In this paper we will look at what the Infancy Narratives in Matthew and Luke may tell us about the way these Gospels were composed. Specifically: We will look at an argument (described more fully here) that the two Infancy Narratives are so different that Matthew and Luke did not know each others’ Gospels.
This claim has broader implications for the way the Gospels were composed, because Matthew and Luke have about 235 verses that parallel each other but that do not have parallels in Mark or John.
We will call these 235 verses “the double tradition,” because it is found in two of the four Gospels.
If Matthew did not know Luke and Luke did not know Matthew, where did the material in the double tradition come from? It represents substantial amount of material that totals more than a fifth of both Gospels, which seems to be too much to attribute to random chance. The most likely answer, therefore, would be that both Matthew and Luke used a now-lost source that scholars have named “Q.”
(NOTE: See here for other parts of my exploration of the Synoptic Problem.)
To appreciate the force of this argument, let’s look at the kind of parallels that we find in the double tradition. It consists both of stories and sayings.
Here’s part of a story that both Gospels have a version of.
And the tempter came and said to him, “If you are the Son of God, command these stones to become loaves of bread.”
But he answered, “It is written, ‘Man shall not live by bread alone, but by every word that proceeds from the mouth of God’” (4:3-4).
The devil said to him, “If you are the Son of God, command this stone to become bread.”
And Jesus answered him, “It is written, ‘Man shall not live by bread alone’” (4:3-4).
Here is some sayings material that they each have a version of.
“Blessed are the poor in spirit, for theirs is the kingdom of heaven.
“Blessed are those who mourn, for they shall be comforted. . . .
“Blessed are those who hunger and thirst for righteousness, for they shall be satisfied. . . .
“Blessed are you when men revile you and persecute you and utter all kinds of evil against you falsely on my account” (5:3-4, 6, 11).
And he lifted up his eyes on his disciples, and said: “Blessed are you poor, for yours is the kingdom of God.
“Blessed are you that hunger now, for you shall be satisfied.
“Blessed are you that weep now, for you shall laugh.
“Blessed are you when men hate you, and when they exclude you and revile you, and cast out your name as evil, on account of the Son of man!” (6:20-22).
In both the story and the sayings material, the phrasing found in each Gospel is a bit different, but the material is in parallel on a verse-by-verse level—Matthew 4:3-4 corresponds directly with Luke 4:3-4, and Matthew 5:3-4, 6, and 11 parallel Luke 6:20-22.
Although there are differences in phrasing and order, it is generally possible to match up the double tradition material in this manner throughout Matthew and Luke.
The way the double tradition material can be paralleled verse-by-verse is striking, and it didn’t have to be that way. One Evangelist could have used the other as a source but so completely rewritten the material that such verse-by-verse parallels wouldn’t appear or would be much less common.
In fact, some might argue that this is what the Evangelist John did—that he took certain stories and sayings from the Synoptic tradition and wrote them in such a different manner that the connection is rarely obvious.
It has been claimed, for example, that his account of the healing of the official’s son (John 4:46-54) is a different telling of the healing of the centurion’s servant (Matt. 8:5-13, Luke 7:1-10; note that in Matt. 8:5 the centurion asks for the healing of his pais—which in Greek can mean either “boy” or “servant”).
Similarly, it has been argued that John’s discourses convey the teachings of Jesus in a paraphrased, literary way that makes specific verse-by-verse parallels to the Synoptics uncommon (though they do exist; e.g., Matt. 10:24, Luke 6:40, John 13:16, 15:20).
The fact verse-by-verse parallels appear in the double tradition, over and over through Matthew and Luke, indicates a form of authorial conservatism: Phrasing and order might be tweaked, but the material still clearly hangs together on the levels of verses and blocks of texts.
Wherever the double tradition came from, it was treated with significant conservatism by Matthew and/or Luke, and that could lead us to expect the same for how one author would treat the Infancy Narrative of the other..
No Verse-by Verse Parallels
The striking thing is that there are no verse-by-verse parallels in the Infancy Narratives—at least no obvious ones as in the previous section.
This can be seen by comparing the verses in which Matthew and Luke describe the one event they definitely both record—the birth of Jesus:
[B]ut [Joseph] knew her not until she had borne a son; and he called his name Jesus (1:25).
And she gave birth to her first-born son and wrapped him in swaddling cloths, and laid him in a manger, because there was no place for them in the inn (2:7).
These verses both describe the same event, but they relate it in very different ways that are utterly unlike the kind of parallels we find in the double tradition.
Lack of Parallels on the Pericope Level
The same thing is true when we compare the Infancy Narratives on the level of blocks of text, or what scholars call pericopes (per-IH-ko-PEES). While the material can be divided different ways, here is one way of looking at it.
MATTHEW’S INFANCY NARRATIVE (31 verses):
- Jesus’ birth announced to Joseph (1:18-25)
- The arrival of the magi (2:1-12)
- The flight to Egypt (2:13-15)
- The slaughter of the innocents (2:16-18)
- The return from Egypt (2:19-23)
LUKE’S INFANCY NARRATIVE (128 verses):
- John the Baptist’s birth announced to Zechariah (1:5-25)
- Jesus’ birth announced to Mary (1:26-38)
- Mary visits Elizabeth (1:39-56)
- The birth of John the Baptist (1:57-80)
- The birth of Jesus (2:1-7)
- The arrival of the shepherds (2:8-20)
- The circumcision and presentation in the temple (2:21-38)
- Return to Nazareth (2:39-40)
- The finding in the temple (2:41-52)
Again, the material is very different, and not just in matters of phrasing or organization. Though both narratives deal with the birth and childhood of Jesus, the topics covered in the two are strikingly different.
These lack of verse-by-verse parallels and the lack of pericope parallels suggest one of two things:
- Matthew and Luke didn’t know each others’ Gospels and wrote independently.
- One did know the other’s Gospel but chose to treat its Infancy Narrative very differently.
We may concede an initial advantage to the first hypothesis since, if one Gospel is dependent on the other, its author obviously thought highly of the work he had in front of him.
If Luke used Matthew then he thought highly enough of the material in Matthew to take a fifth of it into his own Gospel. We might expect him to do the same with Matthew’s Infancy Narrative.
Exactly the same would be true if Matthew used Luke: He used a fifth of Luke’s material, so we might expect him to do the same with Luke’s Infancy Narrative.
This initial advantage is far from insuperable, however. An author does not have to slavishly follow the same procedure in handling each part of his sources. It is perfectly possible for an author to see sufficient value in some parts of his source to include them but not enough value for his purposes to include other parts.
Indeed, we have a control case in Luke’s “Great Omission.” This is a section of Mark’s Gospel that runs approximately 75 verses, from Mark 6:47 to 8:27a. Although Luke borrows a great deal of material from elsewhere in Mark, he simply leaps over this section, apparently because he didn’t think it had sufficient value for his purposes.
This shows that Luke is quite capable of omitting large sections of his sources. In fact, at 75 verses, the Great Omission is more than twice as long as Matthew’s Infancy Narrative, which is only 31 verses. Luke was thus capable of omitting sections of his sources much longer than Matthew’s Infancy Narrative.
In view of this, we can overcome the initial advantage of the independence hypothesis if we can show that there are significant reasons why Matthew or Luke would have treated the other’s Infancy Narrative differently than the material in the double tradition.
Are there such reasons?
The Question of Length
One reason which is easy for moderns to miss entirely, or to dramatically undervalue, is the question of length. In the ancient world, books were amazingly expensive to produce.
There were multi-volume works, such as Tacitus’s Histories and Annals, which together comprised thirty books. However, only the rich could afford to author or own such collections.
As a result, epitomes (abridgments) were very popular in the ancient world. They allowed people to get the gist of a longer work without having to pay the staggering cost to own it. Because epitomes were so popular, they often survived the ages when the original, unabridged works did not.
A well known example is 2 Maccabees, which is an abridgement of a five-volume work by Jason of Cyrene (2 Macc. 2:23). The epitome of this larger work has survived and is in our Bibles today, but the original has perished.
This illustrates the price pressure on ancient authors to keep their works short. If you wanted only the rich to have your work, multi-volume collections were fine, but if you wanted a broader audience—which the Evangelists would have (see Richard Bauckham, ed., The Gospels for All Christians)—then you needed to keep your work to a single volume.
Indeed, there was even price pressure for single-volume works to be shorter rather than longer, since it cost more to author and copy longer ones. Authors of such works needed to find the right balance between content and length, delivering the highest value content for their purposes in the shortest space possible.
This is likely a factor in the popularity of the different Gospels in the ancient world. Using numbers given by Larry W. Hurtado (The Earliest Christian Artifacts, ch. 1), here are the four Gospels ranked from shortest to longest, with the number of surviving manuscripts from the second and third centuries, which is one of our best indicators of how popular they were at the time:
- Mark (1 copy)
- John (16 copies)
- Matthew (12 copies)
- Luke (7 copies)
Even allowing for randomness or “noise” in the number of the copies that have survived, Matthew and John—the Evangelists who wrote middling-size Gospels—seem to have found the sweet spot for the ancient audience, delivering the right combination of high value content and brevity.
Matthew (1071 verses) provided a broad and well-organized representation of the Synoptic tradition, being richer in content than Mark (661 verses with the shorter ending, 678 verses with the longer ending) and both briefer and less expensive than Luke (1151 verses). John (879 verses) was on the short side and provided a wealth of material not found in the Synoptics. It’s no surprise that these proved to be the most popular Gospels in the ancient world.
The full force of the length consideration isn’t felt until you try figuring out just how expensive authoring and copying such works was. While it is intrinsically difficult to do cross-cultural price comparisons, such efforts have been made.
For example, E. Randolph Richards estimates that it would have cost Paul around $2,275 to produce Romans and have one copy to mail and one to retain for his records (Paul and First-Century Letter Writing, 169). Romans contains 433 verses, and if we scale that up for the Gospels, we get these figures:
- Mark: $3,562
- John: $4,618
- Matthew: $5,627
- Luke: $6,047
The production prices would have been even more if (as is likely) the Evangelists had more than one initial copy of their work prepared for distribution, and the costs could have been multiple times the sums involved in making a single personal copy and a single copy for distribution.
In view of these prices, it’s easy to see the motivation the Evangelists had to keep their Gospels short—partly for the sake of their own pocket books but also for the sake of their readers. The longer they wrote, the fewer people would be able to afford their works and the fewer souls would benefit.
Length is likely the consideration responsible for Luke’s “Great Omission.” This is suggested by a look at its contents:
- Walking on the Water (6:45-52)
- People Flock to Jesus (6:53-56)
- The Hand-Washing Controversy (7:1-23)
- The Syro-Phoenician Woman (7:24-31)
- Healing a Deaf Man (7:32-37)
- Feeding the Four Thousand (8:1-9)
- Interpreting the Time (8:10-13)
- “Beware the Leaven” (8:14-21)
- Healing a Blind Man (8:22-26)
The material in this section is not particularly “low value” in and of itself, but it is largely material of the same kind we find elsewhere in Mark (and Luke).
When space is at a premium—and it would be especially for Luke as the author of the longest Gospel—one only needs so many accounts of healings, exorcisms, and multiplications of loaves. It’s easy to see how Luke could have reviewed this section of Mark and decided to skip forward since he was already planning on including parallels to much of this.
This gets us back to the question of how Matthew and Luke selected the material that they did include.
How Matthew and Luke Used Mark
If Mark wrote first then it’s clear that both Matthew and Luke used his Gospel to obtain their general outline. In a sense, they both start with Mark and then supplement it.
They do this in different ways, however. Ninety percent of the verses of Mark are paralleled in Matthew, but only fifty-five percent are paralleled in Luke (B. H. Streeter, The Four Gospels, 160).
Matthew thus had a stronger preference for using material from Mark than Luke did. Matthew’s default position was to include material from Mark unless there was a particular reason not to do so (as there apparently was in the case of ten percent of Markan material).
For Luke, there was a general preference to use material from Mark, but it wasn’t nearly as strong, as he was willing to let forty-five percent of the verses in Mark go without parallel.
How Matthew and Luke Used Their Source for the Double Tradition
It is sometimes argued that virtually all of the Q source must be preserved in Matthew and Luke since the original document is lost. If Q contained much material that wasn’t picked up by the Evangelists, why wasn’t it copied enough to survive?
This argument might be strengthened by an appeal to Matthew, who used ninety percent of Mark. If that’s how he handled Mark, wouldn’t he handle Q the same way?
There are easy rejoinders to this.
First, the idea that Matthew would have treated both his sources the same way is a weak assumption. He may have seen much more value in Mark than in Q and thus only preserved part of Q.
Second, there is the example of Luke, who used only fifty-five percent of Mark. If that’s how Luke treated Mark then we might expect him to treat Q in the same way. This is the flip side of the weak assumption that Matthew would have treated both sources the same.
Third, the only method we have of “identifying” Q material is the fact that it appears in both Matthew and Luke. It’s sheer speculation how the two authors would have treated a Q source, and without knowing how both of them would have treated it, we can’t infer anything with confidence about how much of it they would have used.
Fourth, the argument that if Q contained substantial additional material then it would have survived is weak.
Jesus ministered with his disciples for more than three years, and the Gospels taken together represent only a fraction of the things he said and did. This point is expressly made by John (hyperbolically) at the end of his Gospel:
But there are also many other things which Jesus did; were every one of them to be written, I suppose that the world itself could not contain the books that would be written (John 21:25).
Memory of the majority of things that Jesus did has perished, and we can’t assume that Q would be an exception to this. The vast majority of documents from the ancient world—Christian ones included—are now lost, and the fact that an individual one survived is the exception rather than the rule.
Finally, all of the above assumes that there even was a Q. But suppose there wasn’t? What would that tell us about how Matthew and Luke handled their source for the double tradition?
The answer is straightforward.
If Luke picked up the double tradition material from Matthew then, in addition to selecting a little more than half of Mark for inclusion in his Gospel, he also took 235 verses from Matthew that he thought fit his purposes well. Most of these were taken from Matthew’s large discourses, but since Luke (apparently, on this theory) has less patience for large discourses, he put them at other locations in his Gospel.
On the other hand, if Matthew picked up the double tradition material from Luke then, after making the basic decision to use as much of Mark as possible, he went through Luke and selected 235 verses that he thought were valuable enough for his purposes to include, while still keeping his Gospel a reasonable length. He then integrated most of these verses into his five large discourses.
In either case, one Evangelist selected 235 verses—or about a fifth—of the other Gospel for inclusion in his own. To put the matter another way, one Evangelist “cherry-picked” the other—in the positive sense of selecting the best items for his purposes (not the negative sense of suppressing things he disagreed with).
One can also look at this another way, which results in somewhat different ratios.
Matthew contains 470 verses that are not paralleled in Mark. If Luke used approximately 235 of those then he would have used fifty percent of what remained of Matthew when we take away the Markan material.
Similarly, Luke contains 785 verses that are not paralleled in Mark. If Matthew used approximately 235 of those then he would have used thirty percent of what remained of Luke when we take away the Markan material.
In both cases, the Evangelist would not have a default position in favor of using material from the other Gospel. In Luke’s case, it would be a fifty-fifty tossup as to whether he used material from Matthew, while in Matthew’s case there would be a seventy percent chance he would skip material from Luke.
Whichever way one looks at the cherry-picking, it has implications for our evaluation of how each would have treated the other’s Infancy Narrative.
The Formulas They Would Have Used
One implication is that we can see the formulas that the two Evangelists would have used in composing their Gospels:
- About 365 verses from Mark (55% of the total)
- About 235 verses from Matthew (20% of the total; 50% without Markan material)
- About 550 verses from other sources
- About 600 verses from Mark (90% of the total)
- About 235 verses from Luke (20% of the total; 30% without Markan material)
- About 230 verses from other sources
In both cases, the procedure would have been to produce a shortened version of Mark, supplemented by select material from other sources, one of which was the other Synoptic.
In view of the limited amount that would have been drawn from the other Synoptic, the numerical burden does not fall on the Q skeptic to show why the Evangelist omitted certain material.
The burden would fall on the Q skeptic if there was a bias in favor of including material, but there isn’t. In Luke’s it’s a tossup whether he would include a particular Matthean verse, and in Matthew the odds are that he would not include a particular Lukan verse.
Of course, this looks at the question from a numerical point of view rather than a content point of view. One could still argue that the content of a particular verse would be so compelling that the an Evangelist would have used it, but this has to be argued rather than assumed, and the above numbers indicate the freedom to skip material that both Evangelists would have felt.
(Note: One could argue with the numbers above if one could show that Matthew borrowed a significant amount of material from Luke even though the same material was also found in Mark, or that Luke borrowed a significant amount of material from Matthew even though it was also found in Mark. Determining which version of a verse an Evangelist used—the one found in Mark or the one found in the other Evangelist—would require a significant amount of work that I do not presently have leisure for. The results also would be quite debatable, and they would not change much, since the Evangelist would know that the material was found in both of his sources, making it somewhat arbitrary which version he used. He still would be using only fifty or thirty percent of the remaining verses.)
The Psychology of Cherry Picking
Today, when our knowledge of Jesus is filtered almost exclusively through the four canonical Gospels, every bit of Jesus tradition takes on added value.
Imagine how exciting it would be to have a new story or saying from Jesus that we knew for a fact was accurate. It would be mind blowing!
If we put ourselves in the position of one of the original Evangelists writing a Gospel, it’s easy to imagine that we would include every scrap of Jesus tradition we knew. How could we not? Forget cost and length considerations! To do otherwise would be to risk losing a Jesus tradition for future generations forever!
But the Evangelists were not in the position we are. They had access, orally or otherwise, to many Jesus traditions that have now perished, and—except for John—they may not have had an expectation that there would be future generations. They may have thought that the world would be ending soon and that the memory of the many unwritten things that Jesus said and did would be preserved until the end.
There was therefore less pressure on them to include every Jesus tradition they knew, and this made it possible for them to cherry pick their sources without the debilitating fear that we today would have of losing traditions.
This pressure was also lessened by the fact that later Evangelists knew what the earlier ones had written. They knew that the material was already “out there” in print—that those Jesus traditions had already been preserved in writing. They therefore had less of a psychological need to include every tradition they knew.
Furthermore, as the statement from the end of John’s Gospel intimates, there was a vast pool of Jesus traditions that was still preserved in living memory. The practical realities of book writing, and the corresponding realities of evangelization through books, meant that they had to be selective in what they included.
As Martin Hengel points out regarding Luke (in this case concerning Paul, but the same applies concerning Jesus):
[W]e cannot even claim without further ado, as is the habit of so many scholars today, that Luke only knew what he reported about the early period of Christianity. He certainly knew a good deal more than he put down; when he is silent about something, there are usually special reasons for it. Only by this strict limitation of his material can he ‘put his heroes in the right perspective’ (Earliest Christianity, 36, emphasis added).
The same was true regarding the other Evangelists: They all knew a good deal more than they wrote, and we should not assume that they didn’t know a tradition just because they didn’t record it. The better question is usually why they chose to include a tradition rather than why they chose to omit one.
The assumption that an Evangelist did not know a Jesus tradition just because he doesn’t mention it is absurd given the way the later Evangelists (Matthew, Luke, and John) treated Mark in the composition of their own Gospels. None of them—not even Matthew—preserves every Jesus tradition that Mark does, yet they each knew the Jesus traditions in Mark and deliberately omitted some, in greater or lesser degrees.
When we add to this the facts that there was an even broader pool of Jesus traditions to which the Evangelists had access, and that they were writing under strong pressure to keep their Gospels short, the assumption that silence implies ignorance is more absurd still.
This puts us in a position to look directly at the choices Matthew and Luke would have made regarding the Infancy Narratives.
Both Matthew and Luke wanted to include material about Jesus birth and early life, as is obvious from the fact they included Infancy Narratives. But are there reasons why they wouldn’t use extracts from each others’ narratives the way they would have the double tradition material?
There are, and we’ll look at them from the viewpoint of each Evangelist.
If Luke Used Matthew
Matthew’s Infancy Narrative is only 31 verses. If Luke had chosen to include them, his Gospel would have grown to 1182 verses, representing an expansion of under three percent.
That’s not a big expansion, but it’s also not nothing. Considerations of length could have played some role—but a minor one—in Luke’s decision to omit Matthew’s Infancy Narrative.
What if we consider the content of Matthew’s narrative?
Basically, it consists of two stories. The first deal’s with Joseph learning of Mary’s pregnancy and his reaction (1:18-25) and the second deals with the arrival of the magi and the series of events it sets in motion (2:1-23). This could make it somewhat difficult for Luke to excerpt Matthew without including the whole of one or both stories.
Faced with that choice, he presumably would not have a great deal of interest in recording the first story. Internal indications in Luke strongly suggest that Mary herself was one of his sources (either directly or at a close remove; see Luke 2:19, 51), and he was especially interested in presenting the traditions derived from her.
It could have been difficult to pull away and re-show the situation from Joseph’s perspective, particularly without disrupting the literary rhythms he was establishing with the parallels between John the Baptist’s birth and Jesus’ birth.
Also, given Joseph’s initial intention to divorce Mary (Matt. 1:19), including him in the narrative could cause him to appear in an undesirable, negative light due to comparisons with Zechariah, who initially did not believe (Luke 1:18-20, cf. 1:45).
Regarding the second story, much of it could not be easily excerpted—the flight to Egypt, the slaughter of the innocents, and the return from Egypt to Nazareth make no sense without a discussion of the magi.
Luke could have offered an abbreviated account of the magi’s visit without going into the events their arrival caused. Indeed, some have thought he should have done so given his interest in Gentiles. Robert H. Stein writes:
Why would Luke have omitted such material as the coming of the wise men (Matt. 2:1-12)? Would not the presence of such Gentiles at the birth of Jesus have been meaningful for Luke’s Gentile-oriented Gospel? (The Synoptic Problem, 102).
Stein misspeaks, because the magi did not come at Jesus’ birth. They came up to two years after his birth (Matt. 2:16), and that of itself could provide Luke with a disincentive to mention the visit. Given his interest in providing an orderly narrative (Luke 1:3), he would have needed to indicate a lengthy stay in Bethlehem, which may have been more chronology than he wanted to go into.
Further, he already had the story of the shepherds’ visit, and they were there the night of Jesus’ birth. This tradition presumably came from Mary herself, and Luke was keen to include the traditions he had from her. If he wanted to include that story, he may have considered the visit of the magi less important to record. He would have needed to indicate that the shepherds came and then, a year or two later, the magi arrived.
We have already seen how he recorded the Feeding of the Five Thousand (Luke 9:10-17) but he omits its sequel, the Feeding of the Four Thousand (Mark 8:1-9). In the same way he may have wished to include the initial visit of the shepherds but considered this sufficient to show the miraculous arrival of witnesses, without a need to include the event’s delayed sequel.
Finally, some have argued that Luke may have had other reasons to omit the account. Mark Goodacre writes:
Luke is the only writer other than Matthew in the New Testament to give us a hint of his view of the magi and it is negative—a certain Simon Magus is one of the villains in Acts of the Apostles (8:9-24). Moreover, at least since Conzelmann scholars have been sensitive to Luke’s apparent reticence to have Jesus coming into contact with Gentiles in the Gospel. One only has to witness the lengths to which Luke has gone to keep the Centurion out of Jesus’ sight to see the point (Matt 8:5-13 // Luke 7:1-11) (The Case Against Q, 56).
Personally, I’m more inclined to see Matthew as omitting mention of the centurion’s agents as a way of keeping his narrative of the event uncluttered, but there are still sufficient reasons why, if Luke had Matthew’s Gospel in front of him, he could have decided not to include the material in Matthew’s Infancy Narrative.
Now let’s look at the possibility that Matthew used Luke.
If Matthew Used Luke
The consideration of space would have weighed heavily on Matthew if he used Luke’s Gospel in producing his own. Luke’s infancy narrative is 128 verses long. For Matthew to include it would lengthen his Gospel to 1199 verses, making it the longest Gospel and increasing its volume by twelve percent!
If Matthew had Luke in front of him, he likely wanted to produce something shorter than Luke (since he did), and going even longer would be something he would resist.
Another way of looking at this is by the proportionate length of the Infancy Narratives. Matthew’s is 31 verses long, while Luke’s is 128 verses long. This means that Luke’s Infancy Narrative is more than four times as long as Matthew’s! It’s easy to see how Matthew might have wanted to keep his Infancy Narrative shorter and not devote a large fraction of his whole Gospel to it (as Luke did, with his Infancy Narrative amounting to eleven percent of his whole Gospel).
Further, in keeping with his fundamental choice to only include select material from Luke (twenty percent of it), it is easy to imagine him sticking with his default choice to omit Lukan material when it came to that Gospel’s Infancy Narrative and not lift pericopes from it.
This is particularly the case when we look at the content of Luke’s Infancy Narrative.
First, much of it is taken up with speeches, such as Gabriel’s announcement of John’s birth (1:13-17), Gabriel’s announcement of Jesus’ birth (1:28-33), Mary’s canticle (1:46-55), Zechariah’s canticle (1:68-79), the angels’ announcement to the shepherds (2:10-14), and Simeon’s speech (2:29-35).
Second, much of the material isn’t about Jesus’ birth at all but John the Baptist’s.
Third, the material about John the Baptist’s birth is interwoven with the material about Jesus’ birth in a way that would make it difficult to pull them apart. Much of Gabriel’s appearance to Mary and all Mary’s visit to Elizabeth only make sense if read in light of the John the Baptist birth narrative.
Fourth, Luke spends time narrating how Mary and Joseph did perfectly ordinary things for Jesus that any Jewish parents would do for their firstborn son (2:21-24).
Fifth, Luke relates minor incidents like the encounter with the prophetess Anna (who isn’t even quoted; 2:36-38) and the finding in the temple (2:41-51). As heartwarming as these are, they are not high-priority items, as illustrated by their omission by the other three Gospels.
If you pull out these elements, there is basically nothing left of Luke’s Infancy Narrative, so it is easy to see how a space-pressed Matthew could have looked at Luke 1 and 2 and decided to stick with his default decision to omit rather than include. He has his own traditions about Jesus’ birth that he wants to record, he can relate the important facts about Jesus birth (see the next section) without excerpting Luke, and he knows Luke’s traditions have already been preserved in writing.
Thus far we’ve been looking at the Infancy Narratives through the lens of what is different between them. If not balanced, this can lead to a false impression, because the two narratives also have multiple points in common.
In his book The Birth of the Messiah, Raymond E. Brown notes eleven points shared by the two narratives:
a) The parents to be are Mary and Joseph who are legally engaged or married, but have not yet come to live together or have [sic] sexual relations (Matt 1:18; Luke 1:27, 34).
b) Joseph is of Davidic descent (Matt 1:16, 20; Luke 1:27, 32; 2:4).
c) There is an angelic announcement of the forthcoming birth of the child (Matt 1:20–23; Luke 1:30–35).
d) The conception of the child by Mary is not through intercourse with her husband (Matt 1:20, 23, 25; Luke 1:34).
e) The conception is through the Holy Spirit (Matt 1:18, 20; Luke 1:35).
f) There is a directive from the angel that the child is to be named Jesus (Matt 1:21; Luke 1:31).
g) An angel states that Jesus is to be Savior (Matt 1:21; Luke 2:11).
h) The birth of the child takes place after the parents have come to live together (Matt 1:24–25; Luke 2:5–6).
i) The birth takes place at Bethlehem (Matt 2:1; Luke 2:4–6).
j) The birth is chronologically related to the reign (days) of Herod the Great (Matt 2:1; Luke 1:5).
k) The child is reared at Nazareth (Matt 2:23; Luke 2:39) (pp., 34-35).
What accounts for this material? In his book, Brown makes the following argument:
Since it is generally agreed among scholars that Matthew and Luke wrote independently of each other, without knowing the other’s work, agreement between the two infancy narratives would suggest the existence of a common infancy tradition earlier than either evangelist’s work—a tradition that would have a claim to greater antiquity and thus weigh on the plus side of the historical scale (p. 34).
Brown’s argument assumes that Matthew and Luke wrote independently of each other. Since that is what we are reconsidering, it’s logical to reject this premise and see what the results might be: If one Evangelist had the other’s Gospel in front of him, could that be responsible for these similarities?
It’s difficult to imagine Matthew or Luke being totally dependent on the other for his knowledge of traditions about Jesus’ birth. Such traditions were already out there in the Christian community, and they are reflected elsewhere in the New Testament. For example:
- Jesus is descended from David (Mark 10:47, John 7:42, Rom. 1:3, 2 Tim. 2:8, Rev. 5:5, 22:16, etc.).
- Jesus is from Bethlehem (John 7:42).
- Jesus is “of Nazareth” (Mark 1:9, John 1:45, Acts 2:22, etc.).
It’s difficult to imagine an individual well-informed enough and motivated enough to write a Gospel including an Infancy Narrative not to have done his own research into the question of what happened at Jesus’ birth. Therefore, even if one Evangelist used the other, it’s unlikely that he drew all of the common elements from the other.
It is more likely that each Evangelist knew some or all of the common elements from his own sources and that he included them because they communicated things he wanted his readers to know about Jesus.
However, even if both Evangelists had their own sources for each of the common elements, this does not mean that they worked with no knowledge of the other Evangelist. As Goodacre points out regarding the possibility that Luke knew Matthew:
[K]nowledge of a source is not the same as direct use of a source, and one of the key questions is whether there are any signs of Luke’s knowledge of Matthew in the Birth Narrative. After all, Luke may well have been inspired by Matthew’s account to write his own somewhat different account. If this possibility is taken seriously, the focus shifts away from the lack of extensive parallels between Matthew 1-2 and Luke 1-2 toward the more nuanced question of evidence for Luke’s knowledge of Matthew. In other words, rather than looking at the obvious points of divergence between the accounts, we might ask whether any of the points of contact are sufficiently marked to suggest that Luke may have known Matthew [op. cit., 56].
The same is true of the possibility that Matthew used Luke.
So: Are there indications that one Evangelist knew the other?
Indications of Knowledge?
Though it is not often appreciated, there are indeed signs that Luke knows Matthew’s Birth Narrative. Not only do they agree on matters unique to the two of them within the New Testament, like Jesus’ birth in Bethlehem, the name of Jesus’ father (Joseph) and, most importantly, the Virginal Conception, they even share words in common, including the following key sentence:
teksetai de huion kai kaleseis to onoma autou Iēsoun.
She will give birth to a son and you shall call him Jesus.
kai teksē huion kai kaleseis to onoma autou Iēsoun.
You will give birth to a son and you shall call him Jesus (op. cit., 56-57).
The initial items that Goodacre mentions could be explained other ways. The belief that the Messiah would be born in Bethlehem was widely held, and it even is mentioned in John (7:42), so this was out there in the Christian community. Similarly, anyone within living memory of Jesus’ birth would have been able to find out the names of his parents. And the Virgin Birth is so striking an event that it would have been widely noted in Christian circles.
What about the word-for-word passage that the two share in common? This is certainly not the only time that heaven has directed a child to be given a particular name. In fact, we saw the same thing earlier in Luke, when Gabriel told Zechariah what to name John the Baptist (Luke 1:13).
The same thing has precedents in the Old Testament (e.g., Is. 8:3, Hos. 1:4, 6, 9). Particularly notable are Genesis 16:11, 17:19 and Isaiah 7:14, which in the Septuagint read as follows:
su en gastri ekheis, kai teksē huion, kai kaleseis to onoma autou Ismaēl.
you are with child, and shall bear a son; you shall call his name Ishmael.
hē gunē sou teksetai soi huion kai kaleseis to onoma autou Isaak.
your wife shall bear you a son, and you shall call his name Isaac.
hē parthenos en gastri lēpsetai, kai teksetai huion, kai kaleseis to onoma autou Emmanouēl.
a young woman shall conceive and bear a son, and shall call his name Immanuel.
These are so similar to what we find in Matthew and Luke that it is reasonable to conclude, with Joseph A. Fitzmyer, that:
The message to Mary is couched in rather stereotyped OT phraseology for announcing the conception and birth of an extraordinary child (The Gospel According to Luke (1-9), 346).
Rather than evidence of one Evangelist borrowing this phrasing from the other, it is just as likely that they were borrowing from the Old Testament.
That’s particularly the case with Matthew, who in the next two verses indicates the origin of the angel’s phraseology, stating that the angel’s message was a fulfillment of Isaiah 7:14 (“All this took place to fulfill what the Lord had spoken by the prophet: “Behold, a virgin shall conceive and bear a son, and his name shall be called Emmanuel,” Matt. 1:22-23).
While Luke could have been influenced by Matthew to use this kind of phraseology in his Gospel, the phraseology itself is too common for this to be relied upon. Luke easily could have written independently of Matthew and come up with the same phrasing from the Old Testament parallels.
If the above parallels between the Infancy Narratives are not persuasive, do any exist that are?
I think so. Brown notes that some of the common elements in the Infancy Narratives appear in different forms:
For example, while both Gospels have Jesus’ birth announced by angels, in Matthew the angel speaks to Joseph but in Luke the angel speaks to Mary (op. cit., 34).
This is not the only element of its kind. We may note several elements Brown does not record that mirror each other, in addition to the initial one:
1) Angels speak to both of Jesus’ parents—Joseph in Matthew and Mary in Luke (Matt 1:20–23; Luke 1:30–35).
2) The birth of Jesus is attended by celestial phenomena—a star in Matthew and a host of angels in Luke (Matt 2:2, 7, 9, 10; Luke 2:9-15).
3) These celestial phenomena were observed by others, who were motivated to visit the child and his parents (Matt 2:1-12; Luke 2:15-20).
4) The child’s visitors were of different social statuses (shepherds being of low education and rank and magi being of high education and rank).
5) The child’s visitors were of different ethnicities (the shepherds being Jews and the magi being Gentiles).
Stepping outside the narrow bounds of the Infancy Narratives, we also may also add:
6) Both Gospels include genealogies of Jesus but they are strikingly different in multiple respects (see below).
The way Matthew and Luke mirror each other on these points suggests that one was writing in response to the other. The question is: Why?
One reason might be supplemental intent—that is, one Evangelist knew the other had preserved one set of traditions in writing, and he wanted to preserve additional ones. This kind of intent is demonstrable elsewhere in the Gospels, as when John intentionally supplements Mark (see Richard Bauckham, “John for Readers of Mark,” The Gospels for All Christians; see also here).
However, the way the elements mirror each other suggests that more than just supplemental intent was at work. It has long been noted that:
- Matthew’s narrative focuses almost exclusively on Joseph, while Luke’s focuses almost exclusively on Mary
- Matthew accentuates Jesus’ regal dimension (his genealogy records Jesus’ descent from Solomon and the line of kings that followed him, King Herod being threatened by Jesus’ birth, and the visit of foreign dignitaries seeking to honor the new king) while Luke presents Jesus as a man of the common people (his genealogy records Jesus’ descent from Nathan, Mary praising God for his deeds on behalf of the lowly, and the visit of humble shepherds)
These are significant clues to why one Evangelist may have wanted to respond to the other. The question is: Who was responding to whom?
If Luke was responding to Matthew then he may have found the latter’s emphasis on Joseph and Jesus’ regal dimension not fully to his taste. He then balanced it by using the traditions he had regarding Mary and by bringing out the dimension of God’s compassion through Jesus on the lowly.
If Matthew was responding to Luke then he may have felt that Luke omitted information and themes which would have been important for his audience of Jewish Christians. He may have felt that Luke’s overwhelming emphasis on Mary and his populist themes needed to be balanced for a Jewish audience with an emphasis on Joseph, through whom Jesus would have had legal claim to the Davidic monarchy. He similarly may have felt that the regal aspect of Jesus needed further emphasis, and the traditions he had at his disposal allowed him to accomplish both of these goals.
A Word About the Genealogies
Having mentioned the genealogies of Jesus (Matt. 1:1-17, Luke 3:23-38), it is appropriate to say a few words about them, though they are only ambiguously grouped with the Infancy Narratives. (Matthew’s genealogy could be conceived of either as separate or as part of his Infancy Narrative, while Luke’s is found outside his Infancy Narrative, in his account of Jesus’ ministry.)
Given the well-known differences between these genealogies, including the fact that they trace Jesus’ descent through different lines, many have seen the two as evidence of the independence of Matthew and Luke. Thus Stein writes:
[I]f Luke had before him Matthew’s birth account and genealogy, one wonders if he would not have sought in some way to ‘harmonize’ the one we have in his Gospel with the Matthean version (The Synoptic Problem, 102, emphasis added).
Once again, there are plausible reasons why one Evangelist would choose to include a different genealogy than the one he saw the other using.
If Luke knew Matthew’s Gospel then several things may have leapt out at him regarding its genealogy: (1) It only goes back to Abraham (Matt. 1:1, 17), (2) it omits multiple generations in order to fit a scheme of three, fourteen-generation blocks (Matt. 1:17), (3) it’s right up at the front of the Gospel (Matt. 1:1-17), and (4) it shows Jesus descending from David through Solomon and the line of kings down to Jeconiah (Matt. 1:6-12).
Luke thus may have chosen to include his genealogy to balance each of these: Thus (1) he took his genealogy all the way back to Adam, to make explicit the parallels between Jesus and Adam as sons of God in unique ways (Luke 3:38; cf. Rom. 5:14, 1 Cor. 15:22, 45, 47), (2) he included a fuller list of the generations that is not compressed the way Matthew’s is (though it may be seen as eleven blocks of seven generations; see Richard Bauckham, Jude and the Relatives of Jesus in Early Christianity, 318-321), (3) he placed his genealogy later in the Gospel so that it would not provide the abrupt, contextless start for his Gentile readers that Matthew’s placement of the genealogy right at the front of his Gospel would have, and (4) he recorded Jesus’ descent from David through his son Nathan (Luke 3:31), thus avoiding the line of kings terminating in Jeconiah.
The last deserves special comment. Jeremiah had pronounced a curse upon Jeconiah (aka Coniah, Jehoiachin), indicating that his sons would not be king after him:
As I live, says the Lord, though Coniah the son of Jehoiakim, king of Judah, were the signet ring on my right hand, yet I would tear you off. . . . Thus says the Lord: ‘Write this man down as childless, a man who shall not succeed in his days; for none of his offspring shall succeed in sitting on the throne of David, and ruling again in Judah’” (Jer. 22:24, 30).
Because of the flexibility of the Old Testament concept of “son,” it could be questioned whether the prophecy applied to Jeconiah’s immediate sons or to all of his male descendants, in which case none of them would have a claim to being the Messianic son of David (at least not due to their descent from Jeconiah).
Whether the Messiah could be a son of Jeconiah is disputed in Judaism today, and it may well have been in Jesus’ day.
If so, Luke might have included his genealogy to make it clear that Jesus’ claim as Messiah did not rest merely on his descent from David through Jeconiah; he had a claim to being a son of David and thus a potential candidate for Messiah apart from this.
Or the problem may not have been just Jeconiah, but the entire line of kings from Solomon to David. Bauckham writes:
[I]n the Old Testament prophetic tradition, which both condemned the kings of Judah and expected a renewal of the Davidic monarchy, under a righteous king in the future, the dominant expectation was for a new Davidic king who was not descended from David through the royal line of the kings of Judah. This expectation is classically embodied in Isaiah 11:1: ‘There shall come forth a shoot (ḥōṭer) from the stump of Jesse, and a branch (nēṣer) shall grow out of his roots’ (RSV). The image is of a tree chopped down to a stump. A new shoot grows up from the roots (see Job 14:7–9 for the image). The natural meaning is that the tree of the royal house of David will be cut down in judgment, and the ideal king of the future will be derived, not from the royal line of the kings of Judah, but from the origins of the dynasty, indicated by the reference to Jesse. He will represent, as it were, a fresh start, taken, like David himself, from non-royal stock. If he is a descendant of David at all, then he will have to come of a line of David’s descendants other than the royal line through Solomon and the kings of Judah.
That this is the correct interpretation of Isaiah 11:1 is confirmed by the similar implication of Micah 5:2 (Hebrew 5:1):
But you, O Bethlehem Ephrathah,
who are little to be among the clans of Judah,
from you shall come forth for me
one who is to be ruler in Israel,
whose origin is from of old, from ancient days (RSV).
The new king is to be born not in the royal palace in Jerusalem, but in insignificant Bethlehem, where David’s line began. He will derive not from the royal line of the kings of Judah, but from the ancient origins of the line, from the beginnings of David’s dynasty. Again there is doubtless the intention of going back behind the corruption of the kings of Judah and making a fresh start, comparable with God’s original choice of David himself (Jude and the Relatives of Jesus in Early Christianity, 334-335).
In fact, there appears to have been a tradition that the Messiah would be a descendant of Nathan in particular (Bauckham, op. cit., 347-354).
For each of these reasons, if Luke had Matthew’s Gospel in front of him, he may have been prompted to include his own genealogy, making the descent from Nathan rather than the line of kings ending in Jeconiah clear.
On the other hand, if Matthew had Luke’s Gospel in front of him, several things also would have jumped out about its genealogy: (1) It’s in a very non-traditional, reverse order, (2) it ends in Adam, (3) it’s in a very unexpected place, and (4) it skips entirely the line of kings after David.
Matthew then would have included his own genealogy to balance these: (1) A reverse-order genealogy was extremely unusual for a Jewish genealogy, and Matthew may well have wanted to give his Jewish Christian readers a more standard presentation of the Messiah’s lineage, (2) he may have wanted to relate the Messiah more clearly to the people of Israel and its great historic events (Abraham, David, the Babylonian Exile), compared to the universalist, Lukan genealogy linking the Messiah to the dawn of the whole human race, (3) he may have wanted to put his genealogy of Jesus before his account of the birth, which better reflects the placement of genealogies in the Old Testament and which avoids Luke’s highly unusual placement of Jesus’ genealogy after his baptism, and, finally, (4) for those unfamiliar with or unconvinced by the prophetic interpretations above (a group that may, in fact, have been a majority among ordinary people; “Of course the Messiah is a descendant of the line of Davidic kings! He’s the royal Son of David!”), Matthew may have wanted to make it clear that Jesus did have a claim to being the Messiah via descent through Solomon and the line of Davidic kings.
The fact the two genealogies trace Jesus’ descent from different sons of David is likely explained by ambiguity in Jesus’ day regarding precisely how the Messiah would be descended from David. Indeed, the fact that people had different opinions about this is likely why Jesus’ family (among others) preserved the memory of its descent through both lines—and why the Evangelists felt the need to present both to their audiences.
In view of each of the factors listed above, for both Evangelists the point deliberately would have been not to present the lineage of the Messiah in the same way as the Gospel he had in front of him but to present it in a different way.
A final indication that Matthew and Luke were not writing independently is that they both came up with such similar overall designs for their Gospels. This goes beyond the Infancy Narratives and the genealogies, but it also includes them and so is relevant here.
Both Evangelists saw a promising foundation in Mark, but they wanted to expand it in order to reach particular audiences. The fact that they both expanded it in the same way suggests that one may have been prompted by the work done by the other.
One of the expansions they made was to include post-Resurrection narratives that went beyond the shorter ending of Mark. If Mark originally ended without such appearances or if its original ending had already been lost, then it is easy to understand why they did so. This would be a natural expansion that their audiences would have wanted—as illustrated by the fact that post-Resurrection appearances are also found in John (20:11-21:23), in the longer ending of Mark (16:9-19), and even outside the Gospels in Paul (1 Cor. 15:5-8).
What’s more significant is the fact that they both included Infancy Narratives, and narratives of the kind they did. Considering the possibility that Luke used Matthew, Goodacre writes:
The theory that Luke could not have known Matthew because he does not copy wholesale from his Birth Narrative is not, therefore, especially convincing. Indeed like many arguments for Q, reflection on the evidence can lead in quite the opposite direction, in favor of Luke’s familiarity with Matthew. Perhaps Matthew’s Birth Narrative gave Luke the idea of writing a Birth Narrative of his own; perhaps it was the catalyst for Luke’s identical decision to preface Mark’s Gospel with an account featuring both prenatal (Matt 1 // Luke 1) and postnatal (Matt 2 // Luke 2) stories about Jesus. Because many readers are so familiar with the Birth Narratives, it is easy to assume that prefacing a Gospel with a Birth Narrative is a natural step to take, but neither Mark nor John thought that it was such an obvious thing to do and, all things considered, the presence of a Birth Narrative in Luke is probably a sign that Luke knows Matthew (op. cit., 57).
Or it is a sign that Matthew knew Luke.
In the same way, one Evangelist may have prompted the other to include a genealogy—something no other author of the New Testament chose to do.
The overall design of Matthew and Luke—the fact that they decided to expand on Mark in such similar ways—can thus be seen as further evidence that they were not writing independently.
The argument that the differences in the Infancy Narratives of Matthew and Luke show that they were writing independently of each other is utterly unconvincing.
It rests on the premises that the two Evangelists must have been ignorant of what they did not mention or that one would have found quoting material from the other’s Infancy Narrative irresistible.
Both of these premises are false. As their handling of Mark reveals, Matthew and Luke demonstrably left out Jesus traditions that they were aware of, and there are sound reasons why both could have chosen to omit the material found in the other’s Infancy Narrative. Chief among these reasons are the then-pressing need to save space (particularly for Matthew) and the need to serve the respective Jewish and Gentile audiences they were trying to reach.
Indeed, serving the needs of these audiences is likely the reason why multiple elements of the Infancy Narratives mirror each other, which would not be expected if the accounts were independent. This applies also to the twin genealogies of Jesus, whose inclusion in the New Testament is otherwise very perplexing.
These considerations—as well as the fact that they both chose to compose Gospels that expanded Mark using the same overall design—provide a compelling alternative to the Q hypothesis that must be taken seriously.