“The Marcionite Gospel and the Synoptic Problem: A New Suggestion”

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Matthias Klinghardt

Matthias Klinghardt responded to Mark Goodacre’s 2002 book, The Case Against Q, with an article proposing a Marcionite solution to the Synoptic Problem: “The Marcionite Gospel and the Synoptic Problem: A New Suggestion” published in Novum Testamentum, 2008.

For those of us who like to be reminded, here are the traditional theories on the Synoptic Problem:

The Griesbach or Two-Gospel theory — that Mark was the last gospel to be composed — is a minority view. Recently published proponents are William R. Farmer, Allan McNicol and David Peabody (Klinghardt, p.2).

Arguments for Markan priority — summed up in Goodacre’s book as the case against the Griesbach hypothesis — have persuaded most scholars so for the purposes of this discussion Klinghardt [MK] does not call this into question. It is the major part of The Case Against Q that has proved controversial and that MK addresses. Criticism against Goodacre’s thesis has also come from

MK begins by noting two positive arguments supporting Goodacre’s argument for the Farrer hypothesis (also known as the Farrer-Goulder hypothesis) that Mark alone (without Q) was the primary source for both Matthew and Luke, and that Luke also knew and revised Matthew:

  1. the minor agreements (e.g. both Luke (22:64) and Matthew (26:68) have the mockers of Jesus taunt with “Who is it who struck you?”, but this is not found in Mark)
  2. the hypothetical nature of Q

On the question of the minor agreements MK sides with Goodacre:

As for the minor agreements, Goodacre has a strong point insisting on the principal independence of Matthew and Luke according to the 2DH. This excludes the evasive solution that, although basically independent from one another, Luke knew and used Matthew in certain instances.

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.

This is the reason why Goodacre’s reference to the hypothetical character of “Q” carries a lot of weight. More weight, certainly, than Kloppenborg would concede: he tries to insinuate that Mark is as hypothetical as “Q”, since Mark “is not an extant document, but a text that is reconstructed from much later manuscripts.” This exaggeration disguises the critical point: the hypothetical character of the “document Q” would certainly not pose a problem, if “Q” was based on existing manuscript evidence the way Mark is.

It is, therefore, important to see that these two objections are closely related to each other: They prove that the minor agreements are, in fact, “fatal to the Q hypothesis”.  (my formatting)

But there are problems with thinking that Luke knew Matthew, as MK notes:

  1. Luke gives no evidence of his knowing material that is “special” to Matthew (“M”). (My interjection here: this is an odd way to phrase the problem, is it not? If Luke gave us evidence he “knew” of “M” it wouldn’t be “M”.) Consider also Matthew’s unique account of Pilate’s wife’s dream (27:19) and Peter’s confession and beatitude (16:16-19)
  2. MK: “Then there is the problem of alternating priority: Although in some instances Luke’s version of double tradition material seems to presuppose Matthew, there are a number of striking counter-examples, among which Luke’s wording of the Lord’s prayer or the first beatitude rank highest.
    • The double tradition is the material (circa 200 verses) shared by Matthew and Luke, but absent in Mark. It consists almost entirely of Jesus’ sayings and teachings, and includes most of the Sermon on the Mount and most parables. In addition to these, the double tradition includes a three-verse quotation (Mt. 3:8-10) that is attributed to John the Baptist (the last verse of this quotation also appears in Mt. 7:19, attributed to Jesus) and the story of centurion’s servant (Mt. 8:5-13).
  3. MK: In some cases, the arrangement of double tradition material does not make any sense at all if Luke made use of Matthew as it becomes particularly apparent with the material of the Sermon on the Mount (Matt. 5-7) and its Lukan counterparts. Although these observations carry different weight, their cumulative force renders Luke’s simple dependence on Matthew highly improbable. In light of the double tradition material, one is inclined to suggest a Matthean dependence on Luke rather than the other way round.

Klinghardt acknowledges the strength of Goodacre’s critique of the existing hypotheses, but is forced to concede that his solution to the problem is less convincing.

So we have a situation where sometimes the material in Luke seems to be earlier than what we find in Matthew (e.g. Luke’s “blessed are the poor” against Matthew’s “blessed are the poor in spirit”), and in other times we have Matthean material that looks earlier than what we find in Luke. To resolve this dilemma MK introduces Marcion’s gospel:

There is, however, an additional, yet long neglected text which indubitably belongs in the maze of the synoptic tradition and which, contrary to the hypothetically reconstructed document “Q”, is well attested by ancient sources: the gospel of Marcion, or, more precisely, the gospel which was used by Marcion and the Marcionites (hereafter: Mcn).

Adding Macion’s Gospel to the Question

Marcion’s gospel has never been considered as part of a solution before because it has long been considered to have been an abridged version of the Gospel of Luke. Church Fathers had accused Marcion of deleting passages contrary to his own theology, and scholarship ceased to seriously debate the question from the mid nineteenth century on.

Nonetheless, MK informs us that throughout the 18th and early 19th centuries there was a considerable number of scholars

who proposed the opposite view and claimed that Mcn be prior to Luke, Luke thus being an enlarged re-edition of Mcn. Among them were exegetical heavyweights such as Johann Salomo Semler, Johann Georg Eichhorn, and Albrecht Ritschl. . . . . [T]heir critique of the traditional view has never really been disproved: many cogent reasons for Mcn’s priority to Luke are still valid, which means that in many ways it is much easier to regard Luke as an enlarged edition of Mcn than the other way round. This view was convincingly, yet without any consequences, repeated in the 20th century by John Knox.

MK’s note: Knox reflected on why his theses were never accepted in Jesus, the Gospels, and the Church: Essays in Honor of William R. Farmer)

MK sums up his previously argued case for the priority of Mcn’s gospel to our Gospel of Luke:

  1. All the ancient sources — Irenaeus 1.27.4; 3.12.12; Tertullian 4.6.2-4; Epiphanius 42.9.5-6; 10.3, 5; Adamantius, Dial. 2.18 — admit they are attempting to refute Marcion “on the ground of his own gospel”. That is, if Marcion had really edited the gospel to suit his own theological agenda, he had done a very poor job of it. Tertullian tried to argue that Marcion had deliberately retained passages that contradicted his own views in order to claim that he made no changes at all (4.43.7)!
    • “I am sorry for you, Marcion: your labour has been in vain. Even in your gospel Christ Jesus is mine” (Tertullian 4.43.9).
  2. Who edited whom? Marcion did not complain that catholic Christians changed his gospel after he presumably edited it to reflect Pauline teachings. Rather, he accused them of altering his gospel in order to make it compatible with the Old Testament. Tertullian, similarly, accused Marcion of editing his canonical version of Luke. It thus appears “that Marcion’s assessment as reported by Tertullian might be correct.”
  3. Marcion’s gospel “did not contain any additional, non-Lukan texts”. If Marcion were editing canonical Luke then 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.”
  4. The relation of Luke-Acts poses a problem.
    • Either Marcion found Luke-Acts as a combined text in a New Testament canon, in which case he removed both Acts and the prologues. This would assume that substantial parts of the New Testament canon preceded Marcion which seems improbable given Harnack’s and Campenhausen’s ideas on the emergence of the canon;
    • Or Marcion did not know Luke-Acts as part of a New Testament canon, and only chose to use Luke. But this runs into the problem of the the combined work being written by an anonymous “I” in the prologue. The identity of this “I” in Luke 1:1-4 would not be an issue given the superscription ascribing the work to Luke in the canon.
    • The easy solution, MK says, is that “Marcion’s charge was correct and that a catholic interpolation incorporated ‘his’ gospel into the canonical bible of the Old and New Testament, made some editorial additions and feigned Luke-Acts as a literary unity.”
  5. The differences between Marcion’s gospel and canonical Luke “are best understood as editorial additions in Luke rather than reductions by Mcn.” Most obvious: Luke’s adding to Mcn’s beginning of 3:1a (that admiral place to start a gospel); and the change of sequence of 4:31-37 and 4:16-30 (having Jesus appear in Nazareth before Capernaum).

Similar arguments, especially concerning #5, can be found in Joseph Tyson’s Marcion and Luke-Acts that I have covered in a series of detailed posts on this blog.

It is apparent that the historical Marcion did not create “his” gospel but simply shared an older, already existing gospel. It is labelled “Mcn” here because this particular “Proto-Luke” is well attested to be utilized later by Marcion and the Marcionites.

Testing the Case

To test whether Marcion’s gospel thus does indeed have a rightful place among “the usual suspects responsible for the literary relations between the synoptics”, MK takes a number of examples from Mark Goodacre’s Case Against Q that particularly focus on the relation between Matthew and Luke. Although there is uncertainty about much of Marcion’s gospel, “the general picture is clear enough: in a good number of cases [Tertullian and Epiphanius] explicitly claim certain passages (of Luke) to be present or absent in Mcn.”

Matthew’s additions to triple tradition material not found in Luke

If Luke used Matthew, why did he omit

  1. John the Baptist’s objection to baptizing Jesus in Mat. 3:15?
  2. Jesus’ answer to the Pharisees when they accused his disciples of plucking corn on the sabbath, Matt. 12:5-7?
  3. The full quotation of Isaiah to explain why Jesus taught in parables, Matt. 13: 14-17?
  4. Peter walking on water (Matt. 14:28-31)?
  5. Peter’s confession that Jesus is the Christ and the blessing Jesus gave him (Matt. 16:16-19)?
  6. Jesus’ command to love in his reply to the rich young man (Matt. 19:19)?
  7. Pilate’s wife’s dream and Pilate washing his hands (Matt. 27: 19, 24)?

(These passages are referred to as Matthew’s additions to the “Triple Tradition” material not found in Luke.)

Some of these additions in Matthew would have suited Luke’s agenda admirably, so their omission from Luke stands as evidence that Luke did not know Matthew’s gospel. But the arguments are negative. MK believes that a check with Marcion’s gospel enables a positive argument:

Luke does not have the Matthean additions to Mark, because his main source was neither Mark nor Matthew, but Mcn. All but one of these examples are reported to be part of Mcn. . . .

  • Luke’s abridged account of Jesus’ reply to the Pharisees when challenged over the corn plucking on the sabbath (6:4-5) is what we find in Marcion’s gospel — according to Tertullian and Epiphanius.
  • Peter walking on water is part of the “Great Omission” — the absence of all material in Mark 6:45 to 8:26 from Luke. It should appear between Luke 9:17 and 9:18. Tertullian confirms that Marcion’s gospel also had these verses (9:17 and 9:18) together. So here Luke was following Mcn, neither Mark nor Matthew.
  • Peter’s confession and Christ’s blessing should appear between Luke 9:20 and 9:21. Again Luke was following Marcion’s text here (according again to Tertullian).
  • The restrictive clause of fornication in Jesus’ teaching about adultery and re-marriage is not found in Mcn (Tertullian) and hence not in Luke. Tertullian had to turn to Matthew to argue against Marcion on this point.
  • The story of the rich young ruler is well attested as part of Marcion’s gospel, since his critics used its reference to “God the father” against Marcion. Marcion’s text omitted the command to love found in Matt. 19:19 — and Luke followed Marcion’s text here.
  • Tertullian and others do not give us enough evidence to know what Marcion’s text included about the trial of Jesus.
  • Marcion’s gospel did not contain a baptism scene nor John the Baptist, so Luke could not follow Mcn here. MK suggests that Luke’s omission of Matthew’s claim that Jesus was “fulfilling all righteousness” is not surprising given his own theological agenda.)

(I realize there is a slight mis-match here in my summary between the initial Matthean additions and MK’s responses: but have decided to leave it stand for now — time constraints, etc.)

Special Matthew material (“M”) not present in Luke

Mark Goodacre argues this is not present in Luke because “Luke did not like it” while John Kloppenborg argues that “Luke would have liked it”.

But MK claims that if Luke followed Mcn then he was not “omitting” anything, but simply following this third source.

In the case of the birth narratives, MK agrees with those who argue that Luke did use Matthew here. There are the clear parallels between the virginal conception, the names of Jesus’ parents, the place of birth, etc.

Further, it is possible to show that the direction of influence was from Matthew to Luke — that Matthew’s account was the original one.

The whole logic of the narrative of Jesus being born in Bethlehem makes sense only for Matthew: he knew from Mark 1:9, 24 etc. that Jesus came from Nazareth but nevertheless was interested in depicting him as a descendant of David and did so by locating his birth in “Bethlehem of Judea” (Matt. 2:1, 5-6) whose christological importance is underlined by the formula quotation.

(I still question whether Nazareth was mentioned in Mark and if Matthew was attempting to explain an alternative meaning for the theologically contentious Nazarene epithet.)

So Matthew solved the conflict by bringing in the Herodian slaughter of the innocents, the flight to Egypt, . . . at each step stressing the theme of the “newborn king of the Jews” against the illegitimate (non-Davidic) king with formulaic quotations.

It is, therefore, evident that the Matthean “Bethlehem” is a necessary element in a well-crafted context.

Luke took over Bethlehem but without its central role in the narrative logic. Luke is more interested in the universal circumstances of Jesus’ birth than the Davidic theme.

So Matthew’s birth narrative can no longer be said to be “M” material. It can be seen to have had an influence on Luke.

The alternating primitivity of the double tradition material

One of the main reasons for the development of the two source hypothesis was the way Matthew sometimes appears to be using a more primitive form of a saying while in other cases it is Luke who appears to contain the more primitive version. By adding Q, one could argue that in some cases Matthew was copying from this in its original form and in others he was modifying his source. Ditto for Luke.

But MK offers something else:

On the assumption of Mcn being prior to Luke the observation of alternating primitivity finds a completely different and rather simple solution.

He considers four main examples where Luke seems to have a more primitive text than Matthew.

  1. It seems improbable that Luke would have changed Matthew’s “blessed are the poor in spirit” to “blessed are the poor”. But we learn from Tertullian (4.14.1) that Luke was following Mcn.
  2. The final beatitude in Matthew has Jesus mention suffering evil “on my account” while Luke has “on behalf of the Son of Man”. Tertullian again attests that what we read in Luke was in Mcn.
  3. Matthew’s Lord’s Prayer is more elaborate than Luke’s. Tertullian’s evidence leads us to infer that the shorter prayer was the form in the original Mcn. Further, Mcn had Jesus pray for “the spirit” to come rather than “the kingdom”. This original Lukan passage survived in a few medieval manuscripts before eventually being uniformly “corrected” to conform to Matthew’s version.
  4. Matthew 12:28 says that exorcism of demons was the work of the spirit, but Luke 11:29 attributes the feat to “the finger of God”. Tertullian (4.26.11) informs us that Mcn also had “finger of God”.

This evidence points to Luke using Mcn, and to some of Matthew’s elaborations must be Mathew adding to Mcn, also.

Luke’s presumed re-ordering of Matthean material

The major example here is the Sermon on the Mount. MK finds it hard to accept Mark Goodacre’s explanation that Luke would have broken up well-arranged structure of Matthew 5 to 7 and scattered it over a dozen different places throughout Luke 11 to 16.

But again, including Mcn in the discussion changes the picture completely. . . The overall picture [of Tertullian’s evidence of Mcn] confirms not only Luke’s direct dependence on Mcn but also demonstrated that Matthew collected the material for the composition of the Sermon on the Mount from different places in Mcn.

1. Matt. 5:13 // Luke 14:34-35 (parable of salt):—

2. Matt. 5:15 // Luke 11:33 (parable of light): Tert. 4.27.1.

3. Matt. 5:18 // Luke 16:17 (imperishability of the law): Tert. 4.33.9.

4. Matt. 5:25 // Luke 12:57-59 (on reconciling with your opponent): Tert. 4.29.15.

5. Matt. 5:32 // Luke 16:18 (on divorce and re-marriage): Tert. 4.34.1, 4.

6. Matt. 6:9-13 // Luke 11:2-4 (Lord’s prayer): Tert. 4.26.3-5.

7. Matt. 6:19-21 // Luke 12:33-34 (on collecting treasures):—

8. Matt. 6:22-23 // Luke 11:34-36 (parable of the eye):—

9. Matt. 6:24 // Luke 16:13 (on serving two masters): Tert. 4.33.1-2; Adam., Dial. 1.26.

10. Matt. 6:25-34 // Luke 12:22-31 (on anxiety): Tert. 4.29.1-5.58

11. Matt. 7:7-11 // Luke 11:9-13 (God’s answering of prayer): Tert. 4.26.5-10; Epiph. 42.11.6 (schol. 24).

12. Matt. 7:13-14 // Luke 13:23-24 (the narrow gate):—

13. Matt. 7:22-23 // Luke 13:26-27 (warning against self-deception): Tert. 4.30.4.

The minor agreements between Matthew and Luke within the triple tradition material

To cover MK’s argument here would involve too much detailed complexity. I will mention only a few details here.

The prime example is, of course, the addition of the five words τίς ἐστιν ὁ παίσας σε [“who struck you”] to Mark 14:65 in Luke 22:64 and Matt. 26:68.

These extra words are “well enough attested for Mcn” (Epiphanius, Panar. 42.11.6)

Another case that can be checked against Mcn:

In the Markan version of the pericope about the true relatives (Mark 3:31-5 par.), Jesus is being told that “his mother and his brothers and his sisters are seeking him outside.”65 Luke (8:20) and Matthew (12:47) agree in leaving out the “sisters” (a negative agreement) and in adding that they were “standing outside” (ἔξω ἑστήκασιν). This is exactly what Tertullian read in Mcn. (my emphasis)

And again with the mustard seed:

Similarly, in the parable of the mustard seed, Luke and Matthew use a formulation different from Mark: Mark describes the action of sowing in the passive voice and does not name a subject. Both Matthew and Luke use the active voice, mention the subject and note that the man threw the seed on his own soil. Tertullian, again, attests this very phrase for Mcn.

And finally:

A last example is the annunciation of Jesus’ passion and resurrection (Mark 8:31 par.): Mark dates the resurrection “after three days (μετὰ τρεῖς ἡμέρας),” whereas Matthew (16:21) and Luke (9:22) both give the ordinal number “on the third day (τῇ τρίτῃ ἡμέρᾳ)”, as does Mcn.

MK also explains that Mcn cannot explain all of the minor agreements. I skip the extra details for now. His conclusion, however, Matthew and Luke both used Mcn and that Luke at times did also draw upon Matthew.

Adding Marcion to the picture

Adapting MK’s diagram:

The thick arrows, 1, 2, 3, are the main influences: narrative outline is copied and there is verbal agreement.

The dashed arrows, a and b, indicate additional but minor influences of Mcn on Matthew and Luke, with some influences from Matthew to Luke’s re-edition of Marcion.

Matthew followed Mark closely but at times added some material from Marcion, often editing it in the process:

  • The sermon in the plain (Mcn – to beome Lk – 6:20-49)
  • The healing of the centurion’s boy (7:1-10)
  • John the Baptist’s question (7:18-23)
  • On following Jesus (9:57-62)
  • Commissioning of the apostles (10:1-11)
  • Thanksgiving to the father and the beatitude of the disciples (10:21-24)
  • The Lord’s prayer (11:1-4)
  • Teaching about prayer (11:9-13)
  • Exhortation to fearless confession (parts) (12:2-5, 8-9)
  • Teaching on anxiety (12:22-27, 29-32)
  • Interpreting the times (12:20-21)
  • Parable of the great supper (14:15-24)
  • Parable of the lost sheep (15:3-7)
  • Concerning law and divorce (16:16-18)
  • On forgiveness (17:3-4)
  • Parable of the good and wicked servants (12:41-46)

A detailed discussion will have to await another post if there is enough interest — or I simply invite interested readers to acquire a copy of the article. To mention a few points only, here:

There is a double influence on Matthew within the triple tradition material that now shows up in Matthew and Luke. “These are the major and minor agreements.”

The b line influence from Matthew on Luke includes the birth stories, Jesus’ genealogy, Jesus’ baptism and temptation.

Matthew’s “sign of Jonah” is taken over by Luke, but this was not in Mcn.

One detail Luke copied from Mark was the parable of the wicked husbandmen (20:9-18)

Marcion’s gospel ended with 24:43 “so the ascension and the disciples’ return to Jerusalem are Lukan additions as well.”

Too complex?

MK remarks:

So far, the diagram and its additional explanations should provide the general idea of how the picture changes when Mcn is included in the solution of the synoptic problem. When I discussed this model with my students, they immediately responded that it was too complicated to be convincing (which was somewhat discouraging at the moment). But how complicated is too complicated?

In its favour MK lists

The inclusion of Mcn avoids the methodological weakness of the 2DH/2SH with regard to the minor agreements and the hypothetical character of Q. The alternating primitivity between Matthew and Luke is explained.

The reason for its initial impression of complexity is the result, MK points out, of “a major shift in methodology when compared to 19th century source-criticism.” Instead of simply adding sources on the assumption that the evangelists were mere compilers and editors, this model accounts for them being competent authors with ability to merge and adapt sources to fit their narrative and theological needs.

[W]hereas the 2DH tried to explain the complexity of the data by the addition of two basic sources (Mark + Q), the inclusion of Mcn demonstrates that both Matthew and Luke received their triple tradi- tion material via two different routes: Matthew read Mark directly and in its revised edition in Mcn, and Luke used Mcn both directly and in Matthew’s revised and enlarged edition. Since Luke, as it was demonstrated, also did know and use Mark, Mark was present in all stages of the synoptic tradition. The editorial procedure of both, Matthew and Luke, was not a mere addition of “sources” but a comparison of texts and concepts. This is fully consonant with the insight of the redaction history that the evangelists were ambitious and competent authors rather than mere editors. (my emphasis)

MK concludes with the comment that his paper does not cover all of the implications and consequences of the Mcn hypothesis, but it does serve as an attempt to break new ground for further discussion of the Synoptic Problem.

The following two tabs change content below.

Neil Godfrey

Neil is the author of this post. To read more about Neil, see our About page.

Latest posts by Neil Godfrey (see all)

If you enjoyed this post, please consider donating to Vridar. Thanks!

26 thoughts on ““The Marcionite Gospel and the Synoptic Problem: A New Suggestion””

  1. What is so repellent about Matthew being a rewrite of Luke?
    The temporal incongruities involving Herod and the Quiraneus census were eliminated at the beginning of the story, with some magi stirred in to appeal to the Mithra folks (as opposed to shepherds in Luke who were a sop to the Greeks with its appeal to an ancient Arcadia populated by sheep tenders). “Matthew” revised the end to include a tomb guard to counter the “stolen body” rumor being circulated by Greeks ( later recorded by Celsus) and Jews ( in a Toldoth Jesu precursor text).. Luke then Matthew written and holds together better than Luke, consistent with a later literary clean up. Do this and you gets rid of a need for Q, since the Luke/Matthew overlap is due to the second copying the first. And additional “M” material sprung from “Matthew’s'” imagination.
    Hollywood does rewrites all the time (like story for kids: Willy Wonka and the Choclate factory), so why cannn’t a late 1st century religious propagandist do a creative rewrite for his kids?

  2. Theories of Marcionic priority are not over-complicated once one puts the major apocryphal gospels into the mix. Traditional scholarship assumes that the apocryphal gospels drew from all the canonical ones, but that creates a very complicated diagram indeed. Placing at least some of them prior to the canonical gospels simplifies the picture considerably.

    My Hyper-Synoptic Hypothesis attempts to build upon the work of those like Klinghardt in proposing an even larger view of gospel composition. The latest version is here:


    Of Klinghardt’s proposal, I would say that (1), (2), and (3), are clearly correct, but (a) and (c) are mistaken and (b) is over-simplified. Klinghardt’s largest error is to assume that Matthew drew much of the Q tradition from Marcion. Instead, Q was very real: it was a full-fledged gospel, that we know today as the Gospel of Peter. Marcion used it first, but Matthew used it independently and later, in response to Marcion, in part as a kind of theological counterstrike against him.

    As for (b), the Gospel of the Nazarenes was likely an intermediary between GMt and GLk, though I have not yet demonstrated this.

  3. In my totally amateur way I have suspected for a while that Luke was developed from Marcion. But I also find it curious that although other Gospels are claimed to have been written before Marcion’s Gospel, his seems to be the earliest actual collected canon of Christian scripture. And – based on the similarity of names alone – I cannot help wondering about the relationship between MARCion and MARK.

    “that admiral place to start a gospel”

    Aye, aye, Sir.

  4. I was led to Matthias Klinghardt’s article through a footnote in Robert M. Price’s chapter in “Is This Not the Carpenter?” I’ll tie it in with a broader look at the place of Marcion’s “canon” in the next post on Price’s chapter.

  5. What is it that makes us think we have anything like these documents in their original forms (possibly oral), and that the first versions we have have not been cross-infected with each other and other sources many times. These were tools in a battle of ideas and each time one changed and achieved a seeming example, the school behind another would then change theirs in response.

  6. I don’t think we can trust the witnesses to Mcn in two completely different ways. Klinghardt and the other Mcn-theory supporters would have it that we cannot trust Irenaeus and Tertullian and etc to be correct about Luke being first, and yet we’re also supposed to trust their witness to the text of Mcn simultaneously? I can’t buy that. The only way that would work was if they were all unaware of the correct dating.

    Goodacre, in my view, still has the best current explanation for what we have.

  7. I did a fairly brainless synoptic solution, based on the picture on Wikipedia. My assumption was simply that the writers/redactors would NOT delete anything: if they had a holy document – they would tend to append ALL the extant data onto their new edition.


    This produces a fairly logical flow chart, where there must have been 2 versions of proto-mark, and 1 proto-Luke. Matthew actually ends up being a harmony between proto-Luke, and proto-Mark. Thus Matthew and Luke cannot be dated relative to each other. Matthew may be late. The only anomaly, is the grey stuff (agreements between all the other gospels against Matthew.) my brainless hypothesis predicts that this material must have been objectionable, so got deleted in the harmony.

    The thing I don’t like about including Marcion’s gospel into the picture, is that it may well just be a butchered version of Luke. Here is the reconstructed gospel

    It has all the material which is unique to Luke, against the other synoptics. For me, that pretty much kills the idea that Marcion’s gospel can contain any useful information, because it dates later than proto-Luke. If Marcion was using proto-Luke, then it would be a goal for the skeptics.

    Justin Martyr is clearly using a version of Matthew which includes unique Matthean material. (I’m not sure if he is working from full-version-Luke)

    Cheers and beers.

  8. “The Marcionite Gospel and the Synoptic Problem: A New Suggestion”, by Matthias Klinghardt, seems, to me, to be working in the right direction. However, methinks a further step needs to be taken with Marcion’s ur-Luke 😉

    That step is to face the possibility that Marcion’s ur-Luke is the first of the synoptic gospels. Yes, of course, I’m no expert in all the finer detail in the synoptic debate – what I do is follow storyline development. And that storyline development runs from an ur-Luke to a gLuke of post Antiquities of 93/94 c.e. In between rests an ur-Matthew. gMark is placed, in my system, after gMatthew and prior to gLuke.

    The wider storyline development, prior to an ur-Luke and it’s 15th year of Tiberius, is from that very contentious material now preserved in Slavonic Josephus. A story from prior to the 15th year of Herod the Great. Ur-Luke, the version of Marcion, is the first dated update of the older storyline. The Slavonic Josephus wonder-doer storyline is, like Marcion’s ur-Luke, not a backward reduction i.e. a story taken from a later source minus ‘questionable’ bits, but a story in its own right. And as such, needs, as does a Marcion’s ur-Luke, to be put on the table in any debate over the synoptic ‘problem’.

    What this reordering of the synoptic gospels would suggest is that some ahistoricists/mythicists have to get away from the idea that gMark is a story about a historizing of Paul’s spiritual Christ figure. gMark is late in the developing synoptic storyboard. gMark is not just a handy summary of what has gone before: by removing all of gMatthew’s nativity details, it cleared the road ahead for gLuke and its new ball game of a nativity in the time of Quirinius (around 6 c.e.) Sometimes it is what has been dropped from a story that is just as important as what has been added…

    Oh, just decided to put up a chart on FRDB…;-)


  9. Astrayan, your flowchart made me very enthusiastic.

    But you can make it work even better.

    I have drawn a flowchart on paper myself. I haven’t recreated it on the computer, but I will describe what it looks like.

    It looks like a hexagon with in the middle an extra dot. The hexagon consists of one dot at the top, one at the bottom, one top left, one top right, one bottom left, and one bottom right.

    The dot in the middle represents the triple tradition. This is the material shared by Mark, Luke and Matthew. This is a sort of proto-Mark.

    From the point in the middle goes a line to the dot at the top left, which represents a gospel which consists of the triple tradition plus the material shared by Luke and Mark but which is not present in Matthew. This is basically another proto-Mark (let’s call it proto-Mark+KL ,where KL is the material shared by marK and Luke).

    From the point in the middle goes a line to the dot at the top right, which represents a gospel which consists of the triple tradition plus the material shared by Matthew and Mark but which is not present in Luke. This is another proto-Mark (let’s call it proto-Mark+KM, where KM is the material shared by marK and Matthew).

    At the top is a dot which represents the gospel of Mark as we know it today. A line goes from the dot at the top left to the dot at the top, and another line goes from the dot at the top right to the dot at the top. The gospel of Mark is a merge of proto-Mark+KL and proto-Mark+KM, plus some extra material which is only present in the gospel of Mark (or K-source, where K stands for marK).

    From the dot in the middle goes another line, this time to the dot at the bottom. This represents a gospel which consists of the triple tradition, plus some extra material only shared by Matthew and Luke, but which is not present in Mark. This is basically the triple tradition plus Q. Let’s call this gospel proto-Mark+Q.

    At the bottom left is the gospel of Luke. A line goes from the top left to the bottom left, and another line goes from the bottom to the bottom left. The gospel of Luke is a merge of proto-Mark+KL and proto-Mark+Q, plus some extra material only present in Luke (L-source).

    At the bottom right is the gospel of Matthew. A line goes from the top right to the bottom right, and another line goes from the bottom to the bottom right. The gospel of Matthew is a merge of proto-Mark+KM and proto-Mark+Q, plus some extra material only present in Matthew (M-source).

    So, to sum it up. There are 7 synoptic gospels.

    Gospel 1: A proto-Mark, which is the triple tradition.
    Gospel 2: which is gospel 1 + KL-source
    Gospel 3: which is gospel 1 + KM-source
    Gospel 4: which is gospel 1 + Q-source
    Gospel 5: Mark, which is gospel 2 + gospel 3 + K-source
    Gospel 6: Luke, which is gospel 2 + gospel 4 + L-source
    Gospel 7: Matthew, which is gospel 3 + gospel 4 + M-source

    After I drew the flowchart I felt a moment of eureka. I hadn’t thought before about a solution where Matthew and Luke didn’t use a sayings gospel at all. They used a now lost gospel which is similar to Matthew and Luke, but which is just the gospel of Mark plus some extra material, perhaps some sayings gospels. The Q-gospel doesn’t exist. This lost gospel also had a genealogy of Jesus which both Matthew and Luke used and adapted independently for their gospels.

    It could explain why the reconstructed Q-gospel is largely a sayings gospel plus some very weird out of place narrative elements. This could be because the lost synoptic gospel (gospel 4) has some extra narrative elements added to the triple tradition. Matthew and Luke copied these elements, and that’s why modern scholars who try to reconstructed Q think these weird elements are part of the Q-gospel.

    The lost synoptic gospel could also explain why Matthew and Luke seem to have copied from each other. This is because they simply used the same gospel.

    I don’t rule out that the gospel writers included several sayings-gospels in their own gospels. You could think about an earlier version of the gospel of Thomas, or perhaps some other unknown sayings gospels. But Q as a single sayings gospel doesn’t exist.

    But because the 7-synoptic-gospels theory is so contrived, I now present a simpler solution:

    First there was Mark. Then there was an unknown gospel writer who wrote a now lost gospel, who used Mark and added extra material, now known as Q. Matthew and Luke independently used this unknown synoptic gospel. They perhaps weren’t aware of the gospel of Mark. And they weren’t aware of each other.

  10. If you assume that gospel writers always copy all the material they receive from their sources, and you add Marcion to the equation, then the synoptic problem becomes very difficult. In case of 3 synoptic gospels you need 4 hypothetical lost gospels: 3 versions of the gospel of Mark, and Q. In case of 4 synoptic gospels you need 11 hypothetical lost gospels. You can make it work by imagining a pyramid with 4 sides (a tetrahedron). The four synoptic gospels are in the vertices. You start from the centre of the pyramid which represents the quadruple tradition, the material which is shared by all four gospels. This represents the first lost gospel. Then you create four lost gospels in the 4 triangular faces of the pyramids. All four add triple tradition material to the first gospel, material that is shared by three gospels which are connected to the triangular plane, but not shared by the other gospel. Then you add 6 extra lost gospels by going to the 6 edges of the pyramid. Material is merged by taking material from the two faces which are connected to the edges, and by adding material that is shared by two gospels which are connected to the edges but not by the other two gospels. On the vertices are the final four gospels. They are created by merging the material from the 3 edges which are connected to these gospels, and material is added which is unique to the particular gospels.

    It works like this:

    I will use a code system.

    K = gospel of Mark
    M = gospel of Matthew
    C = gospel of Marcion
    L = gospel of Luke

    When I say for instance KM/CL then I mean all the material that is shared by Mark and Matthew, but which is not present in Marcion and Luke. And when I says for instance MCL/K then I mean all the material that is present in Matthew, Marcion and Luke, but not in Mark.

    And when I say for instance 6 = 2 + 3 + KM/CL then I mean that gospel 6 is a merge of gospel 2 and 3 plus some material that is shared by Mark and Matthew but which doesn’t appear in Marcion and Luke.

    I will use 4 phases.

    Phase 1:
    1 = KMCL (all the material shared by Mark, Matthew, Marcion and Luke)

    Phase 2:
    2 = 1 + KMC/L
    3 = 1 + KML/C
    4 = 1 + KCL/M
    5 = 1 + MCL/K

    Phase 3:
    6 = 2 + 3 + KM/CL
    7 = 2 + 4 + KC/ML
    8 = 2 + 5 + MC/KL
    9 = 3 + 4 + KL/MC
    10 = 3 + 5 + ML/KC
    11 = 4 + 5 + CL/KM

    Phase 4:
    12 = K = 6 + 7 + 9 + K/MCL
    13 = M = 6 + 8 + 10 + M/KCL
    14 = C = 7 + 8 + 11 + C/KML
    15 = L = 9 + 10 + 11 + L/KMC

    This is very complicated. But I can simplify it further. I know that basically all the material in the gospel of Marcion is also present in the gospel of Luke. There seems to be one exception though, and that is that Matthew 15:24 and 15:26 are also present in Marcion, but not in Luke, but let’s call this a minor glitch. Perhaps in the past these verses were also present in Luke.

    Because (nearly) all Marcion’s material is also present in Luke, that means that:

    KMC/L = 0 (zero means that it doesn’t contain information)
    KC/ML = 0
    MC/KL = 0 (so actually this one is not entirely zero)
    C/KML = 0
    KCL/M = KC/M
    MCL/K = MC/K
    CL/KM = C/KM
    KM/CL = KM/L
    K/MCL = K/ML
    M/KCL = M/KL

    So some gospels will disappear.

    It is possible that KL/MC also contains very little information, but I’m not sure. There is not enough information about the gospel of Marcion to determine this. I’ve looked it up and there are some crucial passages of which it is not known if they were present or not, so I can’t answer the question.

    This results in:

    Phase 1:
    1 = KMCL (proto-Mark 1)

    Phase 2:
    2 = 1 + KML/C (proto-Mark 2)
    3 = 1 + KC/M (proto-Mark 3)
    4 = (1 +) MC/K (traditional Q-source minus ML/KC, it becomes optional not to add gospel 1)

    Phase 3:
    5 = 2 + KM/L (proto-Mark 4)
    6? = 2 + 3 + KL/MC (proto-Mark 5, perhaps this one can be removed too)
    7 = 2 + 4 + ML/KC (proto-Matthew)
    8 = C = 3 + 4 + C/KM (gospel of Marcion)

    Phase 4:
    9 = K = 5 + 6? + K/ML (gospel of Mark)
    10 = M = 5 + 7 + M/KL (gospel of Matthew)
    11 = L = C + 6? + 7 + L/KMC (gospel of Luke)

    So then you need only 7 (or 6) additional lost gospels: 5 (or 4) lost versions of Mark, 1 proto-Matthew, and Q. The gospel of Luke becomes a complicated merge between the gospel of Marcion, proto-Matthew, and a proto-Mark (version 5). The last proto-Mark (version 5) may be unnecessary. The gospel of Matthew is a merge of a proto-Matthew and another version of the gospel of Mark.

    It is interesting to notice that in this case it is more economical to assume that Q didn’t contain a version of Mark. So you just need one arrow less. Perhaps the sayings gospel Q existed after all.

    I think this theory becomes too complicated for many people. But perhaps this is how it could have really happened. Perhaps the formation of the 3 synoptic gospels was a complex and chaotic process. Perhaps you can compare a gospel to a bacterium. It multiplies, it mutates, it exchanges genetic information with other gospels etc… Perhaps there was a group/family of gospels which very much looked like the gospel of Mark today. Out of these gospels grew a group of gospels that more or less looked like the gospel of Mark with the Q material added, and some more. And out of this group of gospels grew two new group of gospels which more or less looked like the gospels of Luke and Matthew. Then the gospel of Marcion belongs to the Luke-family of gospels. The gospels continuously exchanged information and that’s what makes it difficult to tell who copied who. Out of this pool of perhaps hundreds of synoptic gospels, the church picked three. This was the start of the synoptic problem.

  11. A little correction here: Matthew 15:24 and 15:26 are NOT present in the gospel of Marcion. Sorry. It was the other way around, these verses were actually present in the gospel Luke, when Tertullian compared Marcion’s gospel with the gospel of Luke. These verses are not present in the gospel of Luke anymore. And these verses weren’t present in the gospel of Marcion.

    Why I’m doing all this is because I’m reading the book by Earl Doherty: Neither Man, nor God. I’m very impressed by what he says about Paul. I find it very convincing. But Earl Doherty also writes a lot about the Q-gospel with all its layers. The problem is, did this gospel actually exist? If it did not, then in reality Earl Doherty talks a lot into a vacuum. Perhaps the different layers of Q are in reality 2 separate gospels, which a predecessor of Matthew and Luke added to the gospel of Mark.

    I also wonder if Doherty has the dating of the epistles of Paul right. Paul mentions king Aretas having control over Damascus, and only Aretas III occupied Damascus. But this was roughly 60 BC! So perhaps the epistles of Paul are much older than we could ever believe.

  12. That is an extremely intriguing theory.

    However, I am extremely skeptical of the notion that Luke used Matthew.
    Did he purposefully give Joseph a different father?
    Omit all those juicy Messianic prophecies?

    Luke’s agreement with Marcion could be the result of one of two things
    1) Luke copied Marcion
    2) Marcion redacted Luke (or a proto-Luke?)

    You are acting like agreement between Luke and Marcion necessarily indicates 1, when it could just as easily be 2.

    I think it would make more sense to say that Matthew and Luke incorporated an infancy gospel source and gave some of their own details than to say that Luke borrowed this story from Matthew.

  13. Did he purposely give Joseph a different father? It’s quite plausible: See http://vridar.org/2012/12/11/is-this-not-the-carpenter-reviewing-chapter-11-lukes-sophisticated-re-use-of-ot-scriptures/

    Omit some of Matthew’s specific prophecies? Quite plausible. Don’t think of Luke somehow trying to “improve” upon Matthew but writing in some dialogue with it — Luke has a quite different perspective of Jesus and his mission. Matthew has gentiles involved in the birth of Jesus, Luke keeps gentiles out of things till Acts. Luke is imitating the stories of Genesis and Judges etc of early Israel.

    I’m open to all options. Yes, Marcion could well have redacted a proto-Luke, too.

    1. Luke’s interpretation of scripture was involved in his geneology, no question. But I still doubt that he would blatantly contradict one of his sources.

      Do you think that the author of Luke also wrote Acts?

      If so, why the different account’s of Judas’s death? The different baptismal formulas?

      1. Genealogies are political-ideological statements. The different genealogies evidence quite divergent theologies. But if a writer chose to write in some dialogue with another work then whether he contradicted or agreed with his partner depended upon his reasons and motivations. John certainly contradicts Mark. Matthew sometimes contradicts Mark, too.

        It may be argued that Luke is a catholicizing gospel — he is attempting to reconcile or correct/make more universal a range of other gospels (not only the canonical ones).

        I tend to think that the author of Acts also redacted an earlier version of Luke. I’ve read a bit, but by no means all I would like. I’m open to learning new things about it all.

  14. Perhaps the sequence went like this:

    1) Proto-Luke was written
    2) Luke used it as a source.
    3) Marcion obtained a copy of proto-Luke through his connections
    as a ship-captain and used it as the basis of his euangelion.
    4) The Orthodox accused Marcion of corrupting Luke.

        1. I meant that the final redactor of Luke — whoever was the final “Luke” — modified an earlier form of “Luke” to rebut Marcion. I think that’s Tyson’s view.

  15. From above: “A last example is the annunciation of Jesus’ passion and resurrection (Mark 8:31 par.): Mark dates the resurrection “after three days (μετὰ τρεῖς ἡμέρας),” whereas Matthew (16:21) and Luke (9:22) both give the ordinal number “on the third day (τῇ τρίτῃ ἡμέρᾳ)”, as does Mcn.”

    Perhaps some clarification is in order. We find in DeBuhn, Jason. The First New Testament: Marcion’s Scriptural Canon, 2013, p. 149, an opposite statement to the above.

    “It (the Evangelion) had ‘after three days’ in agreement with Gk ms D, OL, and Mark 8:31, against ‘on the third day’ found in most manuscripts of Luke and Matthew — the latter being one of the ‘minor agreements’ of Luke and Matthew against Mark, missing from the Evangelion.”

  16. Hi Neil, I just want to highlight another publication from Matthias Klinghardt that might add to the above:
    Marcion’s Gospel and the New Testament: Catalyst or Consequence?, NTS 63 (2017), 318-321
    Klinghardt’s book from 2015 is at the moment only available in German, but acc. to him an English version is in preparation.
    “Das älteste Evangelium und die Entstehung der kanonischen Evangelien: Band I: Untersuchung | Band II: Rekonstruktion, ÜberSetzung, Varianten”
    Best, Frank

        1. Material survey

          • Heil, Uta (2017). “Editorial”. Zeitschrift für Antikes Christentum / Journal of Ancient Christianity. 21 (1): 1–7. doi:10.1515/zac-2017-0000. [“Available online”. Academia.edu]:

          In 2014, Markus Vinzent published Marcion and the Dating of the Synoptic Gospels, and in 2015, three major monographs on Marcion appeared: Matthias Klinghardt, Das älteste Evangelium und die Entste-hung der kanonischen Evangelien, Dieter T. Roth, The Text of Marcion’s Gospel, and Judith M. Lieu, Marcion and the Making of a Heretic: God and Scripture in the Second Century.
          in 2009 two PhD theses were completed, one by Sebastian Moll on The Arch-Heretic Marcion , published in 2010, written under the supervision of Sara Parvis, and the other by Dieter T. Roth under the title “Towards a New Reconstruction of the Text of Marcion’s Gospel: History of Research, Sources, Methodology, and the Testimony of Tertullian,” which resulted in the publication The Text of Marcion’s Gospel in 2015.
          Vinzent’s interest in Marcion did not emerge from New Testament research, but―as a Patristic scholar―from his observations on the second century. It is specifically the question how early Christians thought about the resurrection which led him to Marcion.
          the editors of the Journal of Ancient Christianity have decided to publish a special issue on Marcion. Included are two papers by Jason BeDuhn and Dieter T. Roth, who present the debated issues from their point of view. These are followed by an article on Luke 24 and the resurrected Jesus by Daniel A. Smith,

          • Klinghardt, M. (2017). Marcion’s Gospel and the New Testament: Catalyst or Consequence? New Testament Studies, 63(2), 318-323. doi:10.1017/S0028688516000461

          • Beduhn, J. (2017). Marcion’s Gospel and the New Testament: Catalyst or Consequence? New Testament Studies, 63(2), 324-329. doi:10.1017/S0028688516000473

          • Lieu, J. (2017). Marcion’s Gospel and the New Testament: Catalyst or Consequence? New Testament Studies, 63(2), 329-334. doi:10.1017/S0028688516000485

  17. Per “2SH / 2DH”, are there any errors in the following statements?

    • narrative material that Matthew (A) or Luke (B) share with Mark (C) [i.e. (A⋂C)⋃(B⋂C)]—came from Mark, and everything Matthew and Luke share only with each other [i.e. (A⋂B)⊄C]—came from Q.

    • narrative material that is unique to the Gospel of Matthew is derived from a source named M, and narrative material that is unique to the Gospel of Luke is derived from a source named L.

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from Vridar

Subscribe now to keep reading and get access to the full archive.

Continue reading