When Mark Goodacre asked, “Why Not Matthew’s Use of Luke?”

(A review, by Robert K. MacEwen, of a paper presented at the annual meeting of the Society of Biblical Literature, 18 November 2018)

It was standing room only in room 302 of the Denver Convention Center when Dr. Mark Goodacre, Frances Hill Fox Professor of Religious Studies at Duke University, took his place behind the lectern. Looking around, Goodacre expressed surprise at the size of the crowd. “You do realize this is a session on the Synoptic problem, don’t you?” he asked.

Certainly, Ron Huggins and I, seated in the front row, were not there by accident. We were eager to hear what Goodacre would say in response to our view of the relationships between the Gospels. The Matthean Posteriority Hypothesis (MPH) has often been ignored by scholars, ever since it was first proposed by G. C. Storr in 1786.[1] Therefore, it is gratifying to proponents of the MPH that Goodacre is engaging with their theory. A year ago, Goodacre debated online with Alan Garrow in a “$1,000 Challenge.”[2] More recently, he gave a response to a paper by Garrow at the British New Testament Conference.[3] And now, Goodacre was taking on the MPH in a paper presented at the annual meeting of the Society of Biblical Literature.[4]

Goodacre is today the most prominent advocate of the Farrer Hypothesis (FH)—the view that Luke used Matthew as well as Mark as sources in composing his Gospel. In defending his preferred solution to the Synoptic problem, Goodacre has primarily engaged with scholarship’s dominant theory, the Two Source Hypothesis (2SH), which argues that Matthew and Luke independently used Mark and the hypothetical “Q” document as sources. The MPH, which is the view that Matthew came third and used Mark and Luke as his sources, is the logical third alternative to the other two hypotheses.

Goodacre’s Arguments against Matthew’s Use of Luke

Goodacre began and ended his paper by praising the MPH for its points of agreement with his own theory. He noted that the MPH correctly builds on the priority of Mark, insists on “a literary solution” to the Synoptic problem, and views Q with skepticism (1-3, 22). Naturally, the bulk of Goodacre’s paper was devoted to arguing that Luke’s use of Matthew explains the phenomena of the Gospels better than Matthew’s use of Luke.

Following his introduction, in a section titled, “First Impressions,” Goodacre set out features of Luke that he feels support dating it later than Matthew. These include Luke’s reference to earlier writings about Jesus (Luke 1:1), his use of the first person (Luke 1:1-4 and the “we” passages in Acts)—characteristic of later Gospels, and the historical references he has in common with Josephus.

The next section of Goodacre’s paper was titled, “Matthew’s Redactional Fingerprints”. Here, Goodacre presents two verses in the triple tradition containing minor agreements of Matthew and Luke against Mark (Matt 14:13; 22:27; and parr.), arguing that these are examples in which Luke adopted Matthew’s characteristic wording. Next, he displayed the double tradition pericope John the Baptist’s Preaching (Matt 3:7-10//Luke 3:7-9) and argued that its key elements are distinctive of Matthew’s literary and theological features.

Goodacre’s next section presented his argument from “editorial fatigue” (already well-known to his readers[5]), involving passages in which “an author inadvertently betrays his use of a source by making characteristic changes at the beginning of a passage only to revert to the source’s wording later in the same passage”. Goodacre presented the Parable of the Entrusted Money (Matt 25:14-30//Luke 19:11-27) as “[o]ne of the best examples” of Luke fatiguing in his use of Matthew. He also asserts that there are “multiple examples of fatigue” in both Matthew’s and Luke’s use of Mark, “several cases” of fatigue in Luke’s use of Matthew in the double tradition, but no examples of Matthew apparently fatiguing in using Luke.[6]

The following section is on Luke’s “Knowledge of Matthew’s Literary Structures”. Here Goodacre gave the example of Luke 7:1, where Luke concludes the Sermon on the Plain with a construction similar to those used by Matthew at the end of all five of his major discourses of Jesus. Goodacre’s point being that Luke has, in this single instance, adopted a motif that is characteristic of Matthew.

After this, Goodacre discussed “Matthew’s Failure to Include Congenial Lukan Details”. Here the argument is that, since Matthew includes more information about contemporary political leaders than does Mark, it is surprising that he omits Luke’s list of seven rulers in Luke 3:1-2 (cf. Matt 3:1) if Luke were also his source.

Goodacre’s final section was titled, “What is the Appeal of Matthean Posteriority?” Here he states that MPH proponents make use of two “popular arguments for Q” that are actually invalid “old chestnuts, the argument from order, and the argument from Lucan primitivity”. Regarding Luke’s alleged primitiveness in the double tradition, Goodacre makes three substantive points in response. Regarding the argument from order, he suggests that scholars have been hoodwinked by B. H. Streeter’s “rhetoric” and, in a footnote, refers readers to his earlier discussion of this topic. 

A Matthean Posteriority Response to Goodacre’s Arguments

Having outlined the contents of Goodacre’s paper, I now offer a Matthean posteriority response. Regarding a second-century date for Luke, the arguments for this are hardly conclusive. Luke’s historical references in common with Josephus do not establish that Luke depended on Josephus; such information could have been known from many sources, including hearsay, in the first century. In favor of a first-century date for Luke, it is possible that some very early Christian writings depended on Luke (cf. Luke 10:17 with 1 Tim 5:18 and Luke 24:36-43 with Ign. Smyrn. 3). As for Luke’s similarities to later, non-canonical Gospels, we should not forget that Luke has even greater similarities to Mark and Matthew.

Goodacre’s argument that Luke incorporates Matthean redaction is also not conclusive for his theory; there are also many examples in which Matthew appears to be aware of Lukan redaction.[7] The fact that Matthew uses an expression such as “offspring of vipers” more often than Luke does not necessitate that Luke received it from Matthew. Otherwise, Matthew’s multiple use of Markan items such as the accusation “prince of demons” (Matt 9:34; 10:25; 12:24; Mark 3:22), the proclamation “Repent, for the kingdom of heaven is near” (Matt 3:2; 4:17; Mark 1:15), or the title “son of David” (ten times in Matthew, three times in Mark) would prove that Matthew could not have been dependent on Mark. On the contrary, on a Markan priority view, such Matthean repetitions show that Matthew had the tendency to multiply expressions from his sources that he found congenial.

Helping to confirm this Matthean tendency is an interesting feature of several of the expressions that Goodacre regards as Matthew’s characteristic expressions picked up once by Luke. In each case, it is Matthew’s first use of the expression that is parallel to Luke’s use of it: (1) “offspring of vipers”—Matt 3:7//Luke 3:7; Matt 12:34; 23:33; (2) “weeping and gnashing of teeth”—Matt 8:12//Luke 13:28; Matt 13:42, 50; 22:13; 24:51; 25:30; (3) “you of little faith”—Matt 6:30//Luke 12:28; Matt 8:26; 14:31; 16:8; (4) “And it happened when Jesus finished . . .”—Matt 7:28//Luke 7:1; Matt 11:1; 13:53; 19:1; 26:1. This phenomenon suggests that, in each case, Matthew first encountered the usage in Luke, found it congenial, and chose to use where Luke does and again in other appropriate settings.

As for the alleged Matthean character of John the Baptist’s Preaching (Matt 3:7-10//Luke 3:7-9), this could be explained by Matthew’s desire (worked out later in his Gospel) to show Jesus and John as being in agreement. Certainly, this is what Matthew does in his redaction of Mark, when he puts the initial message of Jesus on the lips of John as well (Matt 3:1-2; Matt 4:17//Mark 1:15).

Goodacre’s discussion of Matt 3:7-10 and Luke 3:7-9 would have been more balanced had he mentioned the different target audiences of John’s preaching in the two Gospels. In Luke 3:7 John warns the crowds in general, while in Matt 3:7 John polemicizes against the Pharisees and Sadducees. Here Luke’s usage appears to be more primitive, while Matthew’s appears redactional (Matthew is the only evangelist who groups the Pharisees and Sadducees together. See Matt 16:1, 6, 11-12; 22:34.).

“Editorial fatigue” is an important argument for the FH, at least if Goodacre is entirely correct in his analysis of the phenomena. More work needs to be done on this issue by a variety of scholars. Questions to be answered include: (1) Is it true that there are no plausible examples of Matthew fatiguing when editing Luke? (2) Could it sometimes be editorial alertness rather than editorial fatigue? That is, could it be Gospel A removing inconcinnities in editing Gospel B rather than Gospel B creating inconcinnities in editing Gospel A?

As for Matthew’s failure to include Luke’s list of seven rulers (cf. Matt 3:1; Luke 3:1-2), Goodacre’s argument here may be suggestive, but it is not strong. It is typical for Matthew to shorten his narratives by deleting material from Mark, including material with inherent interest (e.g., cf. Matt 9:1-8 with Mark 2:1-12; Matt 9:18-26 with Mark 5:21-43). Matthew never mentions political leaders unless they are part of his story. He could have easily omitted Luke’s seven-name list because it had no theological or narrative significance for him.

Roasting the “Chestnuts”

Goodacre views “alternating primitivity” in the double tradition as a poor argument for Q (one of “two old chestnuts”) that MPH proponents have attempted to co-opt for their theory. I agree that Matthew and Luke’s apparent alternating primitiveness is not sufficient to establish the mutual independence of these two Gospels. Goodacre’s strongest argument here is that Luke could have sometimes replaced the wording he found in Matthew with an expression known to him from oral tradition. Of course, this argument is reversible—one could just as well say, on the MPH, that Matthew could have been influenced by oral tradition while using Luke as a source.[8]

Goodacre is on much shakier ground when it comes to his other “old chestnut,” the argument from order. In accusing MPH proponents of substituting “a repackaging of Streeter for an attempt to engage seriously with his critics” (20), he seems to be engaging in the kind of hand-waving he attributes to others. In fact, the matter of order and arrangement of the double tradition material is a serious problem for the FH, because its proponents must explain why Luke used Matthew in ways that make it look as if it was Matthew who used Luke.

Assuming Markan priority, we can see how both Matthew and Luke used their source Mark. For the most part, Luke does not change Mark’s pericope order, does not recontextualize Mark’s sayings, and does not expand Mark’s discourses. Matthew, however, frequently does all of those things in using Mark. The last four of Matthew’s five major discourses were all created around a smaller core of Markan material, expanded by additional sayings material relevant to each discourse’s theme. Since Matthew’s first major discourse, the Sermon on the Mount (Matt 5–7), is a longer version of Luke’s Sermon on the Plain (Luke 6:20-49), it is logical to assume that Matthew expanded this Lukan discourse just the way he did the four Markan ones.

On the other hand, if Luke used Matthew as a source, he would have picked apart Matthew’s sermon and distributed small bits of it into multiple new contexts throughout his Gospel (in Luke’s chapters 8, 11, 12, 13, 14, and 16). Such a procedure would have been exceedingly complex for an ancient writer. Even a writer equipped with a modern word-processor would find such an operation extremely taxing. It raises the questions, “Why would Luke have done this?” and “How could Luke have done this?” Admittedly, there is also complexity in Matthew’s composition of his discourses using multiple sources. Yet it is much easier to envision Matthew gathering material from multiple contexts in order to compose a discourse with a single large theme in mind than it is to imagine Luke breaking up a discourse with multiple small contexts in mind. To say this is not to deny that Luke was a “great literary artist”;[9] it is simply to affirm that he was a normal ancient writer.[10]

The Crucial Issue: Verbatim Agreement

My main disappointment with Goodacre’s paper is that he did not discuss the issue of verbatim agreement among the Gospels, except to affirm that the Synoptic problem is a literary problem. Here, briefly, is why MPH proponents believe that the patterns of verbatim agreement support their theory:

As anyone who has spent time coloring a Synopsis knows, there is extensive word-for-word agreement (1) between Matthew and Mark wherever they have common material and (2) between Matthew and Luke in the double tradition. There is much less verbatim agreement between Mark and Luke. So, Matthew is the common factor wherever we find the strongest verbatim agreement among the Gospels.

Research into the practices of ancient writers has shown that it was unusual for them to copy extensively from their sources at length; they preferred to show their skill and creativity by changing their sources’ wording.[11] In light of this, it is somewhat unexpected if one of the evangelists regularly copied his sources verbatim at length. And it is surprising and problematic if more than one of the evangelists did so. Yet this is what both the 2SH and the FH require. On the 2SH, Matthew was a close copier of both Mark and Q while Luke was a close copier of Q—but not of Mark. On the FH, Matthew was a close copier of Mark while Luke was a close copier of Matthew—but not of Mark. Note Luke’s inconsistency on both hypotheses.

In terms of verbatim agreement, the MPH is the simplest and most straightforward hypothesis. On the MPH, only one of the evangelists, Matthew, is required to have behaved unusually in terms of ancient conventions for using sources. Also on the MPH, neither Luke nor Matthew need be seen as behaving inconsistently in their use of sources. Luke consistently paraphrases from his one source that we know, Mark; we are free to assume that he did the same with his sources that we do not know. Matthew is consistent in closely copying from his two sources Mark and Luke.[12] 


Goodacre, a noted expert on the Synoptic problem, is exceptionally qualified to identify the problems of the MPH. It is worth pausing to notice, therefore, a genuinely remarkable feature of this discussion: Goodacre’s best arguments against the MPH are either weak, readily reversible or inconclusive. And not only that, they fail to address the point that the phenomena of (1) order and arrangement of material and (2) verbatim agreement in the Gospels uphold the MPH and work against the FH. “Why not Matthew’s use of Luke?” is still a great question.

Despite my criticisms, Goodacre deserves appreciation for his paper. He has advanced the discussion of the Synoptic problem by his willingness to engage with the MPH, the often neglected third alternative to the relationship between Matthew and Luke. May the debate continue!

[1] Gottlob Christian Storr, Ueber den Zwek der evangelischen Geschichte und der Briefe Johannis, 2nd ed. (Tübingen: Jakob Friedrich Heerbrandt, 1810), 249-307. The first edition was published in 1786. In addition to arguing that Matthew used the other two Synoptic Gospels as sources, Storr also seems to have been the first scholar to argue for Markan priority. The designation “Matthean Posteriority Hypothesis” was coined by Ronald V. Huggins, “Matthean Posteriority: A Preliminary Proposal,” NovT 34 (1992): 1-22.

[2] Bart Ehrman, Alan Garrow, Mark Goodacre, and Evan Powell, “A $1000 Challenge to Me: Did the Author of Matthew Use Luke?” and subsequent posts, The Bart Ehrman Blog: The History and Literature of Early Christianity, 11-15 December 2017, http://ehrmanblog.org.

[3] Garrow’s paper and Goodacre’s response were presented on September 7, 2018, at the British New Testament Conference, held at St. Mary’s University in Twickenham, UK. For commentary on the Garrow-Goodacre exchange, see Alan Garrow, “$1,000 Challenge—Report from BNTC 2018,” The Problem Page, 12 September 2018, http://www.alangarrow.com/blog/archives/09-2018; and idem, “SBL/AARdvent Calendar: Day 21,” The Problem Page, 14 November 2018, http://www.alangarrow.com/ blog/archives/ 11-2018. The latter blog post features comments by Tim Murray.

[4] I am grateful to Dr. Goodacre for giving me the manuscript of his paper, which he plans to revise for publication in a forthcoming Festschrift. Numbers in parentheses below refer to pages in Goodacre’s manuscript.

[5] See Mark Goodacre, “Fatigue in the Synoptics,” NTS 44 (1998): 45-58; idem, The Synoptic Problem: A Way through the Maze, BSem 80 (London: T&T Clark, 2001), 71-76, 154-55; idem, The Case against Q: Studies in Markan Priority and the Synoptic Problem (Harrisburg, PA: Trinity, 2002), 40-43.

[6] In a footnote, Goodacre acknowledges that MacEwen suggested two examples of Matthew fatiguing in using Luke, but dismisses these as not meeting the proper criteria for editorial fatigue (15n25). For the examples Goodacre dismisses, see Robert K. MacEwen, Matthean Posteriority: An Exploration of Matthews Use of Mark and Luke as a Solution to the Synoptic Problem, LNTS 501 (London: Bloomsbury T&T Clark, 2015), 43n55, 45n61.

[7] See ibid., 27-50.

[8] Garrow prefers to argue that Matthew knew not only Luke, but also Luke’s written source(s) for his “double tradition” material, and sometimes “conflates Luke with Luke’s own source.” See Alan Garrow, “Streeter’s ‘Other’ Synoptic Solution: The Matthew Conflator Hypothesis,” NTS 62 (2016): 215. For a clear, logical explanation as to why alternating primitiveness does not require Matthew-Luke independence, see Garrow’s “Goldilocks” example in his short video The Streeter Less Travelled at http://www.alangarrow.com/mch.html.

[9] Pace Goodacre, The Case against Q, 104, 120.

[10] For a study of the habits and procedures of ancient writers, see Robert A. Derrenbacker Jr., Ancient Compositional Practices and the Synoptic Problem, BETL 186 (Leuven: Leuven University Press, 2005).

[11] John S. Kloppenborg, “Variation in the Reproduction of the Double Tradition and an Oral Q?”, ETL 83 (2007): 63-74; F. Gerald Downing, “Writers’ Use or Abuse of Written Sources,” in New Studies in the Synoptic Problem: Oxford Conference, April 2008, ed. P. Foster et al. BETL 239 (Leuven: Peeters, 2011), 524-36.

[12] For a more detailed exposition of the argument from verbatim agreement, including statistics and charts, see MacEwen, Matthean Posteriority, 50-74. During a question-and-answer time after Goodacre presented his paper, I asked him about verbatim agreement. He replied that he thought this was the best argument in my book. To address it, he suggested that closer copying took place when Gospels were closer to each other in time. Luke copied Mark less closely because he had known it for longer, and worked with it from memory. But he came to know Matthew more recently, so he needed eye contact with Matthew when drawing from it. This eye contact resulted in more word-for-word copying.

In making this proposal, Goodacre presents Luke as going against the grain of ancient compositional practices. Derrenbacker has shown that, when ancient authors closely followed the wording of a source, they also followed that source’s order (See Derrenbacker, Compositional Practices, 116-17). Yet Goodacre’s Luke closely copies from Matthew while radically reordering and recontextualizing Matthew’s material.

Written by
Robert MacEwen

Robert K. MacEwen is a missionary with Cru and an adjunct faculty member at East Asia School of Theology, Singapore. He received his PhD in biblical studies from Dallas Theological Seminary.

View all articles

Your email address has been added

Written by Robert MacEwen