone; and yes, that was considered dishonest, even then). Then he snuck in a bunch of changes and additions. We also know Luke copied Mark verbatim (also without telling anyone who he was copying, when, or why). And then he also snuck in a bunch of changes and additions. That alone establishes that Matthew and Luke are really just redactions of Mark, expansions of the core Markan myth. The same way all other mythic narratives expand and grow over time. But in what Matthew and Luke added, there is a lot of material that’s certainly not from Mark, yet is verbatim identical between Matthew and Luke (as well as a lot of material that’s not identical but similar enough, changed up as needed, just as is the case with how Luke and Matthew used Mark).

By analogy with how Matthew and Luke used Mark, as early as 1800 it was hypothesized that the other material Matthew and Luke shared must also have been copied from a common source, just not Mark. And so that hypothetical (and conveniently lost) source was invented. A century later it acquired the name “Q.”

Wikipedia of course has an extensive entry. And yet it starts right out of the gate with a full-on lie in its very first sentence, describing Q as “a hypothetical written collection of Jesus’ sayings.” Nope. It was not a collection of sayings. Nor simply of Jesus. It included narrative material. Including material expanding narratives found in Mark. And it included narrative material not just of Jesus, but of John the Baptist as well, and things spoken by John, not just Jesus. (And that’s only at the very least; the evidence indicates much more, which I’ll get to later.) So if it existed at all, Q was in fact a full Gospel, not a mere collection of sayings. It even included material from the crucifixion narrative. A fact often denied. Another lie.

In fact all the evidence for Q is 100% consistent with Q being a redaction of Mark, one that added a bunch of material to Mark, expanding things in Mark that were too brief or unsuited for a later author’s tastes or needs. And that means Q sounds pretty much exactly like Matthew. In fact, it’s almost certainly Matthew. Q is literally the least likely hypothesis of any that’s plausible. It should be rotting now at the bottom of the barrel of bad ideas.

Lying for Q

The arguments “for” Q consist of two types: lies and fallacies.

The lies are the most appalling. That any consensus in biblical studies depends on lies is an indictment of the entire field. But even the fallacies call that field into question. And even those fallacies end up producing lies. Essential reading on why Q should be rejected include Mark Goodacre’s “A Monopoly on Marcan Priority?” and Michael Goulder’s “Is Q a Juggernaut?” For an extensive refutation of Q, see Goodacre’s The Case against Q and now (supporting that) Thomas and the Gospels. Further demonstrations by other experts are listed in On the Historicity of Jesus, p. 269 n. 33. But here’s a look at the top five arguments “for Q” and why they should be more embarrassing than respectable.

Lie Number 1: Q Did Not Contain Material in Mark

A common view (espoused by Bart Ehrman) is that everything Matthew and Luke share with Mark came from Mark; everything they share with each other came from Q; and everything they share with no one came from yet other lost hypothetical sources: called M, for the material unique to Matthew; and L, for the material unique to Luke.

This is 100% illogical.

Even if we illogically assume (against all evidence to the contrary) that none of these authors ever made anything up, and thus always “had a source,” since we don’t have Q, we have no way of claiming to know that M and L aren’t just material taken from Q by one rather than both subsequent authors. Just as Matthew and Luke each chose separately to omit material from Mark (e.g. Luke didn’t include Mark 6:45-8:26; Matthew didn’t include Mark 12:41-44; neither Matthew nor Luke included Mark 14:51-52; etc.), so we should expect they chose separately to omit material from Q (even assuming there was a Q). So the fact that something is in Matthew but not in Mark or Luke does not mean it came from M; likewise that something in Luke but not in Mark or Matthew came from L. It may very well have come from Q. And you cannot say otherwise.

But that’s not itself a problem for deciding whether Q existed. It just illustrates that many a consensus in biblical studies is not based on logic.

The real problem is this: how do we know Q did not contain verbatim material from Mark? In other words, how do we know Q wasn’t just an expansion of Mark?

We don’t. Period. Because we don’t have Q. So we don’t know if when Matthew and Luke copy from Mark, that that same material wasn’t also in Q. We don’t even know that when Matthew and Luke copy from Mark, that they were even copying from Mark. If that material was in Q, they may have only copied from Q. In fact, for all we know, they never even read Mark.

The only evidence we have that Luke, for example, was using Mark at all is that sometimes Luke repeats verbatim what’s in Mark, while Matthew changes it. But that only tells us that Luke did not copy Matthew. (And indeed it only tells us he didn’t copy Matthew at that specific point. Because guess what? Often Luke does copy Matthew’s redaction of Mark instead of Mark. Anyone who told you otherwise was lying, or lied to.) And that means they may have both only known of a Q that was an expanded Mark. The Q hypothesis thus entails Matthew and Luke may never have even seen or heard of the Gospel of Mark.

It’s a circular argument to claim that Q lacked Markan material “because” Q only contained the material shared by Luke and Matthew and not Mark, when the only way you can know Q only contained the material shared by Luke and Matthew and not Mark, is because you just arbitrarily defined it that way. This is so phenomenally illogical it should have been laughed out of the field a hundred years ago. And this means we cannot claim Q is an independent source from Mark. Which is the only function Q has ever had in biblical studies. And probably the only reason anyone wants it to be true. Which means claiming Q was written independently of Mark is another lie.

In any other not-fucked-up field of history, if we applied Ockham’s Razor—a valid logical principle—instead of this fallacy of circular argument adopted by all Q defenders, we’d get a different result. Because what is simpler? That Matthew and Luke used two sources one of whom we can only hypothesize the existence of? Or that Matthew used only one source (and made the rest up) and Luke only used Mark and Matthew (and made the rest up)? The latter theory requires no ad hoc hypothetical sources. It relies solely on evidence and texts we actually have. It is therefore the much simpler hypothesis (because the probability of all the facts it rests on is as near to 100% as makes all odds; which is not the case when we start depending on merely hypothesized facts, which for that very reason have a significantly higher probability of being false: see Proving History, index, “gerrymandering” and “Ockham’s Razor”). And on top of that, it turns out, unlike evidence for Q, there actually is concrete evidence Luke copied Matthew (as we’ll see). So we know he did. And that leaves nothing else to explain.

Q thus serves no more function than a Ptolemaic epicycle. Or the proverbial angels pushing the planets coincidentally in exactly the same way they’d move without them.

Q is fallaciously circular. Which means it’s explanatorily useless. Why, then, is it respectable?

Lie Number 2: Luke Can’t Have Used Matthew Because He Never Preferred Matthew to Mark

Christian fundamentalist Daniel Wallace provides the most convenient summary of the arguments any belief in Q still stands on.

And he lies.

Wallace gives six arguments, labeled A through F.

Right out of the gate, in defense of Q against the far simpler and more obvious alternative that Luke just used Matthew, Wallace quotes an earlier scholar, wholly unquestioningly and uncritically, as saying (in defense of argument A):

One of the strongest arguments against the use of Matthew by Luke is the fact that when Matthew has additional material in the triple tradition (‘Matthean additions to the narrative’), it is ‘never’ found in Luke.

That is a lie.

Even Wallace himself (a whole section later) admits that “there are scores of agreements between Matthew and Luke against Mark in terms of grammar and editing,” and he goes on to survey tons of them. Which are only a fraction of them. And he kind of hides the strongest examples. Particularly specific examples of Matthew expanding on or editing Mark, expansions or edits that Luke kept. Luke in fact contains many Matthaean additions to the triple tradition (meaning material common to Matthew, Mark, and Luke).

Quoting Mark Goodacre:

Luke prefers Matthew’s version to Mark’s in several triple tradition incidents: the whole John the Baptist complex (Matt. 3, Mark 1, Luke 3); the Temptation (Matt. 4.1-11 // Mark 1.12-13 // Luke 4.1-13), the Beelzebub Controversy (Matt. 12.22-30 // Mark 3.20-27 // Luke 11.14-23) and the Mustard Seed (Matt. 13.18-19 // Mark 4.30-32 // Luke 13.18-19) among them. On all of these occasions, the parallels between Matthew and Luke are more extensive than those between Mark and Luke.

I give an example myself in OHJ, p. 471:

For example, Mk 14.65a reads, ‘and some began to spit on him, and to cover his face, and to buffet him, and to say to him, “Prophesy!”’, which Mt. 26.67-68 expanded to ‘then did they spit in his face and buffet him: and some smote him with the palms of their hands, saying, “Prophesy unto us, Christ! Who is he that struck thee?”’ Luke 22.63-64 essentially combines Mark with Matthew, repeating the concluding text of Matthew verbatim: ‘and the men that held Jesus mocked him, and beat him; and they blindfolded him, and asked him, saying, “Prophesy! Who is he that struck thee?”’ Except for dropping ‘unto us, Christ’ to economize the passage, the Greek of Luke here is identical to that of Matthew (legontes, Prophēteuson [hēmin Christe]! Tis estin ho paisas se?).

That is far too improbable on any theory other than direct literary borrowing. And so:

Luke then combines this with Mark’s detail that they covered his eyes, which Matthew omitted (or rather altered, having them spit ‘in his face’ rather than cover ‘his face’). Luke thus combined Mark with Matthew, recast mostly but not entirely in his own words, to make what he deemed to be a better passage. That Luke knows the details Matthew added, and even borrows his exact words, is sufficient proof that Luke knew and used Matthew. Note that on no theory of Q is this element of the Passion Narrative a part of Q, so this cannot be explained by appealing to Q. Luke is using Matthew. And if here, so everywhere. There is simply no need of an imaginary Q.

And as I cite there, these agreements are well known in the literature. They are just ignored. For example, see R.T. Simpson’s article ‘The Major Agreements of Matthew and Luke against Mark’ in New Testament Studies 12 (1965–1966), pp. 283-84.

So this claim (Wallace’s argument A) is outright false.

Anyone who tells you Luke never agrees with Matthew against Mark (which agreement proves Luke’s knowledge of Matthew) is either a liar, or the dupe of a liar.

Lie Number 3: Luke Never Borrows Context or Order Unique to Matthew

Two more of Wallace’s arguments are that Luke “never” places “the common (double tradition) material in the same context as it appears in Matthew” and that there is a “complete lack of agreement in order whenever they disagree with Mark” (arguments B and E). In other words, that when Matthew and Luke would hypothetically be using Q or each other instead of Mark, they never put that material in the same place or order. This is supposed to indicate Luke didn’t know where Matthew put it. And therefore that Luke didn’t know Matthew.

But this is a lie.

Of course, even if it were true, it’s not really that strong an argument. Luke and Matthew frequently relocate and reorder material from Mark, so we should expect Luke would treat Matthew the same way. That he does so is therefore not that improbable. Of course, the argument is supposed at least to suggest that certainly if Luke was using Matthew, occasionally he would copy the same order or context invented by Matthew; therefore that he “never” does, is somewhat unlikely if he was in fact using Matthew. Even this weak bit of leverage, however, requires that he absolutely never did. And that’s simply false.

That Luke changes the context of material in Mark as well as Matthew is shown by the anointing scene in Mark 14, which Matthew 26 borrows from Mark nearly identically, yet it is completely relocated and altered in Luke 7. It’s an example of the non-verbatim redaction of Luke’s sources. It is certainly notable that these authors frequently copy their unnamed sources verbatim; but you should never assume that’s the only way they treated their sources. Rewriting them in their own words was also a way they adapted their source material (one more in line with accepted practice of the time). Comparable examples are what Luke does with the Nativity story and with John the Baptist material and the suicide of Judas. So that Luke does this same thing with Matthew is not evidence he didn’t use Matthew for that material, any more than it would be evidence he didn’t use Mark. Both Luke and Matthew did the same thing with Mark. So it was in fact a standard practice in Gospel composition, not something “inexplicable.”

Indeed the main reason almost all the material got changed in order and context is that Matthew and Luke chose completely different structures for their stories, not only from each other, but even from Mark. See how completely Matthew changes order and context even of the material he takes from Mark, and you already have a model for how Luke would change the order and context of the material he takes from Matthew. I show many details of how the Gospels’ artificial choices of structure molded the way they constructed their myths in On the Historicity of Jesus (e.g. pp. 412-13, 418-20, 460-67, 475-80). The result is that we should expect few parallels in order and context, even if Luke adapted material from Matthew. It is therefore not improbable to find exactly that.

And yet, even still:

It’s simply not true that Luke “never” copies Matthaean order or context. As I wrote in OHJ (p. 471):

Luke redacts Matthew’s Sermon on the Mount, conspicuously ‘reversing’ it into a Sermon on a Plain. How do we know Luke is redacting Matthew? Because both speeches are followed by the otherwise-unrelated narrative of Jesus healing the centurion’s son in Capernaum. The latter occurs in Mt. 8.5-13, the Sermon having ended at 8.1 (only a brief healing of a leper lies in between). Luke 6.17-49 redacts that Sermon, and then immediately in Lk. 7.1-10 the centurion’s servant is healed in Capernaum, the story in many respects identical, even down to specific words and phrases (such as revising the centurion’s ‘son’ into the centurion’s ‘boy’, which some translators render as ‘servant’, but in context this is obviously just a different way of saying ‘son’). Likewise, Mt. 4.23–5.1 precedes the Sermon with a general account of Jesus healing many, and Lk. 6.17-19 does the same thing. These parallels are very improbable unless Luke was following and redacting Matthew’s narrative sequence…

Luke 3 and Matthew 3 likewise both place the same new John the Baptist material in the same place and order, almost even verbatim in cases. Likewise Luke 4 and Matthew 4, the temptation scene (another narrative, on top of the baptism and temptation and the Centurion of Capernaum, further exposing the lie that Q was a “sayings” source).

So Wallace’s arguments B and E are also outright false. Luke sometimes does put material in the same context or order as Matthew. Which is actually evidence he used Matthew. By lying about this, evidence that Luke used Matthew is conveniently “converted” into evidence he didn’t. Welcome to Christian apologetics. Ever masquerading as honest scholarship.

Lie Number 4: We Know Luke Would Never Do That

Wallace’s arguments C and D—that if Luke used Matthew, then Luke would never rearrange or simplify Matthew—consist of essentially complaining that Wallace doesn’t know why Luke would have made the choices he did, because Wallace would make different choices. Which certainly proves Wallace and Luke are not the same person. But it proves little else of any interest.

This is actually one of the most common fallacies deployed in defense of Q. There are countless arguments that consist of no actual evidence, not even evidence for the base rates of authorial practices, but simply a repeated armchair argument from incredulity, “But why would Luke do that!?” Really, that a scholar hasn’t thought of why Luke made the decisions he did is not evidence of anything but that scholar’s ignorance or lack of imagination.

There are actually very good reasons why Luke would do every single thing he did. And most scholars simply have baseless assumptions in that regard, for example assuming Luke would have “respected” his sources too much to leave good stuff out or compose his own material for the same purpose, when all the evidence refutes any such fancy: Luke often ignores suitable material in Mark and resorts to his own contrivances instead. We should expect him to do the same thing with Matthew. But even if we couldn’t think of any reasons for Luke’s choices, that still only tells us we don’t know what Luke was thinking. It doesn’t tell us Luke wasn’t redacting Matthew. The evidence that he was is simply far too concrete to conclude otherwise. Everything anyone complains about, regarding how Luke would supposedly have “ruined” Matthew had he done what he did with Matthew, is pretty much the exact same stuff Luke did with Mark. So we already know that’s how Luke treats his sources. It therefore is not evidence against his using those sources.

For example, Wallace laments how Luke ruined the beautiful structure of Matthew’s Sermon on the Mount (and I concur it’s an elegant literary masterpiece: OHJ, pp. 466-67). He concludes Luke surely would never do that, therefore he can’t have known of it. But Luke also “ruined” the equally beautiful structure of Mark’s sea narrative (OHJ, pp. 412-13). Yet Wallace doesn’t then conclude Luke must not have known the text of Mark! The correct conclusion from the evidence is exactly the contrary: this is simply what Luke does with his sources. Even as a matter of Bayesian probability, even if both hypotheses start out equally likely, both produce exactly the same expected outcomes in the evidence, to the same probability. So there is no evidence here. And both hypotheses do not start out equally likely. The evidence otherwise of Luke using Matthew is too vast for coincidence to explain it.

In this case, just as Matthew changed Mark’s elegant structure into his own completely different structure, and inserted five “long discourses” to emulate the Pentateuch of Moses, Luke changed Matthew’s elegant structure into his own completely different structure, and inserted his own innovations in place of Matthew’s. Luke is no more “ruining” Matthew than Matthew was ruining Mark. Luke clearly did not like Matthew’s five long discourse structure. He also didn’t like leaning so much on Moses. Instead, Luke focuses his literary model for Jesus mostly on Elijah and Elisha, reverses almost every message of Matthew’s Nativity narrative (erasing many of the Moses parallels), and spreads the sayings of Jesus more widely across a more engaging narrative, and gets rid of an excess redundancy among them, precisely to avoid ever breaking the action with too long and boring a soliloquy, or being too repetitive. Luke clearly had a very different idea of what is clunky and what is elegant. One just as valid as Matthew’s. There is nothing inexplicable about this at all. It’s actually typical: authors have different aesthetics, different aims. And comparing how Luke changed Mark with how he changed Matthew only confirms the point.

The bottom line is, whether a reading is more primitive or less elegant is a wholly arbitrary aesthetic invention of modern scholars. One that’s again quite circular: it is just assumed Luke always preserves the original form of something, that he wouldn’t simplify passages he thought needlessly verbose (even though all ancient authors often did exactly that; and we can confirm so did Luke), that he wouldn’t break up some material and scatter it into different scenes (even though all ancient authors often did exactly that; and we can confirm so did Luke). But if you don’t circularly assume the conclusion in your premises, it is no longer a given that Luke would never do this. To the contrary, it’s just as possible Luke would do exactly those things, just as every other author did. And when we look at test cases, like how Luke transformed Mark, it’s more than merely possible. It’s confirmed.

Q can only be defended by ignoring these facts. And by replacing facts with the anachronistic aesthetic opinions of modern ivory tower wonks.

There is no genuinely discernible difference between Matthew having embellished a Q statement and Luke having simplified a Matthaean statement. Accordingly, at no point does any scholar “know” one happened and not the other. Therefore, there is no C or D evidence. In the sense of actually existing evidence. It’s just the imaginations of scholars who think they magically know these things when in fact they never do.

And claiming to know something you don’t is lying. The more so when the evidence we have already demonstrates the contrary of what you claim to “know.”

Lie Number 5: We Know Luke Would Have Used More of Matthew

As Wallace argues, “the fact that Luke lacks the ‘M’ material (material unique to Matthew) and, conversely, the fact that Matthew lacks the ‘L’ material, argues that neither knew the other.” You might immediately notice that’s another fallacy of circular argument. We should always expect there will be material in Luke not in Matthew and material in Matthew not in Luke, because we see that’s exactly what happened with Mark: both Luke and Matthew left stuff out; yet still unmistakably used Mark. Therefore, that being the case can never be evidence against Luke using Matthew.

Wallace is aware this is circular reasoning. So he tries to escape that by changing the argument into another “but that’s not what I would have done, therefore Luke wouldn’t have” fallacy:

It should be carefully noted that this is not circular reasoning, though on the face of it it seems to be. As soon as we define ‘M’ as material unique to Matthew, then of course Luke would lack it! But that is not the real point of this argument. Rather, it is that there is so much material in Matthew—and rich material at that—which would in all probability have been utilized by Luke had he known of it, that for him not to have used it strongly suggests that he did not know of its existence.

There is exactly zero evidence Luke would have used any of the material he excluded from Matthew “had he known of it.” We should expect from the evidence of all authors throughout the entirety of human history, including Luke’s contemporaries, that Luke would always have excluded what he thought redundant, unnecessary, or excessive, and things less important to him that took up space he wanted to use for his own material (there being a limited standard size for a scroll; so far as we know the Gospels were not condexed until the next century), and, of course, he would have excluded what he outright didn’t like or couldn’t find a way to fit in. The effect of all that would be exactly the same as what we observe: Luke leaving out a ton of material Wallace loves. But even without any other evidence Luke used Matthew (of which actually there is a lot), it’s already just as likely that Luke didn’t love the material he excluded, as that Luke didn’t know of it. Wallace is not Luke. So what Wallace thinks is rich and wonderful does not constitute knowledge of what Luke thought was. Luke left a lot out of Mark, too, including what’s even called the “Great Omission” (Mark 6:45-8:26). Is that evidence Luke didn’t know Mark? No. Nor is it in the case of Matthew.

Even if we can say there is something in Matthew that “agrees” with something we can see from his Gospel Luke was selling, it is yet another circular argument to insist Luke would have used Matthew to do that, when the only premise we are basing this on, is the fact that Luke liked his own way of doing it better. In every other field of history and literary studies, the conclusion would not be to go against all the other evidence and insist Luke didn’t know Matthew; it would be to conclude Luke obviously felt he knew a better way of making his point. Parallels can be seen in redactions and reimaginings of Shakespeare’s plays: we would never say that because Arthur Laurents left out so many elegant and beautiful bits of Romeo & Juliet in his Westside Story, things so well suited to the latter’s message, that obviously he knew nothing of Shakespeare’s play and must only have had its plot-line on garbled hearsay. No. Trust me. Arthur Laurents read the play. And yes. He thought he could do better than Shakespeare.

Looking at It the Other Way Around

Wikipedia gets one thing right:

Two documents, both correcting Mark’s language, adding birth narratives and a resurrection epilogue, and adding a large amount of “sayings material”, are likely to resemble each other, rather than to have such similar scope by coincidence.

And Matthew and Luke don’t just coincidentally add birth and resurrection narratives with many peculiar parallels (and even some exact verbatim phrases), they also add narratives of Judas’s suicide that are peculiarly similar to each other, and John the Baptist and Temptation narratives that are often verbatim identical, and they both oddly feel a need to include a lengthy genealogy for a man from whom Jesus doesn’t even descend. And even in cases where Luke disagrees with Matthew, such as in his rewrite of the Nativity, it is conspicuously to reverse what Matthew said, as if presciently knowing what Matthew had said (see OHJ, pp. 470-73).

The most obvious explanation of all these strange coincidences is either that Luke is simply freely redacting Matthew (and often being happier with the Gentile-friendly sequence and content of Mark than the overly-Jewish content of Matthew) or Luke and Matthew were both using a “Q” that included genealogies, birth narratives, baptism narratives, temptation narratives, resurrection narratives, suicide narratives, and even a crucifixion narrative (as noted above). But that sounds so much like Matthew, that Ockham’s Razor leads to no more probable conclusion: Luke is simply redacting select content of Matthew into his redaction of Mark. There is no actual evidence for any other hypothesis.

This even explains why most of the content of “Q” consists of “sayings.” Because what we mean by Q is actually just Matthew, and most of what Matthew did to Mark was add five enormous speeches, which are in fact coherent literary products, in Greek, reliant on the Septuagint (a popular Greek translation of the Scriptures). They are not random collections of sayings. They are, like almost all speeches in all stories and histories of antiquity, the inventions of the author: Matthew. And Luke simply didn’t like that literary model. Long, ponderous speeches breaking up the action; and the heavy-handed Moses parallel in having five of them. So he broke them up, changed them up, dropped what he didn’t like or need, and used the rest as he wished. But he also borrowed and adapted lots of other stuff Matthew added, stuff about John the Baptist, the birth of Jesus, the suicide of Judas, and more. Some he took verbatim. Some he rewrote. Just as he did with Mark.

Since the above facts are then added to the even more incredible fact that there are hundreds of cases where Luke follows Matthew’s version of Mark and not Mark directly (in phrasing, grammar, vocabulary, and other elements), the notion that Luke did not use Matthew as a source is absurd. And if Luke used Matthew, there is no evidence left for Q, and no remaining justification for supposing there to have been one.

In future I’ll add here any other arguments for Q anyone thinks I’ve overlooked (anyone may post such to the comments field here, without being a patron, as long as you cite or give the evidence the argument draws its conclusion from). But I’m betting they’ll already fall into one of the five categories I’ve just dispatched as false or ridiculous.


Why is Q still believed, when it can only be defended with fallacies and lies? Why, indeed, is someone normally a skeptic like Bart Ehrman an obsessively fanatical defender of Q? Because of laziness, prestige-management, and stubbornness, perhaps. But it also serves agendas.

Christians like Q because it allows them to claim there were traditions earlier than Mark (even though the Q hypothesis entails nothing of the kind; even if Q existed, there is exactly zero evidence that it predated Mark). And they like it because the alternative entails Luke and Matthew are bald-faced liars, and Christians, um…cannot have that. They’d chop off their own fingers before allowing that to be true.

Non-Christians like it because it allows them to build careers out of guessing what was in Q, who wrote it, why, what their agenda was, and all manner of mythical nonsense that’s the equivalent of speculating on the precise shape of Saint Peter’s nose. In other words, secular bible scholars are essentially playing at being Vroomfondel and Majikthise.

Q is also a woobie that salves cognitive dissonance. The evidence for Jesus is so absolutely terrible, so egregiously fabricated, that historians are painfully grappling with the desperate need for there to have been more and better evidence than there really is. So they love to invent evidence whenever they can. It feels less scary. That’s why they just can’t let go of Josephus, despite all the evidence against Josephus ever having mentioned this Jesus. And that’s why they need Q. It’s the same reason they completely made up M and L, texts for which there is no evidence at all. Because then they can calm down and rest assured Jesus existed, because they “have” “multiple sources,” and therefore there is actually something to study about Jesus and this hasn’t been one fantastic lie all along entailing their entire degrees and careers were a total waste of money and time.

Evidence for this hypothesis is that defenders of Q so readily lie in defense of it, and so readily swallow obvious fallacies to believe it. But even clearer evidence is that they even outright tell the public we “have” these non-existent sources and therefore “know” Jesus existed. Their motive for believing this myth is included in their very assertion of the myth itself.

And this is a perfect model for the historicity of Jesus. All the same reasons Q is believed in, even by secular scholars, when in fact no rational person who believes in evidence-based reasoning should believe in it, are the reasons a historical Jesus is still believed in, even though the evidence for that Jesus is almost as lousy as the evidence for Q. Understand why so many experts believe in Q, and you will understand why they can’t tolerate any criticism of the historicity of Jesus, either. They either lie in defense of Q, or swallow lies in defense of it. They either lie in defense of Jesus, or swallow lies in defense of it. It’s the same phenomenon, the same field.


    1. That requires an enormous and vexatious outlay of time fighting trolls and demagogues to get every change approved. I gave up editing Wikipedia years ago for that reason.

  1. Justin Legault April 27, 2017, 8:39 am

    If the premise of the argument for Q is speculative and hypothetical. Why would this be considered a valid argument? Conjectures are not facts.

    “Well we ‘may’ have a source that was used, we don’t have this ‘source’ either, but it ‘could’ be the reason there’s so many discrepancies within Gospels”.

    I feel like all it is, is believers, trying to reconcile the irreconcilable. Granted if we find evidence you adjust your views accordingly. But you don’t speculate pretending like it has any verisimilitude.

    Unless they (Q supporters) hope that their listeners don’t know anything about it, thus misleading the narrative.

    1. I’ve been asked whether there are any parallels in any other fields of history, where historians rely on “hypothetical” lost sources like Q. I can’t think of any. But if anyone has examples, I very much would love to know and research and compare them.

      The closest thing I can think of is the Documentary hypothesis for the Pentateuch and similar hypotheses (e.g. the Deutero-Isaiah theory). But those don’t operate the same way, so they aren’t really an apt analogy.

      Likewise, we have examples such as Arrian or Tacitus where hypotheses develop as to which of their sources certain sections of their books came from, but there the sources aren’t hypothetical: we don’t have them, but the authors outright tell us they used them. And no one pretends we have the exact wording of those sources, nor is overly confident even in how they were used. So the analogy is again not apt.

      1. Justin Legault April 28, 2017, 8:22 am

        As for Q; none of the Gospels really say they’re using another source (Q) right? What you’re saying is even if they did. We don’t have this ‘source’ nor do we know if it’s actually reliable. So speculation in, speculation out.

        1. None say they are using any sources, except John, who fabricates a non-existent source that isn’t Q (the claimed pre-text of the Beloved, i.e. Lazarus) and Luke, who doesn’t name any of the sources he claims to have used (but does claim to have used more than one written source—which we already know to be Mark and Matthew).

        2. Justin Legault May 9, 2017, 1:29 pm

          Yeah I remember the Lazarus literal interpretation of Luke’s parable version.

          As for Luke, if he’s a so called “historian” he doesn’t write like one if he doesn’t name his sources and does any due diligence. Even if he didn’t use Matt or Mark, the fact that he says he’s using sources but doesn’t name them, it’s still speculative.

          He could even be lying as we know he does (Acts). “A liar who says he has sources but doesn’t name them or demonstrate that they are even reliable” Sounds perfectly fine to me! 🙂 No issues!

  2. Andrew April 27, 2017, 12:50 pm

    You answered the question succinctly in your conclusion: Bible scholars can never, never, never speculate for one second that the gospel writers were creative authors. There MUST always be a “source” for everything they wrote, either “oral history,” a lost document, or whatever. Because once you begin to speculate that Matthew or Luke simply invented a scene or saying in their gospels, that may ultimately lead to the conclusion that everything in the gospels is invented by the authors, and that simply is not an allowable thought in the Jesus Academy, secular or non-secular. There is the great, unspoken anxiety that their religion is based on a myth, and if this line of thinking is actually pursued with rigorous inquiry and honest self-examination, the entire structure will crumble like a house of cards. They cannot let that happen. Thus the fantasy of “Q” is perpetuated, semester after semester, year after year, decade after decade.

    “Q” also allows New Testament scholars to pretend that they have their own “documentary hypothesis,” supposedly going “behind the texts” to reconstruct the ur-text that only they, in their wisdom and analytical power, can detect. That feeling can be powerfully intoxicating to any grad student or wanna-be preacher out there.

  3. Giuseppe Ferri April 27, 2017, 12:55 pm


    Only a curiosity: recently, some scholars (Matthias Klinghardt and Markus Vinzent) are doing a case that proto-Luke (Mcn) is the earliest Gospel (the same Gospel used – or even written – by Marcion).

    Are you open to the idea or do you exclude it a priori as you think that Mark is probably the oldest gospel?

    I listened by someone that you are reluctant to move the creation of the first gospel far beyond the middle of the second century, because otherwise you would have the problem of justifying the rapid expansion of the historicist belief in such a short time.

    How do you reply to this criticism, assuming for sake of discussion that the earliest Gospel was really Mcn ?

    Very thanks

    1. It’s very unlikely for a number of reasons. I’ve never seen any convincing case for that. So far as I’ve seen, they all depend on large systems of unevidenced assumptions. Exactly the wrong way to reach conclusions about history.

  4. David Oliver Smith April 27, 2017, 2:41 pm


    Extremely interesting and enlightening.

    I wrote an article which did not get published in which I performed rank order correlations of triple tradition passages between Mark and Matthew and Mark and Luke. I also did correlations of double tradition (DT) passages between Matthew and Luke. In the triple tradition I separated the passages into two groups – narrative and sayings. I then noted the order of the passages in each gospel. I found that in the total triple tradition Matthew and Mark have a correlation of .87, a very high correlation 1.0 being perfect. With regard to narrative passages Matthew and Mark have a .99 correlation, and with sayings passages Matthew and Mark have a .63 correlation. Most of the reason for the less than perfect correlation in the sayings material is Matthew’s moving Markan sayings material to the Sermon on the Mount.

    Luke and Mark have a .92 overall triple tradition correlation, with a .97 narrative correlation and a .77 sayings material correlation. The higher correlation than Matthew-Mark in the sayings material comes from the fact that Luke’s Sermon on the Plain is shorter than Matthew’s Sermon on the Mount using fewer triple tradition passages.

    Now for the upshot. In the DT Luke and Matthew have a .62 correlation and most of the reason for a less than perfect correlation is that Matthew has 25 DT passages in the Sermon on the Mount and Luke only has 13 in his Sermon on the Plain. I eliminated the 12 DT passages that were not common to the sermons and the resulting correlation between all other DT passages was .84.

    This is a very high positive correlation. One would expect a positive correlation because in general events in all three synoptics occur in relatively the same order, but to get a .84 positive correlation of DT, mostly sayings, material that is easily relocated with respect to context, is good evidence that either Luke copied Matthew or Matthew copied Luke. In other words, there was no Q.

    David Oliver Smith

  5. The reason Q is a good hypothesis is because there are verses in Luke that are more primitive than how they appear in Matthew. Look at the Lord’s Prayer. Why would Luke delete “in Heaven”, “your will be done”, and “deliver us from evil”? Why is it that almost every addition made by Matthew common to Luke is moved to another place in Luke? This is not saying “I would have edited Matthew differently”. This is “No one would edit a manuscript in this way”. There are dozens if not hundreds of examples of text being moved around for no reason.

    Occam’s Razor is thrown around a lot by fans of Farrer but it really does not apply. The ways the gospels were edited were unarguably complex. Period. Trying to cut out hypothetical sources does not make a theory any simper. It just delegates the complexity to the editor’s mind. Why did Luke move hundreds of things around, making his job much harder for no discernible reason? Who knows! No answer necessary! Verse for verse, Q has more explaining power than Farrer.

    And saying that some Biblical scholars use it to posit beliefs back to Jesus is a terrible argument. You know full well that Doherty, Price and the vast majority of mythicists also believe in Q, so why imply Q only exists to further the historicist agenda? I could just as legitimately argue that Farrer exists only to further the Literalist agenda of dismissing all non-canonical sources.

    Another very important thing you missed is that most people believe Q was layered, just like the Gospel of Thomas is. Yes, verses of shared narrative besides sayings are sometimes delegated to the THIRD layer of Q, but it is always assumed that the Sayings Source was originally sayings only. Personally, I accept Q1 and Q2 *might* have come from the same source, but I tend to think that Q3 is another unnecessary attempt at Occam’s Razor. It’s better explained as coming from a different hypothetical source, although something much smaller than Matthew. The fact that scholars used Q to predict correctly that there were Sayings gospels before the Gospel of Thomas was discovered should count for something.

    If you look at Delbert Burkett’s two books on “Rethinking the Gospels”, you will see amazingly well-written technical arguments for why each gospel must have been written using at least 5 different sources each. Neither Matthew nor Luke used canonical Mark but rather, earlier sources also used by Mark. The switching between the sources is what causes all the jumping around and Burkett is able to trace them back by showing how the evangelists tended to go back to where they left off each time they switched back to another source.

    For example, Mark 9:17-27 conflates two different demon possession stories: one is a story of an “unclean spirit” that revolves around the faith of the disciples that Matthew and Luke have, and one is a story of an “unspeaking and mute spirit” that revolves around the faith of the father not known to Matthew or Luke. If Matthew and Luke were just copying from our Mark and/or our Matthew, then that would mean one or both of them would have had to have recognized that the story was a conflation, cut one out, and then ignored the second one.

    Another example provided by Griesbach theorist Thomas Longstaff shows Mark conflating a source common with Matthew and a source common with Luke:

    Mark 14:12a Matthew
    Mark 14:12b Luke
    Mark 14:12c Matthew
    Mark 14:13a Luke
    Mark 14:13b Matthew
    Mark 14:13c-16 Luke
    Mark 14:17-21 Matthew

    There are a lot of these miniature conflations, like Mark 1:32 conflating “When evening came (Matt. 8:16), when the sun set (Luke 4:40)” or Mark 3:8 conflating “across the Jordan (Matt. 4:25) and around Tyre and Sidon (Luke 6:17)”.

    These are conclusive proofs that the complexity of the gospel editing process dictate what we really need are MORE hypothetical texts, not less. Admittedly, the more complex it gets, the less likely we will know for sure the exact evolutionary process, but we at least know that we are closer to the truth than the Two/Four Source Theory and very much closer to the truth than Farrer. Using Occam’s Razor is just an attempt to simplify the process in your mind, not provide an adequate explanation for the complexity in the gospels we have.

    1. That’s all refuted by the evidence of how ancient authors actually redacted and used sources in antiquity. Nothing you describe is even unusual, much less complex. It’s even in fact how we know from school texts authors were taught to redact and use sources (see MacDonald and Brodie on this point). All the things you find bewildering are exactly the kinds of things ancient authors did, even these authors (as evidenced by what they did to Mark). And indeed, we actually know why they did a lot of the things you refer to. You really need to catch up on the literature. But we can’t expect to know it all, because we don’t get to interview them. Not knowing is not an argument for anything. Knowledge does not come from ignorance.

      By contrast, layering Q is a violation of logic. Inventing more and more epicycles with no external evidence only makes the theory more complicated and thus less probable. That confuses “theory” with “evidence.” Bad methodology, bad logic, typifies biblical studies. See my study of that very point in Proving History.

      Burkett’s theories are fanciful and built on piles of possibiliter fallacies that ignore better explanations of the same evidence. A typical work in biblical studies.

    2. Take the one example you give:

      For example, Mark 9:17-27 conflates two different demon possession stories: one is a story of an “unclean spirit” that revolves around the faith of the disciples that Matthew and Luke have, and one is a story of an “unspeaking and mute spirit” that revolves around the faith of the father not known to Matthew or Luke. If Matthew and Luke were just copying from our Mark and/or our Matthew, then that would mean one or both of them would have had to have recognized that the story was a conflation, cut one out, and then ignored the second one.

      That all assumes magical knowledge of what Mark was doing with some other stories, rather than starting with the simpler hypothesis that Mark meant everything he said, that he wrote nothing he didn’t intend to. Mark is not an idiot. Randomly conflating stories is actually the least likely thing he or any author would do. They write on purpose, for a reason. We know this from extensive background evidence of how authors work, especially ancient ones. It’s far simpler (requiring far fewer assumptions and moving parts) to just say Matthew and Luke changed Mark’s story by cutting the parts each didn’t like, and being different men, they disliked different things. It is not possible to argue a more elaborate theory than that without either (a) evidence for that theory other than its circular self-presumption or (b) reducing the probability of the more complex theory in relation to the simpler one. Since you can’t do (a), you are left with (b). And the conclusion is that more probably than not, Burkett is simply wrong. He doesn’t understand how probability works in relation to theory complexity, or what makes one theory more complex than another—or how background evidence relates to probability: e.g. that Matthew and Luke were different men with different values, attitudes, and goals, and therefore would not be expected to always like or dislike or do the same things, is all but 100% certain, and therefore it is presuming otherwise that is improbable, not the other way around.

  6. Keith Johnsen April 29, 2017, 10:09 am

    I read in Burton Mack’s book on Q that one of the criticisms of Q had been there there was no evidence of a sayings Gospel existing in the early church to support what Q scholars determined it would have looked like. He then cites that when Thomas was discovered, that this presented evidence that supports what Q scholars have been speculating. What you present makes a clear and substantial case against it, yet does them seeing a sayings-type Gospel being what it would have looked like, prior to any actual example of one ever existing lend some credence to the possibility since there is now evidence something like that did in fact exist?

    1. I don’t understand the question. It sounds like you are asking something about GThom but I can’t tell what. Goodacre demonstrates GThom is a late redaction into a sayings collection of the prior synoptic Gospels. It is not evidence of any early strand of Christianity.

  7. Gilbert Schwarz April 30, 2017, 4:28 am

    Hi Richard. I’d be interested to read more about the premise that – IIRC – according to Marc Goodacre forms the core of the argument for Q: that Luke and Matthew MUST have written independent of each other (for some reason); only if this is accepted, one actually needs to look for a second shared source besides Mark. I’d be surprised if this idea is well-founded, but if possible could you expand on this a bit more? Who did argue like this at first, and how? Do current Q scholars know that their whole paradigm rests on this arbitrary assumption? Do they still accept this premise? Or do they reject it now? Thanks!

    1. The answers to those questions are given in Goodacre’s and Goulder’s work. See links in the article. They extensively treat of the historiography of Q. And in their articles summarize it in points. In short, Q apologists have been in confusion and not always consistent or honest regarding what assumptions they were actually relying on and when. Some claim they don’t assume Matthew and Luke wrote independently, but then turn around and act like that’s exactly what they assume. And the question does become much muddier when you admit Luke used Matthew, though some Q apologists remain even then.

  8. Grant Willson April 30, 2017, 10:52 pm

    I was interested in you comment:

    “We all know Matthew copied Mark verbatim (without telling anyone; and yes, that was considered dishonest, even then).”

    What is the evidence that plagiarism was considered wrong? I know that Vitruvius distinguished himself from plagiarists but is there other evidence that this was considered “dishonest” in Jewish/Christian tradition? I haven’t seen any discussion of this issue in peer reviewed academic literature and in an email communication with Craig Keener he stated the Gospels were considered open source. Although he didn’t provide any evidence for that assertion.

    Would be interested in your thoughts and direction to robust further reading.Thanks

    1. What is the evidence that plagiarism was considered wrong?

      Extensively presented in Ehrman’s book Forged. So extensively, as to be decisive. More detail even is available in his academic companion to that, Forgery and Counterforgery.

      We always knew this in ancient history, BTW; bible studies has been divorced from the halls of professional ancient historians for ages, so they were able to invent bogus claims in a bubble for apologetic purposes. It’s a great service to see Ehrman finally burst that bubble with actual researched facts on the subject from ancient history.

  9. Alick Wilson May 1, 2017, 1:21 pm

    While I don’t accept ‘Q’, I’ll still accept your offer and post, even though I’m not a patron. I have a question, and a utility you may find useful. You say Luke used Matthew. But Luke 1 & 2 read like later additions, in my view probably added to the original Luke in response to the nativity of Matthew. How can we be sure that Matthew isn’t using Mark and the earlier version of Luke, making Luke version 1 ‘Q’?

    As anyone looked into this possibility before, and would this tie in with Marcion’s shorter version of Luke?

    And the utility if anyone wants to check 2 different texts (such as Mark and Matthew) to see if one copied from another. It uses trigram analysis, and works on any language except HTML. You can opt to remove numbers (handy for Bible passages), as well as quotes and punctuation.


    It doesn’t tell who copied from whom, just how identical the 2 texts are alike. It displays the first text in trigrams, with identical trigrams in a different colour (currently a dull red). This is useful for seeing how large some of these copied sections are, the longer, the more likely they are parts of copied text. One has to be careful with speech, however, as that will provide false positives. Two people independently quoting a third will still have identical text, so simply comparing how different the two texts are is not enough.

    Below that is how many unique trigrams there are, and the lower the number, the more likely it is that one of the texts is based on the other. For large texts, the unique trigrams should be around 95% or higher (Matthew compared to Treasure Island as a test gives 95.4%). Being written in HTML means each text must be 2k bytes or less in size.

    1. That’s handy.

      Some do try to propose a proto-Luke, even Marcion’s, as first, but I have yet to see any plausible theory articulated that explains the redactional trends without ridiculously elaborate ad hoc theories often involving complex yet unevidenced intentions.

      It’s more likely Matthew was the first Gospel than that Luke was, in terms of the models that come from how mythology gets constructed and the well-evidenced (and thus not speculated) intentions of redactors, in general and these in particular. For instance, it’s far simpler and easier to explain how Luke would do what he did to Matthew, than to explain how Matthew would get his text out of Luke. Similarly, though I’m attracted to the Matthaean priority thesis (it would solve a lot of questions), it just doesn’t explain many things, like why Mark would add mistakes to Matthew; whereas that Matthew would fix mistakes in Mark is far easier to explain (which means: far more expected; which means: far more probable). Similar analysis operates from Matthew to Luke, e.g. Matthew could not have magically known all the things Luke added from Josephus, so as to conveniently remove them all, whereas it makes perfect sense that Luke would embellish Matthew’s mythology into a pseudo-history by adding historical color from Josephus (and probably other authors as well, e.g. almost certainly Luke used a historian of the Aegean to lift in all his cultural and political details there, and we just can’t tell because that historian wasn’t preserved and Josephus was).

      One thing to remember from redactional studies in general is that verbatim redaction was actually not the norm in antiquity. It happened more rarely than free redaction, especially over time. Thus, we actually expect less and less verbatim redacting as time goes on. That fits the standard model of Mark, then Matthew, then Luke, then John. Each is a freer redactor of their predecessors than the last.


Add a Comment

I only publish comments by my patrons and anyone who or whose work I discuss in the article commented on. Comments must also follow good etiquette. Those who engage in dishonest, abusive, or harassing behavior may even be banned as commenters and patrons.