All the Lukan doublets discussed below have parallels in Mark, although in all cases there is no corresponding doublet in Mark. Additionally, none of the parallel text in either Matthew or Luke is part of the double tradition.
There are parallel passages in all three synoptic gospels regarding the healing of the man with the palsy, at Mk 2:3-12, Mt 9:2-7, and Lk 5:18-25 respectively, and in all three Jesus tells the man that his sins are forgiven. In all three passages scribes (and Pharisees in Luke) comment that Jesus is blaspheming, and Mark and Luke (but not Matthew) then follow this by asking “who can forgive sins but God alone,” a question asked in Mk 2:7 and Lk 5:21, with a similar question at Lk 7:49. In his Section V regarding formulas Hawkins lists the parallels in Mk 2:5,7, Lk 5:20-21 and 7:48-49 as follows (re-formatted for readability):
ἀφέωνταί σοι αἱ ἁμαρτίαι σου … τίς ἐστιν οὗτος ὃς … τίς δύναται ἁμαρτίας ἀφεῖναι Lk v. 20, 21 = Mk ii. 5, 7 : also
ἀφέωνταί σου αἱ ἁμαρτία … τίς οὗτός ἐστιν ὃς καὶ ἁμαρτίας ἀφίησιν; Lk vii. 48, 49.
thy sins are forgiven thee … who is this … who is able to forgive sins
What is particularly interesting here is the length of the agreement, combined with the fact that the text only occurs in three places in the gospels, with both points suggesting that it perhaps should be considered to be a doublet in Luke instead of a formula. It is also possible that Hawkins does not consider Mt 9:2b to be part of the formula because as Matthew has no parallel to Mk 2:7b / Lk 5:21c “ἀφίενταί σου αἱ ἁμαρτίαι” it may not be long enough on its own to have been counted as such:
On the MwQH the parallels are explainable on the basis that aMatthew saw Mk 2:5-7 and aLuke saw either or both Mk 2:5-7 / Mt 9:2-3, but that aLuke chose to reject most of the parallels at Mk 14:3-9 / Mt 26:7-13 and replace them with his unique version of the story of the woman with the alabaster box that we see in Lk 7:39-49. However, that itself raises the question of where the Lukan version of that passage came from.
None of these verses were in Q, and so on the Mark-Q hypothesis it would appear that there were two different versions of the story, with aMatthew choosing to use the one we see in Mark and aLuke choosing to use the other one, from an unknown source.
On the Early Luke hypothesis Early Luke is the source of the version in Luke, with aLuke preferring this version to the (earlier) one in Mark. aMatthew then saw both versions and chose to use the shorter version in Mark.
None of the major synoptic hypotheses answer the question of the origin of the version of the story that we see only at Lk 7:38-49, nor why aLuke chose this version over that in both Mark and Matthew. As it is unique to Luke it cannot have been in Q but could have been in an ‘expanded Q’ or a different second source such as Early Luke, in which case we would have two versions, one in Mark and one in the second source, with aMatthew using that from Mark and aLuke keeping the one he saw in the second source. In this scenario the reason why Lk 7:48-49 are not part of the double tradition is then simply because aMatthew chose not to include this story.
Lk 9:8 // 9:19 are the only two places in the Bible that refer to the old prophets having risen again. This therefore appears to be a unique addition by aLuke to make the point that the old prophets were no longer alive, perhaps for the benefit of future readers or others who would not know this.
The existence of parallels to Mk 6:15 // 8:28 in both Matthew and Luke, and the doublet at Lk 9:8 // 9:19 is easily explained on any synoptic hypothesis that assumes Markan priority, with for example aLuke copying both halves of the doublet from Mark. Consequently, the only significant issues appear to be why Matthew does not contain parallels to Mk 6:15-16 / Lk 9:8-9, and why aLuke replaced Mk 6:16b with Lk 9:9b.
The latter point is most likely associated with the fact that Luke also has no parallels to Mk 6:17c-29 / Mt 14:4-12a, as explained in Difficult Variants in Text Unique to Mark (Mark 6:14-29 - Herod). It is also pointed out there that while Mt 14:3-14 contains some parallels to Mk 6:18-29 there are many gaps, with there being no parallels to Mk 6:17c, 20b, 21b, 22b, 24b-25a, and 27b, and many differences in those parallels that do exist. It is also the case that there is extreme uncertainty over the wording of Mk 6:16b, with Willker (Mark, TVU 121) recording no less than 13 different variants at the end of this verse and stating: “It is clear that this large number of variants has its cause in a difficult original reading.”
These points, together with the very strange re-location of the Lukan parallels to Mk 6:17ab / Mt 14:3 from their 'expected' location between Lk 9:9 and 9:10 to Lk 3:19-20 instead, suggest that both aMatthew and aLuke saw damage in this area of Mark, with aMatthew having to shorten his parallel Mk 6:16b-29, and aLuke moving his parallel to Lk 3:19-20, and instead ‘filling in the gap’ with Lk 9:9b, which has no parallel in either Mark or Matthew.
This appears to be a simple case of a phrase known to aMark that was used in the parallels at Mt 14:15a and Lk 9:12a, that aLuke then used again at the end of the unique Lukan passage at Lk 24:16-32, an addition the source of which is unknown. However, although there is some overlap in the words used in Lk 9:12 and 24:29, it appears that the references to the end of the day in Luke are simply aLuke reporting early evening in two places, with no suggestion of a connection between the two events.
Lk 9:12a Ἡ δὲ ἡμέρα ἤρξατο κλίνειν Now the day began to be far spent,
Lk 24:29 καὶ ἤδη ἡ ἡμέρα κέκλικεν and by this time the day is far spent.
Both halves of the doublet at Lk 18:18 // 10:25 are in triple tradition passages, and although Mk 10:17d-18 / Mt 19:16b-17a / Lk 18:18b-19 are close parallels, it is only in Luke that there is a doublet. In this case it appears that aLuke chose to repeat his question from Lk 18:18b at Lk 10:25 (i.e. earlier in his gospel) instead of paralleling the different question at Mk 12:28c / Mt 22:36 regarding the first or greatest commandment respectively.
Both Mt 22:35a and Lk 10:25a refer to a lawyer questioning Jesus, about which Willker TVU 312, Mt 22:35, notes that “f1, e (5th CE), 1780, pc, Sy-S, arm” omit the mention of a legal expert or lawyer [νομικὸς], adding:
νομικὸς is a Lukan word and appears in Mt only here. Zahn (Comm. Mat.) thinks that Matthew used νομικὸς because the question concerns the law.
The support for the omission is very strange. If νομικὸς is really an addition, then it must be extremely early. The consequence would be that f1 (almost) alone can preserve the original. It might be noted that the verse is the beginning of a Sunday lection. The same is true for Lk 10:25. Thus the story was well known and a harmonization is likely to occur.
It is also possible that scribes had a problem with the fact that first it is said "one of them" = "one of the Pharisees" (see verse 34) and then it is a lawyer (and not a Pharisee?).
Streeter ("Four Gospels", p. 320) accepts the omission as original. So do Burkitt and Blass.
Willker is making the point that because almost all support for the omission of ‘a legal expert/lawyer’ comes from a single late manuscript family (f1, from the 12th to 15th century), and all early manuscripts extant for the phrase include the word, then if it was an addition it must have been added very early. Although this initially appears to be a straightforward example of a triple tradition passage the issue of the lawyer and other problems with the questions in Mk 12:28 / Mt 22:35 / Lk 10:25 impact the various different synoptic hypotheses:
On the MwQH it appears that aMatthew changed what he saw in Mk 12:28, and then aLuke chose (for whatever reason) to follow Mt 22:35a rather than Mk 12:28, except that he then used the question from Mk10:17 / 19:16 instead. On this hypothesis is no obvious reason why aMatthew would write Mt 22:35a and aLuke would then write Lk 10:25c instead of what they saw in their sources (Mark, and Mark/Matthew respectively).
On the Mark-Q hypothesis the most reasonable explanation for the different questions is that Q 10:25 contained something close to what we see in Lk 10:25, and that aMatthew here chose to follow Q 10:25ab but then Mk 12:28c instead of Q 10:25c-26. In an exchange on 7/15/2018, David Sloan takes up the point of the question being different in all three synoptic gospels:
Most Q scholars think the parallels between Matt 22:35 and Luke 10:25 are due to a combination of coincidence and later scribal assimilation, but there are some exceptions to this. The following have argued that there is a Mark-Q overlap here: B. Weiss, B. H. Streeter, R. H. Fuller, C. H. Tuckett, D. Zeller, J. Lambrecht, D. R. MacDonald, and myself. I presented a paper at SBL in November arguing that almost every time Luke has a passage in its non-Markan position, it is because he is following a non-Markan version of the same story… I think the most thorough treatment of the passage is Jan Lambrecht, "The Great Commandment Pericope and Q" in The Gospel Behind the Gospels: Current Studies on Q, ed. Ronald A. Piper, NovTSup 75 (Leiden: Brill, 1995): 73-96, but Fuller also addresses it at length in Essays on the Love Commandment (pp. 41-56) and Tuckett does too in The Revival of the Griesbach Hypothesis (pp. 125-133).
I have argued for its inclusion on two different occasions. The key evidence I have given for this that no one else (that I am aware of) has considered is the fact that in Luke 10:25-28 a lawyer tests Jesus and then in Q 11:15 the Pharisees test Jesus (though Luke simply says “some” tested Jesus), which leads into Q 11:37-52, where Jesus speaks woes against the Pharisees and the lawyers. Here is an example of Matthew and Luke each leaving remnants of Q that when put together give a picture that neither himself gives. Q seems to have set up the woes, but Matthew loses this by moving these passages to different places in his gospel and changing the woes against the “lawyers” to woes against the scribes and Pharisees. Luke loses the set up by omitting that the second test came from a Pharisee. I believe that Luke 10:25-37 comes from Q. I have long been undecided on Luke 10:38-42.
Sloan’s above point on the assumption of the Mark-Q hypothesis appears sound, however his argument also supports additional sources other than Q. For example, on the MwEL hypothesis Early Luke contained parallels to Lk 10:25-26, and aMatthew followed ELk 10:25ab but not ELk 10:25c-26.
Mk 9:32 / Lk 9:45 are the concluding verses of two of the versions of the second passion prediction. The third, at Mt 17:23, has no parallel to either of these verses and instead reports that the disciples were “exceeding sorry.” Lk 18:34 is the unique conclusion to the Lukan version of the third passion prediction, and as Mk 10:33-34 / Mt 20:18-19 / Lk 18:32-33 is a triple tradition passage it appears that here aLuke is essentially just repeating what he had written earlier. Assuming that aMatthew saw Mark and chose not to include a version of these words at the end of his second passion prediction it is not surprising that he did not do so here either, so the question is why aLuke chose to do so.
Both Lk 2:14 and 19:38 contain δόξα ἐν ὑψίστοις [glory/hosanna ... in the highest] and εἰρήνη [peace]. It appears that in Lk 19:38 aLuke repeated what he had written in Lk 2:14 in preference to what he saw in Mk 11:10 and (possibly) Mt 21:9b, but there is nothing here that points to whether aLuke knew Matthew or not. As Lk 2:14 is in the infancy narrative there is no parallel in Mark, and the nearest equivalent verse in Matthew is very different, so this does not appear to be an example of aLuke writing a parallel to Mk 11:10 first and then going back to add Lk 2:14. Instead, aLuke’s use of glory and peace instead of hosanna suggests that in Lk 19:38 he is simply following Lk 2:14.
There is a clear triple tradition passage in Mk 13:1-8, Mt 24:1-8, and Lk 21:5-11, and consequently there can be no doubt that both Mt 24:2 and Lk 21:6 have Mk 13:2 as a source. The slightly shorter Lk 19:44b appears to be an edited version of Lk 21:6b, except that Lk 19:44b is located before Lk 21:6b, suggesting the possibility that the Sondergut Lk text Lk 19:41b-44 was added to Luke after Lk 21:6. On this point it should be noted that in his Scholion 53 Epiphanius records that in Ev (The Evangelion, attributed to Marcion:
He falsified the section about the ass and Bethpage – and the one about the city and the temple, because of the scripture, “My house shall be called an house of prayer, but ye make it a den of thieves."
This can most reasonably be taken to indicate that Ev did not include Lk 19:29-44, a possibility that is supported by the much shorter text in Lk 19:30-34, 42 in Codex Bezae, the omission of Lk 19:32-34 in G*, 063 and 477, and the great number of variant readings in Lk 19:38, 42 and 45 (Willker, Luke, TVU 308 - 313). As a result, even though there are no mss extant for Lk 19:43-45 that omit this verse it is conceivable that there was a version of Luke that did not originally contain at least Lk 19:42-44, and hence that this half of the doublet is a later addition. This would certainly support the MwEL hypothesis.
On both the MwQH and the Mark-Q hypothesis there is no obvious reason why aLuke would place his parallel to Mk 13:2 at Lk 21:6, presumably after having already placed the earlier half of the doublet at a location with no parallels.
None of these doublets in Luke have a parallel doublet in Matthew, and only one has a parallel doublet in Mark. None of the text is in the Double Tradition, there is no indication that any of these verses were in Q, and with two exceptions they are Triple Tradition verses. Unusually Matthew has no parallel to Mk 9:32 / Lk 9:45, 18:34.
This group of doublets/formulas has none in Matthew, and none of Hawkins numbered doublets in Luke. None of the doublets/formulas have parallels in the double tradition, and consequently on the Mark-Q hypothesis none of this text can have been present in Q as usually defined. However, as discussed in What Exactly is Q? there is no actual ‘barrier’ to Q having been more extensive than is traditionally believed, containing at least some material used by Matthew but not Luke, some used by Luke but not Matthew, and more extensive Mark-Source Overlaps. As a result, the inclusion of this text in a second source that is effectively an expansion of Q cannot be automatically ruled out, and this also applies to other potential additional sources such as the Logia, Early Luke, or Ev.