Q: the Synoptic Sayings Source (Red Letter Edition)

Scholars speak of Q as a primary source for our knowledge of Jesus. Yet few people are familiar with it. You will not find it as a separate book in traditional versions of the Bible. Nor is it in most collections of apocryphal gospels. In fact, until fairly recently there were only a few works that presented it in full, even for professional scholars. It was discovered more than a century before the gospel of Thomas or the Dead Sea scrolls. Yet its publication has been slower than either of these, because of debate over its contents. Much of this debate was due to the way in which Q was discovered. Unlike the excavations at Oxyrhynchus, Nag Hammadi and Qumran, the digging that unearthed Q did not suddenly produce new manuscripts for study. Instead, Q came to light gradually, as the result of refined methods for studying documents already in hand: Matthew, Mark and Luke. Q was not a text that could be easily set beside these works for comparison, but rather a source that had to be extracted from them. The process of separating Q from the surrounding gospel texts has been a long, delicate operation, the scholarly equivalent of innovative major surgery.

Hypothesis and fact.

Since Q was discovered within other works, its existence as a separate source is a hypothesis. Scholars admit this. But that does not make it less certain. A hypothesis is a theory formed to explain a particular combination of observable facts. Much of what we know about reality is based on hypotheses, inferences that enable us to make sense out of and live in the world around us. The sun will rise tomorrow. Spring will come again. Dad can be trusted. I will die. These are all claims that cannot be strictly shown on the basis of concrete evidence. But they are reasonable conclusions from patterns of experience. We accept them as factual because they work. And we organize our lives around them.

Scientific hypotheses differ from everyday expectations in that they are systematically tested and revised to account for all known data. On clear days one may observe the sun rising in the east and setting in the west. This led ancient cosmologists to theorize that the sun revolves around the earth. Most people accepted this because it fit their common earthbound perspective. But a more thorough review of theoretical models led Copernicus to conclude that the experience of sunrise and sunset could just as well be explained if the earth revolved on its own axis. And if the earth was itself moving, despite common opinion, it could also be moving around the sun. This hypothesis challenged both tradition and the average person's worldview. And it made the universe seem more complex than an earth centered cosmos. But it gave astronomers like Galileo a simpler model to coordinate all the patterns of cosmic movement they observed. Still, Copernicus' hypothesis had to survive a period of opposition by church officials. It passed this test because the invention of the telescope allowed any researcher to see things the unaided eye could not. And these new discoveries supported Copernicus. Today the model of a solar system is considered factual enough for scientists to use it to plan space exploration projects that attract public support despite their cost. While no human has yet seen the solar system, few doubt its existence. It is a well-tested hypothesis: a model that accounts for observed data and serves as the basis for workable conclusions. It is certain as long as the facts support it.

The Q hypothesis is similar to that of the solar system. Like the latter it was proposed as part of a new model that researchers needed to explain details that an ancient theory could not. And it met the same type of hostile reaction. But, in spite of continued skepticism in some quarters, it has come to be widely accepted among biblical scholars because it accounts for patterns of material they observe in the gospels better than other models. And it works well enough to base further discoveries about the composition of gospel texts upon it.

The eye-witness hypothesis.

All material comes from a source either directly or indirectly. The gospels present indirect historical evidence of Jesus because Jesus did not personally write them. But the value of indirect evidence depends on its relation to its source. Thus, many gospels were ascribed to people who were close associates of Jesus. Church officials of the second century CE accepted only four of these works as really "apostolic," implying that they preserved tradition that came from the earliest period of Christian mission. But Mark and Luke were not known companions of Jesus. So their information about Jesus could be claimed to be second-hand at best. John and Matthew are names known from lists of the apostles. But John was widely reported to have written last. This left Matthew to claim first place in the churches' canon. After all, an eye-witness' words are generally preferred to admitted hearsay.

Matthew's priority was supported by claims that it was written first. This claim is not based on information in the gospel itself, nor is it credited to anyone who witnessed the composition of the gospels. Though the idea that Matthew wrote first can be traced to the early second century, its origin is unknown. Since the sequence in which the gospels were actually written belongs to the inaccessible past, the theory that our gospel of Matthew was composed before the others cannot be confirmed by direct observation. Thus, whatever its source, this claim remains a hypothesis. And, like any theory, it is valid only if it can account for all known facts, which in this case is the content of the gospels themselves.

Documents and deductions.

The theory that Matthew was the first to write remained untested until about two hundred and fifty years ago, when Christian scholars began to compare the gospels in detail. No one set out to challenge the traditional view. But like astronomers centuries before, biblical scholars discovered that old assumptions did not account for what they saw. Rather than deny these observations, they gradually developed a new model of the relationship of the gospels. Unlike the old model, this was not built on hearsay but on a series of deductions from data in the gospels. Because the evidence is documented, these deductions are able to be demonstrated. That is, they can be tested and shown to be more probable than other explanations of the same data. Since each deduction acts as a base for further research, we may rank them in logical progression (D1, D2, etc.). Together they tell us what we know about the origin of the gospels.

Repeated tales.

From the time the gospels were gathered together, people were aware that their accounts of Jesus were similar in some respects and different in others. This was seen as a sign that they contained independent reports of different witnesses. Each evangelist was thought to report the things he heard and saw in his own words. Yet this hypothesis could not account for the pattern of parallels that scholars found in the narrative portions of Matthew, Mark and Luke. Three eye-witnesses might repeat something they heard in almost the same words. But, if their reports are really independent, each description of anything they saw will differ greatly. Yet stories common to Matthew, Mark and Luke frequently give the same details in the same order. For instance:

Matt 8:14-15 Mark 1:29-33 Luke 4:38-39

When Jesus They left He got up from

the synagogue the synagogue

right away

came and entered and entered

to Peter's house, the house of Simon the house of Simon.

and Andrew

along with James

and John.

he saw

his mother-in-law Simon's mother-in-law Simon's mother-in-law

lying sick was in bed was suffering

with a fever. with a fever, from a high fever,

and they told and they made an appeal

him about her to him on her behalf.

right away.

He He went up to her, He stood over her,

touched her hand. took hold of her hand,

rebuked the fever,

Then she was raised up raised her up

and the fever left her and the fever left her. and it left her.

She immediately got up

and started Then she started and started

looking after them. looking after them. looking after them.

One version may be slightly longer or shorter than another. Grammar, transitions and minor details may vary. But the basic elements are practically identical. Only when witnesses read or hear the same story will their accounts of an event have the same structure and wording. So researchers began to realize that such stories in Matthew, Mark and Luke echo the same basic report.

D1. The synoptic gospels often tell the same tale.

Mirrored outlines

The parallels in the synoptic gospels extend beyond single passages to the skeleton of the works themselves. A string of independent stories in one gospel is often mirrored in another with few major changes. For instance, compare the following sequences:

Matt 8:20-9:26 Mark 4:35-5:43 Luke 8:22-55

Two followers

Jesus calms sea Jesus calms sea Jesus calms sea

Demons enter pigs Demons enter pigs Demons enter pigs

Paralytic cured

Toll-collector called

Jesus eats with sinners

Why not fast? New/old

Jairus calls Jesus Jairus calls Jesus Jairus calls Jesus

Woman's bleeding stops Woman's bleeding stops Woman's bleeding stops

Jairus' daughter raised Jairus' daughter raised Jairus' daughter raised

Each gospel spreads these stories over a period of several days in which Jesus crosses the sea from Galilee to the Golan and back. None gives details of the return trip. The sequence in Mark and Luke is identical. The chance of two writers reporting only the same facts for any extended time is slim, unless they are using the same written record. Matthew, on the other hand, mentions four incidents that Mark and Luke omit here. Yet they both know this material. Each locates it near the beginning of Jesus' career, right after stories that Matthew puts just before this section. Compare:

Matt 7:28-8:17 Mark 1:16-2:22 Luke 4:31-5:39

Jesus calls fishermen*

Jesus unlike scribes Jesus unlike scribes Jesus unlike scribes

Unclean spirit silenced Unclean spirit silenced

Leper cured

Soldier's slave cured

Peter's mother-in-law Simon's mother-in-law Simon's mother-in-law

Evening healings Evening healings Evening healings

Jesus goes out Jesus goes out

Great catch of fish

Jesus calls fishermen*

Leper cured Leper cured

Paralytic cured* Paralytic cured*

Toll-collector called* Toll-collector called*

Jesus eats with sinners* Jesus eats with sinners*

Why not fast? New/old* Why not fast? New/old*

Here we see only two stories in a different order (bold) and two more (italics) that only one gospel recalls in this context. Matthew omits some things in Mark and Luke but reports most of them elsewhere (*). Otherwise the incidents are all in the same order. Neither personal memory nor oral tradition can account for such parallels in large blocks of unrelated material. Substantial agreement in sequence is normal, however, between revisions of the same text. So, scholars conclude, the synoptic gospels are based on a common written source.

D2. The synoptic gospels are edited versions of the same text.

Size and style.

Matthew, of course, was considered the prime candidate for the written source of the synoptic gospels. But this view gradually lost favor as it proved unable to explain the details of Mark and Luke. Matthew's 1039 verses present a lot of material not found in Mark's 678 (traditional count). This is not odd. But Matthew's version of passages they have in common is regularly smoother than Mark's (see the story before D1 above). If Mark used Matthew, it is hard to explain why he not only omitted so much but also reduced the rest to a rougher condition.

Comparison with Luke compounds this problem. Luke's 1150 verses are written in an even more polished style than Matthew. But in passages paralleled in the other two synoptics, Luke is often closer to Mark, even when Mark's version is clumsier than Matthew's (see the example for D1). In addition, Mark and Luke have several passages in the same sequence that Matthew either omits or presents elsewhere (see the outlines before D2). Neither of these facts is easily explained by the theory that Mark and Luke used Matthew as a text. So this hypothesis needs many complicated adjustments to stay in use.

More than two centuries ago scholars discovered a simpler solution by identifying Mark as the written source behind Matthew and Luke. A short, rough account usually precedes a long, polished one. And additions to a written document are easier to explain than omissions or substitutions. Matthew and Luke seldom present the same material in the same narrative context when Mark has no parallel. Where Matthew has something that is not in Mark, the text of Luke also usually differs. In fact, Matthew and Luke seem to have each selected large portions of Mark and inserted other material at random (for instance, the italicized stories in the second outline before D2). This is what usually happens when two people edit the same text independently. So, more and more scholars became convinced that Mark, rather than Matthew, was the written source of most parallel passages in the synoptic gospels. And this new hypothetical model has continued to prove more workable for precise comparison of the gospels than the old one.

D3. Mark was edited differently by Matthew and Luke.

The common residue.

The only trouble with tracing parallels in Matthew and Luke to Mark is that Mark does not have all of the material the others have in common. Only two explanations seem possible: either Luke or Matthew copied these passages from the other; or both got it from somewhere else. In principle it is better to base a case on surviving documents instead of unknown sources. But there are problems in claiming that one synoptic gospel is based on the other two. For Matthew usually echoes Mark's wording more than Luke's, even though Luke's is more refined. For instance:

Matt 9:9 Mark 2:14 Luke 5:27

As As After these events

Jesus was walking along he was walking along, he went out

there,

he caught sight of he caught sight of he observed

a man a toll-collector

Levi son of Alphaeus named Levi

sitting sitting sitting

at the toll booth, at the toll booth. at the toll booth.

one named Matthew,

and he says to him, and he says to him, He said to him,

"Follow me!" "Follow me!" "Follow me!"

Leaving everything behind,

And he got up And Levi got up he got up

and followed him. and followed him. and followed him.

Here Matthew gives the subject a different name but keeps Mark's clumsy colloquial style, oblivious to Luke's grammatical improvements. Luke, on the other hand, generally ignores clarifications that Matthew adds to Mark. For example:

Matt 9:12-13 Mark 2:17 Luke 5:31-32

When Jesus overheard, When Jesus overhears, In response Jesus

he said: he says said to them:

"Since when "Since when "Since when

do the able-bodied do the able-bodied do those in good health

need a doctor? need a doctor? need a doctor?

It's the sick who do. It's the sick who do.
It's the sick who do.

Go and learn

what this means:

'I desire mercy

and not sacrifice.'

For I did not come I did not come I have not come

to enlist religious folks to enlist religious folks to enlist religious folks

to change their hearts,

but sinners!" but sinners!" but sinners!"

Matthew has Jesus quote scripture (Hos 6:6) to counter criticism for eating with non-religious Jews. Here Jesus has prophetic support for his behavior. God himself places humanitarianism before religious observance. Luke, on the other hand, introduces a weaker defense: the religious do not need to repent. Jesus' critics might grant this. But it would still not justify Jesus keeping bad company. Nor does it prevent a retort that, by living with "the sick", he stands more chance of being contaminated than of curing them. If Luke knew Matthew, he should have used his argument.

Editors usually try to improve texts. So it is odd for a writer to prefer a partial account, if a fuller version is handy. And it is even stranger that someone who knew a polished version would favor a rough draft. Thus, it is hard to argue that Luke knew Matthew or vice-versa. For both generally echo Mark. But where Mark is silent, they are seldom in tune with each other.

Some points of agreement between Matthew and Luke against Mark are minor and easily explained. The same grammatical improvement of a story in Mark is probably a matter of coincidence. And random insertion of a few common words in a Markan passage show only that ancient authors did not have a modern scholar's concern to quote texts exactly. Our written gospels are records of oral performances where variation of wording was common. The gospels were then read aloud by and to people who had heard this material before. Familiar words from previous oral performances affect the way people read and repeat texts. This is only natural, since most memory of word patterns is oral. It is easier for a person who has learned a story or saying in one version to repeat that than to copy another exactly. Thus, many word agreements between Matthew and Luke prove only that they knew some of the same oral variations in Mark's tradition. They are not evidence for a different written source.

The case for Q is built on parallels in Matthew and Luke that are more than a paraphrase of Mark. The simplest explanation of these passages is that they are based on some unknown common source. The evidence for this source is concrete. It is embedded in the gospels themselves.

Anyone can identify Q material. All you need is the patience to make a detailed comparison of Matthew, Mark and Luke. If you subtract all passages that are only in one gospel, you are left with the parallels. Then, by subtracting all parallels to Mark, you are left with material shared only by Matthew and Luke. It is this residue that is traced to Q.

D4. Matthew and Luke working separately added material from the same source (Q) to Mark.

It was easier to discover Q than to describe it, because it is found only in pieces scattered in other works. Outside our gospels, no text of Q has yet been recovered. That is not itself surprising. Most ancient writings have been lost. Apart from a few fragments and manuscripts found in excavations, only what was repeatedly copied has survived. Many major works are known to us only from bits and pieces quoted by later authors. What we can learn about these sources has to be reconstructed from the texts we have.

But the Q passages extracted from Matthew and Luke are so varied that it was not immediately clear how, or even whether, they fit together. Various proposals had to be debated to solve the puzzle. While this debate still goes on, enough research has been done to indicate some consensus on several points.

Signs of oral composition.

Differences in Matthew and Luke's presentation of Q material raised questions about the form in which they knew it. Was Q written or oral? A written text is fixed in wording and sequence regardless of its length. It may be revised. But unless it is totally rewritten, the various versions share much of the same wording and structure. By comparing several editions of a work, experts can locate the earliest draft. The use of the same rigid patterns in the synoptic gospels convinced scholars that Matthew and Luke based their gospels on the text of Mark.

Oral tradition is more flexible. It is passed on in relatively brief blocks that are complete in themselves. Common features cause several blocks to be linked in our memories. But these are easily separated, rearranged and interspersed with other blocks of material as the occasion demands. Moreover, oral blocks are like rubber. They may be stretched or compressed or bent as needed and still keep their basic composition.

Q seems to have been more pliable than Mark. Matthew and Luke divide the same Q material into different sized clusters and each inserts it at different points in Mark's outline. Such passages were obviously not context bound in their source. Also, in both gospels common threads appear interwoven with different material, some not found elsewhere.

In comparing such passages, it is not always easy to determine whether one author has omitted material from Q or the other has made insertions. Moreover, the sequence of many Q passages is different in each gospel, more than with those taken from Mark. And in some cases, Matthew and Luke's wording differs so much, it is difficult to determine the original shape of the common core. These are signs of oral composition.

D5. Q was composed of oral units that could be separated.

Structured speeches.

Oral tradition is a grab-bag, always open ended, with unrelated pieces able to be put in or taken out in any order. So, if Q material reached Matthew and Luke only in oral form, reconstruction of the original source is impossible. For it would not have been a unified work. Moreover, if Q was not written down before Matthew and Luke, it is obvious why no separate manuscripts have survived.

To say Q was an oral source sounds like the simplest solution to non-Markan parallels in Matthew and Luke. But it is not the best, since it does not account for close parallels in Matthew and Luke's outline that are not patterned on Mark. These are too complex to credit to oral memory or chance. Compare, for instance, the sequence of common blocks of material in Matthew's sermon on the mount with those in Luke's sermon on the plain:

Matt 5:3-7:27 Luke 6:20-49

Congratulations: Congratulations:

great reward in heaven! great reward in heaven!

Regrets if well off now.

* Salt sayings

* Light sayings

Torah kept, not abolished:

       no anger/ * reconciliation

       no lust / * cut off offender

    * no divorce

       no oaths

Love enemies

    no retaliation     no retaliation

    give without return     give without return

Love enemies

Golden rule

Be like Father Be like Father

Beware public piety of phonies:

     alms / prayer

   * Lord's prayer

     forgive

      fasting

* Treasure in heaven

* Eye & light

* Serve God not wealth

* Don't worry about:

    * food / clothing/ tomorrow

Don't judge Don't judge

     forgive

*give

    repaid measure     repaid measure

* Blind guides

* Disciple like teacher

     speck / log in eye      speck / log in eye

Avoid dogs / pigs

* Response to asking/ seeking

   Golden rule

* Narrow or wide gate

Beware false prophets

      fruit like plant / tree       fruit like plant / tree

      * product like store

Saying & doing Saying & doing

      parable of builders       parable of builders

Matthew's sermon is almost four times as long as Luke's (111 verses vs. 30) and covers a lot more topics. But the skeleton of both speeches (bold items) is the same. On the macro level of common clusters, only three brief sayings (italics) have been transposed. And only the golden rule has been shifted more than a few lines. Matthew and Luke each include sayings that the other omits from this speech. Yet all but one of Luke's additions and most of Matthew's are at least partly paralleled (*) somewhere else in the other gospel. The obvious conclusion is that this is the same speech, which either one or both writers have edited by inserting other material. Q retained much of the flexibility of oral tradition. But it also had enough stability for independent writers like Matthew and Luke to preserve the same basic patterns.

These patterns extend beyond the outline of individual speeches to the sequence of major non-Markan speeches in each gospel. Compare:

Matt Luke

Sermon (on Mount) 5:1-7:34 Sermon (on Plain) 6:20-49

Mission Instructions 10:1-16

On John the Baptist 11:4-19 On John the Baptist 7:22-35

Mission Instructions 10:1-16

Critique of Pharisees 23:1-39 Critique of Pharisees 11:37-54

Eschatological events 24:23-42 Eschatological events 17:22-37

Apart from the transposition of the second and third segments, Matthew and Luke share the same general outline of Jesus' speeches. Such structural similarity is especially striking, when one sees that these blocks of sayings have different narrative settings and internal details in each gospel. It indicates that both writers were using a relatively stable series of speeches.

This reduces the possibility that Q was only an oral source. A speaker can repeat the same series of speeches without cue cards, providing he delivers it often. But a member of his audience is not likely to deliver that string of performances without written notes. Two people are even less apt to duplicate the feat, especially if both did not personally attend the same lecture series. To claim that Q was only oral Jesus tradition, you must first imagine a setting in which Jesus trained disciples to repeat his speeches. Then you have to picture this first generation of performers apprenticing others to take the show on the road. And finally you must explain why neither Paul nor Mark echo this oral material, while two later writers do.

D6. Q preserved speeches in fixed sequences.

Written sayings collections.

The boundary between oral and written composition is not absolute. Words are primarily oral expression, whether they are spoken or written. And much memorable speech is eventually written down. Many written works retain features of material that was formed for oral circulation. This is true of Mark and even truer of most books in the Hebrew Scriptures, particularly those that record the sayings of Jewish sages and prophets. Q is primarily speech material and, so, was probably most like these.

Most Q passages are clusters of sayings ascribed to Jesus. But the current narrative frames have been molded by Matthew and Luke. Differences between them obscure any introduction in Q. Yet, it is clear from the speeches in both gospels that most of Q's sayings were simply strung together without any attempt to provide transition. Jesus is identified as the speaker only at the beginning of long clusters. This is a feature of written rather than oral composition.

Compare the book of Proverbs. Here one finds long lists of wise sayings that are traced to a particular sage only at five points. Ecclesiastes, like most wisdom literature, identifies the spokesman only once. The books of the Hebrew prophets are similar. Oracles spoken on different occasions are often separated by a stock formula, like "Thus says the LORD." But the book names the particular spokesman only at the head of the collection (e.g., Hosea and Micah) or at major seams (Isaiah, Amos, Habakkuk). The bulk of these works was formed orally. But the strings of unrelated proverbs and oracles without mention of setting or speaker are the product of a scribe.

A speaker does not need to introduce his own sayings. So all settings are secondary. But speech is not bound to one setting. A well-formed remark is remembered long after the moment that forged it has been forgotten. A saying has a life of its own. We do well if we can remember who originally said it. And if we are not sure, we tend to credit a good saying to someone whom we know said that sort of thing. Or else we simply use it, without ascription.

So in oral tradition Jesus sayings could survive only if clearly distinguished from the voice of the current speaker. Words can be orally credited to someone else only in small clusters, with the speaker specified. This is what we find in the gospel of Thomas, where individual sayings are punctuated by the monotonous formula: "Jesus said." Since Matthew and Luke quote long strings of Q sayings without ascription, they probably were dependent on the same written collection.

D7. Q was a collection of sayings without setting.
D8. Matthew and Luke knew Q in written form.

Bits and pieces.

Q was the written source of many Jesus sayings for Matthew and Luke. There are similar works ascribed to prophets and sages in the Hebrew scriptures and other ancient literature. And the discovery of the gospel of Thomas confirms that collections of Jesus sayings did circulate in early Christian communities.

But Matthew and Luke share more non-Markan material than just strings of sayings ascribed to Jesus. Other common passages include:

Pericope Matt Luke

(1) Oracles of John the Baptist 3:7-12 3:7-9,16-17

(2) Jesus tested by the devil 4:1-11 4:1-13

(3) Jesus heals a Roman centurion's slave 8:5-13 7:1-10

(4) John's disciples question Jesus 11:2-6 7:18-23

(5) Jesus challenges prospective disciples 8:19-22 9:57-60

(6) Jesus casts out a mute spirit 12:22 11:14

Passages (2)-(6) contain the only non-Markan narrative common to Matthew and Luke. So, they appear out of place in a collection of sayings with little or no hint of setting.

It also seems odd to find a work devoted to recording Jesus sayings opening with pronouncements by someone else (1). All four canonical gospels preface Jesus' public appearance with oracles of John. But each clearly separates the message of John from that of Jesus. Matthew and Luke, however, integrate these non-Markan sayings of John so well into their revisions of Mark's story of Jesus' baptism that it is not at all evident how they were introduced previously.

Also, apart from the recurrence of John in passages (1) and (4), there is no clear relation between these passages. At first glance, they are a hodge-podge of unrelated material: a few prophetic predictions (1), a mythic account of a hero's tests (2), a story of a Jew curing a gentile (3), a string of pronouncement stories about "following" (5), and a brief allusion to an exorcism (6). These hardly seem to have come from the same background, much less belong in Q.

One cannot rule out the possibility that Q had more stories that we can no longer recover. But there is no evidence that it had an extensive narrative framework. Matthew and Luke's dependence on Mark pretty much rules this out. Each follows Mark's story-line and revises it randomly. Their only extended agreement in wording or sequence of narrative is traceable to Mark. So, the source of Jesus sayings that both insert into different places in Mark's outline could hardly have had much narrative context of its own.

On the other hand, it is just as unlikely that Matthew and Luke got these six passages from some source other than Q. The fact that both writers use practically the same Greek wording and arrange these passages in almost the same sequence shows that their source was probably not oral. On the other hand, each item is so different that they certainly were not part of a second unknown written source.

Thus, these passages are best seen as isolated pieces of tradition that were inserted into Q at various stages in its development. As a compilation of sayings, Q could easily be expanded by insertions at almost any point. And the first item is, after all, a string of unsituated sayings. The fact that they are ascribed to John and not Jesus presents no problem. Many ancient sayings collections were expanded periodically by adding sayings from other sages or prophets (e.g., Proverbs and Isaiah). And the gospels of Mark and John also open with proclamations by John the Baptist.

Passages (2)-(5) are basically dialogues with minimal settings. And (6) introduces sayings about demons. So, like the bulk of Q, all are basically speech material. Dialogues are not uncommon in collections of pronouncements. The book of Job provides precedent for introducing the words of a Semitic sage with a dialogue between God and Satan. And several prophetic books insert occasional bits of dialogue and narrative between long strings of oracles (e.g., Isaiah 6-8, Amos 7, Hosea 1-3). Even the gospel of Thomas, which has no narrative interest, includes several dialogues of various lengths (e.g., Thom 6, 13, 61, 99) and three pronouncement stories (Thom 22, 60, 100). So, despite first impressions, none of the non-Markan passages common to Matthew and Luke is really out of place in a collection of Jesus sayings.

D9. Q had no sustained narrative.
D10. Q contained material other than Jesus sayings.
D11. Q resembled ancient wisdom and prophetic collections.
D12. Q was probably compiled in stages.

Overlaps and duplications.

The recognition of Q as a collection composed over a period of time helps to solve the puzzle about sections common to Matthew and Luke that overlap but go beyond Mark. The major passages are these:

Pericope Mark Matt Luke

(1) Preaching of John the Baptist 1:7-8 3:11-12 3:16-17

(2) Jesus tested by the devil 1:12-13 4:1-11 4:1-13

(3) Jesus accused of having a demon 3:22-30 9:32-34 12:10

12:24-32 11:15-23

(4) Mission instructions 6:7-11 10:5-15 9:1-6

10:4-15

(5) Jesus asked for a sign 8:10-13 16:1-4

12:38-42 11:29-32

In each passage Luke's wording is closer to Matthew than to Mark. And in (2)-(5) Matthew and Luke present many of the same sayings, not found in Mark. Such passages have prevented some scholars from recognizing the existence of Q. For they seem to contradict the general pattern of Matthew and Luke revising Mark independently. Here Matthew and Luke share material they could not have gotten from Mark. And their wording and logical structure is too similar to dismiss as a chance oral or editorial variation.

But these passages do not prove that either Mark or Luke knew Matthew. For the omissions and differences in Mark's version are too great to indicate his dependence on the text of Matthew. Luke's version shares more of Matthew's wording. But his arrangement of the sayings in all passages except (1) is not the same. And Luke locates (3)-(5) at different points than in Matthew's gospel.

The pattern of similarities and differences between Matthew and Luke indicates that they got these passages from Q. But the question remains as to where Mark got his versions. Some scholars have suggested that Mark may have known Q. But this creates the problem of explaining why he omitted so much Q material and gave a partial paraphrase of the rest. The greatest obstacle to claiming Mark used Q is that his version of Q passages is often simpler and, thus, probably more original.

The problem is solved, however, by recognizing Q and Mark as independent records of oral tradition. Both knew some of the same sayings, but not all. Mark's version is generally shorter or rougher and, therefore, closer to oral form. Q's clusters are more literary, being longer and often more polished. Matthew and Luke had access to both sources in written form and, so, tended to prefer Q's version of a passage in their attempts to polish Mark. Sometimes this meant replacing Mark's lines (1), other times supplementing them (2). And occasionally Matthew or Luke used both versions, especially with sayings recorded at different places in Mark and Q (3-5). This accounts for the fact that Matthew or Luke sometimes present two slightly different versions of the same saying. Beside the clusters listed above, these gospels contain many doublets of individual sayings. One version is clearly dependent on Mark, while the other is probably based on Q. This pattern of using Q material to improve or supplement passages in Mark shows that both Matthew and Luke regarded it as an important source of Jesus tradition.

D13. Q and Mark preserve variants of some of the same material.
D14. Q and Mark are independent witnesses to earlier oral tradition.
D15. Q is sometimes preferred to Mark.

Order and Edges.

The fact that Q and Mark sometimes present different versions of the same sayings yields a lot more bits of Q material than found simply by subtracting all Markan parallels from passages common to Matthew and Luke. But it also presents a major problem for trying to reconstruct the text of Q. For these sayings often appear in quite different contexts in Matthew and Luke. The question is: where did they originally belong in Q? Did Matthew or Luke rearrange this Q material? Or did both? This is not easy to answer on the basis of only two texts.

A closely related problem is the internal variation in clusters of sayings that Matthew and Luke have in common. Often the difference is simply a matter of sequence: one writer inverts the order of material in the other (for instance, in Jesus' dialogue with the devil). At other points, one includes material the other omits (see the comparison of Jesus' sermons, before D6 above). Here the question is: was material added to Q by one writer or omitted by the other? Or did both reshape Q by adding here, cutting there, and transposing sayings to emphasize different points?

Scholars generally grant that Matthew and Luke both remolded the contents of Q to fit their own needs. So, before Q could be reassembled the original order and extent of each Q saying had to be determined case by case. The difficulty of discovering the exact edges of the pieces of Q and how the whole jigsaw puzzle fit together prevented scholars from venturing to publish reconstructed texts until quite recently.

Fortunately, a general outline of Q is clear where Matthew and Luke present common clusters in the same sequence. Clarification of the fine points has been guided by research into the way in which Matthew and Luke edit Mark. For an editor probably treats two texts in much the same way.

Matthew generally follows Mark's wording and sequence but often inserts material with the same theme to supplement or clarify Mark's text. Sometimes he even rearranges sections of Mark to create larger thematic clusters (see comparison of outlines of parallels to Matt 7:28-9:26 above, before D2). This indicates that Matthew gave priority to Mark's order but an even greater priority to grouping material by topic. In some cases Matthew weaves Q sayings into Mark's outline (e.g., the parables of the mustard and leaven). In other cases Matthew presents a doublet: one version of a saying based on Mark and the other on Q (see the examples before D13). Since Matthew rearranged and supplemented Mark in this way, he probably did the same to Q.

Luke, on the other hand, is often closer to Mark's wording and sequence (see examples before D2 and D4). At several points, however, he omits passages from Mark. Some of these (the parable of the mustard, for instance) are found in Luke's only long digression from Mark's outline, classically dubbed the "great insertion" (Luke 9:51-18:14). This block of material intersperses strings of Q sayings with stories found only in Luke (e.g., the parables of the Good Samaritan and Prodigal Son). When Luke presents a Q parallel to a Markan saying in this section, he does not use Mark's wording. On the contrary, he often omits a Markan saying, if he has a parallel in Q. Moreover, in the great insertion he does not try to weave Q material into Mark's outline but presents it in a block. This pattern of editing indicates that, as much as Luke respected the text of Mark, he favored Q.

This has led current scholars to regard Luke's wording and sequence as generally closer to Q than Matthew. There are some exceptions, where Matthew's version of a passage is obviously more original. And there are other passages where it is equally obvious that both authors have reworked a Q saying, sometimes so much that the original form is unclear. In a few places scholars are still divided over whether a couple of verses belong to Q or not. But these are the types of problems that confront anyone who tries to reassemble fragments of any ancient text. Research and debate on Q is bound to go on. But enough consensus has been reached to present the public with a text that two gospel writers considered a prime source for the sayings of Jesus.

D16. Matthew and Luke adapt Q for their own use.
D17. Matthew tends to integrate Q with Mark.
D18. Luke tends to replace Mark with Q.
D19. Luke is generally closer to Q than Matthew.
D20. The complete text of Q is still uncertain.