Christianity
The Gospel of Mark, Marcan Priority and the Source Q
Abstract
© Dr M D Magee
Contents Updated: 31 October 1998
Gospel Priority
Christian scholarship is a type of treadmill. Christians keep plodding round the same material until they return to the conclusions they had long ago discarded. For over a hundred years, most scholars have been satisfied that the shortest and clumsiest gospel, Mark’s, was the first one, an idea called the “Priority of Mark”.
The hypothesis of the priority of Mark is that Mark wrote, then first Matthew and then Luke each used Mark independently as a source for their own gospels. But Matthew and Luke have a lot of common material not in Mark, mainly sayings of Jesus, but with narrative material too. Where did this common material come from if Matthew and Luke wrote independently? This generated the “Two-Source” hypothesis. Matthew and Luke both used Mark but each had another source, largely of the sayings of Jesus, which is now lost. The lost source was noncommittally called “Qwelle” (“Q”), simply German for “source”.
In the introduction to his gospel, Luke admits that he had sources, saying that “many” including eyewitnesses had declared that which Christians believed. This has been taken to be evidence for his use of Mark and Q as main sources and other lesser ones. Streeter, in 1926, expounded the Two-Source theory fully in his study of the origins of the gospels. Indeed, he went further because some material in both Matthew and Luke are exclusive to each gospel and so must have been the authors’ own work or taken from further sources of which we know nothing. For 60 years, Streeter’s exposition was accepted almost universally.
But Christians never liked Mark being the first gospel because it lacks many of the features that they hold dear, such as Jesus’s appearances and the birth narratives. If these did not appear in the first gospel, the implication is that they were invented for later ones. Christians always wanted Mark to be an abbreviated version of Matthew and Luke as people had once thought. The treadmill having revolved a turn, today Christian polemicists like William R Farmer champion the idea again—in the “Two Gospel” or “Griesbach Hypothesis”.
Its supporters believe the gospels were written in the order Matthew, Luke and Mark. Luke made use of Matthew and Mark made use of Matthew and Luke thus dispensing with Marcan priority and the hypothesised source Q. If Luke did not use Matthew directly, then material common to Matthew and Luke not derived from Mark, must be from the hypothetical source, Q. But, if Q did not exist because Luke had Matthew to hand, Farmer asserts the reasons for Marcan priority are not valid. There are also other theories that dispense with Q and Marcan priority. Here they are summarised.
- The Griesbach hypothesis—Matthew was copied by Luke, and Mark conflated them both—can explain the alternating support in order of pericopes of Mark by Matthew and Luke without appeal to “lost sources”. Its failing is providing a comprehensible account of Mark’s composition, particularly explaining Mark’s major omissions from his sources Matthew and Luke.
- Austin Farrer’s and Goulder’s theory of Marcan priority—Mark without Q—explains the agreements of Matthew and Luke against Mark by accepting Luke’s direct use of Matthew, dispensing with any need for Q, but it cannot explain the order in Mark in relationship to Matthew and Luke as well as can the Griesbach theory. Goulder is not convincing that Q can be abandoned. However, that Matthew used Mark and Luke used Mark and Matthew is more convincing than the previous hypothesis—that Mark used Matthew and Luke.
- The Boismard Multiple Source theory affords an explanation for agreements between Matthew and Luke against Mark but it fails Occam’s Test. It is too complicated to be validated and seems less likely than the Goulder and Austin Farrer theory. Boismard takes us into the realm of conjecture, where everything is possible whereas the scientist tries to keep to the simplest of hypotheses to meet Occam’s razor. Nevertheless, the real history of the gospels will have been complicated.
Farmer’s lengthy article in support of the Two-Gospel hypothesis and attacking the Two-Source hypothesis (Synoptic Problem Seminar at SNTS, Copenhagen, Denmark, 4-8 August 1998) is savage but not convincing. He adopts a high scientific posture but omits much of the context. He pleads that his opponents atomize the gospels in analysing them and says, since the Synoptic Problem is the problem of explaining the agreement and disagreement between the Synoptic gospels—Matthew, Mark, and Luke—it is important to see these three gospels as synoptic, “together in the same view”. Yet, he ignores the background of the early Christians as if it was so irrelevant as to be a distraction.
It is hard to analyze something without pulling it to pieces. Surely, that is an aspect of science. Nowadays, the term used is “backward engineering”. The enemy aircraft is taken to pieces to find out how it can be synthesised—put back together again! No one pretends that an aircraft in bits can fly, but only by taking it to bits can the engineer understand how it flies.
Science has to work in different ways according to the circumstances. Its bases of observation, hypothesis and deduction are constant but the ways of testing hypotheses have to change. In experimental science, an experiment, devised to distinguish between hypotheses, tests them. Unfortunately, in history, palaeontology and textual criticism amongst others, experiments cannot often be carried out and the evidence is fixed pending new physical discoveries. In these cases, scientific criteria cannot be so strict. Christian scholars speak of “proof”, the more fundamental they are, the more affirmatively. Their own personal convictions are proof enough—they are certain, 100 per cent. They often demand the same degree of proof from scientific investigations. It is always impossible. However many times an event is observed there is no certainty it will happen in the same way next time.
We might never know for certain which gospel was first. The evidence is conflicting. The churches always took it to be Matthew, and the only reason that was doubted was because it posed problems in conjunction with our reading of the other synoptic gospels. Unless we unearth an unequivocal statement that So-and-so’s gospel was the first, or an indisputable statue to So-and-so the author of the first gospel, we are unlikely to be able to certainly deduce which of them came first. We have to take the balance of the evidence, and the evidence will be unbalanced to different people to different degrees. Most scholars still use the Two-Source Hypothesis because the other ideas offer even greater problems.
Marcan Priority and Q
Since the point of Farmer’s article is to attack Marcan priority let us, briefly, recapitulate the arguments for it.
- Mark is the shortest gospel. If it is a synopsis of Matthew and Luke, why were much loved verses from both like the Good Samaritan and the Sermon on the Mount, not to mention either birth narrative, omitted while Jesus’s foibles are included? It is more feasible to believe it is both shortest and partial because other traditions had not yet been invented or collected.
- Mark is the least polished and most vivid suggesting its early date as the recollections of an eyewitness—traditionally, Peter.
- Mark’s text is altered in Matthew for good reasons. His Greek was poor, his knowledge of Palestine poor, he was critical of both Jesus and the disciples, and he was often unclear. Streeter argued for Marcan primacy from Matthew’s tendency to improve Mark’s style. Later evangelists expunge Mark’s frank descriptions. If Mark followed Matthew and Luke, it is nigh on impossible to imagine why the developing church would want to adopt such a poor gospel compared with the ones it already had. Sanders and Margaret Davies say that though no one can recover ancient intentions, there are strong objections to accepting Griesbach’s Mark as the third gospel, though they do not rule it out.
- Occasionally Matthew alters Mark then fails to make a necessary later change to match. In Matthew 14:5 Herod was only stopped from murdering John the Baptist because of the crowds who thought him a prophet. Incongruously, in verse 14:9 he was sorry he had to behead him. His sorrow matches Mark where Herod admired John as a good man and just—an Essene. Matthew altered Herod’s admiration to murderous distaste but forgot to alter his sorrow. He was copying Mark. Farmer would reverse this argument saying that Mark was trying to correct Matthew’s error, but Mark’s style is so bad, arguments this way round carry no weight.
- The order of Luke is essentially that of Mark and the order in Matthew is also that of Mark except where he has deliberately grouped items into general discourses by his hero. Once thought to have been a strong argument, critics point out that the order would largely have been followed whoever did the copying, so preservation of order in Mark is not a strong argument for it being original. In other words, the common general order of Matthew, Luke and Mark can be explained by turning Streeter’s argument round. Only someone writing after Matthew and Luke and deliberately combining them can keep the order Luke preserved from Matthew and yet, whenever Luke departed from Matthew, following the order of either one or the other. However, the argument has to be taken in context. If Luke used Matthew, he ignored his special groupings, a strange decision when Matthew was supposedly the first gospel and highly regarded. So, we have to suppose that Luke ignored the special ordering of Matthew and then so did Mark who was copying from both Luke and Matthew. The argument from order is that the priority of Mark is simpler and more reasonable in the context of all three Synoptics—one of Farmer’s bleats!
- Some passages in Matthew and Luke are almost identical. If Mark is basing his choice of extracts from his two master texts on consistency, as the Griesbach hypothesis supporters believe, why has he missed out the most consistent parts?
- If Luke copied Matthew, he sometimes follows Matthew’s words closely but not his order of events and other times he does the reverse. Here also is a curious inconsistency.
A lost source surmised alongside the priority of Mark explains most of the problems of commonality between Matthew and Luke while allowing them such different approaches. Matthew has plainly arranged his material in a non-narrative way to give Jesus several major discourses, the most famous of which is the Sermon on the Mount. Sayings from these discourses appear all over the place in the other two gospels. It is hard to accept that later evangelists would be so cavalier with the work of the original one. Why also should Luke be so respectful of Matthew’s words when he plainly thought his order was wrong. If Luke and Matthew knew nothing of each other’s work but used common sources, such things are not problems.
Different instances in both Matthew and Luke seem to preserve the original story most closely. If Matthew had been the original gospel, it should have a clear predominance of original wordings. On the other hand, why should Mark, copying from the two, miss out passages which are effectively identical in both and therefore apparently doubly reliable. The hypothesis of the source Q explains these puzzles because both evangelists took passages they admired almost verbatim from Q, but Mark had no access to it.
Hoping to be iconoclastic, a supporter of Farmer’s view, Christopher M Tuckett, in the New Interpreters Bible, 1995, notes that nearly all of Mark is paralleled in Matthew or Luke or both, yet all this shows is that some literary relationship exists. It does not prove that the only possibility is that Mark’s gospel was the source of Matthew and Luke. The failure of Matthew and Luke hardly ever to agree against Mark in order and wording does not prove that Matthew and Luke independently used Mark as a source. It only shows that Mark is some kind of “middle term” between the other two in any pattern of relationships.
To Farmer’s glee, Tuckett declares in this “flagship publication” the “weak and inconclusive nature” of some of the arguments for the Two-Source Theory, adding that since Streeter’s arguments for Marcan priority are no longer reliable, work based upon Streeter has to be questioned—most of the work this century!
The situation seems serious. But is it? Is Streeter really demolished? Are the attacks on him well founded or just the perambulations of the Christian “scholars” round their cloistered universe in which the same texts are consumed and regurgitated in different patterns according to the current fashion? Why are the reasons for Marcan priority more “weak and inconclusive” than those of Farmer’s alternative hypothesis?
Criticisms of the Two-Source Hypothesis
Are Matthew and Luke independent as the Two-Source hypothesis requires? Of all the counter arguments to the Two-Source hypothesis, the weightiest is the question of minor agreements between Matthew and Luke—they seem to be prime evidence for literary contact between the two. There are also about six pericopes of about 32 verses present in Mark but absent from Matthew and Luke. One such is the seed growing in secret in Mark 4:26-29. Proponents of the Two-Gospel hypothesis take these passages to be agreements between Matthew and Luke against Mark, and should be added to that category of evidence against the Two-Source hypothesis, as if it is obligatory for editors to use all of the material at their disposal. Farmer argues that advocates of the Two-Source theory have never satisfactorily explained the agreements of Matthew and Luke against Mark.
Though there are arguably many examples of minor agreements, the attitude of Farmer, et al, is that any one of them serves to destroy the idea of the priority of Mark. He cites Goulder, who has taken what Streeter designated as one of his “residual cases” where Jesus is being mocked and has shown that the agreements of Matthew and Luke against Mark cannot be explained except by requiring Q to expand to include a passion narrative. For Farmer, this reduces to an absurdity the chief tenet of the Two-Source theory, that it can most simply satisfy the synoptic data by offering Mark to explain narrative agreement of Matthew and Luke, and “Q” to explain their agreement in sayings of Jesus not in Mark.
Hans-Herbert Stoldt has compiled 272 minor agreements, an apparently impressive number. Many, however, are merely grammatical agreements. If two editors don’t like split infinitives we can get “to boldly go” altered to relatively few alternatives. If both choose to alter it to “to go boldly” is that a minor agreement proving literary contact? Of course not, other than that both are using the same language and have the same idiosyncrasies of style. If Matthew and Luke independently used Mark but corrected him, almost every minor agreement can be explained in an innocent way.
Stoldt’s excessive list actually makes the argument of minor agreements ridiculous. The word “minor” disguises a few quite serious agreements that cannot have come from Mark or from Q. But, though hardly minor, do they necessitate the hypothesis that Luke used Matthew? If he did, the argument leads us back to questions like Luke’s blatant ignoring of Matthew’s arrangement of events. Some have suggested that Matthew or Luke got hold of the other at a late date in writing his own gospel and plagiarised a few stories he thought impressive. Who can say it is not true?
Here, of course, we meet another absurdity of the critics of the Two-Source hypothesis. It is treated as if it alone must be a complete solution to the Synoptic Problem. Yet, anyone knows that the evolution of the gospels took at least a century and probably as many editors. So, there was time for editors to take episodes from one gospel into another, if they thought it illuminating. Other contributions were elements of the oral tradition which emerged, perhaps quite late and became included in slightly different forms in both gospels then were perhaps harmonised to some degree by later editors, and deliberate or inadvertent harmonisation between the two gospels by copyists familiar with both.
Doubtless there will be an outcry—“All this speculation isn’t very scientific!” Perhaps not, but it is probably what happened, and there will never be any way of knowing precisely. The reasonable Two-Source theory is that Mark and Q were the original sources, Mark about 60 AD and “Q” possibly a pre-Christian book of wisdom sayings—Essene, doubtless—but Christianised as the Logia about mid-century and conceivably, even then, associated with the name Matthew. Mark itself probably came in at least two versions—pre-Jewish War and post Jewish War—parts of which were eventually bound together probably in error, giving us the curious double image in the middle of Mark which a diligent editor tried to rationalise but which Matthew and Luke were unaware of. If Matthew and Luke copied parts of this earlier edition of Mark, parts of which were later omitted or changed, their agreements together against our present edition of Mark are explained.
Later, a learned editor improved the Logia “Gospel of Matthew” by incorporating Mark to give us essentially the modern Matthew. Later still, another compiler, known now as Luke, put Mark and Q together in a different style suitable for gentiles rather than Jews. It seems that Mark fell out of fashion and was almost lost because the two newer gospels were preferred, but was eventually saved when the church put together its canonised books. Q was not so lucky perhaps because it was essentially a book of sayings with little or no narrative. It disappeared into history’s dustbin.
Before the gospels became well known and eventually canonised, copyists made piecemeal changes in pious attempts consciously or otherwise to harmonise them. One might ask why the editors did not go the whole hog and just harmonise the gospels instead of writing new ones. Well, Farmer is trying to say they did—the result was Mark, a gospel that was the bare consensus of the other two gospels. If so, the bishops picked a dunce to compile it. Later, there was a complete harmonisation of the four gospels, which was popular for a long time. But some people and some churches preferred the originals. In short, nobody had the authority to replace earlier gospels with harmonised versions, and eventually we finished up with four of them in the New Testament and many others lost or rejected.
The Two-Gospel hypothesis requires the author of Mark to rewrite Matthew and Luke omitting the miraculous birth of Jesus, the sermon on the mount, and the resurrection appearances, but adding the young man fleeing naked, a difficult healing miracle for Jesus and the putative God supposed mad by his family. Farmer merely says that all this needs to be explained if Mark was the third gospel. Why adopt a hypothesis with such serious defects? But the least credible aspect of it is that the church would accept such a poor rendering of the gospels of Matthew and Luke as that of Mark. Not only that but it then added to this simplistic piece of work the legend that it was the gospel spread by the church’s greatest Saint, Peter. If the church really did this then it was the very first pious forgery!
The Direction of Early Christology
Whatever Christians believe, there is no doubt that Jesus, if he lived at all, was a man. From a man he was transformed into a god. That change can be seen in the progression of the gospels. In almost thirty verses, Mark has a more human Jesus than Matthew. If Mark followed Matthew, the Church sanctioned a work in which Jesus was made less god-like than he was in the source gospel. Farmer, quite feebly tries to refute this by saying no one knows the direction of primitive Christology, arguing a high Son of God Christology was well entrenched in the church long before the gospels were written, to judge by Paul’s letters. Paul’s letters themselves have been jumbled up and multiply re-edited before they reached their present state. Furthermore, by Paul’s own evidence, he and Peter were at loggerheads, so why should Paul’s Christology have to be overwhelmingly present in a work that is, according to tradition, Peter’s?
Semitic expressions are found in Mark but not in Matthew. Why should Mark writing for a gentile audience introduce foreign words into his supposed summary of Matthew and Luke? The only explanation again would be that Mark was a deliberate forgery, intended to be attributed to Peter, a simple Aramaic fisherman.
A similar problem is that the Greek text for Jesus’s question from the cross,
My God, my God, why hast thou forsaken me,
conforms to Aramaic orthography in Mark 15:34 whereas Matthew’s parallel Greek text (27:46) conforms to Hebrew orthography. Only Matthew’s Hebrew “Eli, Eli” properly explains the bystander’s observation. “This man is calling Elijah.” Mark’s Aramaic “Eloi, Eloi” seems unlikely to have been mistaken for the name of Elijah.
Farmer believes Mark made the change for some curious reason and so destroyed the credibility of the words of the onlookers. Really, this is an argument for Marcan priority. Mark was, according to tradition, Peter’s interpreter and Peter was an Aramaic speaker. Jesus was well versed in the scriptures as he would necessarily have been as an Essene leader. In his agony, Jesus would have quoted the scripture in Hebrew, as he had learnt it, but Peter would have related the story in Aramaic. Matthew, a Jewish scholar, would have realised Jesus must have quoted Hebrew not Aramaic and therefore restored the original when he corrected Mark. In the 60s of the first century, the church must have had some Aramaic speakers accessible to it, even if most Jewish Christians had by then abandoned Christianity. Why then would it have permitted Mark to make such a blunder in his harmony? “Eloi” in Mark favours the idea that the book records the memories of an Aramaean speaker.
Many of Mark’s readers did not understood Aramaic expressions because he generally translates them (Mark 3:17; 5:41). He translates the Aramaic word “corban” (Mark 7:11) whereas, when Matthew uses a related word elsewhere in his gospel (Mt 27:6), he leaves it untranslated. Farmer believes this shows some of the Greek text of Matthew was first used in circles where Aramaic loan words in Greek were well understood. Possibly! Matthew wrote for Jews, but they would have been worldly Jews in the wider Empire and so would have understood both Greek and Aramaic. Otherwise, it shows that, when Matthew’s gospel was offered to gentile Christians, it did not need “corban” editing out or translating because the gentile Christian audience knew what “corban” was—Mark had explained it, Mark being the earlier work.
Farmer thinks that the use of Aramaic expressions in his gospel shows that Mark had an interest in including such expressions in his text for effect. He explains that the well established use of the Aramaic word “abba” for Father, meaning God, in churches acquainted with Paul’s letters, including the Christian community in Rome, reminded Mark’s readers of Jesus’s origins. So, Mark deliberately wrote the text of his gospel in a bogus Aramaic style to make it seem authentically Nazarene. Farmer is telling us the author of Mark was not honest. Why not believe that Mark wrote what Peter had said and translated it for his Greek speaking audience, as the church always said? Otherwise, the Church is being accused of blatant forgery.
No critic would deny that the gospel of Matthew is more Jewish than Mark, and more Palestinian. Matthew evidently was a Palestinian Jew, though he would likely have been writing in Alexandria or Antioch. It seems odd that Matthew, a Jew, should have taken the clumsy work of the gentile, Mark, to write a Jewish gospel. This can be taken as an argument against Marcan priority, but it is fully understandable if Mark’s work was known as the reminiscences of Peter.
Overlaps between Mark and Q, Farmer’s school tells us, are not allowed in the Two-Source hypothesis. Nor is Q allowed to grow to explain the agreements of Matthew and Luke against Mark. The Two-source hypothesis must maintain the independence of Matthew and Luke. These restrictions are placed on the Two-Source hypothesis because, if it didn’t have them, it would undermine the ground for the Griesbach hypothesis. It is difficult to imagine that Mark and Q did not overlap. Why should the author of Mark not know any of Jesus’s sayings recorded in Q? Mark’s gospel is not rich in sayings but why should some of those which appear in Mark not have also appeared in Q? If Q contained elements of narrative, some of those too might have overlapped.
Moreover, Streeter postulated specific sources M and L for material peculiar to Matthew and Luke. Who is to say for certain that none of this M and L material was not part of Q? The source, Q, is not merely the material common to Matthew and Luke but a different book in its own right. Who knows that Matthew did not omit verses that Luke liked and vice-versa? Parts of M and L therefore originated in Q. Q is understood to be sayings with little or no narrative, but the gospel authors might have changed sayings into narrative. The physical dumbness of the priest Zachariah in Luke’s birth narratives was a metaphorical dumbness in a thanksgiving chant. So, what is to stop Q from expanding, if it explains things?
Once Boismard is followed and it is accepted that one or more gospels existed in more than one edition and that the gospels as we have them may have been dependent on more than one proto or intermediate gospel, the criticisms of the Two-Source hypothesis start to evaporate. Only when forced into its conventional straitjacket, does it fail to explain the agreements between Matthew and Luke against Mark.
Conclusion
The basic Two-Source hypothesis adequately explains almost everything that needs explaining but there are a few rough edges that might never be fettled off. Common sense suggests the historic answer is more complicated than the basic Two-Source hypothesis, but the complete answer is unlikely ever to be certainly known. The Two-Source hypothesis at core is secure, but editorial changes, copyists errors and harmonisations, and later interpolations must have complicated the central idea. These relatively minor problems are being used by the Two-Gospel school to beat the advocates of the Two-Source hypothesis. The Two-Gospel advocates are on the thinner ice, though further discoveries might prove them correct. If so, the earliest bishops will be found guilty of forging a gospel of Peter!
