A favorite argument against the existence of Q is simply that no manuscripts of Q have ever been discovered. No more than this bare assertion is usually made. But a little probing shows that this argument has some serious weaknesses to it. In particular, three come to mind: (1) If Matthew and Luke swallowed up Q in their writings, why would we expect to find any copies of Q? Or to put this another way, Luke says that he used more than one source, presumably more than one written source. If so, why haven’t we found it/them? The fact that we haven’t surely doesn’t mean that Luke was not shooting straight with us, does it? (2) Even the Gospel of Mark has few copies in the early centuries, yet it was endorsed as an official Gospel by Ireneaus. Yet this is a canonical Gospel, which apparently was regarded in some sense as authoritative before the end of the first century, or at the latest in the first decade or two of the second century, because of its association with Peter. Yet if there are only two copies of Mark in Greek before the fourth century still in existence (at least as far as what has been published to date), what chance do we have of finding a non-canonical gospel-source in the early centuries? And as the centuries roll on, the likelihood that such a document would continue to be copied becomes increasingly remote. (3) Apart from having the text of Q, as it has been reconstructed, what other criteria should scholars demand of such an alleged discovery? Do they expect the document to have a title such as “The Gospel according to Q”? That neologism won’t wash. Perhaps just such manuscripts have been discovered but were mislabeled. The burden of this short essay is to examine that possibility.
Eight papyri are known to exist that contain portions from only the Gospel of Luke. On the assumption that Luke is closer to the wording of Q than Matthew is (an assumption that all two-source theorists embrace, I believe), it is at least hypothetically possible that one or two of these papyri are actually manuscripts of Q rather than of Luke. A closer examination, however, severely limits this possibility. The eight papyri are:
P3 (VI/VII): L 7.36–45; 10.38–42
P4 (III): L 1.58–59; 1.62–2.1; 2.6–7; 3.8–4.2; 4.29–32, 34–35; 5.3–8; 5.30–6.16
P7 (III–IV?): L 4.1–3
P42 (VII/VIII): L 1.54–55; 2.29–32
P69 (III): L 22.41, 45–48, 58–61
P82 (IV/V): L 7.32–34, 37–38
P97 (VI/VII): L 14.7–14
P111 (III): L 17.11–13; 17.22–23
In order for these to be manuscripts of Q, we might expect the following: (1) an early date, probably no later than the fourth century; (2) exclusively double-tradition (with Matthew) material; (3) no original-hand markings that identify the text as from Luke (e.g., title, Eusebian canons); (4) the order of the material might be other than what we see in Luke or (a) at least each fragment involves only intra-pericope material rather than inter-pericope material, or (b) the inter-pericope fragments of a single manuscript do not have the same order as is found in Luke; (5) perhaps some slight differences between Luke and this papyrus, with this manuscript displaying a somewhat more primitive text.
In reality, almost none of these expectations is a requirement, though if all are present in a manuscript they increase the positive identification of that manuscript as from Q. The reasons these expectations are not ex hypothesi necessary are as follows:
(1) There is the possibility that Q was copied for centuries; we have some evidence of ‘The Gospel of the Hebrews’—a first-century gospel, most likely (it’s mentioned by Papias)—existing for centuries, judging by patristic comments (see James Edwards’ The Hebrew Gospel and the Development of the Synoptic Tradition [Grand Rapids, Eerdmans, 2009]). If that document can exist, why not Q?
(2) Q can conceivably be found only in Luke rather than in Luke and Matthew. That there are 235 verses found in both Matthew and Luke but not found in Mark gives a solid basis for the existence of Q, but it does not equally argue that Q is no more and no less than these 235 verses. Such a conclusion would, in fact, be bizarre. Analogously, if the only way we could tell if Mark existed was its use by Matthew and Luke in double-tradition material, we would have to say that Mark was about half as long as it really is! Further, we would expect there to be some overlap between Q and Mark, so it can also be found in triple tradition material (the ‘blessed overlap’).
(3) Original-hand markings that identify the papyrus as from Luke is the one piece of evidence that would exclude a particular manuscript. The reason is simple: the original scribe of such a manuscript would have copied out the whole of Luke rather than writing fragments. Of course, if the text is an amulet or some other snippet from the Gospels, the scribe may have claimed that such a text was from Luke mistakenly.
(4) We really have no idea what the order of pericopae in Q was—assuming that it was a single written document (an assumption that has not yet been satisfactorily demonstrated). Thus, if parts of two or more pericopae are found in sequence in a given papyrus, this does not necessarily rule out that the manuscript is a copy of Q. But if multiple fragments from one Luke papyrus seem to be rearranged from the order found in Luke, this could argue for such a document being a copy of Q.
(5) Textual differences, especially if the papyrus in question displayed a more primitive form of text than that found in Luke, would signal the possibility of the papyrus being a copy of Q. But if Luke did not alter the text of Q in a given pericope, then we would expect to see no differences, apart from the usual corruptions, between the established text of Luke and the text of the papyrus in question.
In light of all these points and counterpoints, we would nevertheless conclude that the surest way for scholars to detect a fragment from Q would be for the five guidelines to be in place. Thus, of the eight Lukan papyri listed above, do any meet most or all of these qualifications? The five guidelines will again be enumerated, but this time with the best candidates for each category.
(1) Date: five of the papyri are from the fourth century or earlier: P4, P7, P69, P82 (fourth or fifth), and P111. The rest of the criteria will be examined only for these candidates.
(2) and (4) Of the five earliest Luke papyri, P4 is not restricted to double-tradition material and it also has fragments that encompass more than one pericope (at one point it has text in sequence covering five pericopae in Luke); P69 involves more than one pericope and has parallels in both Matthew and Mark (thus, triple-tradition); P82 involves two pericopae, with the second being in the triple tradition; P111 has one single-tradition pericope followed by a triple-tradition pericope.
The best candidate is P7, which has only Luke 4.1–3, one of the double-tradition texts. Of course, with only three verses, to claim that we have found one of the copies of Q is far more weight than this slender evidence can bear—unless there were strong corroborative evidence.
(3) I have not yet examined P7 to see if there are telltale signs that the original scribe thought that he was copying Luke.
(5) According to the Nestle apparatus, there is nothing out of the ordinary in the text of P7 for Luke 4.1–3.
Altogether, the evidence thus far presented can hardly be said to build confidence that any missing Q fragments have actually been discovered. But at least, ex hypothesi, such a discovery has some reasonable expectations laid on it so that papyrus discoveries yet to come may be examined for whether they supply any evidence of being copies of Q. Still, I’m not holding my breath.
For further reading:
Kurt Aland, Synopsis Quattuor Evangeliorum, 15th edition (Deutsche Bibelgesellschaft, 2007)
Mark Goodacre, The Case Against Q: Studies in Markan Priority and the Synoptic Problem (T & T Clark, 2002)
John Kloppenborg, Q, the Earliest Gospel: An Introduction to the Original Stories and Sayings of Jesus (John Knox Westminster, 2008)
John Kloppenborg, et al., editors, The Critical Edition of Q (Fortress, 2000)
Robert Stein, Studying the Synoptic Gospels: Origin and Interpretation, 2nd edition (Baker, 2001)
Daniel B. Wallace 