The Orchard hypothesis, which holds that Mark is a transcript of lectures that Peter gave on Matthew and Luke, is a unique solution to the Synoptic Problem.
It was proposed in the late 20th century by the British scholar Dom Bernard Orchard, along with other authors, including Harold Riley, David Alan Black, and Dennis Barton.
It can be seen as a variation of the Griesbach hypothesis (according to which Mark composed his Gospel from Matthew and Luke), but it is different enough that it deserves its own treatment.
The view is not common. In fact, it is quite uncommon, but this should not be held against it. The question is not how popular it is but how well the evidence supports it.
Resources on This View
The following resources describe and advocate this view:
Dennis Barton’s web site—ChurchInHistory.org—contains additional materials on the view, including pieces written by Orchard.
I want to personally thank Barton, who provided very kind assistance as I was researching this view.
In what follows, we will principally deal with Orchard and Riley’s book, which is the longest sustained treatment of the position in print.
Stating the View
The basic proposal of the Orchard hypothesis runs along the following lines (page numbers refer to The Order of the Synoptics):
- Matthew wrote his Gospel first, to meet the needs of Jewish Christians (cf. pp. 239-245).
- Luke wrote his Gospel second, based in part on Matthew, to meet the needs of Gentile Christians (cf. pp. 248-250).
- Luke did not publish his Gospel until after Peter had vouched for its accuracy at Rome (cf. pp. 260-262).
- When Luke’s Gospel was brought to him, Peter gave a series of lectures based on Matthew and Luke (cf. pp. 269-272).
- Mark had these lectures transcribed (cf. pp. 269, 273).
- Some Roman Christians demanded copies of the transcripts immediately (cf. p. 274). This amounted to a private publication of Mark’s Gospel, with its original, shorter ending, finishing at Mark 16:8 (cf. p. 272).
- Luke then published his Gospel, its accuracy having been attested by Peter’s lectures (cf. p. 209).
- Later, Mark supplemented his Gospel by providing it with its current, longer ending (i.e., Mark 16:9-20) and published a second edition (cf. pp. 264-265 n. 4, 274).
An Oral vs. Literary Relationship
This view differs from the standard Griesbach hypothesis in that it does not envision Mark being the one to select and combine material from Matthew and Luke. Instead, it envisions Peter playing this role. Mark simply made a transcript of Peter’s oral presentation of this material.
The link between Mark and the other two Synoptics is thus oral rather than literary. This has the potential to avoid the problem (discussed in my piece on the Griesbach hypothesis) of why Mark tends to use more words to describe the same events as Matthew and Luke and why he often combines individual words and phrases from both of them.
If Mark were abridging Matthew and Luke with copies of them in front of him then, like other ancient epitomists, he would be expected to use fewer words—not more—and he would not freely combine wording from the two but would model his wording after one Gospel or the other.
However, if Peter were lecturing from Matthew and Luke, one might suppose that he would have a larger number of words in recounting particular events, and he might more freely mix words and phrases from the two Gospels, based on his memory of what one said when he had the other in front of him.
The fact that the Orchard hypothesis proposes an oral rather than a literary use of Matthew and Luke thus potentially gets around a serious objection to the Griesbach hypothesis, which is one reason it is worthy of independent consideration (for more, see “The Composition of Mark,” below).
Order of Publication and Order of Composition
Another interesting aspect of the hypothesis is that it can be looked at as a combination of the Griesbach hypothesis and the Augustinian hypothesis.
According to the former, the Synoptic Gospels were composed in the order Matthew, Luke, Mark—which is what the Orchard hypothesis proposes.
However, the Orchard hypothesis also proposes that they were published in the order Matthew, Mark (1st edition), Luke—which is the same order proposed by the Augustinian hypothesis.
Normally, the order of composition and the order of publication are not distinguished (it being assumed that they were published immediately upon being composed), but if this distinction is made then it is possible to harmonize the two.
Advocates of the Orchard hypothesis thus might claim the support of patristic texts favoring both the Augustinian and the Griesbach hypotheses.
(Although Orchard appears to propose the idea that Mark’s first, private edition was released before Luke was published, he is not emphatic on this, and it is not an essential element of the proposal. The key is the idea that Mark is based on transcripts of Peter’s lectures, not the specific order of publication. One thus might reject the proposed harmonization of the two sequences without rejecting Orchard’s central hypothesis.)
Two More Advantages?
In The Order of the Synoptics, Orchard suggests two more advantages of his proposal over the Griesbach hypothesis (p. 275).
The first is that it provides a rationale for why there are three Synoptic Gospels: Matthew wrote for Jews, Luke wrote for Gentiles, and Mark was a product of Peter’s validation of Luke vis-à-vis Matthew.
The second is that his account of Peter’s lectures provides an explanation for why the sequence of material in Mark seems to zig-zag between the sequences used for the same material by Matthew and Luke.
To evaluate the Orchard hypothesis, we will look at the four potential advantages just named and then at the individual points of the hypothesis (the claims numbered 1-8, above).
An Oral Advantage?
The fact that Mark typically uses more words to recount the same events that Matthew and Luke cover and the fact that he often fuses the language of the two are both serious problems for the Griesbach hypothesis in its classical form. If Mark had Matthew and Luke in front of him, we simply would not expect to see either of these phenomena on a regular basis, but we do.
If, however, Mark is largely a transcript of Peter’s speeches based on the other two Gospels then this objection is blunted. A speaker might well use more words than a prepared text if he were speaking naturally and occasionally glancing down at the text in front of him.
Similarly, if he knew one of the texts well (and the Orchard hypothesis holds that Peter would have known Matthew’s Gospel for years) then he might well mingle snippets from that text in his oral presentation, even when that text was not in front of him.
Is there a way to test this as it applies to the Orchard hypothesis?
Perhaps. Orchard, et al., propose that at certain identifiable points in his lectures, Peter had Luke’s text in front of him, while on other occasions he had Matthew’s. It thus should be possible to compare these portions and see what kind of intermingling is being done.
If Peter knew Matthew’s Gospel well but had only recently been presented with Luke’s then we would expect a more substantial introduction of Matthean phrasing when he had Luke in front of him and a less substantial introduction of Lukan phrasing when he had Matthew in front of him.
I don’t know of anyone who has conducted an evaluation of Mark with this in mind. Until such time as this test is done, there is no way of knowing whether it would tend to confirm or disconfirm the hypothesis.
Another potential way of testing the hypothesis would involve the number of words used to tell particular stories. While it is true that a speaker, especially an eyewitness like Peter, might expand on the text in front of him, there is the question of whether we would expect this degree of expansion.
One might argue that we should expect even more expansion than what we do see. Unless an eyewitness is (for some reason) following the text in front of him very closely, we might expect him to add many more details than what Mark’s Gospel does.
Unless Peter spent an unexpectedly large proportion of his time looking down at the scrolls of Matthew and Luke and reading them almost word-for-word, we might expect him to elaborate even more—or at least to vary his wording from them more than we see him do.
On the other hand, if he wasn’t looking down constantly during the recounting of a particular incident, we might expect him to summarize it even more briefly than what we see in Mark.
Because of these variables, the construction of an objective test along these lines would be difficult, though it might still be possible.
Regardless of how these tests might turn out, we can say this: Any advantage that the Orchard hypothesis might have over the Griesbach hypothesis is a relative one. At most, it would establish that the Orchard hypothesis is relatively more likely than the Griesbach hypothesis. It would not establish that it is likely in absolute terms.
Put another way: The oral link to Matthew and Luke posited by Orchard could diminish or remove the force of objections to Griesbach, but it would not provide a positive reason to believe the Orchard hypothesis.
A Harmonizing Advantage?
The fact the Orchard hypothesis allows a possible way to harmonize the Augustinian hypothesis and the Griesbach hypothesis by distinguishing between the Gospels’ order of composition and their order of publication at first seems intriguing.
Wouldn’t it be great to be able to appeal for support to both patristic sources supporting the former and the latter? That seems like it would be a net advantage for the hypothesis. But a closer look is less promising.
For advocates of Orchard to claim the patristic texts supporting the Augustinian sequence (Matthew, Mark, Luke) and the Griesbach sequence (Matthew, Luke, Mark) as evidence for their view, there would have to be a causal connection between the way in which the Gospels were composed and published and what is found in the Church Fathers.
In other words: Memory of the fact that Mark was given a private publication before Luke had to be preserved and this memory would have to give rise to the Augustinian hypothesis tradition and memory of the fact that Luke was written before Mark had to be preserved and this memory would have to give rise to the Griesbach hypothesis tradition.
How likely is this?
There are problems on multiple fronts. First, it is not particularly likely that the memory of a quick, private publication of Mark before Luke would have survived and given rise to the dominant patristic view regarding the order of the Synoptics.
If anything, the fact that the composition of Mark was a byproduct of the validation of Luke would have been what stuck in people’s memories, which would have reinforced awareness that Luke had been written first.
The fact that Mark had a few quick copies made of lecture transcripts and then distributed them privately would not have been as significant an event in the life of the Church has the long-awaited publication of the “Gospel for the gentiles.”
The publication of Mark would have been too insignificant by comparison, and too reinforcing of the fact that Luke was written first, for it to be responsible for the tendency of later writers to suppose that Mark wrote before Luke did.
If anything, we would expect it to reinforce a patristic tradition that the Gospels were written in the order proposed by Griesbach, not Augustine.
This leads to a second problem, which is that there is little patristic support for the Griesbach order. As we will see below, there is at most a single quotation from Clement of Alexandria that seems to suggest this, and it may well involve a mistranslation.
Furthermore, the later writers seem to have thought that Mark wrote before Luke, not just that he published (however briefly or privately) before Luke. This means that they would have had to have remembered but misunderstood the way in which Mark came to be. They would have had to have forgotten what motivated the composition of Mark but remembered that it was released to a few individuals slightly before the much more popular Gospel of Luke and then confused the order of publication with the order of composition.
Such a sequence of events is sufficiently implausible that it does not allow the Orchard hypothesis to claim substantial support from Fathers who advocate the Augustinian view.
The most that can be said is that the Orchard hypothesis would, again, enjoy a slight relative advantage over other forms of the Griesbach hypothesis.
However, as we noted, the idea that the Gospels were published in the Augustinian sequence is a secondary aspect of the Orchard hypothesis, not one of its central features, and so one could set this matter aside.
Why Three Synoptics?
Orchard proposed that one advantage of his hypothesis is that it provides a credible rationale for why there are three Synoptic Gospels.
It certainly provides a rationale, but this is of limited value because it is quite speculative and substantially coincides with rival views.
The parts of the rationale that are the firmest—that Matthew wrote more for Jews than Gentiles and that Luke did the reverse—are agreed upon by all, and they are not unique to the Orchard hypothesis.
There is also a danger of overstating the ethnic orientations of these two Gospels. All four Gospels—including Matthew—display significant interest in Gentiles, and Luke contains a significant amount of material that is distinctly Jewish in orientation (e.g., in his Infancy Narrative).
(Indeed, the fact Matthew displays significant interest in Gentiles suggests that it was written after the conversion of a significant number of Gentiles began, which would place it after the A.D. 30-44 period that Orchard proposes for its composition. This dating, however, is another secondary aspect of the hypothesis, not an essential feature.)
The unique contribution that the Orchard hypothesis could make to an understanding of why there are three Synoptics concerns the way Mark came to be.
The idea that Mark is a transcript of lectures that Peter gave to validate Luke relative to Matthew certainly is a rationale for why Mark was written, and in general views that can provide rationales are to be preferred to those that do not, but one has to ask whether the rationale that is provided is the most plausible one available.
As we will see, the rationale provided by Orchard is speculative and not well-supported by either patristic or internal evidence. Further, there are other potential explanations for why there are three Synoptics (e.g., Mark wrote first but was deemed too short and incomplete, so Matthew expanded it to create a longer version with a Jewish orientation and Luke did the same for a Gentile audience).
The fact that the Orchard hypothesis provides a rationale for the composition of three Synoptics thus does not provide significant evidence that it is true.
A Zig-Zag Advantage?
The second advantage that Orchard proposed was that his suggestion provides an explanation for the way Mark’s sequencing of material seems to zig-zag between Matthew’s sequence and Luke’s sequence.
Again, Orchard is correct in that he does provide an explanation for this phenomenon. The Order of the Synoptics describes how Peter would have proceeded to mark up copies of Matthew and Luke to indicate which sections of them he wanted to use so that assistants could roll through them to these points and then hand the scrolls to him while he gave his lectures (pp. 266-272).
However, there is nothing about this that requires a scenario involving lectures rather than a written composition. Mark—or anybody else—could have similarly marked up copies of Matthew and Luke, noting the sections to be used, and then rolled through them in the process of composing a written version of Mark.
Or Matthew could have marked up copies of Mark and Luke, or Luke could have marked up copies of Matthew and Mark.
The fact that the Orchard hypothesis provides an explanation for the zig-zag is thus not of much advantage. There are alternative explanations that work just as well. The zig-zag does not provide evidence that Mark is a transcript of lectures by Peter. This would have to be supported on other grounds.
We now come to the point in our evaluation where we need to look at the plausibility of the individual components of the Orchard hypothesis (numbered 1-8, above).
The Composition of Matthew
Claim #1: Matthew wrote his Gospel first, to meet the needs of Jewish Christians.
We have already noted that Matthew’s Gospel has a Jewish character, though it also shows interest in Gentiles. This is not in dispute. The question is whether he wrote first. Orchard and his co-author Riley attempt to provide both internal and external evidence for this.
The internal evidence they offer is an attack on Markan priority (Lukan priority is assumed to be false, so a rejection of Markan priority would entail Matthean priority). The core of the argument is offered on The Order of the Synoptics, 4-7, where Riley presents a table of the material in Mark with parallel columns showing when that material was used by Matthew or Luke in the same sequence as Mark (relocations of material are not shown on this table).
(Click here to see the table.)
Riley then argues:
[A]t every point where Matthew ceases to follow Mark’s order, whether for a short or longer period, Luke continues in it; and wherever Luke ceases to follow Mark’s order, Matthew in his turn continues in it. There is surely an inescapable conclusion to be drawn from this. If Matthew and Luke were dependent on Mark for the order of events, they must have agreed together that they would do this. Without constant collaboration, the result would be quite impossible. That they followed such a course is incredible, and therefore the conclusion cannot be avoided that the hypothesis that they were dependent on Mark cannot be sustained (p. 7).
This argument does not work. The collusion that Riley argues would have to have existed between Matthew and Luke is an illusion based on the way he has constructed the table.
Since the table lists all of the material of Mark in its middle column, that column is guaranteed to be filled. A particular block of Markan material will then have a parallel in the Matthew (left) column if Matthew chose to parallel that block and if Matthew presented it in the same order. In the same way, a particular block of Markan material will have a parallel in the Luke (right) column if Luke chose to parallel that block and he presented it in the same order.
This means that there are four things we could see on any given row of the table:
- An entry in all three columns, where all three Evangelists used the material and in the same order.
- Entries in the Matthew and Mark columns but a blank space in the Luke one, where Luke didn’t use the material or did not use it in the same order.
- Entries in the Mark and Luke columns but a blank space in the Matthew one, where Matthew didn’t use the material or did not use it in the same order.
- An entry in the Mark column only, where Matthew and Luke either did not use the material or did not use it in the same order as Mark.
All four phenomena are found in the table, and they are consistent with the idea that Matthew and Luke used Mark:
- The most common type of row is the first, which means that on Markan priority, Matthew and Luke would have both made a large number of selections from Mark in which they preserved his sequence.
- The second type of row is the next most common, which is consistent with Matthew having a stronger preference than Luke for including material from Mark and using its sequence.
- The third is next most common. It is consistent with Luke having a weaker preference than Matthew for including material from Mark and using its sequence.
- The fourth is the least common. Its low frequency is consistent with the idea that Matthew and Luke had a strong enough preference for using material from Mark, and Mark’s sequence, that this type of row is rare.
We have no need to suppose collusion between Matthew and Luke to produce the patterns we see. Since the table excludes material that Matthew and Luke might have taken from Mark but relocated, we are certain to have one of the above four phenomena on any given row of the table.
Setting aside the first type of row (where both Matthew and Luke followed Mark) and the fourth (where neither did so), we are guaranteed to have either the second or third. Whenever Luke departs from Mark, we will have the second, and whenever Matthew departs, we will have the third. The alternating pattern that Riley observes is thus a product of the way the table itself has been set up.
You could generate exactly the same kind of pattern by taking a table with all the columns filled in and then rolling a die to randomly blank out cells from the first and third columns (say, blanking a cell every time you roll a 6 for the first column and every time you roll a 5 or 6 for the third column, since Matthew more closely follows Mark than Luke).
There is thus no need to posit collusion between Matthew and Luke to explain the table if Markan priority is true. As long as Matthew and Luke sometimes deviated from Mark, the construction of the table guarantees the pattern that Riley observes.
Riley’s attempt to support Matthean priority by attacking Markan priority thus fails.
What of the external evidence? This is dealt with by Orchard in Part Two of their book.
It is true that by the late second century we have endorsements of Matthean priority and that this view is later dominant. All things being equal, this would be the preferred view.
But not all is equal. In particular, the earliest patristic statement we have on the origins of the Gospels dates to the first-century figure John the Presbyter. Its straightforward interpretation (see below) indicates that Mark was written first.
In view of the fact John the Presbyter likely was one of the authors of the New Testament (whether or not he is identified with John son of Zebedee), this is an extraordinarily important testimony. Coming from the same circle of authors that wrote the New Testament, it has more intrinsic weight than later statements, however popular they came to be (a phenomenon likely driven by the popularity of Matthew’s Gospel rather than its historical sequence).
Orchard and Riley thus do not produce compelling evidence for Matthean priority. The initial point of the Orchard hypothesis thus looks shaky.
The Composition of Luke
Claim #2: Luke wrote his Gospel second, based in part on Matthew, to meet the needs of Gentile Christians.
Again, there is little dispute that Luke’s Gospel has a special concern for Gentiles, and it is generally thought that it was written for a primarily Gentile audience. The questions would be whether Luke wrote second and whether he based his Gospel on Matthew.
The claim Luke wrote second could be inferred from Matthean priority and the view that Mark is a transcript of Peter’s lectures from Matthew and Luke. If those premises are true then Luke had to come second.
If, however, either or both premises are shaky then this inference does not work. We have already suggested that Matthean priority is shaky, and below we will cast doubt on the idea that Mark is a transcript.
It would still be possible to provide support for this view, though. For example, there is the statement attributed to Clement of Alexandria that the Gospels with the genealogies (Matthew and Luke) were “written first” (Greek, progegraphthai; the statement is preserved in Eusebius, Church History, 6:14:6-7).
If that view is true then, unless one advocates Lukan priority, one could infer that Luke wrote second.
However, Stephen Carlson argues that the key Greek verb (progegraphthai) should be rendered “published openly” rather than “written first.” On this view, Clement was claiming that Matthew and Luke were published openly, while Mark was initially written for a group of private individuals, without Peter’s initial knowledge or authorization. (See Carlson’s argument, here.)
The statement of Clement thus may not provide support for the idea that Luke wrote second.
What about the view that Luke used Matthew? This has been advocated by a number of scholars, including advocates of the Farrer hypothesis (a fact Orchard notes; see p. 236). However, the reverse—that Matthew used Luke—has also been maintained by advocates of the Wilke hypothesis, and most contemporary scholars think that Matthew and Luke wrote independently of each other, so there is no scholarly consensus in favor of Luke using Matthew.
To maintain that this happened, one would need to produce positive arguments in favor of the view. If one accepts Matthean priority then, given the similarities of the Synoptic Gospels, the idea that Luke used Matthew naturally results. But if Matthean priority is shaky then, without further argumentation, the idea that Luke used him is shaky as well.
The second point of the Orchard hypothesis is thus not established.
Claim #3: Luke did not publish his Gospel until after Peter had vouched for its accuracy at Rome.
Orchard provides a speculative account of why he thinks this happened. He proposes that Paul recognized that, given its Gentile orientation, Luke’s Gospel would not be well received by non-Pauline churches unless it were vouched for by an authoritative, original apostle, such as Peter. He thus thinks Paul had a role in delaying the publication of Luke’s Gospel until Peter could approve it.
This is possible, though it is quite speculative. Despite this, Peter may well have approved Luke’s Gospel prior to its public distribution.
The reason is that Acts suddenly stops its narrative when Paul is under house arrest for two years in Rome, awaiting trial before Nero. The logical explanation for this is that this is when and where Acts was written. Since Acts is the sequel to Luke’s Gospel, and since both are addressed to the same individual (Theophilus; Luke 1:3, Acts 1:1), it is likely that Luke also wrote his Gospel at Rome during the same two years.
Early Christian authors indicate that Peter ministered at Rome for an extended period of time and that Paul and Peter were martyred there at approximately the same time. It is thus quite possible that Peter was at Rome when Luke-Acts was being written.
This seems confirmed by the way Peter dominates the early chapters of Acts. If Luke had stopped writing with chapter 12, the book could have been titled “The Acts of Peter.”
There is also a marked difference in the sources that Luke had available to him for the first half of the book. This is shown by comparing the general description of travels in the first half with the detailed descriptions of how long and by what routes Paul travelled in the second.
Given the shift in Luke’s sources and the prominence of Peter early in Acts, it is very likely that Peter himself was one of Luke’s sources when he was writing Acts. If so, then it is also likely that Peter was one of Luke’s sources for writing his Gospels and that Peter stands behind some of the material found in Luke but not in Mark.
Peter’s participation in Luke and Acts could itself be construed as evidence of his approval of them. Even apart from this, if Luke’s Gospel was published at Rome then it would be inevitable that people would ask Peter’s opinion of it—and that it would have been dealt a severe blow if Peter disowned it.
It is therefore probable that Peter gave Luke’s Gospel some form of approval, either before or shortly after its publication in Rome. The third point of the Orchard hypothesis therefore looks very reasonable, though perhaps not for the reasons that Orchard himself proposed.
The Composition of Mark
We now come to the core of the Orchard hypothesis and what sets it apart from the Griesbach hypothesis.
At this point we should note the implications of the fact that the Orchard hypothesis is different from the Griesbach hypothesis. The latter proposes that Mark is based on a written conflation of Matthew and Luke, while the latter proposes it is based on an oral conflation of them.
That is important.
It means that some of the arguments used with respect to the Griesbach hypothesis do not work—or work as well—when applied to the Orchard hypothesis.
For example, it has been argued that it is very implausible that an ancient author would stitch together two passages in Matthew and Luke on a phrase-by-phrase basis. That kind of thing is possible with word processors, but the ancient world lacked even writing desks. It would be far more likely for Mark to base his version of a passage on either Matthew or Luke, perhaps including a word or phrase from the other, but not alternating between them frequently.
As discussed above, this argument is less acute if we suppose Peter was familiar with both Matthew and Luke and was mentally combining them on the fly as he delivered lectures, looking down for reference at the text of one or the other Gospel in front of him.
Similarly, given the length of Matthew and Luke, Mark makes little sense as an epitome. In the ancient world, epitomes were produced for long, multi-volume works, but Matthew and Luke are both one scroll long, and Mark is not that much shorter than either.
Also, if Peter were giving lectures, he might skip some material in the sources in front of him that he originally meant to cover (a phenomenon known to almost all givers of speeches), accounting for at least some of the omissions of important material we would not expect an author to omit if he were carefully selecting which bits of Matthew and Luke to incorporate in his new, written work.
This is not to say that the arguments commonly used with respect to the Griesbach hypothesis have no value with respect to the Orchard hypothesis. They may have some force—or be reconfigurable in a way that has force—but it is important to give the Orchard hypothesis its due and not simply dismiss it as if everything that applied to the Griesbach hypothesis applied to it.
Claim #4: When Luke’s Gospel was brought to him, Peter gave a series of lectures based on Matthew and Luke.
There is absolutely no external testimony to this, which means in the first place that we have introduced a new speculative element into the mix.
How likely is it?
As an eyewitness of Jesus’ ministry, Peter almost certainly preached from his memories on the large majority of the occasions when he discussed the events in question. He would have had no need to preach from written texts.
Having said that, it is possible that he could have given a series of lectures from Luke if, as we have argued, Luke was available to him at Rome. It is also possible that he could have preached a series from Matthew if this Gospel were also available to him. However, these are possibilities and not probabilities.
The Orchard hypothesis goes a step further by proposing that during a single series of lectures Peter alternated between Matthew and Luke and that he alternated between the two during the course of the individual lectures. That is, in a single session, he alternated between reading from a scroll of Matthew and a scroll of Luke.
It is difficult to see why an eyewitness of Jesus’ ministry would choose to do this.
The choice would only be rational if Peter had some very specific reason to do so, and Orchard proposes one: that Peter was intending to vindicate Luke vis-à-vis Matthew.
If that’s what he was doing, however, we would expect to see evidence of it in the written record of these lectures (i.e., Mark). For example, we would expect comments like, “Although Luke places this account in a different location than Matthew, it is still fundamentally accurate,” or “Matthew is placing this story here in order to make a particular point, but Luke’s placement is chronological.”
Comments of this nature are natural in any talk on how to harmonize the Gospels, yet there are no comments of this sort in Mark.
If Peter’s lectures were meant to be a vindication of Luke—whether by harmonization or in some more general sense—then it is striking that there are no references to the texts upon which Peter is lecturing in the written record.
This is very implausible. Any speaker attempting to vindicate one source with respect to an already established source is certain to make reference to the two sources he is discussing, and yet Mark makes no mention of either source!
The proposal that Peter was lecturing to vindicate Luke thus is not established and we revert to the antecedent probability that it is very unlikely that an eyewitness like Peter would choose to do a series of lectures in which he read alternatingly between Matthew and Luke in the course of individual talks.
Claim #5: Mark had these lectures transcribed.
How likely is this?
The antecedent likelihood is not great. For a start, the number of people who were trained in stenography and thus capable of taking dictation was very small.
The Christian community in Rome c. A.D. 60 was also small, likely including only a few dozen members and no more than one or two hundred at the very most.
In Romans 16:3-15, Paul greets twenty-four individual Christians and five groups (e.g., the church that meets in Aquila and Priscilla’s house, “those who belong to the family of Aristobulus,” “those who belong to the family of Narcissus”). It is unlikely that Paul knew all of these Roman Christians personally since most people did not travel and Paul had not yet visited Rome. Yet it is likely that he is trying to be as complete as possible in his greetings, given the sensitivity of the letter. He most likely relied on Tertius (Rom. 16:22), who was probably a visiting Christian scribe from Rome, to flesh out his knowledge of the Roman churches. Romans 16 thus may represent a near-complete representation of the size of the Christian community at Rome at this time, in which case it would be only a few dozen individuals.
Given that, it is likely that none of the Christians at Rome were able to write in shorthand. Tertius was able to write and serve as a scribe, but that is not the same thing as being able to take dictation at the speed a public lecture is given. For that, training in special, stenographic signs is needed.
There were people in Rome trained in shorthand, but they were probably not Christians, and they would have needed to be hired to serve as stenographers for a dodgy group like the Christians.
It is very questionable whether a non-Christian scribe would be willing to take a dictation assignment from such an iffy group as the Christians, who were advocating a king rival to Caesar (Matt. 27:37, Mark 15:26, Luke 23:38, John 19:12, 19-20) in Rome itself!
Even if a scribe trained in stenography (Christian or not) was available and willing to take the job, it would have cost money to hire him, and it would have cost a great deal.
Even after the cost of the initial shorthand transcription was paid, the scribe would then need to prepare a version in readable Greek. If such a Greek version were not made then the only people who could read it would be those trained in the same system of stenographic signs as the transcriptionist.
Based on E. Randolph Richards estimates for the cost of producing Paul’s letters (see Paul and First-Century Letter Writing, 169) and scaling the numbers up for a work the size of Mark’s Gospel, it would have cost the ancient equivalent of approximately $1,400 just to have a single copy of the resulting work made in Greek.
In view of all this, there would have needed to be a powerful reason to have a particular set of Peter’s talks transcribed, if a transcriptionist were even available.
After all, people in Rome had the opportunity to hear Peter preach all the time. Mark, in particular, had the opportunity to hear him preach on a regular basis. As his interpreter, Mark likely heard Peter preach so much that he could reconstruct large amounts of this material by memory. Of all people, Mark had no need for a transcription of a set of Peter’s lectures.
So why transcribe this set of lectures?
Presumably, on the Orchard hypothesis, because it was expected to deliver something of special value, beyond what was found in Matthew and Luke—and sufficiently beyond that to justify the costs involved.
If that was the expectation, Peter spectacularly failed to deliver on it in his lectures.
First, Mark’s Gospel contains almost nothing that is not found in either Matthew or Luke, and what it does have is of low value (e.g., the parable of the growing seed, the mention of Jesus’ relatives responding to the claim that he was beside himself, the mention of the man who ran away naked) compared with the material it omits (the Lord’s Prayer, the Infancy Narratives, the Sermon on the Mount, the resurrection accounts).
Second, Peter’s proposed vindication of Luke vis-à-vis Matthew would have been done in such a way that he never even mentioned the two sources he had before him, much less did he provide commentary on their differences and how the two might be understood in light of each other.
Ultimately, Mark’s Gospel doesn’t contribute anything that Peter couldn’t have provided in just five words if asked about the merits of Matthew and Luke: “Yes, they are both good.”
Indeed, knowing the fantastic costs associated with transcription, it’s hard to imagine that Peter—knowing what he was planning to do in the proposed lectures—wouldn’t have simply told Mark to save the money and not burden the church’s treasury with a project of such de minimis value.
If Matthew and Luke already existed, paying such sums to transcribe the set of lectures Peter allegedly gave would have seemed frivolous and irresponsible.
(In fact, the costs associated with producing Mark—by whatever means—were such that it would make no sense to pay them if Matthew and Luke already existed. This is itself an argument for Markan priority. See here.)
The antecedent probability that Mark would have had the proposed lectures of Peter transcribed is thus very low.
But an event’s antecedent probability does not trump strong after-the-fact evidence that it occurred. So what posterior evidence is there?
Advocates of the Orchard hypothesis have appealed to elements of orality in Mark’s Gospel (i.e., traces that it is based on orally performed material rather than the material being composed exclusively in writing). They have also appealed to John the Presbyter’s statement that Mark’s Gospel was based on Peter’s preaching.
However, neither of these elements provides the support that is needed for this point.
In the first place, the oral elements in Mark’s Gospel are equally accounted for whether Mark was a transcript of Peter’s speeches or whether it was based on Mark’s memories of Peter’s sermons. Either way, it would be based on oral performance, and traces of that could (and did) remain in the final text.
In the second place, John the Presbyter does not say that Mark had Peter’s talks transcribed. What he says is:
Mark, having become the interpreter of Peter, wrote down accurately, though not in order, whatsoever he remembered of the things said or done by Christ. For he neither heard the Lord nor followed him, but afterward, as I said, he followed Peter, who adapted his teaching to the needs of his hearers, but with no intention of giving a connected account of the Lord’s discourses, so that Mark committed no error while he thus wrote some things as he remembered them. For he was careful of one thing, not to omit any of the things which he had heard, and not to state any of them falsely (Eusebius, Church History 3:39:15).
Not only does John the Presbyter stress that Mark’s Gospel is based on Mark’s memories (not a transcript), he also indicates that the speeches on which Mark is based were delivered with “no intention of giving a connected account of the Lord’s discourses” (Greek, logia, which may refer to stories about the Lord rather than sayings given by the Lord). Yet the speeches would have been a connected account of traditions concerning Jesus if Peter was giving a series of lectures harmonizing Matthew and Luke.
John the Presbyter’s statement that Mark’s Gospel is based on Mark’s memories of Peter’s preaching—not on Matthew and/or Luke—thus supports the idea that Mark was written first.
The earliest evidence we have, and from a likely New Testament author, is thus that Peter gave various, unconnected sermons about Jesus’ ministry at different times and, at a later date, Mark composed his Gospel based on his memories of them.
We thus do not have posterior evidence capable of overcoming the extremely low antecedent probability that the lectures would have been transcribed (if they even occurred).
Claim #6: Some Roman Christians demanded copies of the transcripts immediately; this amounted to a private publication of Mark’s Gospel with its original, shorter ending (when the women flee from the tomb).
It is difficult to imagine Peter giving a series of lectures on Matthew and Luke and then suddenly stopping when he got to the resurrection accounts. The narrative momentum of the story would make the lectures seem extraordinarily incomplete and unsatisfying if Peter suddenly stopped at that point, without recounting the joyous and climactic vindication of Jesus.
Further, the differences between how Luke records these compared to Matthew’s version would have been one of the things the Roman Christians would have been most interested in hearing about. If Peter was trying to vindicate Luke vis-à-vis Matthew then he should have talked about how their resurrection accounts should be understood in light of each other.
Nor can one appeal to the proposal (sometimes made by Orchard advocates) that Peter was only commenting on things from Matthew and Luke that he was an eyewitness of. That might explain why he didn’t cover the Infancy Narratives, but it would not explain why he didn’t include the resurrection appearances, for he was an eyewitness—as Luke’s Gospel explicitly states (Luke 24:34; cf. 24:12).
It is thus very hard to imagine Peter not covering the resurrection narratives in his lectures and thus that this is the explanation for Mark’s shorter ending.
However, we may set aside the question of whether the first publication of Mark’s Gospel included only the shorter ending found in early manuscripts. As noted above, this is not a key aspect of the Orchard hypothesis.
What is more important is the idea that Mark’s Gospel was originally published when Roman Christians demanded copies of the transcript of Peter’s lectures on Matthew and Luke.
How likely is that?
Even if we assume that such a set of lectures took place then the antecedent probability does not appear high.
After all, Matthew was already in circulation, and Luke existed. Mark adds nothing of substance to these two, and that correspondingly diminishes the desirability of having a transcript of the lectures (especially if you hear Peter preach all the time).
If obtaining a copy were as easy as hitting “Print” is today, then some might have wanted a copy for archival purposes, but this does not apply when a single copy in Greek costs the equivalent of $1,400!
It boggles the imagination to envision a group of Romans paying for the multiple copies of the transcript that would be needed for Mark to survive the ages when they could have gotten so much more for their money by having copies of Matthew and Luke made instead.
Indeed, Mark could even be open to charges of swindling people if he allowed them to use their money in this way rather than pointing out the advantages of getting copies of Matthew and Luke.
The survival of Mark is vastly more explicable if it was the first Gospel written and had established a reputation as a sacred and inspired document—and thus one worth copying—before Matthew and Luke appeared.
The antecedent probability that people would have demanded copies of the transcript, given the existence of Matthew and Luke, is thus low.
What about after-the-fact evidence?
According to Clement of Alexandria (c. A.D. 200):
The Gospels containing the genealogies . . . were written first (Greek, progegraphthai).
The Gospel according to Mark had this occasion: As Peter had preached the Word publicly at Rome, and declared the Gospel by the Spirit, many who were present requested that Mark, who had followed him for a long time and remembered his sayings, should write them out.
And having composed the Gospel he gave it to those who had requested it.
When Peter learned of this, he neither directly forbade nor encouraged it (Eusebius, Church History, 6:14:6-7).
Whether progegraphthai is taken to mean “written first” or, as has been discussed, “published openly,” the first part of the statement is consistent with the Orchard hypothesis, since it both holds that Matthew and Luke were written first and that Mark had an initially private publication for those who requested it.
Clement agrees that Mark was approached by Roman Christians who wanted a record of Peter’s preaching, but the account is not that given by the Orchard hypothesis.
Clement specifies that they made their request of Mark because he “had followed him [Peter] for a long time and remembered his sayings.”
As with the quotation from John the Presbyter, we again have Mark being based on the Evangelist’s memories of what Peter preached over “a long time” and not being a transcript of a specific set of lectures.
The quotation from Clement thus contradicts the Orchard hypothesis, and we do not have posterior evidence sufficient to overcome the low antecedent probability that people in Rome would have wanted and paid for transcripts of the lectures, given the existence of Matthew and Luke.
We now come to two lesser claims, neither of which is essential to the Orchard hypothesis.
Claim #7: After the publication of Mark, Luke published his Gospel, its accuracy having been attested by Peter’s lectures.
We agreed above (see Claim #3) that Peter likely approved Luke’s Gospel at Rome, though we also argued (see Claim #4) that it is unlikely he gave a set of lectures on Matthew and Luke of the kind proposed by the Orchard hypothesis.
Still, given the scenario proposed by Orchard advocates, what should we make of the claim that Luke would have been published after Mark’s initial, private publication?
This is possible, though it is a weakly-supported claim.
In terms of antecedent probability, we must confront the fact that Luke’s Gospel already existed. Multiple copies of it probably had been made before Peter’s lectures. At a minimum, Luke would have had a copy, Peter would have had a copy (that he then marked up according to a frequent proposal by Orchard advocates), Theophilus likely would have had a copy, and so would Paul. Given this, there easily could have been other copies in circulation.
Even supposing that there weren’t, Luke’s Gospel was ready to be duplicated by the scribes available to the Christian community in Rome. For Mark’s Gospel to come out first, the scribes would have had to be diverted from copying Luke next to making copies of Mark’s transcript (with no split among the scribes allowing some to copy Luke and some to copy Mark).
Given the probable low demand for copies of the latter (see Claim #6)—and the fact that Mark would need to be transcribed from shorthand symbols into Greek and then edited, whereas Luke was already in Greek and ready to go—this is unlikely.
In terms of posterior evidence, we have multiple early sources attesting to the Augustinian order of composition (Matthew, Mark, Luke) and only a possible attestation from Clement of Alexandria to the Griesbach order (Matthew, Luke, Mark). This is entirely too slender a basis on which to propose that the Griesbach order is that of composition and the Augustinian order is that of publication.
In the first place, the sources we have assume that the order of composition was the same as the order of publication. The texts that support the Augustinian hypothesis indicate that the Gospels weren’t just published in that order but that they were composed in it. They thus contradict the Orchard hypothesis.
In the second place, the gap between the original publication of Mark and that of Luke would have been so slight (a few days or weeks at most) that it is extremely unlikely that memory of the difference between the two sequences would have been preserved and given rise to the later traditions.
The Orchard hypothesis’s proposed explanation for harmonizing the Augustinian and Griesbach sequences is clever, but it is too clever by half.
Claim #8: Later, Mark supplemented his Gospel by providing it with its current, longer ending (i.e., Mark 16:9-20) and published a second edition.
This claim is highly speculative. It presupposes that Peter’s lecture series did not include coverage of the resurrection appearances, which is quite improbable (see Claim #5).
It would be more likely, if Peter preached such lectures, that he covered the Resurrection in some detail, that this was transcribed, and that the original ending of Mark was then lost and its deficit supplied at a later date by an unknown party.
This could have been anybody. It did not require special knowledge that was unique to Peter or Mark. Indeed, the longer ending of Mark appears to largely parallel material found in Matthew, Luke, John, and Acts.
The fact it is in a different style also points to a different author.
And we have no patristic testimony attributing the earlier, shorter version to Mark and then attributing the longer ending to him as well. There is thus no patristic evidence for this claim.
What we said above regarding the cleverness of Claim #7 applies to the Orchard hypothesis as a whole: It is a very clever proposal. But it is too clever by half.
Its advocates sometimes observe that it accounts for a large number of data points, including ones that other proposals must set aside.
This is true. Other proposals must set aside some data points and conclude they are simply in error. Thus:
- Advocates of Markan priority and the Griesbach hypothesis must both set aside patristic testimony to the Augustinian hypothesis.
- Advocates of the Augustinian hypothesis and the Griesbach hypothesis must set aside John the Presbyter’s statement regarding the composition of Mark.
- Advocates of the Markan priority and the Augustinian hypothesis must set aside the “written first” interpretation of Clement of Alexandria.
By “set aside,” we do not mean that they must simply dismiss these data points. They can and should provide arguments for why particular patristic claims are not accurate.
But some selection among these claims is inevitable, because the data we have available to us is not uniform in what it says. It is messy, and scholars must weigh and select which bits of it they think are accurate and which are not.
The Orchard hypothesis is sometimes presented as if it gets around this problem by explaining all of the data, but this is inaccurate.
While the Orchard hypothesis gestures at each of the data points mentioned above, it does not succeed in incorporating them because it changes their meaning. Specifically:
- It only harmonizes the Augustinian and Griesbach orders by interpreting one as the order of publication and the other as the order of composition, when this distinction is not made in the patristic sources.
- It takes John the Presbyter’s statement pointing to Mark being a transcript of a specific set of Petrine lectures rather than the Evangelist’s memories of Peter’s preaching.
- It takes Clement of Alexandria’s statement as pointing to a Roman request for copies of a transcript of a specific set of Petrine lectures instead of a request that he write a new work based on his memories of having heard Peter preach for “a long time.”
This phenomenon can be described different ways. One could say that the Orchard hypothesis reinterprets or adjusts the patristic claims in order to incorporate them, but while it gestures at these data points, it does not incorporate them into the theory without adjustment.
It thus must do what every account of Synoptic origins does: Take the patristic claims as in some measure accurate and as in some measure inaccurate. The parts it deems inaccurate the Orchard hypothesis sets aside just like other proposals do. It thus is not any more comprehensive with regard to the evidence than the other proposals.
Indeed, the fact that it is sometimes represented as explaining the evidence without also noting that it adjusts the meaning of the patristic testimony is worrisome.
Ultimately, the Orchard hypothesis fails to convince. The core of the hypothesis is the proposal that Mark is based on a set of lectures Peter gave on Matthew and Luke, and the claims made regarding this core set of propositions are implausible.
Even granting for purposes of argument that Matthew and Luke were written first (something that can be vigorously argued), it is antecedently very unlikely that:
- an eyewitness like Peter would have given a set of lectures on Matthew and Luke, alternating between them in the course of a single lecture (Claim #4)
- Mark would have had such a set of lectures transcribed (Claim #5), and
- there would have been a demand for multiple copies of the transcript, given its high cost and the existence of Matthew and Luke (Claim #6).
Further, it is precisely with respect to these claims that the Orchard hypothesis must adjust the patristic testimony (from John the Presbyter and Clement of Alexandria) in order to support itself.
It could even be said that the Orchard hypothesis misrepresents the patristic data. However, we may view this more charitably by saying it partially incorporates the data (by acknowledging John and Clement’s claims that Mark is somehow connected with the preaching of Peter) and partially rejects it (by setting aside the way in which those sources say it is related to Peter’s preaching—i.e., based on Mark’s memory).
However, the fact that it must reject the means by which these sources connect Mark’s Gospel to Peter’s preaching means that they do not offer the Orchard hypothesis positive support. It must simultaneously accept and reject significant claims from these sources, resulting in no net support.
In sum: The core of the Orchard hypothesis involves a set of claims that are antecedently improbable and not supported by the posterior, patristic evidence we have available to us.
It thus represents a clever and noble effort to account for the data, but one which ultimately does not succeed.