The Synoptic Problem

Matthew, Mark and Luke are called the synoptics because they can be set out in parallel columns and seen together, from the Greek synopsis. The arrangement itself goes back to Johann Jakob Griesbach's printed synopsis of 1776, and the problem became unavoidable the moment the three were laid out that way, because what the columns show cannot be explained by three men independently remembering the same events.

Begin with quantity. By the usual counts, Mark has about 661 verses, of which roughly 600 have a parallel in Matthew and roughly 350 in Luke. The figures shift depending on how you count a parallel, but the shape does not: almost all of Mark is somewhere in Matthew, and more than half of it is in Luke. Beyond that triple tradition there are around 235 verses shared by Matthew and Luke alone, almost all of it teaching rather than narrative. Then there is material unique to Matthew, including the infancy narrative and several parables, and material unique to Luke, including his infancy narrative, the good Samaritan and the prodigal son.

Quantity alone proves little; people who witness the same event report the same event. What forces the question is the character of the agreement. In the triple tradition the three agree in Greek wording at roughly the fifty per cent level, including in the ordinary connecting particles and participles that no two people would choose alike by chance. They agree in unusual vocabulary. And they agree in oddities.

Three oddities are worth having by heart, because every solution has to account for them. First, in the healing of the paralytic all three break the same sentence at the same point to insert the same aside, he said to the paralytic, producing the same broken Greek construction in all three. Second, in the discourse on the Mount of Olives both Mark 13:14 and Matthew 24:15 contain the words let the reader understand, an instruction addressed to someone holding a text and not to a crowd on a hillside, reproduced in two gospels. Third, Mark 1:2 introduces a composite quotation as written in Isaiah the prophet when the first half of it comes from Malachi 3:1; Matthew and Luke at the corresponding point both quote only the Isaiah portion, and both use the Malachi words later, at Matthew 11:10 and Luke 7:27, in almost identical wording. Asides, awkward grammar and a shared handling of a difficulty are the fingerprints of documents, not of memory.

Then there is order. Matthew and Luke both follow Mark's sequence for long stretches. Each departs from it at various points, but they rarely depart in the same way at the same place. Mark is always in the middle: agreements of Matthew and Luke in order against Mark are scarce. Against that stands a smaller but stubborn body of evidence known as the minor agreements, places inside the triple tradition where Matthew and Luke agree in wording against Mark. The most discussed example is at the mocking: Mark 14:65 has some of those present say only Prophesy, while Matthew 26:68 and Luke 22:64 both add the same five Greek words, tis estin ho paisas se, which the BSB renders Who hit You. There are several hundred such agreements, most of them tiny, a few of them not.

The case as its defenders make it. Mark is shorter, rougher and harder, and the differences run in a direction that is easier to explain one way than the other. Mark's Greek is full of redundancies, of the historic present, and of Aramaic words that then need translating; Matthew and Luke have the same episodes in tidier Greek with the redundancies gone. Mark says Jesus could not perform any miracles at Nazareth, apart from a few healings, and Matthew says he did not do many. Mark has Jesus ask why do you call Me good and Matthew has him ask why do you ask Me about what is good. In each case the Marcan form is the one a later writer would want to soften, and it is difficult to see why anyone reading Matthew would produce Mark's version instead. Karl Lachmann's essay of 1835 on the order of narratives, and then Heinrich Julius Holtzmann's work of 1863, put this argument into the form that carried the field, and B. H. Streeter's The Four Gospels in 1924 gave it the shape most students still meet.

The Q part of the hypothesis is a separate argument and should be assessed separately. If Mark came first, the roughly 235 verses of teaching shared by Matthew and Luke still need an origin. The two-source answer is a lost collection of sayings, labelled Q, generally taken to stand for the German Quelle, source; the siglum came into use in German scholarship in the later nineteenth century. The three main arguments for it are these. The shared material sits in almost completely different places in the two gospels, which is strange if one writer had the other's book in hand and was free to keep its arrangement. The primitivity alternates: sometimes Matthew's version of a saying looks earlier and less developed, sometimes Luke's, which is hard if either is simply revising the other. And both gospels contain doublets, the same saying twice in slightly different forms, as though each writer found it in two sources and used both. Defenders also point out that a sayings collection is not an exotic thing to postulate; the Gospel of Thomas, found at Nag Hammadi in 1945, is a collection of sayings with no narrative, which shows the form existed.

The difficulties, stated at full strength. Q is a document nobody has ever seen. No manuscript, no fragment, no ancient writer mentions it, and a hypothesis whose central object is invisible carries a burden that a hypothesis using only extant texts does not. The minor agreements are a genuine problem: if Matthew and Luke worked independently from Mark, they should not agree against him hundreds of times, and Streeter's explanations of them, by appeal to textual corruption, coincidence, and an overlap between Mark and Q, struck many readers then and since as a list of expedients rather than an explanation.

There is also a logical difficulty at the heart of the argument from order, and it is worth naming precisely because it is often skipped. Butler pointed out in 1951 that the pattern in which Matthew and Luke never agree in order against Mark shows only that Mark is the middle term of the three. A middle term can be the source both others used, or the document that combined the other two. The order argument by itself cannot tell you which, and treating it as a proof of Marcan priority is sometimes called the Lachmann fallacy. Defenders of Marcan priority generally concede this and rest the weight on the direction of the wording changes instead.

Finally, the sociological point cuts both ways and should be stated plainly. The two-source hypothesis is the majority position and has been for well over a century; most commentaries, most lexica and most introductory textbooks assume it, and a student will meet it presented as settled. That majority is a fact about the discipline. It is not by itself an argument, it was formed in a particular nineteenth-century German setting, and a view can be dominant because it was taught to everyone who now teaches. A view can also be dominant because it is the best explanation anybody has produced, which is the usual reason views become dominant, and the bare fact of the majority does not tell you which of the two is the case here.

The case as its defenders make it. Austin Farrer's 1955 essay On Dispensing with Q proposed keeping Marcan priority and deleting the invisible document: Mark wrote first, Matthew used Mark, and Luke used both Mark and Matthew. Michael Goulder developed it through the 1970s and 80s, and Mark Goodacre's The Case Against Q in 2002 gave it its fullest modern statement. Its first claim is economy. Every document in the theory is one you can hold in your hand, and the double tradition is explained without postulating anything.

Its second claim is that it solves at a stroke the problem the two-source hypothesis handles worst. If Luke read Matthew, the minor agreements are not a difficulty at all; they are exactly what you would expect, the traces of Luke's second source showing through where he follows Matthew rather than Mark. Those five Greek words need no explanation beyond Luke having read them in Matthew. Farrer's supporters argue that the accumulated weight of several hundred such agreements is decisive, and that the two-source hypothesis pays for a tidy account of the double tradition by making a mess of the triple.

It also brings its own positive evidence. Goodacre's argument from editorial fatigue observes that a writer copying and altering a source tends to keep up the alteration at the start and then lapse back into the source's wording, leaving inconsistencies. He finds such lapses in Matthew and Luke against Mark, which supports Marcan priority, and also finds them in Luke against Matthew in double-tradition material, which points to Luke's use of Matthew. Farrer supporters further argue that Q, as reconstructed, has quietly acquired a community, a theology, a set of redactional layers and even a Galilean home, which is a remarkable amount of scholarship to build on a document whose existence is the thing being argued.

The difficulties, stated at full strength. The hardest is Luke's treatment of Matthew. If Luke had Matthew in front of him, he took the Sermon on the Mount, one of the most effective pieces of prose in ancient literature, and broke it into fragments scattered across ten chapters. He ignored Matthew's infancy narrative and wrote a completely different one. He did not take over a single one of Matthew's distinctive additions to Marcan material, such as Peter walking on the water or the coin in the fish's mouth. He dropped the great commission. Defenders answer that Luke had his own compositional plan and that a writer may well dismantle a source he thinks ill-arranged, and they point to Luke 1:1 as evidence that Luke thought earlier attempts needed improving. Critics answer that this is not editing, it is vandalism, and that no one has produced a convincing account of Luke's purposes in doing it.

A second difficulty is the alternating primitivity of the double tradition. On the Farrer view Luke is always the later writer, so Luke's versions should consistently look secondary to Matthew's. Many readers judge that they do not; the Lucan form of some sayings and of the Lord's Prayer looks earlier. Farrer supporters reply that judgements of primitivity are soft, that a shorter form is not automatically an earlier one, and that the criterion has been applied with an eye on the answer. That reply has force, and it also shows how much of this debate turns on judgements that cannot be made mechanical.

The case as its defenders make it. On the two-gospel or Griesbach hypothesis, Matthew was written first, Luke used Matthew, and Mark wrote last, working from both and abbreviating. Griesbach argued this in 1789, it went quiet for a century and a half, and William Farmer's The Synoptic Problem in 1964 revived it as a serious scholarly programme. A related older view, the Augustinian hypothesis, keeps Matthew first but has Mark abridge Matthew and Luke come third.

Its first argument is external. Every early witness who says anything about order puts Matthew first. Papias, quoted by Eusebius, says Matthew compiled the oracles in the Hebrew dialect and that Mark wrote down what Peter preached, not in order. Irenaeus places Matthew first. Clement of Alexandria, again through Eusebius, reports a tradition that the gospels containing the genealogies were written before the others, which puts both Matthew and Luke before Mark. The two-source hypothesis requires that this unanimous early testimony be mistaken about the one thing it actually addresses. Defenders of Matthean priority ask why a modern reconstruction should be preferred to the only witnesses who were anywhere near the events.

Its second argument is internal, and it inverts the usual reasoning. On this view Mark is the conflator, and that explains the phenomena better than Marcan priority does. The minor agreements cease to be an embarrassment and become the expected result: wherever Mark departs from the shared text of Matthew and Luke, the other two are left agreeing against him. The order pattern is explained too, since a writer working with two documents in front of him would naturally follow one where they diverge and both where they agree, leaving himself permanently in the middle. And on this view Mark's roughness is not primitiveness but the ordinary result of compressing two polished texts at speed for a Roman audience, while adding vivid detail to the scenes he keeps.

The difficulties, stated at full strength. The central objection is about Mark's omissions. If Mark had Matthew and Luke open in front of him, he left out the Sermon on the Mount, the Lord's Prayer, both infancy narratives, the beatitudes, the good Samaritan, the prodigal son and almost all the resurrection appearances, while expanding the individual healing stories he did keep with extra detail. An abbreviator who lengthens the miracles and discards the teaching is an author who requires explaining, and critics judge that no satisfying explanation has been offered. Farmer's answer, that Mark wrote for a specific catechetical setting where the teaching was already in use, is a real answer, but few outside the camp have found it sufficient.

The second objection concerns direction of wording. Griesbach supporters have to hold that Mark changed did not do many miracles into could not perform any miracles, and why do you ask Me about what is good into why do you call Me good, that is, that Mark systematically made the tradition harder and rougher. That is not impossible; harder readings can be created as well as smoothed away, and Mark's whole presentation of the disciples is unsparing. But it runs against a general instinct about how traditions develop, and that instinct is doing a great deal of the work on the other side too. It is worth noticing, as Farmer's critics and defenders both eventually do, that the criterion of the harder reading is a rule of thumb about scribes, not a law about authors.

Settled, in the sense that essentially no one working on the texts disputes it: there is a literary relationship among the three. Shared asides to the reader, shared broken sentences and fifty per cent verbal agreement in Greek are not produced by three independent memories. Somebody used somebody's writing. Also settled: the early church preserved three overlapping accounts rather than suppressing the duplicates, and the writers felt free to arrange, compress and rephrase.

Not settled: who used whom. The two-source hypothesis holds the majority and explains the double tradition's independent arrangement and alternating primitivity well, at the cost of an unseen document and a weak account of the minor agreements. The Farrer hypothesis explains the minor agreements with nothing left over, at the cost of a Luke whose treatment of Matthew is hard to make sense of. The Griesbach hypothesis honours the external testimony and turns the order pattern into positive evidence, at the cost of a Mark whose omissions are hard to make sense of. There are further options in play: the Augustinian order, Matthean posteriority in which Luke precedes Matthew, defended recently by Robert MacEwen, and various multi-source models with lost intermediate editions. Each of these is held today by people who know the Greek text far better than either of us.

There is also a live minority position that the whole question is malformed, because it assumes a world of writers at desks copying scrolls when the reality was a world of continuous oral performance in which written texts were unstable and interacted with living memory. Scholars working on orality and social memory, including Werner Kelber and James Dunn, argue that some of the variation everyone is trying to explain by copying is the ordinary variation of a story told many times by people who knew it. This does not replace a source theory, since the verbatim asides still need explaining, but it does suggest that a clean stemma of documents may be the wrong shape of answer.

Here is the practical consequence, and it is smaller than the volume of the debate suggests. No doctrine held by any Christian tradition depends on the answer. Whether Luke had Matthew or a sayings collection or both, Luke tells you in his own first sentence that he used sources and worked over them, and the church that received his book knew that and received it anyway. Inspiration, on every account of it that the tradition has actually held, works through ordinary human processes of writing, including research and revision.

So the discipline to practise is this. When you read a commentary, notice which solution it assumes, because almost every commentary assumes one silently and the assumption shapes its exegesis of every parallel passage. When you hear an argument, ask whether it turns on data both sides accept or on a judgement about what an author would plausibly have done, because the second kind is where nearly all the real disagreement lives. And when someone tells you the problem is solved, ask them to state the strongest objection to their own view. If they cannot, they have not finished reading.