Prolegomena to Proto-Luke

E Bruce Brooks, University of Massachusetts at Amherst
SBL, San Diego, 17 November 2007


It has several times been suggested that the Gospel of Luke was composed, not in one stage, but in two. Reconstructions of the first stage, usually called Proto-Luke, have varied considerably. Given that variety, I take the matter as open for discussion, and I will approach it de novo, using only basic philological methods as I understand them. I will also rely on what I call trajectories, lines of development which are likely to occur in the growth of any religious tradition, or any tradition at all. One of the constants in tradition growth is aggrandization, which tends to increase respect for the founder, his family, or his close associates. Take the case of Mary. In Mark, she and Jesus's brothers try to restrain him. He refuses to see them, and identifies as his family those who believe in him. Things are quite different in Matthew, where Jesus is miraculously conceived, and still more so in Luke, where John the Baptist is also miraculously conceived, and recognizes Jesus as the Messiah, not on the banks of the Jordan, but from the womb. These are the kinds of development that we may expect to find in the history of any movement. It would be extraordinary if they went in the direction Luke > Matthew > Mark. The probability is all the other way. This is my reason for assuming, in what follows, that among the Synoptics, Mark precedes Luke, and that Matthew comes in between.

Thus we might argue if Luke was written at one time. The possibility that it was written in more than one stage requires that we consider the matter afresh. To that consideration I now proceed.


I will begin by focusing on several places where Lukan segments with parallels in Mark seem to be out of place in our present Lukan narrative. Hypothesis #1 will be that this material originally stood in Markan order, and that the original order was later changed. It is these two stages, the Markan-order Luke and the changed-order Luke, that I propose to distinguish as Proto- and Deutero- Luke.

Fitzmyer (1/73) notes seven places where a Lukan unit is out of order with respect to Mark. Some of them are shown on Handout §1. Fitzmyer also suggests reasons why Luke might have preferred a different order. Bultmann explains other differences in the same way. Very good. But there are two possibilities: either Luke rearranged his material in the act of selecting it from Mark, or else he first put it in Markan order and later changed it. I believe we can decide between these possibilities. Consider these examples:

1: The Preaching at Nazareth

In Lk 4:16-37, Jesus preaches at Nazareth, and people ask why he did not do miracles of the kind he had earlier done at Capernaum. But at that point in Luke's story, the Capernaum miracles have not yet been narrated. We might say, Well, that detail was in the story as Luke got it from Mark. But there is no such feature in Mk 6:2, where the audience do not mention Capernaum. Or perhaps, we might say, Luke expects his hearers to know of Jesus's miracles in Capernaum without being told, and does not feel a need to introduce them to Capernaum. That possibility can be tested. When Luke's narrative first comes to Capernaum, in Lk 4:31, what does it do? Answer: It locates Capernaum: "And he went down to Capernaum, a city of Galilee." I think we must take Luke's word for it, that his audience needed to be narratively introduced to Capernaum. If so, then the second explanation fails. I cannot think of a third. I conclude that Lk 4:16-30 is out of place. It needed to follow, at whatever distance, something like the end of the Capernaum miracle narrative, which concludes, "And reports of him went out into every place in the surrounding region" (Lk 4:37). It is exactly this which would make the present Lukan Nazareth story work right. As that story now stands, at least in this respect it does not work. It was written for a different position in the Lukan narrative than the one it presently occupies. This requires the assumption of a two-stage composition process within Luke

2: The Calling of Simon

In Lk 4:31-37, Jesus teaches and heals at Capernaum, and afterward, at 4:38, enters "Simon's house." It seems we should know who Simon is. But who is he? He is introduced later, when he is called to be Jesus's disciple. This occurs in Lk 5:3, "Getting into one of the boats, which was Simon's, . . ." This too is abrupt, but it is consecutive. The call to Simon follows. The visit to Simon's house would read easier if the call to discipleship had come before it, not after it, so that Jesus might plausibly go to his disciple Simon's house after preaching in the synagogue. As it is, Jesus functions in Lk 4:31-37 as something of an interloper, who goes in the house without an invitation, and begins healing people.

We now have three units, which in Markan order are the Calling of Simon, the Preaching at Capernaum, and the Preaching at Nazareth. (see Handout §1). In Luke, the order is the exact reverse. We have seen that certain difficulties in the narrative arise from the Lukan order. They do not arise in the corresponding narrative in Mark. They are a problem that has been created at some point within the process which resulted in our Luke.

3. Healing of the Leper

In Mk 1:39, Jesus leaves for the "next towns," and goes "throughout all Galilee" preaching and healing. In Mk 1:40 there follows the story of the healing of a leper. No place is mentioned, but there is no trouble if we imagine the story as taking place somewhere in Galilee, and next, indeed, we have in Mark 2:1 "and when he returned to Capernaum after some days." All is geographically consecutive. In the corresponding Lk 4:43, Jesus departs for "the other cities," and the next line informs us that they were not Galilean cities: Jesus "was preaching in the synagogues of Judea." At some distance there follows the story of the Leper, Lk 5:12-16, and in it, Galilee is mentioned as though it were a different place, consistent with the idea of a Judean setting. The story indeed begins, "While he was in one of the cities," picking up the same word as in Lk 4:43, and implying a Judean setting. So far so good: The continuity is Judean, but it holds. But between these two places there now stands the Calling of Simon, and that explicitly takes place in Galilee. The Lukan geography is therefore flawed. The problem is not in the Leper story itself, but in the fact that something comes between it and its seeming introduction. That something is the Calling of Simon, and here is another sign that the Calling of Simon as it stands is out of sequence. There would be no problem if Luke's story units were consistently arranged in Mark's order.

If we rearrange the Lukan units, as now written, to bring them into Markan order, we will eliminate the difficulties here noted. Will new ones arise? That experiment is carried out for the Lk 4-5 part of the material, and it is successful. We note in passing that Luke has introduced the Calling of Simon with a preaching from a boat segment, which he probably borrowed from Mk 4:1. Accepting that as part of the Call to Simon, and letting Mark determine the sequence otherwise, we get Handout §2. It will be seen that no problems arise. Luke has fine-tuned the Markan narrative in ways that increase the reverence shown to Jesus and the status accorded to Peter, and has somewhat diminished the prominence of Galilee in Jesus's career, but he has not gotten himself into narrative trouble.

The first test of the hypothesis has thus been successful. Now for a second. What happens if we relocate the Nazareth episode back to the position in Luke which it would have had if Luke had been following Markan order? In Handout §3 that experiment is tried on the relevant portions of Lk 8-9, with the result that here also, no problems arise. The Nazareth episode fits smoothly back into the place we had conjectured for it in Proto-Luke. The conjecture is to that extent confirmed.


Hypothesis #1, at least in these parts of Luke, thus gives us something that is plausible for Luke at one point to have written: an imaginable Proto-Luke. What could have been his motive for changing it?

It will be evident that Proto-Luke is constantly concerned to trim Mark of vivid but superfluous details, and more importantly, to increase the reverence with which the people in the narrative treat Jesus. We may note that in Mark the response of Simon to his calling by Jesus is simply, to come. We are left to suppose, and the commentators do suppose, that he recognized something compelling in Jesus, but it is not clear to the reader what that was. If instead Simon was among the spectators at Jesus's preaching and exorcism in Capernaum, he will be better prepared to regard Jesus as something out of the ordinary. Moving the Calling of Simon after the Preaching at Capernaum provides that preparation. His response becomes more intelligible psychologically.

The Nazareth case brings in larger issues. In Mark, Nazareth is the only story where the people generally, and not the Pharisees or the Temple authorities, challenge Jesus. The large movement of Jesus's career in Mark is from an initial success with the people, to a later death at the hands of the authorities. By moving the one scene of popular opposition to a prominent place at the beginning of Jesus's career, Luke casts a shadow over Jesus's popular success, and produces a more consistently oppositional picture. Would this have suited an agenda he is likely to have held? Yes; the associated work Acts ends (Ac 28:28) with an explicit statement that the Jews have definitively rejected Jesus, and God will take his message of salvation instead to the Gentiles; "They will listen." If this new Acts message conflicted with that of the previously written Gospel, Luke would be in trouble. By moving the Nazareth unit to the head of his Gospel, Luke reduces the trouble, and Luke-Acts as a whole now portrays the working out of the theme of Jewish rejection, rather than introducing it as a new idea halfway through. It might, then, have been to accommodate the message of Acts that Luke has moved his Nazareth scene to the place where we now find it in his Gospel.

Can we find confirmation for this possibility? Yes, in the long passage in that scene which has no counterpart in Mark, and which is also unmotivated in the Nazareth story itself, Jesus mentions the notion that people might not believe in him because they have known him since childhood; this is the Markan theme slightly transposed. But then Jesus goes on to say, with citations from Scripture, that God will show his favor, not to Israel, but to other peoples, including the Syrians. This enrages his hearers, and they attempt to kill him, thus prefiguring the enraged mob to whom Pilate appeals, when Jesus actually is killed. All these resonances have been noted by the commentators, among them Creed (1960).

I now take the next step, by suggesting that the moving of Nazareth, and the insertion of this Rejection of Israel theme into it, were done at the same time, and also at the time when Acts was added to what was previously the Gospel of Luke. If we had to deal only with the Calling of Peter, a sufficient reason for Deutero-Luke would be to improve on the motivations of characters such as Simon. But the Nazareth case requires a different solution. If all of Luke's relocations were made at one time, and of that we cannot be sure, it was probably the time when Luke's original Gospel was extended by the addition of Acts.


Hypothesis #3 is that at one point Luke came under the influence of Matthew. When might that have been, in the Proto/Deutero-Luke sequence? Proto-Luke seems, so far, to be discussible in purely Markan terms. What about Matthew?

I noted above that canonical Luke must follow canonical Matthew because some Lukan passages are later, in developmental or trajectory terms, than the corresponding parts of Matthew. The question is: how firmly situated are those passages in Luke?

Luke's Birth and Infancy narratives are clearly meant to be improvements over those in Matthew, and the same is true of the universalist Lukan genealogy, as distinct from the merely Davidic genealogy in Matthew. Luke wants to give Jesus to the world. Also, and some of us with rivals in our own fields will understand this, he wants to show old Matthew, "This, my learned friend, is the way to write a Gospel." Certain otherwise puzzling departures from Matthew in Luke, like the notorious case of the Sermon on the Mount, are easily explained in this way. Luke may not always achieve a superior effect, but he is determined at least not to acknowledge his rival by copying the same effect.

If then the Lukan Birth and Infancy narratives were written with knowledge of Matthew's attempts in that direction, when were they added to Luke?

To start with, it can easily be shown that they are not original. Thus, if we eliminate the Birth and Infancy narratives (Lk 1-2) from Luke, we find that the Luke begins, at 3:1, with a perfectly satisfactory introduction:

In the fifteenth year of the reign of Tiberius Caesar, Pontius Pilate being governor of Judea, and Herod being tetrarch of Galilee . . . the word of God came to John the son of Zechariah in the wilderness, . . .

So might Proto-Luke have once begun. Note that it makes a point of telling us who John was, and thus ignores the intimate introduction given in our present Lk 1. Nor later, when Jesus comes on the scene, is there any sign that John knows who he is. Jesus just emerges from the crowd; the Baptism is not even described. Only later (Lk 7:18-22), does John send to ask if Jesus might be the Messiah. I take it as highly probable that the Lukan Birth narrative, which already answers this question, was not yet in place when these passages were written.

Later in that opening scene, however, we find clear signs of influence from Matthew. The Markan version passes quickly over John's preaching, but Matthew, followed as it seems by Luke, gives it in detail, including the unforgettable "brood of vipers" remark. We also find here at least a hint of what I have identified as the Deutero-Luke agenda, the Rejection of Israel. Thus Mt 3:8:

Bear fruits that befit repentance, and do not presume to say to yourselves, We have Abraham as our father, for I tell you, God is able from these stones to raise up children to Abraham. [Lk 3:8 is virtually identical]

If from the stones, then God should have no trouble finding among the Gentiles those worthy to inherit his ancient promise. If Luke got this passage from Matthew, and seemingly he did, he may well have picked up the Rejection of Israel theme at the same time. This would suggest that all the changes we have been describing may have been part of a single event; the revision of a modest early Gospel attempt into something more like our Gospel of Luke.


It is very convenient for us when ancient authors work with well-defined units: adding them, taking them out, moving them around. And sometimes they do. But they may also fix up the margins a bit, or rewrite to fit the final context. This is harder to detect. Sometimes Luke rewrites; see Handout §4. One adjustment is minor: the change in Lk 4:42 to account for the fact that after the Calling of Simon has been moved further down in Luke's narrative, it cannot be the disciples who seek Jesus early in the morning, so they are replaced by "the people." The other is major: Jesus's rejection of Israel and the rage of the Nazareth crowd. I have pointed out that this is an insertion. But if we remove it, we do not get the original story, we get a story with no ending at all. That particular rewrite has replaced any previous text. So it goes; not all the problems are easy, and not all the changes can be completely restored.


The conclusion is that we have a Luke in two stages. They are summarized as a tentative hypothesis in §5 of the Handout.


So far my exposition. The questions that might arise are so obvious that I will ask the first few myself.

Q1. Very nice, but why should I adopt your model in place of the one I have been using. Will it do things that mine won't?
A1. Yes. It makes clear several things which previous commentators have left somewhat in a mist. Christopher Tuckett in 1996 asked, Is Luke for or against the Gentile Mission? He notes the message of Acts, but he goes on to point out that Luke in his Gospel very rarely hints at a Gentile Mission. Then he lists the exceptions, as follows:

1. Simeon in the Temple foretells "a light to lighten the Gentiles." This we have attributed to Deutero-Luke.

2. The opening quote from Isaiah is extended to include "All flesh shall see the salvation of God." I didn't mention it, but this is another Lukan transposition of Mark: Luke puts the quote
second and John first, and then he gives both of them greater sonority by introducing them this way: "And the word of God came to John the son of Zechariah in the wilderness." Again a change
in Markan order is accompanied by the Gentile motif.

3. The Nazareth scene. It is already on our Deutero-Luke list.

I think we have a pattern here. Tuckett's "exceptions" all belong to Deutero-Luke.

Q2. I'm not buying this until you show that it applies to more than the Lk 4-5 segment you have given us on the Handout.
A2. I already mentioned Lk 8-9 as amenanble to this interpretation. Reading further in Tuckett:

"Luke duplicates the story of the mission of the disciples, so that alongside a mission of the 12, there is a mission of the 70 [Lk 10:1-16]. It is possible that the latter prefigures the Gentile Mission, with the number 70 perhaps intended to correspond to the total number of the nations of the world."

I think we might now be entitled to put that stronger. So also with Tuckett's next example, the second sending out for wedding guests, Lk 14:16-24. It thus seems that the later parts of Luke will also fit the hypothesis.

Q3. But Tuckett himself points out that Luke reduces Jesus's contact with Gentiles. Doesn't this refute your argument?
A3: No. Luke intentionally locates the literal Gentile Mission, as opposed to its symbolic prefigurement, in the time after Jesus. This is in accordance with his master plan for Luke-Acts. For the logic of that plan, see Conzelmann, "Mitte der Zeit."

Q4. You rely on a "Mary trajectory" of increasing respect for Jesus's family, but Luke does not fit that trajectory. In Lk 11:27-28, Jesus says to someone who calls Mary blessed, "Blessed rather are those who hear the Word of God and keep it." Yet Lk 1 makes Mary blessed above all women. Two different stages of the supposed trajectory are present in Luke, so there can be no trajectory between the Gospels.
A4. Lk 8:19-21 is Luke's more gentle rewrite of the harsh Mk 3:31-35, "Whoever does the will of God is my brother and sister and mother." Consistently with that stance, in Lk 11:27-28, which has no Synoptic parallel, Luke is reacting negatively to an emerging Mary cult. Thus far Proto-Luke. In Deutero-Luke, by contrast, Luke has found a way to join the Mary trend. There is evolution, between the two stages, but there is no conflict at any one stage.

Q5. You seem to be assigning unique Lukan material to both Proto- and Deutero-Luke. Doesn't this require positing two L sources, L1 and L2, and aren't you then multiplying sources?
A5. No. I am multiplying authors, both of whom are the same person. As to the source concept as such: There was no source L. Luke was not a scribe. The Gospel of Luke is not a failed copy of the Gospel of Mark. Nor is it a dutiful conflation of every scrap of text available to Luke. Luke has his own ideas, which are equally clear when he is rewriting Mark, and when he is writing on his own. The thing is, his ideas changed. Reading Matthew, when at some point he later encountered it, obviously hit him hard. It pushed him into bigger thinking, in which the traces of his previous thinking are still apparent. The result is universally acknowledged to be a masterpiece. If Luke needed to reach his masterpiece in two steps instead of one, so what? Let those who have never rewritten their SBL paper cast the first stone.

Calligraphic Separator

Postscript 2010. The view taken here of Acts and Luke must be revised to take account of the theory, due to Torrey, whose division (but not whose scenario) I consider convincing, that Acts was composed in two parts, the first ending at Ac 15:35. This necessitates a three-layer theory of Luke, not the two-layer theory suggested in this paper. (EBB)

All lectures and abstracts posted on this site are Copyright © by their authors.

 Back to Lectures Page

6 Sept 2010 / Contact The Project / Exit to Lectures Page