Home‎ > ‎The Synoptic Problem‎ > ‎

MwEL: A New Synoptic Hypothesis

What is the synoptic problem?

Much of the text of the gospels of Mark, Matthew, and Luke is very similar to, and in some places identical to, text in either one or both of the others, and for this reason they are collectively referred to as the Synoptic (or ‘seeing together’) Gospels, or just the ‘synoptics.’ The existence of the parallel text in the synoptic gospels raises the issue of how the gospels came to exist in this form, and a very good general statement of the problem (usually referred to as the synoptic problem), with suggested solutions, is given by Stephen C. Carlson on the Synoptic Problem Website:

The synoptic problem is an investigation into the existence and nature of the literary interrelationship among the first three "synoptic" gospels. Matthew, Mark, and Luke are called the synoptic gospels, in contrast with John, because they can readily be arranged in a three-column harmony called a "synopsis." Unlike John, the synoptic gospels share a great number of parallel accounts and parables, arranged in mostly the same order, and told with many of the same words. Any proposed solution to the synoptic problem, therefore, must account for these literary similarities among the synoptics, not so much in terms of their factual content, but in the selection of that content, the arrangement of the material, and wording of the parallels.

The synoptic problem is not yet ‘solved,’ in the sense that it is not yet possible to fully explain either the content or the order of all the parallels on the basis of any of the current synoptic hypotheses. This document is not intended in any way to be a critique of all the hypotheses, although some of the problems with some of the hypotheses are discussed, but instead it discusses how a combination of modified forms of two of the currently most accepted hypotheses combines their strengths, while avoiding their major weaknesses.

The various hypotheses generally involve copying and/or editing between Mark (Mk), Matthew (Mt), and Luke (Lk), possibly supplemented by additional hypothetical source material, typically either earlier material (oral and/or literary) from which one or more of the synoptics was derived, or a document containing material common to two of them. The hypotheses can be conveniently split into two main groups: those that propose that Mark was the first of the three synoptics to be written and was known to and used by the authors of both Matthew and Luke (known as Markan Priority); and those that propose that Mark was last (known as Markan Posteriority).

The current majority opinion is that Mark was the first of the synoptics to be written, with Matthew and Luke following, probably in that order, and there are a group of hypotheses that have this general form (For arguments against Markan Priority see The Synoptic Problem from Daniel B. Wallace andThe Argument from Order’ in The Present State Of The Synoptic Problem by William R. Farmer). Within this group the existence of the parallel material in Matthew and Luke but not Mark (often referred to as the Double Tradition) is usually accounted for by the author of Matthew (aMatthew) knowing Luke, the author of Luke (aLuke) knowing Matthew, or by both having had access to a hypothetical additional document containing the original of the parallel material. Most people who favor the idea of such a hypothetical document (usually referred as Q, see The Current State of Q by Nancy R. Heisey for details) believe it consists (largely) of sayings of Jesus, and it is defined in such a way as to provide a solution to various difficult issues concerning the Double Tradition. For example, in some places two different versions of the same parallel text, known as Doublets (see also Doublets in the synoptic tradition) exist in either or both Matthew or Luke. In other places the text in Matthew sometimes appears to be earlier (more primitive) than that in Luke, while sometimes Luke appears to be more primitive than Matthew. The latter issue is commonly referred to as ‘alternating primitivity,’ (although bi-directional primitivity would be a better term) and is described in Fallacies at the Heart of Q by Mark Goodacre.

Limitations of Current Hypotheses

A note on nomenclature: The terms 'theory' and 'hypothesis' seem to be generally regarded as interchangeable when it comes to discussing possible synoptic solutions, even though in most fields they do have distinct meanings. As a result, some of the suggested synoptic solutions are called theories, and some hypotheses. As 'theory' generally denotes a tested, well-substantiated, unifying explanation for a set of verified, proven factors (and no proposed solution has so far reached that stage), this discussion will in general use the term hypothesis. However, when referring to a specific suggested solution it will use the name by which the theory/hypothesis is generally known.

In whatever way the existence of the parallel material is explained, none of the current synoptic hypotheses is able to explain all the data, whether it is the doublets, alternating primitivity, the minor agreements of Matthew and Luke against Mark, the Mark-Q Overlaps, or something else. Even though the addition of a fourth source document (Q) does appear to solve several issues that hypotheses that limit themselves to just the three synoptic gospels cannot, it appears impossible to define Q in such a way as to solve them all. However, this does not mean there is no solution, or that any actual solution must contain more than four documents. Instead, it is possible that the Q adherents are on the right track by suggesting an additional document, but that they are wrong as to what it contained. What Exactly is Q? points out that what is generally considered to be ‘Q’ is just one possible ‘second source’ (SS) used by the authors of Matthew and Luke in addition to Mark, and that any hypotheses in which a source referred to Q is used is just an example of a continuum of hypotheses in which Matthew and Luke have two sources, which can collectively be referred to as ‘Mark and Second Source,’ or MaSS, in which the form and content of the second source may vary according to the details of the hypothesis.

An unstated, but nevertheless potentially important, problem is that when the texts of Mark, Matthew, and Luke are compared and relationships between either two or all three are considered, it is almost always done so by using the Greek text of the canonical versions of the gospels. It should be clear that this assumes primacy of the Greek versions over the Old Latin, Syriac, or any other non-Greek version. While this is perhaps justified by the extant mss evidence, there is a much more important issue to consider even if this assumption of primacy is correct, which is that we do not know what differences there may have been between the text that each of the synoptic authors saw in their source material, and the UBS/NA texts that we use today that have been constructed from mss written 100 or more years later. Unfortunately, this issue is usually ignored (or at least glossed over) when discussing synoptic relationships. As Peter Head points out;

… scholars accept the solution from outside their primary discipline as the means to help solve their difficulties. Synoptic specialists tend to think that textual critics have sorted out the text of the NT (which textual critics do not think). It is difficult to dispute the accuracy of Parker’s comment that “the study of the Synoptic Problem as normally conducted includes the agreements between practitioners that the text of Nestle-Aland is, to all intents and purposes, what Matthew, Mark and Luke originally wrote.”

It should also be noted that synoptic hypotheses generally do not take account of possible parallels in other gospels, in particular John and Thomas (although finding the Gospel of Thomas did significantly impact the general view of Q), with it being deemed preferable to find a solution within the confines of just Mark, Matthew, and Luke. John S. Kloppenborg warns us of the problems raised by these simplifying assumptions in the introduction to his ‘Synoptic Problems: Collected Essays:’ 

Simple reflection on the state of our knowledge of the transmission of early Christian texts should tell us that it is dazzlingly improbable that Matthew had direct access to Mark, or that Luke had Matthew’s authographic copy. At best they had copies (or copies of copies), which at the very least were subject to the ordinary alterations introduced by copyists, and perhaps more major alterations such as additions, deletions, and rewordings as well. If the latter is the case, then the arguments that we invoke for testing hypotheses cannot take the form of simple deductive testing. That is, if the Synoptic Problem were a logical puzzle on the same level of the proposition that “all swans are white,” where the discovery of even a single black swan would refute that proposition, it would be a simple matter to refute, for example, the FH’s claim that Luke derived all of his “double tradition” material [from - sic] Matthew by pointing to a single instance where Luke’s version of a double tradition saying or story was clearly earlier than Matthew’s version. But such is hardly the case. For we have neither the autographs of Matthew or Luke, nor is it reasonable to suppose that there was a hermetically sealed conduit between Matthew and Luke such that no alternate memories, performances, or information could affect Luke’s re-performance of a Matthaean unit. Though the fictions of simple synoptic diagrams are heuristically useful, we must not be beguiled by their simplicity and assume that the relationship between the two gospels was a simple one.

This problem can be seen in the fact that several hypotheses rely on the existence of oral sources or ‘traditions’ (or even changes 'sourced' purely from the mind of a gospel author) to supply text that we see in a synoptic gospel that cannot (according to the hypothesis) have been present in whatever source texts are specified in the hypothesis itself. For example, some hypotheses include a source unique to aMatthew (generally known as M), and some include an equivalent source (L) used by aLuke, at least parts of which are assumed to be oral. The problem with this is that by their very nature, oral sources are nebulous, and so can be hypothesized to supply whatever text is necessary to solve a problem. Even some hypotheses that do not formally allow for such sources do so in practice. For example, although the Farrer Theory (FH, or Farrer-Goulder Theory, or Mark without Q Hypothesis – MwQH) defines Mark and Matthew as the only sources used by aLuke, its current chief advocate Mark Goodacre allows for oral sources as well, a position on which Paul Foster takes him to task in ‘Is It Possible to Dispense with Q?’:

Yet is this not the "thin end of the wedge" for those who advocate Markan priority, but non-Q solutions to the synoptic problem? Goodacre vigorously protests that his theory is not susceptible to such a charge. He asserts, "Some Q sceptics feel a little uncomfortable with this scenario since it might at first sight appear to allow Q to creep in through the back door. Is this, to use another image, a kind of 'closet Q', believing in a form of the Q hypothesis but not owning up to it? I don't think so." Despite this declaration of not reinventing Q in a different guise, it was precisely in order to escape the necessity of hypothetical sources or traditions that Farrer and Goulder framed and maintained the theory of Luke's use of Matthew. Without sticking to this hard line the theory loses its appeal, since it results in the multiplicity of hypothetical sources with which this theory is trying to dispense.

Another problem that is not generally recognized as a synoptic issue is that we do not know whether, for example, aLuke saw a copy of Mark that contained everything that was in the copy of Mark seen by aMatthew. Kloppenborg takes up this issue in Composing Matthew by Recomposing Q: The Composition of Matt 23–25:

From a historical point of view, it is rather unlikely that the text of Mark hypothetically used by Matthew in the late first century CE was identical to the Markan text of Nestle-Aland, based as it is on an early third century manuscript (P45), the fourth-century Vaticanus, and their successors. This means that there will always be some slippage between what Matthew did to his sources, and how many of those changes are visible to us. Nor is it safe to assume that the text of the sayings gospel Q used by Matthew was either the same as that produced by the authors of Q, or that used by Luke. Observation of the transmission of the early papyrus copies of early Christian texts makes plain the extent of variation from one copy to another. The level of variation and transformation is likely to have been greatest at the early stage of the transmissional process, before Mark was recognized as an authoritative text.

This is significant because Luke does not include any parallels to Mk 6:45-8:26 (a gap generally referred to as The Great Omission), arguably because the copy of Mark seen by aLuke did not include this text, as described by Mark Allen Powell in ‘What are They Saying about Luke?’:

Another question scholars must consider when they examine Luke’s use of Mark is the status of the evangelist’s Markan text. In Lukan studies, Mark 6:45-8:26 is sometimes called “the big omission” and Mark 9:41-10:12 “the little omission,” because Luke does not include any material from these sections. Unable to explain such lapses, some interpreters have proposed that Luke’s copy of Mark was defective or incomplete.

In addition, the endings of both Matthew and Luke have almost nothing in common with Mk 16:9-20 (the long ending of Mark), and it is possible that this was due to both seeing a copy of Mark in which these verses were not present. Also, a small amount of the text of canonical Mark is not present in either Matthew or Luke, and again, one possible explanation is that this text was not in whatever copy of Mark was seen by either aMatthew or aLuke (see Fatigue in Mark – or Damage to Mark?). Despite these issues, most synoptic hypotheses do not interact with possible different versions of Mark.

However, a bigger problem is the very strong evidence suggesting that Luke went through one or more significant developmental phases, of which the most notable is that Luke appears to have originally begun at v. 3:1, omitting chapters 1 and 2 in their entirety. There are several other things that also point to canonical Luke not being the original form of the text, for example:

  • The genealogy in Lk 3:23-30 appears to be out of place, coming as it does after John and the baptism, rather than before it; 
  • The references to Nazareth and Capernaum in Luke chapter 4 appear to have originally been in the opposite order, and Lk 4:14b-15 appear to have been added to Luke only after Nazareth was swapped with Capernaum; 
  • The incident regarding Simon’s mother-in-law related in Lk 4:38-39 takes place before Jesus has even met Simon, so suggesting a different original order, as seen in Mark and Matthew; 
  • In Luke there are two significantly different variants of the Lord’s prayer. The shorter variant (found in e.g. P75, B) appears to be original, while in the longer a number of phrases found in Mt 6:9-13 appear to have been interpolated into the text. 
  • There are many different extant variants of Lk 22:17-20, with a strong possibility that Luke 22:19b-20 was not originally in Luke; 
  • The existence of the Western Non-Interpolations (see The Western Textsuggests that the text of an early version of Luke may also have been shorter in other places, particularly so in chapter 24.

Although none of the above conclusively points to either Mark or Luke existing in different versions, the evidence is strong enough that ideally, each of the various synoptic hypotheses should be split into four ‘sub-hypotheses,’ with different combinations of longer and shorter versions of Mark and Luke, and re-evaluated. Although it is unrealistic to expect this to happen in all cases, it can perhaps can be done in the case of the two current (2020) synoptic ‘front runners:’ The Two Source hypothesis (2SH, or Mark-Q theory); and the MwQH.

Both these hypotheses assume Markan priority, but while the former (an ‘instance’ of the generic MaSS hypothesis) accounts for the Double Tradition material by having both aMatthew and aLuke know and use material from an additional source (Q), the latter accounts for it by having aLuke know and use Matthew instead. However, it should be noted that although the purpose of the MwQH is essentially to show that aMatthew and aLuke need not have had a common written source other than Mark, this does not preclude it from having been the case. On the other hand, because of the way the Mark-Q theory is defined, this hypothesis requires aMatthew and aLuke to have not seen each other’s gospel. In “On Dispensing with Q?: Goodacre on the Relation of Luke to Matthew,” Kloppenborg states:

Goodacre argues that given Luke’s direct knowledge of Matthew the supposition of a sayings source is simply unnecessary. This is indeed the right way to frame an argument against Q, which is not a hypothesis on its own, despite those who tirelessly refer to ‘the Q hypothesis’. Rather, Q is a corollary of the hypotheses of Markan priority and the independence of Matthew and Luke, since it is then necessary to account for the material that Matthew and Luke have in common but which they did not take from Mark. The case for Q rests on the implausibility of Luke’s direct use of Matthew or Matthew’s direct use of Luke.

Kloppenborg’s above suggestion of “the implausibility of Luke’s direct use of Matthew” is surprisingly easy to demonstrate when evaluating the Double Tradition Doublets in Mt 10: The Mission of the Twelve, and the argument is so straightforward that it is surprising that the MwQH, in which aLuke does directly use Matthew, is nevertheless still considered to be one of the two ‘front runners.’ However, the argument is only valid against the MwQH, because on that hypothesis Luke’s only sources are Mark and Matthew, and the “case for Q” (i.e. for the Mark-Q hypothesis) requires that the only sources of Matthew and Luke are Mark and Q, in which case by definition neither aMatthew nor aLuke even knew the gospel of the other.

Matthew 10 is almost totally devoted to the mission of the twelve. In Mt 10:1-4 Jesus calls and names them, in Mt 10:5-6 he tells them who go to and in Mt 10:7-14 what to do there, and in Mt 10:15-23 Jesus tells them about the problems they will face. Mt 10:24-40 then contains several aphorisms providing context to what Jesus wants them to do. Despite this quite well-defined structure, on the assumption of the MwQH hypothesis aLuke went to extreme measures to split up Matthew 10 and place parallels to almost all of it in seven different chapters of his own gospel, as follows: 

            MARK                 MATTHEW             LUKE1          LUKE2
            Mk 3:13b-19a    Mt 10:1-4                Lk 9:1b          Lk 6:13-16

            Mk 6:7               Mt 10:5a                 Lk 9:1a           Lk 10:1
            ---                      Mt 10:5b-6              ---                   ---
            ---                      Mt 10:7-8                Lk 9:2             Lk 10:9b,a
            Mk 6:8               Mt 10:10a,9            Lk 9:3a-b        Lk 10:4b,a
            Mk 6:9               Mt 10:10b               Lk 9:3c           Lk 10:4c
            ---                      Mt 10:10c               ---                    Lk 10:7
            Mk 6:10             Mt 10:11                 Lk 9:4              Lk 10:8
            ---                      Mt 10:12-13            ---                    Lk 10:5-6
            Mk 6:11a           Mt 10:14                 Lk 9:5a            Lk 10:10
            Mk 6:11b           ---                           Lk 9:5b            Lk 10:11a
            Mk 6:11c           Mt 10:15 // 11:24    ---                     Lk 10:12
            ---                      Mt 10:16                 ---                    Lk 10:3

            ---                      Mt 10:17-19            Lk 12:11-12
            Mk 13:11-13      Mt 10:20-23            Lk 21:15-19
            ---                      Mt 10:24-25            Lk 6:40
            Mk 4:21-22        Mt 10:26-27            Lk 12:2-3
            ---                      Mt 10:28-33            Lk 12:4-9
            ---                      Mt 10:34-36            Lk 12:51-53
            ---                      Mt 10:37-38            Lk 14:26-27
            ---                      Mt 10:39                 Lk 17:33
            ---                      Mt 10:40                 Lk 10:16
            ---                      Mt 10:41-42            ----

The parallels to Matthew 10 show that it is structured in distinct ‘sections’. Following a few verses with parallels in Mk 3:13b-19a / Lk 6:13-16, Mt 10:5-16 has parallels in Mk 6:7-11 and Lk 9:1-5 that suggest that Mt 10:5-16 is an ‘expansion’ of the versions in Mark 6 and Luke 9. However, there is also a longer parallel in Lk 10:1-12 (i.e. more like Mt 10:5-16), except that the order of the parallel text in Luke 10 is very different. There is then a parallel to Lk 10:16, and two verses with no parallel. 

The second section of Matthew 10 appears to have been created in a different way to the first, as it has very few parallels in Mark, and it has been used differently as well, having no parallels in Luke 9, and instead has parallels spread over eight chapters of Luke, so that most of this portion of Matthew 10 is part of the Double Tradition. Even more surprising is the fact that on the MwQH it appears that aLuke reused and rearranged almost the whole of Matthew 10 in a process that involved three intertwined passes through the text: 

  1. He first spread parallels to the instructions in Matthew 10:1-22 across Luke 6, 9, 10, 12, 21, and replaced Mt 10:23 by Lk 21:19;
  2. He then ‘backtracked’ in the text he was creating, spreading parallels of the aphorisms in Mt 10:24-39 across Luke 6, 12, 14 and 17;
  3. Finally, he backtracked again to place a parallel of Mt 10:40 at Lk 10:16, finally ignoring Mt 10:41-42.

To achieve this on the MwQH presupposes an almost bizarre amount of planning by aLuke, knowing how he wanted to split up Matthew 10, and where he wanted to place the pieces relative to other text he was going to include, before even beginning to write any of these sections of his gospel. These actions appear so extraordinary that it is almost impossible to envisage aLuke rearranging the contents of Matthew 10 in this way so as to achieve the observed result, let alone creating two parallels of the same verses from Matthew 10 with the second in a completely different order. This puts a very large question mark against such a procedure as dictated by the MwQH, and it also brings to mind the ‘crank’ comment from B.H. Streeter in a section of ‘The Four Gospels’ entitled The Document Q’:

Sir John Hawkins once showed me a Greek Testament in which he had indicated on the left-hand margin of Mark the exact point in the Marcan outline at which Matthew has inserted each of the sayings in question, with, of course, the reference to chapter and verse, to identify it; on the right-hand margin he had similarly indicated the point where Luke inserts matter also found in Matthew. It then appeared that, subsequent to the Temptation story, there is not a single case in which Matthew and Luke agree in inserting the same saying at the same point in the Marcan outline. If then Luke derived this material from Matthew, he must have gone through both Matthew and Mark so as to discriminate with meticulous precision between Marcan and non-Marcan material; he must then have proceeded with the utmost care to tear every little piece of non-Marcan material he desired to use from the context of Mark in which it appeared in Matthew—in spite of the fact that contexts in Matthew are always exceedingly appropriate—in order to re-insert it into a different context of Mark having no special appropriateness. A theory which would make an author capable of such a proceeding would only be tenable if, on other grounds, we had reason to believe he was a crank.

However, none of the above prevents the reverse, i.e. with aMatthew gathering together small pieces of text from multiple chapters of the text in what we know as Luke (possibly together with small portions of Mark) in order to construct what we see as Matthew 10, although even here aMatthew would have to contend with merging what he would have seen in Luke 6, 9, and 10. There is also the possibility they some of these pieces of text that we see in were not in their Lukan form when aMatthew created his gospel, and instead they could have been in another source document (or documents). However, this could not have been Q as usually defined because of the significant number of Markan parallels shown above, unless all of those parallels are considered to be Mark-Q overlaps.

Despite (or perhaps because of) the above problem with the MwQH, before considering how earlier or otherwise different versions of either Mark or Luke might affect these two hypotheses, it is necessary to consider any other previous hypothesis that also assumes Markan priority, and that does already take account of other versions of Mark or Luke. The Synoptic Problem Website identifies the following three such, all of which are variations on the Mark-Q theory, and so are within the scope of the MaSS hypothesis:

  • Mark-Q Theory with proto-Luke (Streeter): aMatthew and aLuke had ‘special sources’ (Documents or other sources referred to as 'M' and 'L' respectively). Matthew was created by combining M, Mark, and Q. Q was combined with L to create proto-Luke, which was then combined with Mark (and possibly a separate ‘infancy’ source: see Luke Chapters 1 and 2) to create Luke. (Matthew <= Mark+M+Q, proto-Luke <= L+Q, Luke <= Mark+proto-Luke) 
  • Mark-Q Theory with proto-Mark (Weisse): Mark is an edited version of a non-extant proto-Mark (or Ur-Markus), which was also used by aMatthew and aLuke. (Mark <= proto-Mark, Matthew <= proto-Mark+Q, Luke <= proto-Mark+Q) 
  • Mark-Q Theory with deutero-Mark (Abbott): A revised (but no longer extant) version of Mark was used by both aMatthew and aLuke instead of Mark. (deutero-Mark <= Mark, Matthew <= deutero-Mark+Q, Luke <= deutero-Mark+Q)

We can say the following about these hypotheses:

  • In the only hypothesis that has an early version of Luke, Streeter's proto-Luke has no effect on Matthew, and from the synoptic point of view is therefore not much more than a detail in the development of canonical Luke;
  • In the two hypotheses that have an additional version of Mark (proto-Mark and deutero-Mark), both aMatthew and aLuke use these additional versions, not the canonical ones;
  • None of these hypotheses have either aMatthew knowing Luke or aLuke knowing Matthew, i.e. Matthew and Luke are independent (which is not surprising given their use of Q instead).

None of these three hypotheses is widely regarded today, but the last one is particularly interesting, since from its position in the synoptic diagram deutero-Mark could possibly be considered to be a proto-Luke or proto-Matthew instead, depending on how different deutero-Mark was from Mark, and how similar it was to either Luke or Matthew. However, without knowing the actual content of deutero-Mark, it is not possible to tell how close deutero-Mark would have been to any of the synoptic gospels.

Although other versions of Mark and Luke have been taken into account in various synoptic hypotheses (and those assuming both Markan priority and involving Q are described above), there are almost no hypotheses that do so and assume Markan priority without Q, e.g. variations on the Farrer Theory (aLuke used Mark and Matthew) or the Wilke Model (aMatthew used Mark and Luke) that involve proto- (earlier) or deutero- (later) versions of any of the synoptic gospels. Given that there is significant evidence for an earlier version of Luke (as described above), but essentially no evidence in the text of Matthew (but see The Gospel of the Ebionites) for an earlier version of Matthew, the rest of this document will consider the impact that an earlier version of what we know as Luke (not necessarily written by the author of the 'final' version of Luke) would have on both the Farrer Theory (MwQH) and the Mark-Q Theory. It should be noted that adding an earlier version of Luke to the Wilke (or Mark-Luke) model has no effect on aMatthew (who still sees the final version of Luke), and so this hypothesis need not be considered.

Markan Priority With Early Luke (MwEL)

The MwEL hypothesis can be thought of as a combination of the MwQH and the Mark-Q theories, and is also a particular 'instance' of the MaSS hypothesis. It assumes Markan priority, has two sources common to Matthew and Luke: Mark and a second source, in this case Early Luke instead of Q, but also has aLuke knowing Matthew. So, on the MwEL hypothesis aMatthew and aLuke each knew both Mark and Early Luke, and (as also on the MwQH) aLuke also knew Matthew, but aMatthew did not know Luke

These relationships are shown in the synoptic diagram on the right, in which the circles represent the texts of the three synoptic gospels and that of Early Luke (eLk), and the arrows show the direction of the flow of the text from one to another as a result of the actions of their respective authors. This also indicates the dependencies, i.e. the text or texts upon which each depends, and which in the MwEL hypothesis can be represented as:

Early Luke <= Mark
Matthew    <= Mark + Early Luke
Luke          <= Mark + Early Luke + Matthew

(Note that if eLk and its connections to the gospels are removed then the diagram becomes that of the MwQH, and instead if the links from Mk to eLk and from Mt to Lk are removed and eLk is renamed Q then the diagram becomes that of the Mark-Q theory.)

In common with other synoptic diagrams, the MwEL hypothesis and synoptic diagram does not deal with the issue of how the synoptic gospel deemed to be the earliest (in this case Mark) was created, as the object of a synoptic hypothesis is to identify the dependences between each pair of synoptic gospels, except that any dependences between either two or all three of the synoptic gospels and any hypothetical text included in the hypothesis are also shown in the diagram.

The MwEL hypothesis is related to the 1967 hypothesis of H. Philip West Jr., in which he suggests “that Matthew used Mark and a primitive version of Luke,” and is consistent with the conclusions of my previous stylometic analysis (which finds evidence of a non-Markan second source for Matthew and Luke) which itself owes much to the related 2002 analysis from Dave Gentile.

This hypothesis can also be seen as a variation of the Holtzmann-Gundry Three Source Hypothesis (3SH, or Mark-Q-Matthew model), and is similar to Bruce Brooks’ Luke A/B/C model, with Early Luke occupying the same 'synoptic space' as Q in the 3SH and Luke A in Brooks’ model. In addition, Klinghardt's synoptic suggestion has Marcion's Gospel of the Lord as an early version of Luke, so being similar to the MwEL hypothesis (This last possibility is discussed below, although it should be noted that the MwEL hypothesis does not in any way rely on Marcion's Gospel being Early Luke).

Finally, the MwEL hypothesis also has some affinity with Alan Garrow’s Matthew Conflator Hypothesis (MCH), in which Matthew and Luke have both Mark and another document (smaller than Q) as sources, with aMatthew knowing Luke (known as Matthean Posteriority). Although at first sight this appears to be a version of the Mark-Q theory, the link from Luke to Matthew in the MCH acts in a similar way to the link from Early Luke to Matthew in the MwEL hypothesis. In Is It Possible to Dispense with Q?, Foster comments on the use of Luke by aMatthew:

It is extremely surprising that the theory of Matthew utilising both the gospels of Mark and Luke has not attracted the widespread attention Farrer's proposal has engendered. Perhaps part of the reason for this is the Lukan preface, for whereas Luke tells readers that he has drawn on many sources, Matthew is silent in relation to his source material. Yet there may well be another reason. While the advocates of a Matthew-dependent Luke have complained that their theories have often not been taken seriously, they have been quick to scorn the reverse point of view, namely that of a Luke-dependent Matthew.

Ropes stated categorically that "The third possibility, that Matthew is dependent on Luke for these sayings, may, for a variety of reasons, be dismissed, although the idea is sometimes advanced." Farrer, by contrast, simply assumes that the dependence must be in the order of Luke utilising Matthew and not the reverse. He states, "we can conceive well enough how St. Luke could have read St. Matthew's book as it stands, and written the gospel he has left us."

Yet, the alternative possibility, that of Matthew having read Luke's book, is never considered by Farrer, instead it is dismissed by silence. By contrast Goodacre does not damn the alternative theory with silence, instead he has a single polemical sentence to dismiss consideration of Matthew's knowledge of Luke. In very bald terms he states, "The theory that Matthew has read Luke ... is rarely put forward by sensible scholars and will not be considered here."

This comment seems to reflect different standards. Goodacre bemoans the fact that the theory of Lukan dependence on Matthew rarely receives the examination it deserves and is marginalized by critical scholarship. Yet this appear [sic] to be the way Goodacre treats the alternative, especially by implying that those who propose such theories do not fall into the category of sensible scholars.

It is indeed surprising that Matthean posteriority seems to have been largely overlooked, although in 1967-68 it was revived by H.P. West in A Primitive Version of Luke in the Composition of Matthew, in 1992 by Ronald Huggins in Matthean Posteriority: A Preliminary Proposaland now by Garrow with the MCH. Both the MwEL hypothesis and the MCH provide the power seen by many in allowing aMatthew to use both Mark and a version of Luke, so eliminating the need for much of a Q-like source. It should also be noted that under the MwEL hypothesis the Lukan preface did not exist in Early Luke and could not therefore have been seen by aMatthew, so removing any reason why he might have felt it necessary to emulate Luke and mention his sources as well.

Markan Priority and Second Source (MaSS)

The MwEL hypothesis should perhaps have been named the MaSS (Markan Priority and Second Source) hypothesis, as the use of the term 'Early Luke' is not intended to suggest that there is necessarily any form of link between Early Luke and Luke other than Early Luke being a source (perhaps the primary source) for what we see in Luke today, and that Early Luke was also a source for what we see in Matthew. As previously suggested, this could be taken to indicate that there is no distinction between Early Luke and Q, except that Q, at least as defined by the IQP (the The International Q Project), can by definition only exist in a synoptic hypothesis in which aMatthew and aLuke worked completely independently. In contrast, Early Luke does not require this stricture, and so (for example) is not required to be the source of all the Double Tradition material seen in Matthew and Luke. However, in other respects Early Luke and Q could be quite similar, and some of the arguments for or against one also apply to the other. For this reason it is sometimes useful to consider a hypothesis in which both Matthew and Luke depend on both Mark and a 'generic' Second Source (or SS), the exact definition of which is to be determined by investigation of the relationships between the texts of Mark, Matthew, and Luke.

The Text of an Early Version of Luke

Although it is suggested above that the Lukan preface did not exist in an early version of Luke, it is not intended that this proposal detail the entire text of any early version of Luke. Instead, it provides a framework for the text, and constrains the text in various areas (e.g. by comparison with Mark, Matthew, and Luke), while allowing many of the details to be explored later. The lack of a full text of an early version of Luke at this stage should not be regarded as detracting from the MwEL/MaSS hypothesis, as in an exactly similar way it took the IQP a decade, using a "team of about forty scholars," to complete a reconstruction of Q, long after it was first hypothesized (and this is by no means the final word regarding the text of Q). Instead, such information as is provided here regarding the content of an early version of Luke is sufficient to show the viability of the MwEL/MaSS hypothesis on the assumption that an early version of what we know as Luke did exist. Although it does not prove that there was an early version of Luke having the characteristics described here, the extra synoptic explanatory power that results from adding an early version of Luke to the MwQH, or by adding a link from Matthew to Luke (so allowing Q to be replaced by an early version of Luke) in the Mark-Q theory, adds significantly to that possibility.

SS: Early Luke, Deutero-Mark, Ur-Matthew, or (perhaps) Q-Lite?

Because the generic Second Source (SS) is, like Q, a hypothetical document its text can be defined to be whatever is needed to best support its function in a synoptic solution. One consideration is that, as with deutero-Mark in Abbott’s synoptic suggestion, SS could be considered to be a later version of Mark rather than an earlier version of Luke, depending on whether it is closer textually to Mark or to Luke. Using the same logic SS could be considered to be an early form of Matthew (e.g. an Ur-Matthew), and as SS is a second source for both Matthew and Luke it seems to fulfill the requirements for being a Q-Lite (But not Q, as in the MwEL/MaSS hypothesis it does not have to contain all the Double Tradition text) - See What Exactly is Q?. 

Although the MwEL/MaSS hypothesis has SS as a source for both Matthew and Luke, it does not consider the possibility that SS might be an Ur-Gospel also serving as a source for Mark. As B. Ward Powers comments in ‘The Progressive Publication of Matthew, An Explanation of the Writing of the Synoptic Gospels:’ 

There is a complete lack of evidence for an all-embracing Ur-Gospel containing all that is found in the three Canonical Gospels. Chapter 10 [A Quick Look at Some Other Ideas] shows that it cannot be accepted that if such a Gospel had existed it would have been passed over by the church in favor of our Matthew, Mark, or Luke (which on this hypothesis would each be only a partial replica of its contents), and certainly it cannot be accepted that it would have disappeared without mention and without a trace.

It should be noted that this argument appears to also be fatal to the idea of an all-embracing Maximal-QPowers continues: 

Other versions of an Ur-Gospel theory in which an Ur-Matthew or an Ur-Markus is postulated, cannot be as conclusively eliminated. The crucial issues are why such a theory is needed (is it an unnecessary hypothesis?) and what data can be offered that requires it as an explanation.

It is not immediately obvious that there is any need for an Ur-Matthew, as Matthew (as we see it) shows little sign of having been created in multiple stages, notwithstanding the references to The Gospel of the Ebionites. However, this is not so the case with Luke, as there is textual evidence (see the Evidence for Early Luke below) within canonical Luke for an early version of Luke, and so this hypothesis will below refer to Early Luke, rather than SS, Late (or deutero) Mark, Ur-Matthew, or Q-Lite.

Nevertheless, as the dual purposes of adding Early Luke/SS to the MwQH are to avoid the need for Q while at the same time minimizing the known problems of the MwQH (such as the rather convoluted actions required of aLuke on that hypothesis), we are constrained by the need to avoid adding unnecessary complications, and so must try to minimize the ‘textual distance’ between Early Luke/SS and each of the synoptic gospels (In a somewhat analogous way to which the text of Q is constrained by the text of Matthew and Luke as both depend on it).

Note: One possible complication is that on the MwEL/MaSS hypothesis as defined above aLuke can obtain the text of Mark (or text parallel to that in Mark) via four possible paths: Directly from Mark; from parallel text in Matthew; from parallel text in Early Luke/SS; and from parallel text in Matthew derived from parallel text in Early Luke/SS. Provided that everything from Mark that we see in Luke was transmitted to aLuke via one or other of the latter three paths, then aLuke might not have needed to have seen Mark at all. This possibility is discussed later.

When discussing whether any particular piece of text may or may not have been present in Early Luke/SS, one obvious consideration is to take account of which of the synoptic gospels contain that text (or a parallel of it). For example, because on the MwEL/MaSS hypothesis Early Luke/SS has only Mark as its source, it seems intuitively more likely that Early Luke/SS would contain text that is in Mark, than that it would contain text that is in Matthew but not Mark. The following sections therefore discuss the likelihood of text of being present in Early Luke/SS depending on which of the synoptics contain it (on the assumption that there were no other hypothetical sources apart from any specifically referred to).

Other (non-Synoptic) Sources

In both the explanation of the MwEL hypothesis and the synoptic diagram above, Mark is indicated to be a source (indeed, the only source) for Early Luke, even though Early Luke also contains material not in Mark. In a similar way, on the MwQH Matthew contains material that could not have come from Mark, and Luke contains material that could not have come from Matthew, but the MwQH synoptic diagram does not suggest any sources for this material. These are examples of a simplification often used in descriptions of synoptic hypotheses, in which sources (of any kind, either oral or literary) other than the synoptic gospels are often not explicitly noted. In some cases an extra source is identified, such as in the Mark-Q hypothesis where Q is shown as a source for both Matthew and Luke, and, of course, in the MwEL/MaSS hypothesis Early Luke/SS performs the same function. However, this still leaves open the question of whether there might have been one or more non-synoptic sources for Q or Early Luke, and, indeed, what sources or sources were used by aMark, or by aMatthew or aLuke for the unique portions of those gospels.

Because there is no way that we can know what traditions and/or oral sources might have been known to or used by the authors of any of the synoptic gospels it is impossible to identify what these sources might have been, so their possible existence is generally assumed but not directly indicated in a synoptic diagram. In most cases the same applies to non-extant literary sources, although the possible existence of some (e.g. the Gospel of the Hebrews, The Logia, Q, Early Luke/SS, and various other Proto, Deutero, or Ur gospels) is at the heart of some synoptic hypotheses, including, of course, this one.

The Other Possible Sources of Text in Early Luke

The following sections identify the categories of text that, on the MwEL hypothesis, might have been present in Early Luke. This takes account of whether that text either is or is not in Mark, and so whether Early Luke might have been the source of that text as used by aMatthew and/or aLuke, but does not try to identify any non-Synoptic sources as just described.

Text in Matthew, and Luke but not Mark (Double Tradition)

In any synoptic hypothesis that depends on Markan priority the text shared by Matthew and Luke but not Mark is perhaps the most interesting, because its source is highly dependent on the hypothesis. On the MwQH the double tradition was formed by the actions of aLuke creating parallels of text he saw in Matthew that did not exist in Mark, and on the Mark-Q theory it was created by aMatthew and aLuke both writing parallels of non-Markan material they saw in Q. In both these hypotheses there is only one copying ‘mechanism’ through which the double tradition was created, but on the MwEL hypothesis there are two: aMatthew wrote parallels of non-Markan text he saw in Early Luke that we also see in Luke, and/or aLuke wrote parallels of text in Matthew (and possibly Early Luke) that had no parallel in Mark.

As indicated in The Making of the Double Tradition a large portion of the text of the Double Tradition is in the same order in both Luke and Matthew (but not the same locations relative to the Markan text in those gospels), and where there is a hypothesized second source for that text the order in that source largely follows the order of Luke. On the MwEL hypothesis this indicates that the order of the Double Tradition text in Early Luke is largely the same as that in Luke, i.e. that the hypothesized Early Luke does indeed have a relationship with Luke.

Although in the Mark-Q hypothesis the Double Tradition in Luke largely follows the order of Q, and there are portions of Matthew in which the Double Tradition also follows the order of Q, a significant portion of the Double Tradition in Matthew is ‘out of sequence’ compared with Luke, and various odd actions by aMatthew (including ‘working backwards’ through the double tradition) have been suggested to account for it. However, on the MwEL hypothesis the same oddities can be seen to be the result of aLuke copying material that aMatthew had previously added to his gospel, and re-arranging that material to suit his own purposes.

As a little over half of the Double Tradition material in Luke is in the same order, but not in the same position, as it is in Matthew, it is reasonable to hypothesize that this material was in the same order in an Early Luke, and both aMatthew and aLuke took this material and used it in their respective gospels in the order in which they found it, although they placed their Markan material in different orders. Early Luke may have also contained other Double Tradition material that aLuke chose to keep in the same order but aMatthew did not. However, a significant portion of the Double Tradition material may have originated in Matthew, with aLuke choosing to use it by fitting it into the Early Luke ‘framework’ by changing the order of this (originally) Matthean material as he saw fit.

It has been noted, for example in Synoptic Tradition in the Didache Revisited by Dr. Aaron Milavec, that some of the text in the Double Tradition has significant similarities to some of the ‘Two Ways’ section of the Didache, and indeed it has been suggested that Matthew was a source for the Didache. However, as Didache 1-5 may in fact pre-date Matthew it could instead be a source for some of the Double Tradition material in Matthew that was then used by aLuke.

Text in Mark, Matthew, and Luke (Triple Tradition)

The next consideration regarding the text of Early Luke is that it should probably include a version of all the passages with parallels in Mark, Matthew, and Luke (Triple Tradition). Although if such a passage was not in Early Luke then aLuke could obtain it from either Mark or Matthew, and aMatthew from Mark, we would then need to explain not only why aeLuke (the author of Early Luke) chose to exclude it, but also why both aMatthew and aLuke chose to use it even though it was not in Early Luke. It is also possible that having a version of a Triple Tradition passage in Early Luke as well can explain otherwise difficult problems with doublets and alternating primitivity between Matthew and Luke (see below).

Whatever Triple Tradition passages should or should not be in Early Luke, there is also the question of their order, as the order of the Triple Tradition passages is not always the same in all three synoptics. As Professor Barry Smith writes:

The order of pericopes in the triple tradition is similar. Although they agree at times in not having a pericope found in Mark, when they depart from Mark's order, Matthew and Luke do not do so in the same way. Rather, when Matthew departs from Mark's order, Luke supports it, and, when Luke departs from Mark's order, Matthew supports it. (They do agree, however, in not having some Markan material.) This means that Mark is the middle term in the relationship between the three: Mark is closer to Matthew and Luke than they are to each other.

While the above is true, it obscures the fact that Luke departs from Mark’s order far less often than does Matthew, and it is also worth noting that the order in Q (as usually hypothesized) also largely follows Luke, to the point where since the mid 1980s verses in Q have usually been referenced using the parallel in Luke, e.g. Q 13:34-35 means the Q parallel to Lk 13:34-35. In ‘The Reconstruction of Q and IQP / CritEd Parallels’ Frans Neirynck wrote:

It is not my intention here to contest the use of Lukan chapter and verse numbers in the citation of Q-passages. That this practice was so readily accepted has to do with the assumption that the sequence of Q is better preserved in Lk.

To many people it seems more likely that aMatthew would collect together and re-order material from either Mark or Luke than that aLuke would break-up the discourses in Matthew. By the same token it seems unlikely that aLuke would re-order Early Luke if the triple tradition material that was also in Early Luke was in the same order as in Matthew. As Luke supports the order in Mark when Matthew departs from it, it is reasonable to hypothesize that any viable candidate for Early Luke would also support the order in Mark for any text that has parallels in Mark.

Text in Mark but not Luke

Text in Mark but not Luke (whether or not it is in Matthew), is more problematic, although no more so than on the MwQH or the Mark-Q theory, as we have to explain why either aLuke, or both aMatthew and aLuke, saw this text in Mark but did not use it. If the text was not in Early Luke then it is easy to explain aLuke, and possibly aMatthew, not using it, since it would have only been present in one of their sources. If the text was in both Mark and Early Luke then it is harder to explain why either aMatthew or aLuke would ignore text they saw in both sources, so the better option is to exclude this text from Early Luke. However, this does then raise the question of why aeLuke would exclude this text (e.g. The Great Omission), and one possibility is that this text was not included in Early Luke because of damage to the copy of Mark from which it was derived (see Fatigue in Mark – or Something Else?).

Text in Mark and Luke, but not Matthew

Where Matthew does not contain text that is in Mark, aLuke could not have seen it in Matthew, and must therefore have obtained his version of the text either from Mark (e.g. Mark 1:35-38, 9:38-41) or Early Luke. If all of this text that was not in Matthew was in Early Luke then it is possible that aLuke did not need to have seen Mark, but if any of it was not in Early Luke then aLuke must have seen Mark. If the text was in Early Luke then aMatthew saw it in both his sources (Mark and Early Luke) making it hard to explain why he did not use it, whereas if it was not in Early Luke then aMatthew simply made a choice to go with the shorter of the two (from Early Luke), while aLuke made the opposite choice, and went with the longer (from Mark). On this basis the latter seems preferable, even though it requires aLuke to have used both Mark and Early Luke.

Text only in Matthew (Sondergut Matthew)

There is text in Matthew that is not found in either Mark or Luke, usually termed Sondergut (or Special) Matthew, that by definition cannot have come from canonical Mark or canonical Luke. This text must therefore have come from another source or sources, and in some synoptic hypotheses this is made explicit by including a source for Matthew typically termed ‘M.’ However, this should not be taken to mean that M was a single document, but instead it refers to an unknown number of sources, possibly both oral and literary.

Early Luke need not contain any of the Sondergut Matthew text since (by definition) it was not in Early Luke’s source (Mark). However, there is no actual barrier to having at least some of this text present in Early Luke, since in this case Early Luke would simply be acting as a source for Matthew outside the known synoptic tradition, i.e. as 'part of’ M. This means that we are free to include in Early Luke text that we see as Sondergut Matthew material if there is a legitimate reason why aLuke would have seen it in both Early Luke and Matthew but not used it in Luke.

However, on the assumption that Early Luke was the primary source for Luke, unless there is some other pressing reason why it should be included in Early Luke, excluding it is the more parsimonious option. In this case aLuke would have seen that only Matthew contained what we see as the Sondergut Matthew material, and chose not to include this text, perhaps because it was in only one of his sources. There is of course another alternative, which is that aLuke saw and used a version of Matthew to which some or all of the Sondergut Matthew text had not yet been added. However, in the absence of evidence for a version of Matthew that did not include the Sondergut Matthew text, this possibility will not be discussed here..

Text only in Luke (Sondergut Luke)

As with the M text we only see in Matthew, Early Luke need not contain any of the text we only see in Luke, since it was in neither Early Luke’s source (Mark), nor Matthew, which on the MwEL hypothesis depends at least in part on Early Luke. However, in the same way that Early Luke could contain some ‘M’ text, it could also contain some ‘L’ text (L is a hypothetical source of the Sondergut Luke text used in some synoptic solutions). Therefore, as with Sondergut Matthew we are free to include Sondergut Luke text in Early Luke if there is a legitimate reason why aMatthew would have seen it in Early Luke but not included it in Matthew. In this case we may reasonably assume that Mark was the primary source for Matthew so it may be easier to see why text in Early Luke but not Mark would not have been used by aMatthew. However, we know that aMatthew did not include all of Mark, so it would not be unusual for aMatthew to have excluded ‘L’ text if he saw it in Early Luke. Nevertheless, excluding it from Early Luke is the simpler option.

Text only in Mark

Text we see only in Mark is, by definition, in neither Matthew nor Luke. As the MwEL hypothesis assumes Markan priority then this implies that both aMatthew and aLuke saw this text in Mark, but nevertheless both chose not to include it in their respective gospels. If this text was in Early Luke then we would have the odd situation where aMatthew and aLuke both saw the text in Mark and Early Luke, but still chose to exclude it. However, if the text was not in Early Luke then aMatthew and aLuke would have just seen the text in Mark but not Early Luke. aMatthew may well therefore have chosen to follow Early Luke and omit the text, and if so then aLuke would have seen the text in only one of his three sources, making it likely that he would also omit the text. On this basis it is more reasonable to assume that Early Luke did not include this text.The question of why aeLuke would not have included this text is not part of this discussion.

Text not in Mark, Matthew, or Luke

It is possible for there to have been text in Early Luke that is not in any of the synoptic gospels. In the same way that we accept that aMark, aMatthew, and aLuke may have had sources other than the synoptic gospels themselves (of which the most well known is Q), so too may aeLuke have had sources other than Mark. It is perfectly possible for aMatthew to have seen this text in Early Luke and chosen to omit it because it was not in Mark, and aLuke then having chosen to omit it because it was in Early Luke but not Mark or Matthew. Although there is no reason to hypothesize such text, its existence in some other document does not on its own exclude that document from being a candidate for Early LukeHowever, the more text with no synoptic parallel we see in any candidate document for Early Luke, the more likely it is that the candidate actually post-dates the synoptic gospels.

To summarize, the starting point for determining the possible content of Early Luke is to assume the following:

  • Triple Tradition text: Present in Early Luke, but not necessarily identical to text in Luke. Determined on a case-by-case basis; 
  • Double Tradition text: Some in Early Luke, some not. Determined on a case-by-case basis; 
  • Text in Mark and Matthew but not Luke: Probably not in Early Luke; 
  • Text in Mark and Luke but not Matthew: Assumed to be in Early Luke unless shown otherwise;
  • Sondergut Matthew text: Unlikely to be in Early Luke. Although aLuke had seen it in Matthew but had chosen not to use it, if he had seen it in both Early Luke and Matthew it is more likely that he would have used it. It is therefore simpler to assume that it was not in Early Luke; 
  • Sondergut Luke text: Some possibly in Early Luke, and used by aLuke but not aMatthew. Although it is not necessary to exclude all of the Sondergut Luke text from Early Luke, it may be easier to assume that it was not in Early Luke; 
  • Text in Mark only: Not in Early Luke.
  • Text not in Mark, Matthew, or Luke: No need to hypothesize any such text in Early Luke.

As suggested above, these are not hard and fast rules, and in particular the existence of known variants and the versional evidence (e.g. the Old Latin, Syriac) may give rise to various exceptions. Note that these rules do not suggest that there ever was an Early Luke. Instead, they simply indicate that if Early Luke did ever exist, then it was most likely of a form described by these rules.

It should also be noted that although the MwEL hypothesis does not depend on whether aLuke knew anything regarding aeLuke, or even whether they were actually the same person, the case for aLuke using text from Early Luke in preference to parallel text from Matthew would be strengthened if aeLuke and aLuke were the same person. However, the discussion of this issue is outside the scope of this document.

Evidence for Early Luke

It must also be borne in mind that in the MwEL hypothesis Luke is a development of Early Luke, and therefore any valid text of Early Luke must also take account of the evidence in canonical Luke itself that suggests the existence of an early version of Luke, as noted above, and detailed below.

Luke Chapters 1 and 2

There is strong evidence (see Luke Chapters 1 and 2) to suggest that Luke did not originally contain chapters 1 and 2. We therefore need to consider what (if anything) Early Luke might have contained instead at this point. There are four possibilities:

  1. Early Luke contained an infancy narrative different to that in both Matthew and Luke, containing the basics from which both were developed
  2. Early Luke contained an infancy narrative similar to that in Matthew; 
  3. Early Luke contained an infancy narrative similar to that in Luke; 
  4. Early Luke did not contain an infancy narrative, i.e. it also began with what we know as Lk 3:1.

The first three possibilities all suffer from essentially the same problem: Given whatever form of infancy narrative aMatthew and/or aLuke saw in Early Luke, why did either or both alter it so much that what we see now in Matthew and Luke are so different? The fourth possibility given above avoids this problem, with neither Mark nor Early Luke containing an infancy narrative, but then requires that both aMatthew and aLuke wrote their own versions.

The Mark-Q theory does not suffer from the problems raised by the first three possibilities, but only because on this theory neither aMatthew nor aLuke saw each others gospel, and Q is specifically defined as not containing an infancy narrative, leaving both aMatthew and aLuke to write their own versions. However, on the MwEL hypothesis, even if Early Luke did not contain an infancy narrative, aLuke still saw the version in Matthew, and apparently rejected it in favor of his own story. One possible answer is to consider the curious fact that both Matthew and Luke actually do have infancy narratives, even though Mark does not. If Matthew and Luke are independent (as for example in the Mark-Q theory), then why did both aMatthew and aLuke, completely independently, decide to add an infancy narrative to their respective gospels in the first place? On the other hand it would make perfect sense if one saw the infancy narrative written by the other, and decided that his gospel also needed one.

On the MwEL hypothesis aLuke would have seen the infancy narrative in Matthew, so why would he not just embellish it, instead of completely re-writing it? On the assumption that neither Mark nor Early Luke contained an infancy narrative, it is not unreasonable for aLuke to regard that in Matthew as being largely an invention designed to connect Jesus with Old Testament prophesies, and perhaps to want to replace it with something else, based on his own sources. On the other hand, if Early Luke did contain an infancy narrative, then it would be natural for aLuke to have based his own one on what he saw in his primary source, i.e. Early Luke. However, in this case aMatthew would then also have seen the version in Early Luke, but because Mark (his primary source) did not contain an infancy narrative he would have then had to have rejected what he saw in Early Luke, and instead create an infancy narrative more to his own liking. On balance, it seems more reasonable to see Early Luke following Mark here by not having an infancy narrative (i.e. being without any equivalent to Luke chapters 1 and 2), and so having aLuke react only to the one he saw in Matthew.

There is also the question of whether Lk 1:1-4 would have been present in Early Luke. If it was present then we would need to explain why aMatthew removed it (perhaps only a minor problem), but more importantly, the only known written source on which Early Luke is based is Mark, rather than the ‘many’ delivered to aLuke as indicated in Lk 1:1-2. Both points suggest that Early Luke did not contain Lk 1:1-4.

The Genealogy

As is well known, the genealogy in Matthew and Luke partly run in opposite directions, with Luke going backwards in time, and while in Matthew the genealogy follows Mary’s lineage, in Luke it follows Jospeh’s. In addition, while Matthew begins with the genealogy and follows it with the infancy narrative and the baptism, Luke (ignoring Lk 1:1-4) begins with the infancy narrative and the baptism, followed by the genealogy. While it would make sense for a genealogy to follow the baptism if there was no preceding infancy narrative, it would be far less so if the infancy narrative was present as well. On this point, in a blog post entitled Did Luke Originally Have Chapters 1-2?, Bard D. Erhman writes:

The genealogy that is given in ch. 3 doesn’t make sense if the Gospel already had chs. 1-2. The genealogy is given *after* the baptism. But the natural place for a genealogy is at the point in which a person is *born* (since the genealogy traces the bloodline up to the time of birth), not at the point of baptism (as a 30 year old!). Without chs. 1-2, however, the genealogy makes sense at the baptism, since it is at the baptism that Jesus is made the son of God according to the voice from heaven, and so immediately afterward the genealogy is given, in which Jesus’ family line is traced not only to Adam (so that he is the son of Adam) but from Adam to God (so that he is the son of God).

On the assumption that the genealogy in Luke has not been moved its position makes it unlikely that Early Luke contained Lk 1-2 at the time that the genealogy was written.

Nazareth and Capernaum

In Luke Jesus preaches in Nazareth before Capernaum, while this does not happen in either Mark or Matthew. Given the evidence of both Mark and Matthew, and the inconsistencies in Luke, e.g. with Jesus in v. 4:13 referring to things he had done in Capernaum before (according to the narrative) he had even been there, it is highly likely that in Early Luke Jesus preached in Capernaum before Nazareth (See Capernaum or Nazareth First?). However, this does not necessarily mean that the Nazareth and Capernaum passages were ‘swapped’ wholesale. Instead, as in Adv. Haer. IV.23 Irenaeus reports that the reading from Isaiah in Lk 4:17 originally occurred in Capernaum, not Nazareth, it seems likely that just the names of the locations may have been reversed in Luke, and not the actions themselves.

It therefore seems likely that Early Luke would have Capernaum before Nazareth, but it is not so clear whether the reading from Isaiah would have been present in Early Luke or not. However, the fact that this passage is not in either Mark or Matthew does suggest that it was also not in Early Luke (otherwise why would aMatthew have excluded it?), which means that it was added to Luke after Early Luke, but (from Irenaeus’ comment) before the names of Capernaum and Nazareth were swapped. This in turn suggests that there may have been another interim version of Luke (not seen by aMatthew, and so not affecting the synoptic problem) between Early Luke and canonical Luke.

Luke 4:14b-15

This text is only required in canonical Luke because Nazareth is placed before Capernaum (see Well Known in Galilee - Twice), and so it would not have been present in Early Luke.

Simon’s Mother-in-Law

In Lk 4:38-39 Jesus heals Simon’s mother-in-law, and then in Lk 5:3 Jesus (apparently) meets Simon for the first time, when he is washing his net. Both events take place in all three synoptics, and so almost certainly would be present in Early Luke, but in Mark and Matthew Jesus meets Simon (casting a net rather than washing it) before healing his mother-in-law. Either the two incidents were present in Early Luke in the Mark, Matthew order and were reversed later, or they were already reversed in Early Luke. Given the Nazareth – Capernaum reversal just referred to, it is more likely that the ‘Simon’ reversal took place at the same time, i.e. that in Early Luke the two incidents were in the Mark, Matthew order.

The Great Omission

Most of Mk 6:45-8:26 is paralleled in Mt 14:22-16:11, with the events in these verses in Matthew appearing in the same order as in Mark. However, at the corresponding place in Luke (v. 9:18), there appears to be no trace of these 75 contiguous verses from Mark, a feature that is known as The Great Omission. Assuming that aLuke knew Mark, there are basically two hypotheses as to why these verses have no parallels in Luke: Either aLuke chose to exclude the whole block of text (for whatever reason), or the copy of Mark that he saw did not contain it.

The majority opinion favors the former explanation (supported by the fact that all ms of Mark that are extant for this part of Mark include at least some of these verses), even though there is no other instance of aLuke apparently excluding such a long section of Mark. It also suggests a remarkable aversion by aLuke to all the events in these verses, because even though (at least on the MwQH) he had a ‘second chance’ to include some or all of them on seeing the parallel versions in Matthew, he still chose not to do so. However, if the verses were not in the copy of Mark seen by aLuke, then from his perspective they would have no more support than the other verses unique to Matthew that he chose not to include, and so there would have been no specific reason why he would have selected for inclusion these verses from Matthew but not others.

On the MwEL hypothesis aLuke could potentially have seen a version of these verses in Early Luke as well as in Mark. On this hypothesis there are three possible authors who saw and used Mark: aMatthew, aeLuke, and aLuke, and the arguments for whether aeLuke would have seen the text of the Great Omission in Mark differ depending on whether aLuke saw it in Mark or not:

1.     aLuke’s copy of Mark contained the text of the Great Omission
On this scenario aLuke saw both the text in Mark and the parallel text in Matthew, but still chose to exclude it from Luke. Whether or not Early Luke also included parallel text is then unlikely to be an issue: aLuke didn’t want the text however many versions he saw. On this basis we are free to include it in Early Luke, so avoiding any issue regarding the possibility of there being a copy of Mark seen by aeLuke that did not contain the text of the Great Omission.

2.     aLuke’s copy of Mark did not contain the text of the Great Omission
On the MwQH aLuke did not see the text in Mark, and then chose to not include it when he subsequently saw the parallels in Matthew. On the MwEL hypothesis aLuke would also have seen the text in Matthew but not in Mark, but may or may not have seen it in Early Luke. If it was also in Early Luke (which by definition was an early copy of Luke, and so most likely aLuke’s primary source) then aLuke would probably have included it, while if it was not in Early Luke then aLuke would have excluded the parallels he saw in Matthew. On this basis the text of the Great Omission should be excluded from Early Luke.

This then raises the question of whether aeLuke would have seen the text of the Great Omission in Mark. If aeLuke did see it then his copy of Mark most likely contained the same text as that seen by aMatthew. However, if this was the case then aeLuke chose not to include it for unknown reasons, although possibly similar to those that aLuke had for not including the parallel text from Matthew. In this scenario aLuke and aeLuke cannot be the same person (one saw the text in Mark while the other did not), so the chances of both seeing the text but rejecting it seem very slim, and the possibility that aeLuke saw the text in Mark can be rejected.

That leaves us with the possibility that neither aLuke nor aeLuke saw the text of the Great Omission in Mark. If aLuke and aeLuke were actually the same person then this needs no more thought, but if not then we now have two different people seeing (possibly) the same actual ms of Mark, or (much less likely) two mss of Mark with exactly the same ‘missing’ portion of text.

The most straightforward possibility is that aeLuke and aLuke both saw a copy of Mark essentially the same as that seen by aMatthew, i.e. a copy containing the text of the Great Omission. The alternative is that aeLuke and aLuke both saw the same copy of Mark, a copy missing the text of the Great Omission. Which alternative is more likely can only be determined by resolving the whole question of the origin of the Great Omission, which is outside the scope of this investigation.

The Lord’s Prayer

Lk 11:2-4 contain the Lukan version of the Lord’s Prayer. The parallel in Matthew, at Mt 6:9-15, is significantly longer, and is generally regarded as an expanded version of the one in Luke. However, the Lord’s Prayer in Luke exists in two forms. In the KJV Lk 11:2-4 read:

And he said unto them, When ye pray, say, Our Father which art in heaven, Hallowed be thy name. Thy kingdom come. Thy will be done, as in heaven, so on earth. [11:2]
Give us day by day our daily bread. [11:3]
And forgive us our sins; for we also forgive every one that is indebted to us. And lead us not into temptation, but deliver us from evil. [11:4]

The highlighted text is found in the majority of mss of Luke, but the shorter version without the highlighted text is considered to be original, being found in P75, B, L, and some other important mss. For example, Metzger comments on Lk 11:2:

In view of the liturgical usage of the Matthean form of the Lord’s Prayer, it is remarkable that such a variety of early witnesses managed to resist what must have been an exceedingly strong temptation to assimilate the Lukan text to the much more familiar Matthean form. It is not surprising, therefore, that the great majority of witnesses read Πάτερ ἡμῶν ὁ ἐν τοῖς οὐρανοῖς, as in Mt 6.9.

As described by Peter Head, Streeter saw these differences as proof that the Matthean and Lukan versions of these verses must have had their origins in different sources: 

Streeter appeals to text critical factors in connection with both Q and the Minor Agreements, but the basis for both those is his conviction that the assimilation of the texts of the Gospels to one another “is the commonest of all forms of error”… Streeter brings this to bear on the reconstruction of Q, and especially the question as to whether Q contained the Lord’s Prayer. Essentially he argues that “in view of the immense pressure of the tendency to assimilate the two versions of this specially familiar prayer”, we should follow those texts of Luke which are least harmonised to Matthew, not only B for the most part, but also 700 and 162 for the reading “Thy Holy Spirit come upon us and cleanse us” (a reading known from Marcion and other Fathers). Once the differences between the Lukan and Matthean versions are recognised it becomes unlikely that they were both derived from the same written source; since in Matthew the Lord’s Prayer occurs within the middle of an M block; and since in Luke the Lord’s Prayer occurs within the middle of an L block, the most natural conclusion is that the different versions of the Lord’s prayer come from these special sources.

However, as noted above, it is possible that at least some text of the Lord’s Prayer that Streeter considered to have originated in either M or L could in fact have been present in Early Luke, e.g. on the MwEL hypothesis the shorter text in L may well have preceded both Matthew and Luke. Other sources agree with Streeter that these shorter readings in Luke are likely to be original, with, for example, the NET noting the following regarding Lk 11:2: 

Most mss, including later majority (A C D W Θ Ψ 070 Ë13 33vid Ï it), add ἡμῶν ὁ ἐν τοῖς οὐρανοῖς (Jhmwn Jo en toi" oujranoi", “our [Father] in heaven”) here. This makes the prayer begin like the version in Matt 6:9. The shorter version is read by Ì75 א B (L: + ἡμῶν) 1 700 pc as well as some versions and fathers. Given this more weighty external evidence, combined with the scribal tendency to harmonize Gospel parallels, the shorter reading is preferred.

Bob Waltz makes a similar observation regarding harmonization: 

But it is the internal evidence that is absolutely decisive. The longer reading is, of course, that found in Matthew 6:9, and in Matthew there is no variation. Equally important, every one of these copyists must have known his paternoster, and they would all know it in Matthew's form (since it is at once fuller and earlier in the canon). If they found a short form in Luke, they would inevitably have been tempted to flesh it out. And under no circumstances would they ever have removed the longer words. Thus it is morally certain that the short form is original (here and in the several other expansions found in the Lukan version of the Lord's Prayer).

A similar apparent harmonization from Mt 6:10 occurs at the end of Lk 11:2. The NET again: 

Most mss (א A C D W Θ Ψ 070 Ë13 33vid Ï it) read at the end of the verse [following “may your kingdom come”] “may your will be done on earth as [it is] in heaven,” making this version parallel to Matt 6:10. The shorter reading is found, however, in weighty mss (Ì75 B L pc), and cannot be easily explained as arising from the longer reading.

Also, in Lk 11:4: 

Most mss (א1 A C D W Θ Ψ 070 Ë13 33 Ï it syc,p,h) add “but deliver us from the evil one,” an assimilation to Matt 6:13. The shorter reading has better attestation (Ì75 א*,2 B L 1 700 pc vg sa Or). Internally, since the mss that have the longer reading here display the same tendency throughout the Lord’s Prayer to assimilate the Lukan version to the Matthean version, the shorter reading should be regarded as authentic in Luke.

Bezae has some unique additions in Lk 11:2. After “when ye pray,” it adds the following:

do not babble repetitiously like the pagans, for they think that because of their many words they will be heard. But when you pray,

Then, in place of “Thy kingom come” Bezae reads: Let thine kingdom come upon us.

As these variants only exist in one extant ms, and only the first addition has a parallel (in Mt 6:7), it is unlikely that this text is the source of Mt 6:7. Apart from these differences specific to Bezae, the other differences suggest that the longer variants in Luke are assimilations to, and thus later than, Matthew. This then presents a problem, since if the original version of Luke was also later than Matthew, it is hard to explain why aLuke would have removed these short pieces of text. However, on the MwEL hypothesis, if Early Luke contained the shorter text then it is easy to see how aLuke could have kept that text, while aMatthew expanded and changed what he saw in Early Luke. Later copyists then expanded the version in Luke by harmonizing the text to Matthew. 

However, we do not see any parallel to Mt 6:13b in Luke. The NET notes: 

Most mss (L W Θ 0233 Ë13 33 Ï sy sa Didache) read (though some with slight variation) ὅτι σοῦ ἐστιν ἡ βασιλεία καὶ ἡ δύναμις καὶ ἡ δόξα εἰς τοὺς αἰῶνας, ἀμήν (“for yours is the kingdom and the power and the glory forever, amen”) here [at the end of Mt 6:13]. The reading without this sentence, though, is attested by generally better witnesses (א B D Z 0170 Ë1 pc lat mae Or). The phrase was probably composed for the liturgy of the early church and most likely was based on 1 Chr 29:11-13; a scribe probably added the phrase at this point in the text for use in public scripture reading (see TCGNT 13-14). Both external and internal evidence argue for the shorter reading.

The absence of any parallel to these words in any extant ms of Luke may simply suggest that aLuke never saw them in Matthew, but even if he did, their absence in Early Luke could have made it very unlikely that he would use them.

The Last Supper

Another order difference can be found in Lk 22:17-20, where the verses in Luke in the majority of mss contain the sequence ‘cup-bread-cup,’ whereas in both Mark and Matthew it is just ‘bread-cup.’ There are actually several different variants of these verses in Luke, in most of which some or all of Lk 22:19b-20 (which have close parallels in 1 Cor 11:24b-25) are not present. In addition, in some variants the text that is present follows the Mark/Matthew order. Overall, there appears to be no consensus as to whether the longest or the shortest variant is original. However, when considered as a ‘trajectory’ from Mark => Matthew => Luke, it is clear that the shorter variants in Luke are closer to Mark and Matthew than the longer. Therefore, the text in Early Luke would have been similar to that in Mark and Matthew, with the longer variants (including text possibly obtained from 1 Cor 11) being created later.

The Resurrection Accounts

The final chapters of the synoptic gospels exhibit an interesting pattern. The early verses (Mk 16:1-8, Mt 28:1-8, and Lk 24:1-9) are broadly similar, and are clearly parallel accounts of the same events. Although some of the details vary, all three resurrection accounts describe a number of female followers of Jesus coming to his burial place, seeing the stone rolled away from the entrance, finding that Jesus’ body had gone, being told by one or two men (or an angel) that Jesus was risen, that (in Mark and Matthew) they would see him in Galilee, and then leaving. However, at this point the three accounts diverge, with each synoptic having a very different ending.

The very similar first halves of each of the resurrection accounts in Mark, Matthew, and Luke are easy to explain on the basis of a common literary source, but the very different second halves show almost no literary commonality, with significant material unique to Matthew, and even more that is unique to Luke. However, there is one big difference: Although Matthew and Luke contain complete accounts, the origin of the second half of the account in Mark is disputed. Although most extant mss of Mark end with vv. 16:9-20 (known as the long ending of Mark), the majority opinion is that this ending is secondary, and that Mark originally ended at v. 16:8 (a few mss have another, much shorter, ending, which in almost all cases is followed by the long ending as well).

This immediately suggests a reason for the lack of commonality seen in the second half of the resurrection accounts: The first halves of the accounts in Matthew and Luke are based on that in Mark, but Mark contained nothing upon which aMatthew and aLuke could base the second halves of their accounts, leaving them to find other sources for their material.

On the Mark-Q theory there is no other source, as not only is Q defined as not having a resurrection account, but neither aMatthew nor aLuke knew each other’s gospels. While this accounts for the differences between the second halves of the accounts in Matthew and Luke, it does not explain why there are nevertheless three significant parallels. On the MwQH the parallels are easily accounted for by aLuke knowing and using Matthew, but this then makes it very hard to explain why almost everything else in the second half of the accounts is material unique to either Matthew or Luke.

In contrast, on the MwEL hypothesis both the similarities and differences are easily explainable as a consequence of Mark originally ending at v. 16:8, and Early Luke containing a resurrection account that did not have most of the material unique to Luke that we see in Lk 24:10-53, but instead had close parallels to what we see as Mk 16:9-20. aLuke then simply added his own material to the ending he saw in Early Luke, while aMatthew (not seeing any ending in Mark, his main source), rejected most of what he saw in Early Luke in favor of his own unique additions. Mk 16:9-20 was then later created by making relatively simple changes to the ending of Early Luke that were needed to blend the two accounts. For details see The Long Ending of Mark.

Doublets in Matthew and Luke

Both Matthew and Luke contain doublets, which are essentially two different versions of the same passage, one version of which appears to originate in Mark, and the other from, presumably, a second source (A useful list of doublets can be found here http://www.textexcavation.com/doublets.html, and see also the Introduction to Doublets).The need for a second source is a major reason for the popularity of synoptic solutions that hypothesize an additional source for both Matthew and Luke, such as Q.

On the MwQH, although Luke has two sources (Mark and Matthew), Matthew only has one (Mark). This makes it impossible to account for doublets in Matthew other than as redactional activity by aMatthew, or by suggesting that he had some other (unspecified) written or oral source that overlapped with Mark in places. Neither of these explanations is satisfactory, because neither can be either supported or falsified, and a similar problem affects hypotheses in which Luke has only one source, such as the Griesbach (or Two-Gospel) hypothesis. Alternatively, although on the Mark-Q theory both Matthew and Luke have Mark and Q as sources, the fact that this hypothesis is symmetrical as to the sources makes it hard to explain why aMatthew and aLuke created different doublets from these identical sources, other than by hypothesizing different redactional choices.

In contrast to both the MwQH and the Mark-Q theory, the MwEL hypothesis provides a simple and natural explanation for the existence of different doublets in Matthew and Luke: For doublets in Matthew one of each pair is taken directly from Mark by aMatthew, while the other is derived from a different version of the text that aeLuke created from Mark. aLuke also has two main sources: Matthew and Early Luke (potentially three if he saw Mark), with the use of Matthew allowing for the creation of different doublets in Luke. This hypothesis also explains doublets in Luke where only one version exists in Matthew, because here one of the versions is based on what aLuke saw in Matthew, while the other is based on something he saw in Early Luke that was not used by aMatthew.

Examples of doublets in Matthew and/or Luke are given below.

The missions of the twelve and the seventy (Doublet in Luke)

In Lk 9:1-5 Jesus calls together the twelve disciples and sends them out to preach, and there is a similar passage in Mt 10:5-14. Jesus gives brief instructions to the twelve regarding how to equip themselves for their missions, and what to do in the houses of the cities they enter. Later, in Lk 10:3-16, Jesus gives similar, but more extensive, instructions to the seventy(-two) disciples before they set out on their missions. These verses have no parallel in Mark, but Matthew 10 contains parallels similar to these more extensive instructions in Luke 10, except that Matthew does not contain a sending of the seventy disciples, and instead these instructions are given to the twelve. It is clear that these three different sendings are related, but in what way? The parallels (marked by //) are as follows:

 Mark                Luke                        Matthew
                                 Lk 10:3            Mt 10:16a
Mk 6:8a    Lk 9:3a // Lk 10:4a          Mt 10:9
Mk 6:8b    Lk 9:3b // Lk 10:4b          Mt 10:10a
Mk 6:9      Lk 9:3e // Lk 10:4c          Mt 10:10b
                                 Lk 10:5-6         Mt 10:12-13
                                 Lk 10:7            Mt 10:10b
Mk 6:10    Lk 9:4   // Lk 10:8            Mt 10:11
                                 Lk 10:9            Mt 10:8a,7
Mk 6:11a   Lk 9:5a // Lk 10:10-11a   Mt 10:14
Mk 6:11b                   Lk 10:12          Mt 10:15    [Mk 6:11b – Possible assimilation]
                                  Lk 10:16a       Mt 10:40a

The sending of the twelve appears to have originated in Mark 6, while the sending of the seventy appears to be a version of something that on the MwQH originated in Matthew 10, or comes from SS, e.g. Q or Early Luke. The significant point here is the doublets in Luke, and when these and their parallels are isolated it is easier to see the pattern that they form:

Mk 6:8a      Lk 9:3a // Lk 10:4a        Mt 10:9
Mk 6:8b      Lk 9:3b // Lk 10:4b        Mt 10:10a
Mk 6:9        Lk 9:3e // Lk 10:4c        Mt 10:10b
Mk 6:10      Lk 9:4   // Lk 10:8           Mt 10:11
Mk 6:11a    Lk 9:5a // Lk 10:10-11a  Mt 10:14

In Mark, Matthew, and both Luke 9 and 10 these verses are in order (but not necessarily adjacent), strongly suggesting that these orders are original (i.e. they have not been rearranged), and with other verses being added at a later stage to form what we see in Matthew 10 and Luke 10. Even more notable are the parallels Mk 6:8-11a and Lk 9:3-5a, a strong indication that, assuming Markan priority, Lk 9:3-5a have Mk 6:8-11a as their source. It is also clear that the whole of Mt 10:7-16a and Lk 10:3-12 are strongly related, but as both Matthew and Luke contain additional parallels (when compared with Lk 9:3-5a) it appears either that one is a development of the other, or both are developments of another source. There are three ways to look at this, depending on the synoptic hypothesis being assumed, but (assuming Markan priority) in all three Lk 9:3-5a depends on Mk 6:8-11:
  1. On the Mark-Q hypothesis the versions of this passage in Mt 10:7-16 and Lk 10:3-12 both depend on a longer version of the sending in Q, and because Q must have included parallels to Mk 6:8-11 they are Mark-Q Overlaps. These are places at which Mark and Q contain parallel text and so ‘overlap’ in places (contrary to the Mark-Q hypothesis itself), and the multiple overlaps raise the issue of whether Mark and Q were independent or not.
  2. On the Farrer theory / MwQH there is no second source, and instead aMatthew saw Mk 6:8-11 and expanded it to form Mt 10:7-16, while aLuke created Lk 9:3-5a from what he saw in Mark 6, but decided to also merge the Markan text with the longer version he saw in Mt 10:7-16 (re-ordering it in the process), to form Lk 10:3-12.
  3. In any hypothesis which includes a non-Q SS that is allowed to overlap with Mark (e.g. the MwEL hypothesis), a natural corollary is that the author of SS can obtain what we know as Lk 9:3-5a from Mark 6 (as also on the MwQH). aMatthew can then expand this into Mt 10:7-16, which aLuke can re-order and use as Lk 10:3-12.

Delivered Up (Doublet in Luke)

Mk 13:11 contains the following:

But when they shall lead you, and deliver you up, take no thought beforehand what ye shall speak, neither do ye premeditate: but whatsoever shall be given you in that hour, that speak ye: for it is not ye that speak, but the Holy Ghost.

This verse is paralleled in Mt 10:19-20, and at two different places in Luke: Lk 12:11-12 // 21:14-15, the former of which is a close parallel to Mark:

And when they bring you unto the synagogues, and unto magistrates, and powers, take ye no thought how or what thing ye shall answer, or what ye shall say: [Lk 12:11]
For the Holy Ghost shall teach you in the same hour what ye ought to say. [Lk 12:12]

However the latter, while still a parallel, is much less close to the original, omitting approximately half of Mk 13:11, and adding text at the end of Lk 21:15 that has no parallel in Mark.

Settle it therefore in your hearts, not to meditate before what ye shall answer: [Lk 21:14]
For I will give you a mouth and wisdom, which all your adversaries shall not be able to gainsay nor resist. [Lk 21:15]

On the MwQH it is hard to explain why aLuke created two different versions of this text, one of which is significantly different to the parallels in both Mark and Matthew. On the Mark-Q theory it is possible that Lk 21:14-15 come from Q, but in this case there is no explanation for why aLuke included something from Q that aMatthew did not. On the MwEL hypothesis Lk 21:14-15 is a version of Mk 13:11, and it is easier to see why aMatthew might choose the version in Mark (his primary source) over that in Early Luke, while aLuke would choose to include both the Mark/Matthew version and the altered/relocated one from his primary source, Early Luke.

The Sign of Jonah (Doublet in Matthew)

Mk 8:12 reads: “... Why doth this generation seek after a sign? verily I say unto you, There shall no sign be given unto this generation.” Mt 16:4 contains a parallel to which a ‘rider’ (in bold) is added: “A wicked and adulterous generation seeketh after a sign; and there shall no sign be given unto it, but the sign of the prophet Jonas. And he left them, and departed.

Mt 12:39 also contains a parallel similar to that in Mt 16:4: “But he answered and said unto them, An evil and adulterous generation seeketh after a sign; and there shall no sign be given to it, but the sign of the prophet Jonas.” This is then followed in Mt 12:40-42 by an explanation of the sign. Lk 11:29 then follows Mt 12:39, reading: “... he began to say, This is an evil generation: they seek a sign; and there shall no sign be given it, but the sign of Jonas the prophet,” followed by the same explanation as in Mt 12:40-42, although with the text in a different order.

On the MwQH it is easy to see how Luke could contain a parallel to what aLuke saw in Mark, but not how aMatthew (who only had one source) created two in Matthew, and on the Mark-Q theory it would seem to require a considerable degree of overlap between the two sources. However, on the MwEL hypothesis this situation simply requires a parallel in Early Luke similar to that in Mt 16:4, but re-located to a different point in aeLuke’s narrative. On this scenario aMatthew would have seen both Mk 8:12, and the parallel in Early Luke containing the reference to Jonas. He then chose to include both in Matthew, but also, perhaps for consistency, chose to include the OT reference to Jonas in both places, to the first of which he added an explanation. aLuke then chose the longer version from Matthew, but some scribes/copyists of Luke also knew the version from Early Luke, accounting for the variants in Lk 11:31-32 found in some mss.

The Little Omission (Doublet in Matthew)

The term “The Little Omission” is sometimes used to refer to the fact that none of Mk 9:41-10:12 has any parallel in Luke. However, most of these verses do have parallels in Matthew, including Mk 9:43, 45, 47:

And if thy hand offend thee, cut it off: it is better for thee to enter into life maimed, than having two hands to go into hell, [Mk 9:43a]
[into the fire that never shall be quenched:] [Mk 9:43b]

And if thy foot offend thee, cut it off: it is better for thee to enter halt into life, than having two feet to be cast into hell, [Mk 9:45a]
[into the fire that never shall be quenched:] [Mk 9:45b]

And if thine eye offend thee, pluck it out: it is better for thee to enter into the kingdom of God with one eye, than having two eyes [Mk 9:47a]
[to be cast into hell fire:] [Mk 9:47b]

The ends of each of these verses, shown above enclosed in [], are considered to be part of the longer Byzantine text, and most critical editions do not include these words. In addition to these longer endings, the Byzantine text also includes these verses: 

Where their worm dieth not, and the fire is not quenched. [Mk 9:44, 46, 48]

As stated above, Mk 9:43-48 have no parallels in Luke, while Mk 9:43, 45, 47 have two different sets of parallels in Matthew, one of which just refers to hell, while the other refers to the fires of hell:

And if thy right eye offend thee, pluck it out, and cast it from thee: for it is profitable for thee that one of thy members should perish, and not that thy whole body should be cast into hell. [Mt 5:29]
And if thy right hand offend thee, cut it off, and cast it from thee: for it is profitable for thee that one of thy members should perish, and not that thy whole body should be cast into hell. [Mt 5:30]

Wherefore if thy hand or thy foot offend thee, cut them off, and cast them from thee: it is better for thee to enter into life halt or maimed, rather than having two hands or two feet to be cast into everlasting fire. [Mt 18:8]
And if thine eye offend thee, pluck it out, and cast it from thee: it is better for thee to enter into life with one eye, rather than having two eyes to be cast into hell fire. [Mt 18:9] 

Assuming Markan priority Mk 9:43, 45, 47 would be expected to have been seen by both aMatthew and aLuke, so any valid synoptic hypothesis needs to be capable of explaining why Matthew has two different versions of these verses, while Luke has none. Even though the differences between the doublets in Matthew suggest perhaps that one is derived from the shorter text in Mark, and the other from the longer Byzantine text, this does not explain the lack of any parallels to Mk 9:44, 46, 48 in Matthew, nor does it suggest how aMatthew might have seen both the longer and shorter text of Mk 9:43, 45, 47.

This is a particularly difficult point for the Mark-Q theory, since by definition Mt 5:29-30 and 18:8-9 (with no parallels in Luke, and so not being part of the Double Tradition) cannot have had parallels in Q, and so there is no additional written source in which aMatthew could have seen a second version of these verses. With regard to aLuke, it is possible for him to have simply decided to not include parallels to Mk 9:43-48 for a reason that made sense to him, but it is also conceivable (as some believe) that he simply did not see any of Mk 9:41-10:12 in his copy of Mark, but neither of these possibilities can be either verified or falsified.

The MwQH faces even greater difficulties over this doublet. Again, there is the problem of aMatthew creating two different versions of these verses from Mark (his only written source), but on this hypothesis aLuke would have seen the doublet in Matthew whether or not his copy of Mark included Mk 9:41-10:12, making it even harder to explain why he would not include any parallels in Luke.

On the MwEL hypothesis the situation regarding aMatthew seems straightforward: Mark contained the original of these verses, Early Luke contained a second version, and aMatthew included parallels of both in Matthew. However, aLuke would then have seen the doublet in Matthew, the version in Early Luke, and (possibly) the original in Mark, so why would he not have included at least one parallel in Luke? One possibility is that having seen at least three versions he was simply unsure of which one(s) to trust, and this would be even more likely if his copy of Mark did not include these verses. Alternatively, if Early Luke did not contain a parallel of these verses (for whatever reason), then the situation is effectively the same as on the MwQH.

None of these hypotheses provides a ready explanation for both the doublet in Matthew and the lack any parallel in Luke, even if aLuke did not see Mk 9:41-10:12 and (on the MwEL hypothesis) also parallels to these verses in Early Luke. This problem appears to be tied up with the variants in Mk 9:43-48, and it is conceivable that aMatthew saw the different variants in two different copies of Mark. Alternatively, and perhaps more likely, one of the variants came from an oral source to which aLuke had no access.

Take up your cross (Doublets in both Matthew and Luke)

In ‘Excavating Q,’ Kloppenborg notes:

In several instances, Matthew has small collections of sayings that are dispersed in Luke. Nevertheless, Matthew reproduces the sayings in Lukan order, as if he had scanned Q, collecting sayings that he thought were related and might fit together well. The clearest instance of this is found in Matt 10:24-39, comprising ten Q sayings:

Matt 10:24-25   Disciple and teachers         Luke 6:40
Matt 10:26        Revelation of the hidden     Luke 12:2
Matt 10:27        What is said in the dark      Luke 12:3
Matt 10:28-31   Do not fear                          Luke 12:4-7
Matt 10:32-33   Confessing Jesus               Luke 12:8-9
Matt 10:34        Casting fire on the earth     Luke 12:49
Matt 10:35-36   Families divided                 Luke 12:51-53
Matt 10:37        Hating one’s parents          Luke 14:26
Matt 10:38        Carrying the cross              Luke 14:27
Matt 10:39        Saving/losing one’s life       Luke 17:33

After commenting on why it is unlikely that aLuke might instead have broken up what was a Matthean sequence in Q and “scattered them throughout Q” (surely meaning Luke here?), Kloppenborg states that:

It should be obvious that it is much simpler to suppose that Matthew collected and organized sayings than to think that Luke broke up originally unified clusters and used the debris in such a counterproductive manner.

While this on the face of it this seems reasonable, Kloppenborg omits the fact that several of these verses (both in Matthew and Luke) are one half of a doublet, with the other halves having parallels in Mark that suggest a different original ordering in another source (SS):

     Mk           Mt (Mk)       Mt (SS)      Lk (Mk)       Lk (SS)
Mk 8:34a     Mt 16:24a                                          Lk 14:25       Saying …
                                        Mt 10:37                        Lk 14:26       Hate (or love less than me) one’s parents
Mk 8:34b     Mt 16:24b // Mt 10:38      Lk 9:23b // Lk 14:27       (Not) Carrying the cross
Mk 8:35       Mt 16:25   // Mt 10:39      Lk 9:24   // Lk 17:33       Saving/losing life
Mk 8:36       Mt 16:26a                        Lk 9:25                           How is a man helped?
Mk 8:37       Mt 16:26b                                                               Exchanging a soul

                                        Mt 10:24a,25a               Lk 6:40         Disciple and teachers
                                        Mt 10:26                        Lk 12:2         Revelation of the hidden
                                        Mt 10:27                        Lk 12:3         What is said in the dark
                                        Mt 10:28a,c-31              Lk 12:4-7      Do not fear, sparrows
                                        Mt 10:32                        Lk 12:8         Confessing Jesus
Mk 8:38                           Mt 10:33       Lk 9:26   // Lk 12:9         Being ashamed / denying Jesus

                                        Mt 10:34,36,35               Lk 12:51-53 Families divided

Matthew 16:24b and Luke 9:23b both contain almost identical parallels to Jesus’ words from Mark 8:34b: “If any man will come after me, let him deny himself, and take up his cross, and follow me” (Luke adds “daily” after “cross”), and assuming Markan priority these parallels are easily explained. However, there are two other looser parallels of these verses, both of which are the ‘counterpoint’ to Mark 8:34b, with Jesus stating that not taking up your cross and following him prevents you from being worthy of him (Mt 10:38) or being his disciple (Lk 14:27), neither of which have a parallel in Mark.

On any hypothesis with no SS, e.g. the MwQH, this would require aMatthew to have done one of two things:
  • Redacted Mk 8:34b to create the counterpoint at Mt 10:38, but then later in his gospel written Mt 16:24b, the close parallel to Mk 8:34b; or 
  • Created and added Mt 10:38 AFTER already having written the parallel at Mt 16:24b.
Both alternatives seem unlikely, and both also require aLuke to have decided to include his own parallel to the counterpoint even though he did not see one in Mark. Without any SS these actions (by Matthew in particular) are very hard to understand, even when allowing that aMatthew and aLuke are free to edit their sources in whatever way they choose. This is of course a problem for the MwQH, in which there is no SS.

In contrast, on the assumption of a second source then it is easy to see that SS is likely to be the source of the counterpoint: On the Mark-Q theory Mt 10:38 // Lk 14:27 originated in Q, while on the MwEL hypothesis they originated in Early Luke or (much less likely, for reasons just discussed) Mt 10:38 came from aMatthew and aLuke created his own version. Given the other parallels also shown above the obvious inferences are that:
  • Mt 16:24-26 and Lk 9:23b-26 are derived from Mk 8:34b-38 (with the unanswered question from Mk 8:37 not asked in Luke); 
  • Mt 10:37-39, 24-33 and Lk 14:25-27, 17:33, 6:40, 12:2-9 are derived from SS. 
On the Mark-Q theory (where SS is Q) this represents a significant Mark-Q overlap. The lack of parallels to Mk 8:37 // Mt 16:26b in both Luke and Q may simply mark a difference between sayings in Mark and Q, but even allowing such differences it is still odd that the parallel to Mt 10:39 is located at Lk 17:33. In a private exchange David Sloan wrote:

I also think one of the strangest things about the current state of Q scholarship is that the location of Q 17:33 is dictated by Mark’s placement of its parallel after Mark’s parallel to Q 14:27. So you will see in the Critical Edition of Q that these verses occur together. I would rather think that Luke 14:28-33 follows the same source as Luke 14:26-27, 34-35 (as I argue in my JSNT article), and that Q 17:33 belongs within Q 17:20-37.

Sloan is pointing out that the IQP’s Critical Edition of Q contains the sequence below, in which Q 14:27 and Q17:33 follow the order of Mk 8:34b-35:

Q 14:27      Taking One’s Cross (Mk 8:34b)
Q 17:33      Losing One’s Life     (Mk 8:35)
Q 14:34-35 Insipid Salt               (Mk 9:50)

The Markan parallel to Q 14:34-35 (Salt) is Mk 9:50 (not Mk 8:36), so there is nothing following Q 17:33 that suggests it should be located in its current position. Instead, it would be natural to locate it between Q 17:30 and 34. Although this would move Q17:33 closer to the end of Q, the Matthean parallels to Q 14:27 and 17:33 (Mt 8:34b and 35) would still be in the same order. This is significant, because although there is no reason to believe that Q (or any SS) would maintain the same order as Mark for passages they had in parallel, as reported above Kloppenborg states that:

… if one does not measure sequential agreement of these Matthew-Luke materials relative to Mark, but relative to each other, approximately one-third of the pericopae, accounting for almost one-half of the word count, are in the same relative order.

This is not unlike what we see with the Triple Tradition material, with some of the Matthew-Luke pericopae agreeing in order, and some not. Whether or not Q and Early Luke would follow the same order is not known, but as Q is usually hypothesized to follow Lukan order (with very few exceptions, e.g. Q 17:33), it is convenient to suggest that any early version of Luke would also follow Lukan order, at least for the Double Tradition passages.

Agreements of Matthew and Luke Against Mark

In Triple Tradition passages there tends to be a considerable degree of agreement between Mark and Matthew, and also between Mark and Luke, but less agreement between Matthew and Luke. On the assumption of Markan priority this can be reasonably interpreted as both aMatthew and aLuke using material from Mark, but selecting or changing the material in different ways to suit their own purposes. However, there are other places in the gospels where Matthew and Luke agree, but where Mark has something different.

In some places Matthew and Luke have no text corresponding to something in Mark (Sondergut Mark), in some Mark has no text corresponding to something common to Matthew and Luke (the Double Tradition), and in others Mark has text different to something common to Matthew and Luke. At a basic level the possible mechanisms for these agreements (again assuming Markan priority) are easy to identify:

  1. aMatthew knew and used Luke or aLuke knew and used Matthew; and/or
  2. Matthew and Luke both had access to one or more common sources other than Mark (e.g. Q, deutero-Mark, or Early Luke).

However, in the second possibility we are leaving open the question of the source or sources of Q (which by definition is independent of Mark), and in the first possibility the source or sources used by aLuke and aMatthew respectively. In addition, there is also the question of why both Matthew and Luke would choose to use text from another source in preference to that in Mark. For the Double Tradition text the answer is straightforward, because by definition there was no text in Mark, and so (we assume) both aMatthew and aLuke felt free to insert text from elsewhere. This argument basically still applies where Matthew and Mark contain a common passage that is substantially changed compared to that in Mark (major agreements) but becomes more and more strained as the degree of change reduces to no more than a few words, e.g. where Matthew and Mark apparently made the same trivial change to a passage in Mark (minor agreements), including omitting the same few words of Mark (negative minor agreements)

If aMatthew and aLuke independently copied some passages from Mark and some from one or more different sources we would have no specific reason to believe that they treated Mark and the other source(s) differently, so we would expect that the percentage of verbatim agreement to be the approximately the same for each source. As the first case describes what we call the Triple Tradition and the second the Double Tradition, this can be tested, and the results show that that the percentage of verbatim agreement in the Double Tradition is actually higher than in the Triple Tradition (see table below). To suggest that this is due to aMatthew and aLuke being ‘more careful’ regarding their use of this other source than their use of Mark is hard to defend (and if anything we might expect the opposite), but is what we would naturally expect to see if either aMatthew or aLuke used the other’s gospel, as follows:

  • Where aMatthew and aLuke independently use a source (whether Mark, Q, or anything else), then each will make some copying errors or changes that the other does not, and when Matthew and Luke are compared then the errors or changes made by both will appear as differences between the two parallel texts (If both make an identical error or change then this will not appear as a difference).
  • Where aMatthew uses a source, and then aLuke uses Matthew, the differences between Matthew and Luke will be just those copying errors or changes made by aLuke (and the reverse applies if aMatthew uses Luke).

On this basis if the Double Tradition is a result of either aMatthew or aLuke copying from the other, we should expect to see up to two times the rate of differences between Matthew and Luke in the Triple Tradition than in the Double Tradition, with something less than that if both forms of copying took place (e.g. if aLuke used both Matthew and another common source). 

An alternate way of describing this is that if the Triple tradition has a higher percent of differences between Matthew and Luke, then it will have a lower percent of verbatim agreements, and that is what we see. The following figures are taken from a table originally published in Once More: Statistics and Q by Charles E. Carlston and Dennis Norlin: 


Triple Tradition


Double Tradition

















Words of Jesus








Misc. Words
















These figures show a consistent pattern: that there is greater verbatim agreement (fewer differences) in the text of Matthew and Luke in the Double Tradition than in the Triple Tradition, so pointing to either aMatthew copying from Luke, or aLuke copying from Matthew, instead of both just copying from Q or another common source for the Double Tradition. Although these specific numbers have been challenged, the basic result still stands, providing support for synoptic hypotheses that allow for copying from Matthew to Luke or from Luke to Matthew, while not excluding copying from another source common to both Matthew and Luke as well.

Mark-Q Overlaps

In section 4 of Fallacies at the Heart of Q Goodacre explains the meaning of the term ‘Mark-Q Overlaps,’ but then in his blog he makes a case for deprecating the term, and instead using neutral expressions such as "major agreements between Matthew and Luke against Mark" or "triple tradition pericopae where Mark is not the middle term." However, not least because “middle term” itself does not mean the same to everyone, I will continue to use the term ‘Mark-Q Overlaps’ here.

According to the Mark-Q theory, Mark and Q are independent. However, as the term implies, ‘Mark-Q Overlaps’ refers to text common to both Mark and Q (i.e. where they overlap), that is deemed necessary to explain how aMatthew and aLuke (who on the Mark-Q theory worked independently) nevertheless in otherwise Double Tradition material incorporated common portions of text (some ‘minor’ some ‘major’) also present in Mark, into their respective gospels. In The synoptic problem: a way through the maze, Goodacre comments that: 

The argument from Luke's ignorance of Matthew's additions to Mark runs into insurmountable problems: The examples given are weak: Luke's omissions are quite natural when one looks at them in line with his redactional interests. The argument is based on a fallacy: wherever Luke features Matthew's additions to Mark, these are placed in the category 'Mark-Q overlap' and, as far as this argument is concerned, they are ignored.

Not everyone agrees with Goodacre (for example, how do we know Luke's "redactional interests?"), and a good example of a potential Mark-Q overlap can be found very early in all three synoptic gospels. All contain a passage describing John baptizing people in the Jordan, prior to baptizing Jesus. While the passage in Mark is quite short (Mk 1:2-8), it is significantly longer in both Matthew and Luke (Mt 3:1-12 and Lk 3:2-17). Not only that, but the verses in both Matthew and Luke that have no parallel in Mk (Mt 3:7-10, 12 and Lk 3:7-9, 17) contain a distinctly Matthean saying: “O generation of vipers” (which appears in Mt 3:7, 12:34, and 23:33, but in Luke only in Lk 3:7), followed by several references to fire, also considered to be Matthean in origin.

On both the MwQH and the MwEL hypothesis the interpretation of this text in both Matthew and Luke is simply that aMatthew used and expanded Mk 1:2-8, while aLuke then further expanded the version he saw in Matthew, by adding Lk 3:5-6, 10:16a, and 18:21a. On the MwEL hypothesis Early Luke most likely just contained the text from Mark, because if Early Luke had contained any of what we now see in Lk 3:5-6, 10:16a, and 18:21a then we would expect to see some of this in Matthew.

On the Mark-Q theory aLuke did not know Matthew, and there must therefore be another reason why both aMatthew and aLuke added virtually identical text to a passage they both saw in Mark. This non-Markan text could of course come from Q (which both aMatthew and aLuke used), but in order for both to insert this text into identical positions in the narrative they got from Mark, Q must have also contained the same surrounding text from Mark, so overlapping with Mark. However, as soon as you allow for Mark and Q to have overlapped, then there has to be a mechanism to allow for this, and there are four possibilities:

1. the author(s) of Q knew Mark;
2. aMark knew Q;
3. aMark and the author(s) of Q both used an earlier hypothetical source; or
4. the text came from a common oral tradition that both aMark and the author(s) of Q knew.

None of these possibilities are really satisfactory. The first two imply a direct dependency link between Mark and Q that is not made explicit in the Mark-Q theory, and with such a link in place the whole rationale for Q as a source independent of Mark is undermined. The third explanation adds a second hypothetical source behind the original hypothetical source (Q), and as this violates the principle of Occams's (or Ockam's) Razor that "Entities are not to be multiplied beyond necessity" it can be dismissed forthwith. The final explanation, 'a common oral tradition,' can be neither proved nor disproved, and is not much more than a ‘Get Out of Jail Free’ card. More rigorously, Klinghardt points out that:

Methodologically, it is not permissible to develop a theory on a certain assumption and then abandon this very assumption in order to get rid of some left over problems the theory could not sufficiently explain. The methodological inconsistency of this solution would be less severe, if “Q” existed. But since “Q” owes its existence completely to the conclusions drawn from a hypothetical model, such an argument flies in the face of logic: it annuls its own basis.

In short, the Mark-Q overlaps may solve some problems with the Mark-Q theory, but then cause others that cannot be resolved. At issue is a problem that seems to be rarely discussed outside what I will call the 'Q community,' and that is What Exactly is Q? This may seem a somewhat esoteric issue, but in fact is fundamental to the question of how any non-Markan source common to Matthew and Luke fits in to a synoptic hypothesis.

Minor Agreements

There are many other examples of agreements between Matthew and Luke against Mark, and Goodacre lists several of them, covering a continuum ranging from:

“... pure triple tradition passages which feature Minor Agreements, to Mark-Q overlap passages which feature major agreements between Matthew and Luke against Mark, to double tradition passages where Luke is dependent solely on Matthew.”

Because of the overlaps all the examples given by Goodacre are problematical for the Mark-Q theory, whereas both the MwQH and the MwEL hypothesis allow naturally for all the required dependencies. The Mark-Q overlaps (the "major agreements between Matthew and Luke against Mark") are discussed above, while at the other end of the scale are the minor agreements between Matthew and Luke against Mark, i.e. those places in which both aMatthew and aLuke appear to have made the same small (sometimes trivial) change to the text of Mark. In his web-page on the Two Source Hypothesis (2SH), Stephen Carlson writes: 

The minor agreements pose a special dilemma for the 2SH, because they are suggestive of a literary connection between Matthew and Luke outside of either Mark or Q, calling into question the relative independence of Matthew and Luke.

For example, a few scholars explain the minor agreements by Luke's use of Matthew in addition to Q and Mark (3SH). The problem is that the modern argument for Q requires Matthew and Luke to be independent, so the 3SH raises more questions than it solves, namely, how to establish Q if Luke is dependent on Matthew. Other scholars keep Q while acknowledging the force of the minor agreements to attribute the minor agreements to a proto-Mark, such as the Ur-Markus in the Markan Hypothesis (MkH) that was adapted by Mark independently from its use by Matthew and Luke. Still other scholars feel that the character of the minor agreements suggests that they are due to a revision of our Mark, called deutero-Mark. In this case, both Matthew and Luke are dependent on deutero-Mark, which did not survive the ages.

Therefore, the minor agreements, if taken seriously, force a choice between accepting pure Markan priority on one hand or the existence of Q on the other hand, but not both simultaneously as the 2SH requires.

Carlson’s first sentence above essentially hits the problem on the head: In places where aMatthew and aLuke appear to have made the same small change (for no obvious reason) to the text of Mark, then (on the assumption that one author choosing a trivial change by the other in preference to what was in Mark is very unlikely) why did they both make the same change to the text, unless they had another source that read that way? Two of the possibilities suggested above by Carlson are essentially the same: That aMatthew and aLuke both had access not only to Mark, but also to a text that was related to Mark, and that was slightly different to Mark in the areas of the minor agreements (How close to Mark this other text was in other places is not part of this argument). 

In 1899 John Hawkins discussed the agreements in: ‘Horae Synopticae: Contributions to the Study of the Synoptic Problem - Second edition.’ In his Appendix III he refers to: “… a considerable amount of matter, chiefly of discourse, [which] is found in Matthew and Luke, while it is absent from Mark.” He then notes that there are 58 other agreements where different text is present in Mark, discusses and largely dismisses most of them, but then states that for others that: 

… in these sections there are certain other alterations from, and additions to, the Markan narrative, as to which it seems almost impossible that Matthew and Luke could have accidently concurred in making them. In these cases at least the changes seem to be owing to some influence, direct or indirect, of a common source, and not to the independent judgment of two compilers.

Hawkins then lists 21 instances of these agreements, and continues: 

If this evidence is regarded as sufficient to prove that in at least 21 sections - and it is reasonable to suspect in others also - a common source has supplied Matthew and Luke with variations from and additions to the Marcan narrative which apparently forms the basis of these 58 sections, then the difficult question arises, What was the nature of this source?

After discussing the possible nature of this source, he concludes that some agreements may be due to copying from either Matthew to Mark or vice versa, but also that: 

… it appears to me now that others of them, and perhaps the majority, may be best accounted for by Dr. Sanday’s suggestion that they are due to the use by Matthew and Luke of ‘a recension of the text of Mark different from that from which all the extant MSS. of the Gospels are derived’.

However, others disagree with Hawkins, with for example in 1909 Cuthbert Turner suggesting instead that these agreements are the result of harmonization: 

St. Mark was, from an early date, the least familiar of the three Gospels, and therefore his text had, as a rule, even where it was the fullest, less influence over the other two than they had on his and on each other’s. But the text of Matthew influenced the scribes of Mark and Luke, the text of Luke the scribes of Mark and Matthew, to an almost incredible extent.

The problem with this is that harmonization is not in itself an explanation – it does not identify why scribes would do this. Instead, it assumes a particular scribal behavior on the basis that ‘harmonization’ can be explained as the result of such behavior, with hypotheses as to why scribes would do this being created to support this idea. For example, owing to the (very common) suggestion that the text of Mark or Luke has been harmonized to Matthew because scribes were more familiar with Matthew, harmonization is very infrequently challenged in any particular instance. However, when a non-extant deutero-Mark (or Early Luke) is allowed to enter the picture, many of the ‘harmonizations’ can be seen to instead be simply due to aMatthew and aLuke choosing text from different sources, or, in the case of the agreements against Mark, both choosing from deutero-Mark/Early Luke rather than Mark itself.

Alternating Primitivity

In Fallacies at the Heart of Q Goodacre defines “alternating primitivity” as follows:

The argument from "alternating primitivity" is a key element in the standard case for the existence of Q. For if sometimes Matthew and sometimes Luke has the more primitive wording in the double tradition, this might well seem to be a sign that both were dependent on a prior document.

He then continues by raising difficulties with this argument:

The difficulties can be divided into two categories: (a) problems with the means by which scholars arrive at the conclusion that Luke's material is more original than the Matthean parallels and (b) problems with the assumption that greater primitivity in Luke would necessitate the existence of a Q document.

As Goodacre suggests, but does not make explicit, the term "alternating primitivity" is not meant to imply that the originality of Matthew and Luke actually reverses in a regular pattern, but just that sometimes Matthew seems more original than Luke, while at other times Luke seems more original than Matthew. For this reason "bidirectionality" is more appropriate, although "alternating primitivity" will still be used here as being the commonly used term. There is however a more serious problem with Goodacre’s definition, as he couches it purely in relation to Q. Although it may be considered a standard test for Q, it should be a standard test for any synoptic hypothesis. That is, if Matthew in some places seems earlier (more primitive) than Luke, while in others Luke seems earlier than Matthew, then any valid synoptic hypothesis (not just Mark-Q) must provide a mechanism whereby this could take place. This then would be a problem for any synoptic hypothesis that (like the MwQH) does not provide such a mechanism. As Klinghardt points out:

Whereas Goodacre’s criticism of the 2DH is convincing, his attempt to understand Luke in direct dependence on Matthew is not: The observation that in some cases Luke seems to be earlier and in other instances Matthew seems to be earlier, cannot be explained with the help of a simple “Benutzungshypothese” (the proposal of MwQH) but necessarily requires an additional source. Thus the Janus-faced character of the double tradition is one of the strongest arguments for the 2DH: The assumption of “Q” seemed to solve this problem of mutual influence in the double tradition. For want of an alternative text that could explain this problem of mutual influence in the double tradition, many scholars seem to put up with “Q” in spite of the apparent weaknesses of the 2DH.

Perhaps not surprisingly, Goodacre argues that alternating primitivity is hard, if not impossible, to prove:

In short, the difficulty is that scholars have routinely confused issues of literary priority with issues over the relative age of traditions. The theory of Luke's literary dependence on Mark and Matthew does not necessitate the assumption that his material is always and inevitably secondary to Matthew's and Mark's. 

Even though Goodacre argues that Q is not necessary to account for the fact that Luke is in places more primitive than Matthew, he argues that instead, oral sources are likely to account this phenomenon:

Just as most of us do not deny the likelihood that Luke interacted with oral traditions when he was working with Mark, so too we should not think it odd that he might have interacted with oral traditions when he was working with Matthew.

This appeal to oral traditions raises problems for Goodacre, as Foster notes:

A significant point at which Goodacre diverges from the earlier formulation of the Farrer-Goulder form of the theory is in his allowing the possibility that other traditions, apart from Mark and Matthew, influenced the Lukan narrative. In effect this possibility is an attempt to address the charge that at times Luke has the more primitive form of a Double Tradition pericope, and this appears inexplicable on the assumption of Lukan dependence on Matthew. Thus he states,

Not only has the extent of Luke's supposed primitivity been greatly overestimated, based partly on misconstrued assessments of the presence of Matthean language, but even on occasions where Luke does show possible signs of primitivity, this is only evidence for Q if one is prepared to deny a role to the living stream of oral tradition in the composition of Luke's gospel. [The Synoptic Problem: A Way Through the Maze, Mark Goodacre]

Here Goodacre is aware that such a possibility stands in direct opposition to the Farrer-Goulder form of this theory, which vigorously denies any source material outside of the fourfold gospel canon. Quite sensibly Goodacre acknowledges the reality that the writing of a gospel did not instantaneously supplant the stream of traditions that had fed into its composition. However, apart from the words of institution (Lk. 22:20) and the Lord's Prayer (11:2b-4), there are no examples provided to illustrate where Luke is drawing upon a pre-Lukan tradition rather than creating material de novo.

For instance, is the Lukan Passion Narrative tradition or redactional formulation, what about the distinctive parables of the third gospel, the infancy narrative and the unique genealogy? If Goodacre is willing to admit that alternative sources are being used in some of these cases, (and this seems to be the implication of his statement about "other traditions"), how is he able to determine with such confidence that such material was oral and not written?

Although it should be obvious that all the synoptic authors could have had non-written sources, i.e. information they received orally, whether in ‘interviews,’ overheard conversations, from preachers, etc., any resort to ‘oral tradition’ to solve a synoptic issue should raise a red flag, as it indicates an inability to find a non-oral solution within the constraints of the synoptic hypothesis under discussion, and, as previously noted, can neither be proved or disproved. Therefore, solutions that do not need to invoke ‘oral traditions’ to solve synoptic issues are preferable to those that do.

Despite Goodacre’s statement above, many people do find evidence of ‘non-oral’ alternating primitivity between Matthew and Luke. For example, Burkett comments on the following:

… Luke probably preserves the more original form of the sign of Jonah (Luke 11:30/Matt 12:40)… Likewise, in the Beelzebul debate, Luke’s “finger of God” is probably more original than Matthew’s “Spirit of God” (Matt12:28/Luke 11:20)… So too in the pericope on faithful and unfaithful slaves, Luke’s “unbelievers” is probably more original than Matthew’s “hypocrites” (Matt 24:51/Luke 12:46). 

… if Luke has a more original form of the tradition than either Matthew or Mark, such instances speak against every form of the theory that Luke used Matthew. Since Matthew provided the triple tradition in the Griesbach hypothesis, and Mark did in Farrer’s theory, the Gospel of Luke should never have the earliest form of the triple tradition in any of these theories. Yet there is good reason to think that it does.

Other evidence of alternating primitivity can be found in variants in Luke that are typically attributed to assimilation to either Mark or Matthew. For example, the story of the woman with the issue of blood in Lk 8:42b-48 has a very similar parallel at Mk 5:25-34, and a much shorter one at Mt 9:20-22. Several mss of Luke (including P75, B, and Bezae) also have shorter variants of the story, omitting various details from Lk 8:43-45. On the MwEL hypothesis a likely scenario for these otherwise difficult variants is that the shorter variant in Luke originated in a predecessor of Luke (an Early Luke), and that aMatthew based his shorter version on that in Early Luke, but later assimilation from Mark resulted in the longer Byzantine variant in Luke.

Then, in Mk 9:7 and Mt 17:5, the voice from the cloud refers to Jesus as: “my beloved son.” Most mss of Luke also have this reading at Lk 9:35, although P45, P75, 01, B, L, Θ, Ξ, f1, 579, 892, 1241, 1342, pc, Lat(b, c, e, f, q, r1, vg), Sy-S, Sy-Hmg, Co, arabMS have “my chosen one” instead. The latter phrase is considered to be original, with “my beloved son” being an assimilation to Mark or Matthew. However, while on the MwQH there is no obvious source for “my chosen one,” on the MwEL hypothesis it can be seen as the original reading in Early Luke.

Although there may be doubt as to whether alternating primitivity does exist, there is no doubt that synoptic hypothesis that do not allow for Matthew and Luke to have a common source (Early Luke/SS) for the double tradition text have a problem when compared to those that do. While both the Mark-Q and MwEL/MaSS hypotheses do allow for it, the MwQH does not, and on this basis, is the least-preferable of the three.

Matthew’s Additions to Triple Tradition Text

One problem that affects synoptic theories in which aLuke knows both Mark and Matthew (such as the MwQH and MwEL hypothesis), is that in a number of places Matthew adds to what is otherwise triple tradition material. When defending the Mark-Q theory, Christopher Tuckett writes: 

Luke never appears to know any of Matthew's additions to Mark in Markan material. Sometimes, in using Mark, Matthew makes substantial additions to Mark, cf. Matt. 12.5-7; 14.28-31; 16.16-19; 27.19, 24. If Luke knew Matthew, why does he never show any knowledge of Matthew's redaction of Mark? It seems easier to presume that Luke did not know any of these Matthean additions to Mark and hence that he did not know Matthew. 

Although not stated exactly this way, what Tuckett is asking is: If aLuke knew Matthew, why did he not use what we know as Sondergut Matthew (and the question can be reversed if aMatthew knew Luke)? Goodacre defends this charge in Fallacies at the Heart of Q, a defense on which Kloppenborg comments:

What the objection normally has in view are the Matthean additions to Markan pericopae in Matt 3.15; 12.5–7; 13.14–17; 14.28–31; 16.16–19; 19.9, 19b; 27.19, 24, all of which Luke lacks. Two of these offer no difficulty to the MwQH: Matt 14.28–31 (Peter’s maritime outing) and 19.9 (Matthew’s qualification of the divorce prohibition with mē epi porneia) are additions to Markan pericopae that Luke omits entirely. Goodacre does not comment on the Matthean additions in 12.5–7; 13.14–17; 19.19b; 27.19, 24, and instead focuses his defence on Matthean additions to Mark at Matt 3.15 and 16.16–19.

Kloppenborg then provides an explanation for Mt 3:15, but with regard to Mt 16:16-19 states that:

Goodacre invokes Farrer’s notion of certain elements of Matthew being ‘Luke pleasing’. Thus Luke omitted some elements of Matthew because they were not ‘Luke pleasing.’

In other words, there were some places where aLuke simply chose not to use aMatthew’s words because he did not like them, but, as Kloppenborg notes, this: “merely renames the problem; it does not offer an explanation.” In other words, it does not explain why aLuke did not like them. Goodacre does suggest that this is part of a redactional pattern in Luke, but Kloppenborg is not convinced, and argues that the idea of aMatthew using ‘Luke displeasing’ words cannot be used in the case of Mt 12:5-7 (because aLk employs a similar argument at v. 14:4-6), and 13:14-17 (because aLuke has similar quotes from Isaiah elsewhere). Kloppenborg also points out that:

Pilate’s wife’s dream (Matt 27.19, inserting into Mark 15.10–11) and Matthew’s hand-washing scene (27.24, added to Mark 15.15) would have served Luke’s purposes admirably.

Despite the various explanations from Goodacre, it remains that case that there are a number of places where, on the MwQH, there is no obvious reason why aLuke would have omitted these Matthean additions to the text in Mark. However, on the MwEL/MaSS hypothesis the situation is very different, and there is a simple and natural reason why aLuke did not include any of the above additions to Mark. One of the main points is that his second source (SS) preceded Matthew, and hence the author of SS did not see Matthew. Consequently, it would be impossible for any Matthean additions to the text of Mark to be in SS, and it then follows that, if SS was aLuke’s primary source, he would have used what he saw in SS (and perhaps Mark) in preference to Matthew.

Unscrambling the Egg With a Vengeance

Many people regard the order of events in Matthew to be greatly preferable to that in Luke, in particular suggesting that aMatthew’s Sermon on the Mount is a tour de force that is unmatched by anything in Luke, so raising the question: If aLuke used Matthew, how do you explain him spoiling aMatthew's order? Of course, not everyone agrees that Luke’s order in some way ‘spoils’ Matthew’s, as Foster reports: 

Goodacre sees the Q theory as an attack on the artistry of the Lukan narrative. By comparison he states, "adherents of the Farrer theory, in denying themselves the expedient of the Q hypothesis for accounting for every peculiarity in Luke's order, are inevitably more inclined to look to Luke's literary skill as a means of explaining the narrative development of his gospel." However, Goodacre is incorrect to imply that proponents of the 2ST suggest that there is no narrative thread or literary artistry in the third gospel. The problem for supporters of the 2ST is not the narrative continuity in Luke's gospel, but the implausibility of the evangelist unpicking double tradition material from Matthew's gospel. This becomes more problematic since such a procedure is unknown in ancient literary documents.

In Frequently Asked Questions on the Case Against Q Goodacre answers this charge:

'Matthew's order' is precisely that, Matthew's order and it is straightforward to see why Luke would have wanted to alter it. Whereas Matthew's order is more wooden, with its five great edifices (5-7, 10, 13, 18, 24-25), Luke has a plausible, sequential narrative. In the words of Luke Johnson, his narrative is 'essentially linear, moving the reader from one event to another ... Instead of inserting great blocks of discourse into the narrative, Luke more subtly interweaves deeds and sayings' (Anchor Bible Dictionary IV, 405- 6).

Despite Goodacre’s defense of aLuke’s order, many people see the differences in order as a particular problem for the MwQH, and indeed any synoptic solution in which Luke depends on Matthew. For example, in defense of the Mark-Q theory, Kloppenborg comments on how (on the MwQH), aLuke apparently broke up coherent blocks of text in Matthew into small pieces that were then spread around Luke:

A major challenge for the MwQH concerns its necessary supposition that Luke reordered the Matthean ‘double tradition’. Famously, Luke has drastically abbreviated Matthew’s Sermon on the Mount in Luke 6.20–49 and distributed other portions of it throughout Luke 11–16. In other instances, Luke would have disengaged ‘Q’ sayings from the Markan setting that Matthew had given them, and used them elsewhere. For example, Matthew uses the saying about having faith as a mustard seed in a Markan pericope (Mark 9.14–29 || Matt 17.14–21), but Luke, who uses the Markan pericope in Markan sequence (Luke 9.37–43), relocates the saying to Luke 17.6. The phenomenon is observed with Matthew’s twelve thrones saying (19.28), inserted into a Markan pericope that Luke also uses (Matt 19.23–30 || Mark 10.23–31 || Luke 18.24–30); Luke, however, moves the saying to 22.28–30. R. H. Fuller once characterized Luke’s procedure on the MwQH as ‘unscrambling the egg with a vengeance’, and it is this phrase that Goodacre chooses as a title for his Chapter 4.

Goodacre’s own defense of what appear to be very peculiar choices made by aLuke when adding to his gospel material that originated in Matthew (so creating the double tradition) relies largely on what he perceives to be rational choices made by aLuke. For example, where Matthew had relocated material in Mark but then added to it, Luke chose to keep the order as in Mark, and so re-located the Matthean additions accordingly. In other places, Goodacre argues that aLuke simply prefers shorter thematic units than aMatthew. However, aLuke himself appears quite happy with long discourses, e.g. in Lk 12:1-13:9, and in particular Lk 14:7-17:1, in which Jesus talks to various different groups of people with there being no apparent change of venue. If this is aLuke unscrambling aMatthew’s egg, then after having done so, he re-scrambled it in a different order.

Together with alternating primitivity, “unscrambling the egg” is perhaps the most serious problem facing the MwQH. However, on the MwEL/MaSS hypothesis aLuke can be seen to be making perfectly reasonable decisions. In this case his primary source for the Double Tradition material is not Matthew, but SS, his second source. Consequently, even if aMatthew uses non-Markan material from SS and re-orders it to suit his purposes, if aLuke uses the same material then he is likely to use it in the order in which he saw it in SS, not that in Matthew. Although aMatthew may have scrambled the SS egg, aLuke does not have to then unscramble it, because the egg he used (SS) was never scrambled in the first place.

Absence of Evidence of Matthew in Luke 

In the webpage The Existence of Q, Peter Kirby lays out a number of arguments for the existence of Q as used in the Mark-Q hypothesis: 

Independence in the Special Material
  • The Infancy and the Resurrection
Independence in the Triple Tradition
  • Disuse of Matthean Additions in the Triple Tradition
  • Absence of Matthew's Use of Mark
  • Absence of Matthean Redaction in the Triple Tradition
Independence in the Double Tradition
  • Absence of Matthean Redaction in the Double Tradition
  • Primitivity of the Double Tradition in Luke
  • The Order of the Double Tradition in Luke
  • Different Markan Contexts for the Double Tradition
  • The Evidence of Doublets in Matthew and Luke

After presenting evidence from several scholars in support of these arguments, Kirby concludes: 

I do not pretend to have achieved a certainty but only a probability in favor of the Q hypothesis. Perhaps each of these arguments are surmountable, but each points in a certain direction, and the cumulative force of these arguments lead me to favor the Q hypothesis as the best explanation of the synoptic data. The presence of the minor agreements is the only one very serious argument against the Q hypothesis, and it has been successfully addressed in detail by writers from Streeter to Neirynck. An assessment of the totality of the evidence indicates a balance in favor of the Q hypothesis, and thus it is my working hypothesis.

Several of the arguments are discussed above (e.g. the issues of primitivity, order, and doublets), and almost all the remainder are arguments that the lack of Matthean text, or a particular feature of the Matthean text, in Luke is best explained by Matthew and Luke being independent. For example, Kirby quotes from "Towards the Rehabilitation of Q" by F. G. Downing, who concludes that: “the Lukan omissions of pure Mark from his rendering of material similar to that which Matthew has conflated with Mark" presents difficulties for the MwQH. Regarding the Beelzebul controversy in particular, Downing writes: 

On Dr. Farrer's argument, we have to suppose that Luke sat down (or stood) with Matthew's and Mark's works before him. He must have then, we have suggested, decided to follow Matthew (he has only three Markan words not in Matthew, and two in another context). But for some incomprehensible reason, he decides not to follow Matthew throughout, but to follow Matthew only where the latter has added new material to Mark or has largely altered him. He notes that one and a half sentences exactly quote Mark, and so omits them. It is not that he is going to use them somewhere else. He just arbitrarily excludes them, in one case actually in favor of writing his own version (verses 21-22): so it is not even that he finds the Markan material repetitive. It is not that he objects either, to Mark as such, for on Dr. Farrer's thesis, Luke does not know (as we have noted) that the B material is not basically Mark, but slightly emended; and he includes this, quite happily. All that he excludes is the material in Mark that Matthew obviously saw fit to include pretty well as it stood!

It seems very much more sensible to assume that Luke did not know Matthew's use of Mark, and in fact here reproduced his own version of the B and C (= Q) material, with no reference either to Matthew or Mark.

This argument, as with several others presented by Kirby, is couched in terms of the absence in Luke of evidence that aLuke knew and used Matthew. Strictly speaking, these are not arguments in favor of the Mark-Q theory, but arguments against the MwQH, essentially on the basis that if aLuke knew Matthew then he would have included in his gospel text from Matthew that is not seen in Luke. However, even if we allow that aLuke’s actions in apparently omitting Matthean material seem, at least, unusual, there is the problem that the assumption that these arguments support the Mark-Q theory is itself based on the assumption that the MwQH, in which aLuke’s only sources are Mark and Matthew, is the only alternative.

On the MwEL/MaSS hypothesis the situation is very different from the MwQH, because both aMatthew and aLuke also have Early Luke/SS as a source. For aMatthew Early Luke/SS is a second source for at least some Markan parallels, as well as being the source of material added by aeLuke to his Markan source (which when used by aMatthew becomes Double Tradition material). In this hypothesis aLuke now has two sources for at least some of the Triple Tradition and Double Tradition material, and if he so wishes he can therefore choose: 

  • ‘More primitive’ versions of material from Early Luke/SS rather than later versions from Matthew;
  • To exclude additions by aMatthew on the basis that they were not in Early Luke/SS, and
  • To include text in the order seen in Early Luke/SS rather than that in Matthew.

The arguments presented by Kirby therefore can be used against the MwQH, but can only be used to support the Mark-Q theory if the MwEL/MaSS and other similar hypotheses are ignored.

Could Marcion’s Gospel of the Lord be Early Luke?

Most people who have commented on Marcion’s Gospel of the Lord (Marcion) have believed it to be a later, edited, version of Luke. Fewer have believed Luke to be an expanded version of Marcion, and fewer still have considered that if the author of Luke had Marcion as a source, then so too may the author of Matthew. One consequence of this is that when considering whether the text of Marcion with parallels in Luke might be distinct from the parts of Luke with no parallels in Marcion, very few people have considered that the text with parallels in Luke may have been taken either directly from Marcion, or from aMatthew’s version of the text of Marcion. In addition, the text of Luke having no parallel in Marcion may have come from aMatthew’s text or be an addition by aLuke himself.

It is well established that, textually, Marcion is almost entirely a shorter version of Luke, and the idea that it might be dated to before Luke is not in any way new, dating back at least to J.S. Semler in the 18th century, and to the views of Albrecht Ritschl and Ferdinand Christian Baur in the 19th century. Although the orthodox view (that Marcion in an edited version of Luke) was then strongly defended, in 1936 Paul-Louis Couchoud supported Baur’s views when he published an article entitled: ‘Is Marcion’s Gospel One of the Synoptics?’ Dieter T Roth writes: 

He recognized the merits of Harnack’s work on Marcion’s Gospel, though pointing out that it was “not perfect” and was influenced by Harnack’s own convictions. Couchoud concluded that Marcion’s Gospel was very similar to Streeter’s and Taylor’s Proto-Luke and that a methodical comparison of the Gospels and Luke would reveal the former as original, and the latter as corrected and considerably amplified.

The discussion has since swung back and forth several times, and more recently, on March 23, 2015, on his blog Alan Garrow was asked the following question regarding his Matthew Conflator Hypothesis (MCH):

It's so refreshing to hear an argument on the subject laid out with such unobfuscating rigor... At least the best case since Klinghard's. I wonder how well your hypotheses might be harmonized; that is, what if we consider Marcion's gospel in place of GLuke?

Garrow answered: 

The first question to determine, I think, is whether it is probable that Matthew used a version of Luke. If this does seem likely then there is a follow-up question: which version of Luke did Matthew use? If Matthew used Marcion's Gospel (or some other form of Proto-Luke) then this could explain some of Matthew's omissions from Luke. Having Luke composed in two stages - both before and after Matthew - might also explain some instances of Alternating Primitivity. So, incorporating Marcion into the MCH could offer some useful refinements. The downside of taking this route is that it brings in fresh elements of speculation and complexity. My preference at this stage, therefore, is to work with the broad notion that Matthew used something similar to canonical Luke - while accepting that there is always scope for greater complexity.

Although Garrow recognizes that Marcion could be an early version of Luke, he does not (at this point) explore this possibility. Neither, it seems, did either Harnack or Roth, and in this Roth is influenced by Harnack, as he comments that: “The text of Marcion’s Gospel as reconstructed by Harnack, with its continuous text and copious documentation, quickly became the standard reference for subsequent scholarship,” and notes that Harnack observed that: 

… Marcion’s Gospel text reveals a strong influence of Matthew and Mark, both in readings that are elsewhere attested in the “Western” textual tradition and in otherwise unattested readings. Following this observation Harnack noted that he considered it highly unlikely that Marcion himself was responsible for these harmonizations and that therefore Marcion possessed a text that had already been harmonized to Matthew and Mark. Yet, he did not pursue the import of this fact other than simply to observe in a footnote that it is of great significance for the history of the canon.

Not only did Harnack not “pursue the import of this fact,” but it appears that neither did Roth, as neither he nor Harnack seem to have realized that the observation that: “Marcion’s Gospel text reveals a strong influence of Matthew and Mark” is in reality not a fact. Rather, it is an interpretation of the fact that the text of Marcion has similarities to that of Matthew and Mark, but which on its own does not suggest the direction of any influence. As a result, both Harnack and Roth miss the reasonable inference that this supposed 'fact' also supports the idea that Marcion was not an edited version of Luke, but instead not only preceded Luke, but may also have preceded Matthew, and so instead have influenced Matthew. This view is explored in more detail in Marcion's Gospel Today.

As mentioned above, in The Marcionite Gospel and the Synoptic Problem: A New Suggestion, Mattias Klinghardt suggests that the inclusion of Marcion in a synoptic solution avoids the problems of both the 2SH and MwQH. The Abstract to Klinghardt’s article reads:

The most recent debate of the Synoptic Problem resulted in a dead-lock: The best-established solutions, the Two-Source-Hypothesis and the Farrer-Goodacre-Theory, are burdened with a number of apparent weaknesses. On the other hand, the arguments raised against these theories are cogent. An alternative possibility, that avoids the problems created by either of them, is the inclusion of the gospel used by Marcion. This gospel is not a redaction of Luke, but rather precedes Matthew and Luke and, therefore, belongs into the maze of the synoptic interrelations. The resulting model avoids the weaknesses of the previous theories and provides compelling and obvious solutions to the notoriously difficult problems.

Contrary to the majority opinion, Klinghardt is placing Marcion earlier than either Matthew or Luke, and therefore potentially capable of being used as a second source by both aMatthew and aLuke. In this respect it functions exactly as Early Luke in the MwEL hypothesis, so raising the question: Could Marcion actually be Early Luke? An obvious hurdle is that, as just indicated, the majority opinion is that Marcion is an edited version of Luke, and so later than Luke. However, when the text (as far as we can reconstruct it) of Marcion is compared with that of Luke in isolation, i.e. without being influenced by the almost universal views of Tertullian, Epiphanius, et al. of aMarcion as a heretic, it is clear that textually Luke is more likely to be an edited and expanded version of Marcion than Marcion being a ‘cut down’ version of Luke. However, even taking that as given, it is still necessary to determine whether the text of Marcion could have been a second source for both Matthew and Luke by testing whether it meets the constraints on the text of Early Luke identified above. Only if it does could we say that Marcion might be Early Luke.

There are a number of obvious similarities between Marcion and Early Luke: Marcion lacks Luke 1-2 and also the genealogy; Nazareth and Capernaum are reversed in Marcion; and it has no mention of Lk 4:14b-15 and 38-39. In addition, there is no reference to the text of the Great Omission having been in Marcion, which also has short versions of both the Lord’s Prayer and the Last Supper. Finally, from the writings of both Tertullian and Epiphanius, Marcion appears to omit several verses of the resurrection account. In addition Marcion does not contain any material that is absent from the synoptic gospels, as Klinghardt points out:

The gravest objection against Marcion’s assumed redaction of Luke is the fact that Mcn [Marcion] obviously did not contain any additional, non-Lukan texts: According to the traditional view, Marcion’s assumed editorial alterations would only have consisted of abridgments but not of enlargements, not to speak of any substantial additions. With respect to what we know about editing older texts within the New Testament and its literary environment this procedure would be unique. There is not a single example of a contemporary re-edition of an older text that did not support its editorial concept by including additional material. The supporters of the traditional view have duly and with great surprise noted the uniqueness of Marcion’s assumed redaction but did not take this hint seriously enough to rethink their presuppositions.

In her review of The Text of Marcion’s Gospel by Roth, Judith Lieu comments on the impact that Marcion being earlier than Luke would make:

Such a possibility might seriously disrupt one of the fundamental principles — or, rather, hypotheses — on which much New Testament study is built. Often early in their introduction to the critical field, generations of students have been taught to observe and to analyze the interrelationships between the so-called Synoptic Gospels, Matthew, Mark and Luke: most adopt the majority position that both Matthew and Luke drew on Mark, with additional shared material between Matthew and Luke being explained either by further literary dependence (usually of Luke on Matthew) or, more commonly, by their recourse to a lost common source (‘Q’); a minority might opt for the older principle of Matthean priority. The exercise, however, depends on the assumption that the texts to be studied, compared, and explained are in all that is essential — allowing for known textual variants — those known through the subsequent canonical manuscript tradition. Inserting Marcion’s Gospel early into the equation would be far more significant than the discovery or the removal of Q; it would demand rethinking the theological as well as the literary processes in the formation of a threefold (or fourfold) Gospel tradition and the timescale against which they took place.

Given the potential impact described by Lieu, it is perhaps not surprising that, for many, the heavily biased statements of Tertullian, Epiphanius, et al. regarding aMarcion (see these Conclusions) outweigh what their discussions of the actual content of Marcion tell us about Marcion itself: That the text of Marcion is almost a strict subset of that in Luke. Although Marcion does not contain everything in Luke, Luke contains virtually everything in Marcion, with the only exceptions being a few short phrases reported by Tertullian to be in Marcion that we see in Matthew instead of Luke (which can be explained as aLuke simply choosing not to include some text from Marcion). If Luke followed Marcion then it would be natural to expect that at least some material that we know as Sondergut Luke would not be in Marcion, and that is indeed the case. However, according to both Tertullian and Epiphanius there is also some material in both Mark and Luke (and in some cases also Matthew) that was not in Marcion.

On the MwEL/MaSS hypothesis any material in both Mark and Luke (whether also in Matthew or not) would by default also be expected to be in Early Luke/SS, and by the same token (assuming that Marcion preceded Luke) that material would also be expected to exist in Marcion. It is therefore necessary to explain why some of this material was not in Marcion. One possible explanation is that Marcion was derived from an earlier version of Mark that also did not contain the material, but because this requires another hypothetical text an explanation that does not require an earlier version of Mark is preferable. To find such an explanation, each case in which material in both Mark and Luke did not exist in Marcion must be examined to see whether (assuming that Marcion preceded Luke) there is a valid reason why it was not in Marcion.

That examination is undertaken in Is Marcion's Gospel Based on Mark?where it is shown that even if Marcion pre-dated Luke, there are valid reasons why Marcion does not include some passages present in both Mark and Luke, as summarized here:

  • A significant portion of the content (often complete verses) of some of the otherwise triple tradition passages having no parallel in Marcion is actually material we see as unique to Luke, and hence would not be expected to exist either in Early Luke, or in Marcion if it preceded Luke.
  • Similar to the above, some of the content (again, sometimes complete verses) we see as being unique to Mark, and therefore also should not be required to be in Early Luke or Marcion.
  • Some of the text in Luke contains small (in some cases trivial) differences from the parallel text in Mark and Matthew. If in these places Early Luke followed Mark then aLuke changed what he saw in Early Luke, but if instead aeLuke changed what he saw in Mark and aLuke then followed Early Luke, we might expect to see the same changes in Matthew. Neither explanation is completely satisfactory, suggesting perhaps that in these places there may have been no text in Early Luke, as we also see in Marcion.
  • In other triple tradition passages the text in Mark is very similar to, and sometimes almost identical to, the parallel text in Matthew, while the parallel text in Luke is significantly different, with no obvious explanation in some cases. In some the ‘framework’ of the passage in Mark/Matthew is used in Luke, but changes to the text significantly alter the message conveyed by the passage.

Given that on the MwEL hypothesis Mark is a source for Early Luke, in these cases we would reasonably expect that where Early Luke contains text parallel to that in Mark, it would be similar to the text in Mark. Similarly, we would expect the text in Luke to be similar to the parallel in Early Luke, although perhaps with some changes originating in Matthew where Matthew also has parallel text (i.e. in triple tradition passages). However, what we actually see here is more complex, with either there being significant differences between Mark, Matthew, and Luke, or in some places Mark and Matthew being very similar, with Luke being different.

As on the MwEL hypothesis Early Luke provides a second ‘Markan’ source for both Matthew and Luke, those places where either Matthew or Luke differ significantly from parallel text in Mark suggest that aMatthew or aLuke either ignored this second source at these points, or perhaps that Early Luke simply did not contain parallel text at these points. This means that the lack of some triple tradition passages in Marcion does not prevent Marcion from being a candidate for Early Luke. However, this does not mean that Marcion must be Early Luke, but only that Marcion meets the criteria for Early Luke, and so could be Early Luke. As Klinghardt concludes:

Finally, it is clear that this paper only intends to open the window for further discussion: I am fully aware that I am far from seeing all the implications and consequences of this suggestion, neither within the realm of the traditional issues of the synoptic problem nor the historical consequences that lie beyond it. But since this model provides a solution of the contentious issues of the present debate, it may help to break the deadlock in which the discussion of the synoptic problem seems to be caught for too long now.


Perhaps the most interesting point that can be made regarding the synoptic problem is that, after more than two centuries of effort by many eminent minds, we still have no agreed solution. Kloppenborg does not appear to view this as a problem, as he states: “Every Synoptic theory has to accommodate anomalous data.” If by this Kloppenborg is suggesting that it doesn’t matter if a hypothesis cannot explain everything, then I disagree. If instead he is just indicating that no current hypothesis is capable of explaining all the data, and therefore is not a complete solution to the synoptic problem, then I agree.

And yet, it is plainly obvious that there is a solution, because the problem is rooted in reality: The synoptic gospels do exist, and anyone who so chooses can read them, or at least, can read what we believe them to have been at the time of their formation. And there perhaps is the nub of the problem: we are trying to find a solution without having all the information to hand, based on what we believe to have been the text of the synoptics at the time of the interactions among them, because we do not know what it was, or how many ‘drafts’ or earlier versions of one of the gospels might possibly have been seen by the author of another. Consequently, we cannot easily allow for potentially significant variants for which we lack extant manuscript evidence, or (for the most part) for there to have been earlier or otherwise different editions or versions of any or all of the synoptics.

Nevertheless, there have been various suggested solutions to the synoptic problem that have found favor over the decades, and until very recently the Mark-Q theory seemed close to being the ‘winner.’ However, the notion of the solution being dependent on a totally hypothetical document (Q), for which no evidence has ever been found, and invented purely to try to solve the synoptic problem, was (and still is) too much for some people to stomach. In addition, as described above, the problem of the Mark-Q overlaps highlights a fatal flaw in the Mark-Q theory, since once it is allowed that Mark and Q overlapped at some points, then the whole basis on which the independence of Mark and Q depends falls apart, and it then becomes impossible to define what Q contained.

As a result of problems with the Mark-Q theory a strong defense of the Farrer Theory (the MwQH) was mounted (with Mark Goodacre featuring prominently, as witnessed by the number of references to his work here), leading to the current position in which the Mark-Q theory and the MwQH can be regarded as joint front-runners. Even so, this does not on its own mean that either hypothesis is necessarily the basis of the solution. As Foster writes in the conclusions of ‘Is it Possible to Dispense with Q?’:

Goodacre is the latest in a line of scholars who posit a solution to the synoptic problem that holds to Markan priority, but refutes the existence of Q. The examination of the theory, as he formulates it, has been shown to fail to answer many of the objections of scholars who hold to the Two Source Theory. Rather than reiterate those arguments, two issues are raised which might help to generate further discussion and clarify the objectives of various scholars who propose competing solutions to the synoptic problem.

First, it would be helpful if supporters of the Farrer theory could explain why Lukan dependence on Matthew is so much more appealing than the theory of Matthean posteriority. Such an explanation would assist in clarifying important methodological issues. Second, it would be beneficial if proponents of the Farrer theory could tell those who believe in Q what evidence they would require to be convinced of that document's existence.

Perhaps they would require the discovery of an extant manuscript, but maybe such a requirement would be a little unfair. It would be similar to those who do not think the Farrer-Goulder theory is plausible saying that they would only accept this alternative if a document was discovered written by Luke that explicitly acknowledged dependence on Matthew (or a Matthean document acknowledging Lukan dependence, if Huggins' theory was correct). Instead, probably most proponents of the Two Source theory would be more likely to adopt one of the alternative positions if clear evidence of Matthean redactional material was present in Luke's account, or Lukan redaction in Matthew, if the latter were drawing on Luke as a source.

Moreover, those who hold to the Two Source theory would be grateful for a more plausible account of the literary activity of Luke in removing the non-Markan material from Matthew's gospel, instead of the cacophony of voices that propose that Luke was critical of Matthew, either for theological, narratival or pastoral reasons. So, in conclusion, an open question is addressed to the defenders of "Markan priority, non-Q" type solutions to the synoptic problem: Is there any level of evidence that is internal to the literary accounts of the gospels that would be considered plausible in showing that a document which is no longer extant stood behind Matthew and Luke as a common source?

Supporters of each hypothesis see problems with the others. For example, the agreements (both major and minor) between Matthew and Luke against Mark cause problems for the Mark-Q theory, and the ‘re-ordering’ of the text of Matthew by aLuke, and the alternating primitivity between Matthew and Luke, cause problems for the MwQH. Although both sides can explain away these issues, to do so involves invoking what are seen as dubious (even fatal) arguments by the other side, for example: the need for ‘Mark-Q Overlaps’ to complete the Mark-Q theory; and the idea that alternating primitivity depends on a ‘value judgment’ and does not really exist, and so on the MwQH does not need to be explained. As it stands, the current position looks like a stalemate. However, the two apparently quite different hypotheses are not as far apart as they might at first sight appear.

One of the key characteristics of the Mark-Q theory is the independence of Matthew and Luke. However, it is this very independence that leads directly to the major problem with this hypothesis, which is to explain the agreement of Matthew and Luke against Mark in material otherwise deemed to have been obtained from Mark. The solution to this problem actually modifies the Mark-Q theory (although it is not couched in these terms) by allowing Mark and Q to have material in common: the so-called Mark-Q overlaps. This in turn implies either that aMark knew Q but used just a small part of its text, or that the author(s) of Q used just a small part of Mark. Given that Q (as usually defined) contains mainly non-Markan sayings material, then (unless Q actually was multiple documents, or there were multiple versions or layers) the latter seems more likely. Then, depending on how much of Q was material sourced from Mark, and how much of Luke is material sourced from Q, it is not hard to see that Q could perhaps be considered to be either a development of Mark, or even an early version of Luke. In comparison with the Mark-Q theory, the MwEL hypothesis makes this possibility explicit by replacing Q by a different second source (here referred to as Early Luke), and avoids the need for common Matthew-Luke material in this source by adding a link from Matthew to Luke.

Unlike the Mark-Q theory, the MwQH provides an explicit link between Matthew and Luke by having aLuke use Matthew as one of his sources. However, there is no corresponding path by which aMatthew could use any of Luke as one of his sources. As a result, the MwQH has no ready explanation for the phenomenon of alternating primitivity between Matthew and Luke, requiring adherents of the hypothesis to ‘explain it away’ as being simply due to different interpretations of the text. Unfortunately, the doublets in Matthew cannot be similarly explained away, and as on the MwQH Matthew has only one source (Mark) this is a problem. The MwEL hypothesis solves this problem by the expedient of adding an early version of Luke known to aMatthew, providing a ready explanation for doublets in both Matthew and Luke, and also alternating primitivity between Matthew and Luke.

For those who hold to Matthean posteriority as the best explanation for the difference in order between Matthew and Luke, the MwEL hypothesis allows aMatthew to have seen and used a shorter, earlier, version of Luke, while still allowing the final version of Luke to contain material that originated in Matthew. This avoids the problem with which hypotheses such as the MCH are faced when trying to explain why so much of Luke was ignored by aMatthew, by specifying that the infancy narrative, much of the resurrection narrative, and other parts of the Sondergut Luke material were not in Early Luke.

As previously indicated, a second source such as Early Luke occupies the same ‘synoptic space’ as a Late Mark (with the only differentiating factor being how close the text is to either canonical Mark or Luke), and there is at least one earlier synoptic hypothesis that includes a later form of Mark, as Carlson notes on his Two Source Hypothesis web-page: 

Under Markan priority, the triple tradition is derived from a narrative source that resembles Mark and that both Matthew and Luke used. In the present form of the 2SH, that source is Mark 1:1-16:8. Variations of this source include the supposition of an early form of Mark called Uk-Markus or proto-Mark, a revised form of Mark, deutero-Mark, or both, but these possibilities are only supported by a handful of active scholars.


Stlle [sic] other scholars feel that the character of the minor agreements suggest that they are due to a revision of our Mark, called deutero-Mark. In this case, both Matthew and Luke are dependent on deutero-Mark, which did not survive the ages.

Carlson then concludes the same page by stating:

The hypothesis of a deutero-Mark within the framework of a four-source hypothesis (Mark, Q, L and M) is extremely sensible. It is the most economic theory that explains all we know. Why has it not been accepted more widely?

Carlson sees that including a deutero-Mark (a second source logically similar to Early Luke) in a synoptic hypothesis makes a lot of sense, and in his suggestion Matthew and Luke now have three possible common sources (Mark, deutero-Mark/Early Luke, and Q). However, as shown here in the MwEL hypothesis, by the same token the addition of a deutero-Mark/Early Luke is also an "extremely sensible" change to the MwQH, so that Matthew has two 'synoptic' sources (Mark and Early Luke), and Luke has three (Mark, Early Luke, and Matthew).


As initially stated above:

The MwEL hypothesis can be thought of as a combination of the MwQH and the Mark-Q theories, and is also an 'instance' of the MaSS hypothesis. It assumes Markan priority, has two sources common to Matthew and Luke: Mark and a second source, in this case Early Luke instead of Q, but also has aLuke knowing Matthew.

Evaluating a combination of the MwQH and Mark-Q hypotheses (each of which appears to offer both advantages and disadvantages compared to the other) seems so obvious and un-controversial that it is perhaps surprising that no in-depth assessment of such a hypothesis appears to have been undertaken before (although similar hypotheses have been touched on in passing). It appears that this may simply be due to the entrenched positions of the defenders of both Q and a link from Matthew to Luke as an explanation for the Double Tradition, who see any encroachment from the other hypothesis as attacking their territory. Instead, the reality is that as shown here, the combination of the two hypothesis simply builds on both, especially when it is acknowledged that Q (as defined in the Mark-Q theory), could have contained parallels to large portions of (perhaps most of) Mark.

In comparison with the Mark-Q theory (assuming a link between Mark and Q implied by the Mark-Q overlaps) the MwEL hypothesis adds a link from Matthew to Luke, and in comparison with the MwQH the MwEL hypothesis adds a second source (SS) for both Matthew and Luke that is derived from Mark. By doing so the MwEL hypothesis combines the explanatory power of both the Mark-Q theory and the MwQH in a way that overcomes the objections to both by positing an Early Luke that is “a document which is no longer extant [that] stood behind Matthew and Luke as a common source.” (Foster)

That the MwEL hypothesis is similar to the 3SH or Mark-Q-Matthew hypothesis should not be a surprise. The difference is that the MwEL hypothesis explicitly recognizes that by allowing Matthew to be a second source for Luke the ‘Q’ in the 3SH is no longer Q. Bird calls it ‘Q-lite,’ but this is only to distinguish it from Q. In reality the removal of the restrictions on the text of this source imposed by the need for Matthew and Luke to be independent allow much greater flexibility in the definition of this second source, to the point where it is valid to see it as possibly a deutero-Mark, an Early Luke, or perhaps even a text that Marcion saw and referred to as The Gospel of the Lord. For an in-depth examination of this final point see The Synoptic Gospels and Marcion.


Abakuks, Andris: The synoptic problem and statistics, September 2006

Bigg, Howard C: The Present State of the Q Hypothesis, 1988

Bird, Michael F:  The Gospel of the Lord: How the Early Church Wrote the Story of Jesus and The Holtzmann-Gundry Solution to the Synoptic Problem (Three Source Hypothesis). Also Goodacre, Mark: Mike Bird on Luke's use of Matthew and Q

Bratcher, Dennis: The Gospels and The Synoptic Problem: The Literary Relationship of Matthew, Mark, and Luke

Brooks, E Bruce: The Luke A/B/C model

Burkett, Delbert Royce: Rethinking the Gospel Sources: Volume 2: The unity or plurality of Q, SBL, 2009

Carlson, Stephen C: The Synoptic Problem Website

Cassels, Walter Richard: Supernatural Religion: The Synoptic gospels (continued). The Fourth gospel

Couchoud, Paul-Louis: Is Marcion’s Gospel One of the Synoptics, Hibbert Journal 1935

Davidson, Paul: From ‘Is That in the Bible?’: Did Luke Know and Use Matthew? The Parable of the Talents/Pounds as a Test Case, and Jesus and the Beelzebul Controversy: A Devilish Synoptic Puzzle

Derrenbacker, Robert A, Jr, and Kloppenborg Verbin, John S: Self-Contradiction in the IQP? A Reply to Michael Goulder, 2001

The Didache

Downing, F. G.: Towards the Rehabilitation of Q. New Testament Studies, 11, pp 169-181, 1965

Dunn, James D. G: Jesus in Oral Memory: The Initial Stages of the Jesus Tradition, University of Durham

Early Christian Writings.com: The Existence of Q

Eilers, Marco: The Problem with Mark-Q Overlap

Farmer, William R: The Present State Of The Synoptic Problem, 1998

Foster, Paul: Is It Possible to Dispense with Q? , 2003

Friedrichsen, Timothy A.: Critical Observations on a Team Effort: Beyond the Q Impasse – Luke’s Use of Matthew

Garrow, Alan: Streeter's Other Synoptic Solution: The Matthew Conflator Hypothesis (MCH)

Gentile, David: A statistical approach to the synoptic problem

Godfrey, Neil (Vridar): Blog posts: Marcion enters the Synoptic Problem, Marcion's Gospel, its character and contents, and  “The Marcionite Gospel and the Synoptic Problem: A New Suggestion”   

Goodacre, Mark: The Case Against Q: Studies in Markan Priority and the Synoptic Problem and The synoptic problem: a way through the maze.

Online: The Case Against Q, Ten Reasons to Question Q, A Monopoly on Marcan Priority? Fallacies at the Heart of Qand Beyond the Q Impasse or Down a Blind Alley.

See also Mark Q Overlaps I: Terminology, II: Major Agreements Between Matthew and Luke, III: Minor Agreements between Mark and Luke, IV: Back to the Continuum, V: the degree of verbatim agreement, VI: The Direction of Dependence

The Gospel of Q at The Nazarene Way

Goulder, Michael: Luke: A New Paradigm

Gundry, Robert Horton: Matthew: A Commentary on His Handbook for a Mixed Church Under Persecution, Second Edition, 1994

Hawkins, John C: Horae Synopticae: Contributions to the Study of the Synoptic Problem, Second edition, 1899

Head, Peter M: Textual Criticism and the Synoptic Problem, New Studies in the Synoptic Problem, Oxford Conference, Part 1, April 2008

Head, Peter M. and Williams P.J: Q Review, Tyndale Bulletin 54.1, 2003

Heisey, Nancy R: The Current State of Q, TIC TALK 39, 1997

Himes, Paul A: Review of Markus Vinzent, Marcion and the Dating of the Synoptic Gospels, 2015

Hogan’s Blog: A Critique of the approach and evidence of the Q theory in light of the Synoptic Gospels. MA assignment, 2012

Huggins, Ronald V: Matthean Posteriority: A Preliminary Proposal, 1992 

Just, Felix: The Synoptic Problem

Kelhoffer, James A: Miracle and Mission: The Authentication of Missionaries and Their Message in the Longer Ending of Mark, 2000

Kirby, Peter, The Existence of Q

Klinghardt, Matthias: The Marcionite Gospel and the Synoptic Problem: A New Suggestion, Novum Testamentum 50 (2008)

Kloppenborg, John S: Synoptic Problems: Collected Essays, 2014, which includes On Dispensing with Q?: Goodacre on the Relation of Luke to Matthew, 2003, also Composing Matthew by Recomposing Q: The Composition of Matt 23–25, in An Early Reader of Mark and Q, edited by Gilbert van Belle and Josef Verheyden. Biblical Tools and Studies, vol. 21, 187–215. Leuven: Peeters, 2016.

Kok, Michael J.: Euangelion Kata Markon, The Case For and Against Q

Lieu Judith: Marcion and the Ideology of Texts Comments on Dieter Roth’s ‘The Text of Marcion’s Gospel,’ 2015

Lindemann, Andreas (Ed): The Sayings Source Q and the Historical Jesus, 2001

Mattila, Sharon Lea: A Problem Still Clouded: Yet Again – Statistics and “Q”

McNicol, Allan J, (Ed) with Dungan, David L., and Peabody, David B: Beyond the Q Impasse: Luke’s Use of Matthew, 1996

Milavec, Aaron: Synoptic Tradition in the Didache Revisited, Center for the Study of Religion and Society University of Victoria

Neirynck, F: IQP and the Critical Edition of Q - The Sayings Source Q and the Historical Jesus, Andreas Lindemann (Ed)

NET: The New English Translation 

Parker, D.C: The Living Text of the Gospels, Cambridge, Cambridge University Press, 1997, p. 115

Peterson, Jeffrey: Order in the Double Tradition and the Existence of Q, Austin School of Theology

Poirier, John C. and Peterson, Jeffrey (Ed): Marcan Priority Without Q: Explorations in the Farrer Hypothesis , 2015

Powell, Mark Allen: What are They Saying about Luke?, 1989

Powers, B. Ward: The Progressive Publication of Matthew, An Explanation of the Writing of the Synoptic Gospels, 2010 (Also in dissertation form)

Roth, Dieter T: Marcion's Gospel and Luke: The History of Research in Current Debate (2008), Towards a New Reconstruction of the Text of Marcion’s Gospel: History of Research, Sources, Methodology, and the Testimony of Tertullian (2009), and The Text of Marcion’s Gospel (2015).

Sanders E.P, and Davies, Margaret: Studying the synoptic Gospels

Sloan, David B: Q as a Narrative Gospel, 2015

Smith, Barry D: The Synoptic Problem, Crandall University

Smith, Ben: Doublets in the synoptic tradition

Smith, Daniel A: The Trouble with Q, 2012

Smith, Mahlon H: The Canonical Status of Q, 1998

Streeter, B.H: The Four Gospels: A Study of Origins. Treating of the Manuscript Tradition, Sources, Authorship, and Dates

Text Excavation.com: The Synoptic Problem  

Theopedia.com: The Synoptic Problem

Tuckett, Christopher: Q and the History of Early Christianity: Studies on Q, Edinburgh: T & T Clark, 1996, pp. 7-8, and From the Sayings to the Gospels, Mohr Siebeck, 2014

Turner, Cuthbert Hamilton: The study of the New Testament, 1883 and 1920

Turton, Michael: Is Mark Q?

Vinzent, Markus: Marcion and the Dating of the Synoptic Gospels (Studia Patristica Supplements), 2014

Wallace, Daniel B: The Synoptic Problem

Waltz, Robert: The Encyclopedia of New Testament Criticism (online) or in PDF form

Watson, Francis: Q as Hypothesis: A Study in Methodology

West, H. Philip Jr.: A Primitive Version of Luke in the Composition of Matthew, 1967

Wikipedia.com: Synoptic Gospels

Willker, Wieland: A Textual Commentary on the Greek Gospels: Vol. 2b The various endings of Mk, and the Western Non-interpolations

The Gospel of Q 

Page created  Feb 11 2014.