Return to Main Page -O- Order The Case Against Q (US) / UK

Is Q a Juggernaut?

by Michael D. Goulder
University of Birmingham, Birmingham B15 2TT, United Kingdom

This article first appeared in Journal of Biblical Literature 115 (1996), pp. 667-81. I am grateful Professor Goulder and to the Society of Biblical Literature for permission to reproduce it on the Mark Without Q web site. Reference to the original page numbering is placed here in bold type with square brackets.

[This paper requires the free Scholars Press fonts SPIonic (Greek) and SPTiberian (Hebrew). You can download these fonts quickly and easily from the Scholars Press Fonts site. If you have Windows 95/98, simply download the fonts into the file C:/Windows/Fonts. If you have any problems with this, let me know at]

In the last twenty years the pursuit of Q has become a considerable industry. Not only do articles and books on Q appear in flood, (1) with competing versions of the original text, (2) but a great concourse of scholars is at work to produce a definitive edition. (3) Many Ph.D. students are labouring in the same vineyard. The studies are often concerned to find not the final version but the phases to the final version, the theology of the different editions, the different groups within the early church that produced these editions, and so on. The study is not merely of the Q hypothesis but of the hypothetical earlier forms of Q and the even more hypothetical settings in life that gave birth to them.

So much piling of Pelion on Ossa might itself be cause for alarm, but what further raises the anxiety level is the confidence with which this is being done. Q is a hypothesis that has stood for a century and a half, but it is not an unchallenged hypothesis, and Ronald Piper notes, in his introduction to a recent book of essays on Q, that there is considerable hesitation in Britain over whether Q [668] ever existed. (4) For example, Christopher Tuckett's article in a recent symposium is entitled "The Existence of Q," (5) and David Catchpole opens his Quest for Q with a fifty-nine-page chapter entitled "Did Q Exist?" (6) The live alternative to Q is in essence that proposed by Austin Farrer in 1957: (7) Mark wrote first; Matthew wrote an expanded version of Mark; and Luke used and adapted both earlier Gospels. Under this theory Luke either copied or rehandled the "Q" verses in Matthew, so the lost source can be dispensed with. Such a position requires a detailed defense, and I have offered such myself in a two-volume, eight-hundred-page work, Luke: A New Paradigm (1989). (8) Q is now hardly defended in the University of Oxford.

R. A. Piper notes that the scholarly communities have different concerns in different countries, (9) and although, for example, my Luke has had a respectful reception in Britain and in Belgium, (10) it is alarming to note that so eminent a Q scholar as Dieter Lührmann has apparently never heard of its hypothesis! (11) This situation gives rise to fear that Q is a juggernaut that cannot be stopped but is fated to consume the lives of generations of Ph.D. students who will in the course of time become professors and direct its further ineluctable progress. For experience, if not common sense, ought to disabuse us of the simple notion that the truth will conquer with time. People, including scholars, are not purely rational beings. They hold opinions for a multitude of reasons -- respect for their teachers, attachment to the familiar (especially the familiar that is held by most other people), religious predilections, unwillingness to think that they have been wrong for years (especially when they have committed themselves publicly), the enormous volume of material on the subject that no one can read, to name a few. (12) None of these considerations is disreputable. [669] But whereas in a scientific subject a paradigm shift is possible because new irrefutable evidence may come to light, new evidence in an arts subject is rare, and so are paradigm shifts. (13)

In view of this apparent logjam it may be helpful to offer a brief discussion of the issues; and as they are set out simply and clearly by Tuckett in the essay referred to above, we may take this as the basis of the argument. I appreciate Tuckett's work: he reads the text carefully; he responds to the debate with acuity; and he is unfailingly fair and courteous.

1. Ancient Sources

There is no reference to Q in any ancient source. Tuckett comments (1) that this is unremarkable -- we know there was a Gospel of the Hebrews and a "prior" and a "tearful" letter to the Corinthians, all of which have been lost; and (2) just as Mark's Gospel was eclipsed by the advent of the fuller Matthew and Luke, so we might expect the fuller Matthew and Luke to have eclipsed Q. These response are rather weak. The Gospel of the Hebrews was a significant document treasured by Jewish Christians, and we know it existed because we hear about it, even from its opponents. Q is supposed to be the sane, going through several editions, and (as Tuckett is keen to stress) with a theology different from Luke and Matthew. So we might have expected Q to be at least referred to, as the Gospel of the Hebrews was, if not preserved, as Mark was, with its independent theology -- especially as it is ex hypothesi older than the canonical Gospels and must have enjoyed enormous (probably apostolic) prestige.

It should be mentioned in this connection that Q was known until the 1890s as the Logia and is still often called in German the Logienquelle. It was first conjectured by K. Credner and C. H. Weisse because they thought that it was testified to in early writings, by Papias. (14) Thus, the Q hypothesis rests in part on a misunderstanding; I do not think that anyone today maintains that Papias meant by ta\ lo/gia a sayings document such as Q. [670]

II. The Minor Agreements

Tuckett correctly says: "At one level, the Q hypothesis is simply a negative theory, denying the possibility that one evangelist [Luke] made direct use of the work of the other [Matthew]." (15) It is for this reason that the Minor Agreements (MAs) have been so much discussed recently. (16) There is an apparently irresolvable rump of small agreements between Matthew and Luke against Mark, especially in the passion story, where ex hypothesi there was no Q. One of these, the later Gospels' common addition of the five words ti/j e0stin o9 pai/saj se; at Mark 14:65 is particularly acute: Frans Neirynck's famous apologia postulates that every manuscript of Matthew has been corrupted. (17) But a more straightforward explanation would be that Luke simply carried these words over from Matthew: so why is this not accepted? We should then have no Minor Agreements problem anywhere.

It is not accepted because if Luke knew Matthew here, he knew him everywhere, and he could have taken all the Q material over from Matthew's Gospel directly. And so the main reason for postulating Q would have disappeared. Neirynck disputes this, quoting in his essay in Piper's Gospel from an earlier article of his:

If [dato non concesso] Lukan knowledge of Matthew would be the conclusion to be drawn for the minor agreements, then Luke would have used Mark notwithstanding his knowledge of Matthew, and the inference could only be that elsewhere, where Luke is using another source, a similar subsidiary influence of Matthaean reminiscences can be expected. (18)

This is true but irrelevant. It is true that [on my theory] Luke knows Matthew and slips Matthean reminiscences into his version of Mark and that if he also knew Q we might expect him to use Q with Matthean reminiscences. But Luke's use of Mark is a fact (or generally accepted as one), while Q is a mere postulate. The question is why Q is to be postulated, and the answer given by Tuckett is, correctly, because Luke cannot have "made direct use of [Matthew]." (19) But it now appears that Luke did make direct use of Matthew, in the MAs. So Q is undermined -- though not of course discredited: the questions [671] still have to be answered about why, if Luke had Matthew's Gospel, he handled it as he did.

III. Matthean and Lukan Vocabularies

A key issue in any reconstruction of Q is the establishment of agreed vocabularies favoured by Matthew and Luke. For instance, in the temptations Matthew has the devil take Jesus ei0j th\n a9gi/an po/lin (4:5), while Luke writes ei0j 0Ierousalh/m (4:9); but both expressions are favoured by their writers, Matthew introducing ei0j th\n a9gi/an po/lin redactionally in 27:53, and Luke showing a marked preference for the Semitic 0Ierousalh/m (2/0/27 + 36 [= occurrences in Matthew/Mark/Luke + Acts]). So the modern editor of Q has to balance his judgment which wording is the less characteristic of the respective evangelist and opt for that, or otherwise conjecture a third possibility. The reconstruction of Q is based on its vocabulary being unlike that of Matthew and Luke.

I argued that this procedure was based on a fallacy. Three of the best known Matthean-favoured language forms are o0ligo/pistoj, e0kei= e1stai o9 klauqmo\j tw~n o0do/ntwn, "And-it-came-to-pass-when -jesus-had-completed-all-these-[words]"; and each of these occurs once in a parallel passage in Luke, and so [on the Q view] in Q. Nor is this an accident: in the very first Q pericope, the Baptist's preaching, come phrases -- "You offspring of vipers!," "make fruit," "cut down and cast into the fire," "gather his corn into the garner/burn the [chaff]" -- all of which recur, some often, in Matthew's Gospel where there is no Markan or Lukan parallel. Two possible conclusions would seem to follow from this: either Q and Matthew are the same person, as I think, or at least Qs style is indistinguishable from Matthew's. If the latter, then reconstructions based on the criterion that Q's style differs from Matthew's are valueless.

Tuckett makes two responses to this. (1) He claims that I am implying that all Matthew's redactions are his own creations and that this is "dangerous nonsense": Matthew is often taking over standard Jewish expressions like "the kingdom of heaven." (2) He asserts that my argument could be used to prove that Mark used Matthew. For instance, Matthew is fond of calling the Pharisees hypocrites, and Mark uses the word once in 7:6, with a parallel in Matt 15:7. So just as I am maintaining that Luke carried over single uses of o0ligo/pistoj, and so on, inadvertently, so might I argue that Mark carried over the single use of u9pokritai/, inadvertently. The same would apply to to/te, which Matthew has ninety times, with some few parallel uses in Mark. So my criticism can be reduced ad absurdum. On the other hand, it would be possible for Matthew to have found an expression in a single instance in Q and used it a number of times on his own account.

The first point is again true but irrelevant. I do not at all think (nor have I [672] ever voiced) such dangerous nonsense as that Matthew created his favoured expressions de novo. But Tuckett's own instance proves my point. "The kingdom of heaven" is a standard Jewish phrase, which is common in Matthew (32/0/0 + 0), and it is immediately struck out by every reconstructor of Q as being likely MattR: the more general "kingdom of God" is always preferred. So when a Matthean phrase appears in Q-wording differing from that in Luke, it is explained on the basis that Qs style differs from Matthew's; and when such a phrase appears in Q-wording identical with Luke, it is explained on the ground that Q's style is very similar to Matthew's.

The second point is, however, a fallacy. Hypocrisy is a regular charge in NT religious tirade and is used by most NT authors; (20) to/te is one of the commonest of all NT words. Of course it is easy to pick out one or two such expressions from any pair of authors and show that it would be fallacious to argue that the earlier took it from the later. But this is not what I am arguing. My argument is that the occurrence of so many rare and striking phrases in Q that are also favourite expressions of Matthew destroys the principal criterion for reconstructing Q. If Q was a Jewish-background author who, like Matthew, whitewashed the disciples (not no-faith but little-faith), was keen on hell (weeping and gnashing of teeth, cut down and cast into the fire), used Matthew's regular linking clause ("And it came to pass when Jesus had finished. . ."), and his animal invective ("You offspring of vipers") and other whole clauses, why should we not think that Q spoke of "the kingdom of heaven"? The whole basis of any edition of Q has been taken away. I cannot of course, as Tuckett says, infer that Luke has carried these phrases over from Matthew rather than Matthew having fancied them in Q and reused them; I have just avoided the illogicalities mentioned -- rejecting the entire manuscript tradition to save a Minor Agreement, building a reconstruction on a contradiction -- and found the resulting alternative, that Luke used Matthew, rather plausible.

IV. Prior Lukan Forms?

An important argument in favour of Q has been that although the Matthean form of a saying seems often to be prior, so also does the Lukan one: famous instances of the latter are claimed to be the Lord's Prayer and the Beatitudes, and Tuckett and I have publicly discussed the latter, (21) with (to put it at [673] its least) no clear result that the Lukan form was the earlier. If the Lukan form were ever clearly shown, even in a single instance, to be the earlier (and Luke, ex hypothesi, wrote after Matthew), then it would seem that Luke was using a pre-Matthean form and must have used a pre-Matthean source, that is,Q. (22) One general point should be made about this side of the debate, and I will then comment on the instance Tuckett takes up several times in his essay, Luke 11:49//Matt 23:34.

The general point is again about a flaw in the Q method. Matthew writes rather a stylized Greek, with quite a number of stereotyped phrases, some of which have just been discussed, whereas Luke writes some of the "best" Greek in the NT, with a large vocabulary and a willingness to vary his use of particles, sometimes surprisingly introducing a word on his own (e.g. u9pa/gein ) which he avoids elsewhere. Hence, reconstructions naturally tend to see the Lukan wording as earlier, because it is less easily identifiable as Lukan. One example of many is discussed in Neirynck's article in the Piper volume: (23) lusitelei= is defended as Q by A. Polag in Luke 17:2, although J. Schlosser suggested that it might be "ein von Lk ausgesuchtes, gut griechisches Wort." (24) This richness of Lukan writing puts a slant on the whole argument and tends to make its conclusions uncertain; so that it becomes difficult to be confident that any Lukan form of a saying must be prior.

Tuckett offers an instance of a "decidedly un-Lukan" verse as a test instance, which is helpful:

Matt 23:34 dia\ tou~to i0dou\ e0gw\ a0poste/llw pro\j u9ma~j profhta\j kai\ sofou\j kai\ grammatei=j . . .

Luke 11:49 dia\ tou~to kai\ h9 sofi/a tou~ qeou~ ei]pen, 0Aposte/llw ei0j au0tou\j profhta\j kai\ a0posto/louj

He argues that Luke does not elsewhere speak of God's Wisdom, apart from 7:35, which is itself taken over from Q, on his theory. Matthew, however, has Jesus speak in the present, and this is in line with his equation of Jesus with divine wisdom elsewhere. I had replied that Luke has interpreted Matthew's profhta/j as Old Testament prophets, and so required God, not Jesus, to be the speaker, with h9 sofi/a tou= qeou= suggested by Matthew's sofou\j; but Tuckett finds this "weak and uncompelling": [674]

Why should Luke have interpreted the "prophets" of Mt 23:34 as OT figures? They are (in Matthew) clearly figures sent out by Jesus, and Luke certainly knows of Christian prophets.... why did Luke feel obliged to introduce a periphrasis for God, and why this particular periphrasis? Elsewhere Luke has no compunction about talking of God himself sending prophets. (25)

The answer to the first question is the Matthean context, which is concerned with OT prophets: "you build the tombs of the prophets . . . the blood of the prophets . . . you are the sons of those who killed the prophets . . . that upon you may come all the righteous blood shed on earth from the blood of righteous Abel to the blood of Zechariah" (Matt 23:28-35). Luke follows this theme in 11:47-48, the verses immediately preceding 11:49. He then thinks to combine the sending of these [OT] prophets [by God] with the sending of the NT emissaries mentioned in Matt 23:34; only he does not fancy Matthew's Jewish-type Myrpwsw   Mymkx and replaces them with the more universal a0posto/louj, itself suggested by Matthew's a0poste/llw. He has not (of course) misunderstood Matthew's clear phrasing in 23:34: he has interpreted it, turning Matthew's profhta/j into OT prophets in line with the context.

This interpreting then entails changing the sender from Matthew's Jesus ("I") to God, since it was God who sent the OT prophets -- and for that matter the NT ones too. I do not know that Luke "felt obliged" to use a periphrasis for God here, but it is not a problem to understand his motive. He has followed Matthew in linking the sending to the tomb building with dia\ tou=to but whereas Matthew had a logical and chronological sequence in mind -- Jesus sent the Christian mission [in part] so that the Jewish people should continue their rejection and meet their deserts in 70 -- Luke faces an illogicality with a reverse order. God has now to send the OT prophets and NT apostles from divine foreknowledge, indeed predestination, of Jewish rejection. So Luke writes h9 sofi/a tou= qeou=; as we might say, "So the providence of God said . . ." We find the identical doctrine, the divine hardening of the Jews, in Romans 9-11, and Paul sums the matter up with a doxology using the identical word: "O the depth of the riches and sofi/aj and knowledge qeou= (11:33). So Luke's eye was caught by Matthew's sofou/j; Luke sometimes uses periphrases for God -- his hand or his finger; and there are texts implying divine wisdom given to Stephen and others in Acts (6:3, 10). But the central reason for his introducing h9 sofi/a tou= qeou= here is his need to invoke the Pauline predestination doctrine, for which this is the natural expression. It must, as Tuckett says, remain a matter of judgment which of two theories is the more convincing; but it does not seem to me possible to think of Luke 11:49 as a text where the Luke-knew-Matthew theory is in any sense implausible.


V. Falsification

I had introduced K. Popper's notion of falsification as a means of pointing to the evasiveness of Q theorists, with the honourable exception of "hard-liners" like Neirynck and Tuckett. (26) Q, as represented by the latter, consisted of [at least] the matter common to Matthew and Luke in the ministry, but without a passion story. Hence, if we could find a single passage in the passion story where it was clear [= overwhelmingly probable, in a nonexperimental study] that Luke knew Matthew, we should have cut the ground from under the hypothesis. So Q would be falsifiable: and Popper had said that only falsifiable hypotheses were useful, because if there were no test, we could never find out if they were right or wrong. "All swans are white" is a useful hypothesis because it would be falsified by the discovery of one black swan.

The point is significant because when an apparently clear test is in evidence, in the form of the Minor Agreement ti/j e0stin o9 pai/saj se, most defenders of Q disappear into the smokescreen. Either there was an earlier lost form of the text (Ur-Markus), or a later lost form of the text (Deutero-Markus), or another version of the story (Nebenquelle), or an oral version, or multiple layers of tradition, or Luke had Matthew as well as Q. All of these possibilities are open to defenders of Q; but if they take them they make the theory unfalsifiable, and so not useful -- we can never know if it is true or not. One great virtue of Neirynck and Tuckett is their willingness to stand their ground. They are willing to maintain that every manuscript of Matt 26:68 (and all the versions and patristic citations) are derived from a corrupted text. I am, of course, happy with this situation, because I do not think that the scholarly community will be willing to countenance such a conclusion (especially as there is not just one text involved). I am hopeful that when the situation is recognized, it will be seen that the Q hypothesis is either false (Neirynck, Tuckett), or disreputable (the "soft-liners"), and will be abandoned.

Tuckett counterattacks with the claim that my own theory is no more falsifiable than his, because "no theory about the Synoptic Problem is falsifiable in the strict sense." (27) If the strict sense means disproof such as we can supply by experiment in science this is so: I cannot disprove the claim that all manuscripts of Matt 26:68 are corrupt, but I am quite content to have my opponents rely on a defense that I believe to be generally quite unacceptable. The same ought to be true of my hypothesis. If Luke did not know Matthew, then my attempts to explain nineteen thousand words of his Gospel on the theory that he did ought [676] to seem grossly implausible sooner or later. Luke 11:49 is not a good example of such a claimed implausibility. (28)

It was urged against Popper that the theory of the circulation of the blood was unfalsifiable but useful (and indeed true). But this is not the case. If we were to find one human being whose blood did not circulate (one black swan), the theory would be discredited. As we have not found such, the circulation theory seems well corroborated, and so likely to be true. Perhaps the Luke-expounded-Matthew theory is like that.

VI. The Order of the Lesser Matthean Expansions in Luke

Tuckett says, "A strong part of the argument for the existence of Q has always been the fact that (a) Luke never uses any of Matthew's additions to Mark in Markan material, and (b) Luke never has the Matthean material in the same context relative to the Markan material." (29) The point is a little overstated: Luke does use Matthew's additions to Mark in the Baptist's preaching in the same context (Matt 3:7-10, 12; Luke 3:7-9, 17), and similarly with the temptations (Matt 4: 1-11; Luke 4:1-13). These exceptions take much of the force out of the argument. Put more neutrally, Luke may combine Markan and Matthean material in narratives, where there cannot plausibly be duplication: sometimes considerable combinations (the Baptist's preaching, the temptations, the Beelzebul healing -- "Mark/Q overlaps"); sometimes inconsiderable (the "Minor Agreements"). But such combinations are the exception, not the rule.

The explanation I offered for this is that Luke adopted a policy of using his sources in blocks; and whatever theory is held, the presence of blocks is a fact. From Luke 4:31 to 6:19 (with small, irrelevant exceptions), Luke follows Mark as his source, and again from 8:4 to 9:50, and (with more considerable insertions) from 18:14 on: the remaining blocks contain occasional echoes of Mark, but the source being followed is non-Markan. Now Matthew has rewritten the Markan stories and often has added a few lines of his own -- Peter's walking on the water (Matt 14:28-31), for example, or the Petrine commission (Matt 16:17-19). The question is therefore why Luke should not have included these smaller Matthean expansions if he knew them. The obvious answer seems to be that it was easier simply to follow the source in front of him over a series of pericopae and to go over his second source, Matthew, when he was handling that. One may easily think of reasons for Luke's leaving out the two additions [677] referred to; Acts shows Luke as an admirer of Paul and sees Peter as the rock of the church by strengthening his brethren (Luke 22:32), not pontificating. But more generally, Luke probably had small writing space, and no synopsis: finding the Matthean parallel would be tiresome, and the procedure I have suggested is entirely believable. (30)

Tuckett raises two objections. First, what about Matthew 24-25? Here Luke has "gone through Matthew very carefully, marking off those extra bits of Matthew" (31) which he has included in Luke 12, 17, and 19, from the Mark 13 discourse, which he has included in Luke 21. "Clearly Luke's 'block' policy has been rather different in Mk 13/Mt 24-25 than elsewhere!" (32) The point is true, but has little force. Luke has used Mark 13 for a number of years before Matthew wrote, and he is used to following the Markan order (in sermons in church, in teaching, or whatever other ways); so he means to use it in its "Markan block" position immediately before the passion (Mark 14-15; Luke 22-23). The chapter describes a series of things to come, with a brief warning at the end, all of which Luke transcribes with suitable changes in chap. 21, but without any of the Matthean additions, exactly in line with his normal block policy. The Matthean additions at the end, Matt 24:37-25:46, and minor expansions in 24:23-28 he takes at his convenience in his Long journey section: mostly in the Matthew section, 9:51-18:14, but the pounds parable, Luke's version of Matt 25:14-30, in a suitable Markan context, Luke 19:11-27. So he has followed the same general "block policy" as elsewhere; only instead of adding in some of the Matthean additions later, like the leaven parable (Matt 13:33; Luke 13:20-21; not in Mark 4), he takes them earlier. Whether he isolated these additions by knowing Mark 13 so well or by using a pen or some other method is beyond our knowing.

Tuckett's second objection is the alleged inconsistency of supposing that Luke did not refer to Matthew when following a block of Mark, while at the same time including Minor Agreements so small as a0po/ for u9po/ and the omission of two articles, all in Luke 9:22. This is a misunderstanding. Any member of the clergy preaching from a Synoptic text is likely to cite the most familiar form of the text by reminiscence: someone preaching, say, on Mark 8:31 -- but having not looked the piece up -- may cite the similar wording of Matt 16:21 (= Luke 9:22). But when preaching on the ten virgins, the preacher looks Matthew 25 up and gets it right. Coincidences of wording may occur either by [678] accidental reminiscence or by deliberate exposition: I am arguing the first for Luke's MAs and the second for his use of the Matthean expansions of Mark.

VII. Luke's Diminution of the Matthean Discourses

A long-standing question has been why, if Luke knew Matthew, he should have broken up Matthew's well-ordered discourses, especially the Sermon on the Mount. I noted that Luke also much abbreviated Mark's parable discourse, Mark 4:1-34 (the first part of which appears in Luke 8:4-18), and argued that Luke felt that long discourses were indigestible and wasteful, especially hundred-verse discourses such as Matthew 5-7. He therefore cut up his teaching material into short sections, normally a dozen to twenty verses, signaling the breaks with a rubric such as a question from a disciple or the crowd, or a brief scene in the synagogue or a dinner.

Tuckett criticizes this on the ground that such questions may not constitute a real break --for instance, Peter's question in 12:41 in the sustained teaching section 12:22-53. But then I am not alleging a "real break," only that Luke's insertion of the question serves the purpose of dividing an overlong section. More significantly, Tuckett points out that Luke has retained a long unit in Luke 21:5-38 (par. Mark 13) and notes that Matthew has transferred part of this chapter to Matthew 10, and Luke could have done something similar. Furthermore there are some pretty lengthy sermons in Acts, not least Stephen's speech in Acts 7.

The objection here is a failure of imagination. In Mark 13 Jesus is outlining the three phases of future history: first the beginning of birth pangs, then the great tribulation, and finally the coming of the Son of Man. No evangelist could break this up without doing violence to the topic; Matthew has indeed taken the persecution of the church in Matthew 10, but then he has substituted similar material of his own in 24:9-14. One needs a longer span to cover the things to come in their sequence; and the same is true of Stephen's short history of the Jewish people in Acts 7. But there is an important difference between the Lord's teaching and the apostolic preaching: Jesus' words are the deliverance of the Son of God and are pearls beyond price, while a Petrine or Pauline sermon is merely good currency.

The force of my argument may be seen from modern church use. Prayer books across the spectrum of churches provide a Gospel reading for each Sunday and Holy Day, normally of about the length I have suggested for Luke -- a pericope or a short teaching section. The Sermon on the Mount is not read as one lection. But in Holy Week an exception is made, and there may be a seventy-five-verse Gospel from Matthew 26: Holy Week is special. So Luke may be thought to have provided a sequence of lessons of digestible length, but a fuller chap. 21 when needed (and chap. 15 when the Spirit took him). The [679] Gospel takes three minutes, the sermon ten; but then the vicar, like the apostles, is not so inspired as the Lord and needs a bit more time.

VIII. Luke's Ordering of the Major Matthean Expansions

In general the "Q material" follows the same order in Matthew and Luke -- the Baptist's preaching, temptations, sermon, centurion, mission discourse, Beelzebul, signs, eschatology --though with some notable variations, and these are particularly striking after Luke 13. Since Matthew collects the material in topics into his discourses, Vincent Taylor was able to argue that Luke preserved the order of Q, which Matthew then improved: (33) he did not feel obliged to account for the detailed order of Q itself. I took Matthew's careful order as primary, and this left the question of the order of Luke.

For the detail I must refer the reader to my Luke: but one part of the solution offered is criticized by Tuckett, which I will comment on. By chap. 13 Luke seemed to have reached Matthew 25 ("the door was shut -- Lord, open to us -- I know you not"), having left out considerable portions of Matthew 16-24. In 13:31-35 he virtually transcribes Matt 23:37-39, "Jerusalem, Jerusalem. . . ." In 14:1-14 he expounds the oppressive rulings of lawyers and Pharisees (Matt 23:2-4), their liking for chief seats at dinners (Matt 23:6), and the exaltation of the humble over the proud (Matt 23:12). In 14:15-24 he tells the parable of the great dinner (cf. Matt 22:1-14). In 14:28-32 he has the parables of the tower builder and the embassy sent to make peace, features occurring in the tenants parable in Matt 21:33-40, with the casting out of the salt in place of the unready guest of Matt 22:13. In 15:11-32 he gives the parable of the two sons, a more elaborate form of Matthew's parable of the two sons in Matt 21:28-32. In 16:1-13 he tells the parable of the steward who remitted the tenants' debts, a theme continuous with the parable in Matt 18:23-35: Peter is told to forgive all offenses and is contrasted with the king's servant, who refused to forgive his debtors. Luke 16:14-31 takes up another element in Matthew 18, the torments of the rich man in hell, while also drawing momentarily on the forbidding of remarriage (16:18; Matt 19:9) and the perils of riches (Matt 19:24). Luke 17:1-10 reproduces material on offenses, forgiveness, and faith found in Matt 18:6-21 and 17:20. Luke 17:20-18:8 speaks of the signs that will precede the coming of the Son of Man, material that combines elements of Matt 16:4-28 with Matthew 24.

The suggestion of this little sequence is that, having reached the end of the Matthean teaching in a first cursory run, Luke went back up his Matthean scroll and took the topics in order as he came to them. There are nine sections here in [680] the reverse order of Matthew 16-23, with the Markan material broadly omitted. In a rhetorical flourish I commented, "Not one diamond shall be lost from the Matthean tiara: all must be included." Tuckett objects to this: (1) Often Luke provides not the Matthean diamond but a substitute, for example, the casting out of the salt in Luke 14:35 for the casting out of the unready guest in Matt 22:13. (2) Some diamonds do seem to have gone missing, like the labourers parable in Matthew 20. (3) Sometimes Luke is said to be including elements that Matthew had taken over from Mark, like the tower and embassy, which are to be included in Luke's Markan section, in Luke 20:9-17. (4) Sometimes there are leaps that do not follow the order: for example, the divorce saying in Matt 19:9 and the rich man of Matt 19:24 are out of sequence and should, on the theory, come before Luke 16:1-13, based on Matt 18:23-35.

Scholarship is often a matter of judgment, and I must leave it to my colleagues whether I have forced the text or Tuckett has been pedantic. When I spoke of diamonds, I was comparing Luke to a master jeweler, breaking up a Victorian piece to reset it, polishing and cutting the stones to more modern taste. Often he does not care for Matthew's emphases and restyles and regroups his predecessor's sayings: he is less willing than Matthew to see Christians sent to hell and is not the last reader to be scandalized by Matt 22:13-14; the murmuring of Matthew's Pharisees/labourers, their protest of long service and their expectation of better treatment all recur in Luke 15, but without the obvious unfairness of Matthew 20. Perhaps Luke had added the attractive little parables of the tower builder and the embassy sometime when preaching on the tenants and inserted them in chap. 14 despite his normal practice. Perhaps he did not think twentieth-century scholars would be calling him to account if his eye caught a line out of his general reverse-order strategy. Master jewelers take responsibility only for producing works of art, not for following handbook rules without modification.

IX. Paired Animal Imagery

Tuckett is writing an article, and he cannot be expected to comment on every argument I have used; but I have introduced one novel and, I think, cogent argument about which I notice that defenders of Q regularly keep silence. Caroline Spurgeon, in a classic treatment, Shakespeare's Imagery, contrasted Shakespeare's use of images with Marlowe's, Bacon's, and other contemporaries' and showed how often a particular use of images will recur in a particular author. (34) The use of imagery may be a more certain guide to authorship even than the vocabulary. Images are a basic element to the thinking of an individual mind.


The Gospels are full of imagery, but one striking set of images is animals; and we find that animal images often occur in pairs, the animals frequently being in some way symbolic. There are ten such pairs in the Gospel tradition:

Give not what is holy to dogs, and cast not your pearls before swine.

Or if he asks for fish, will he give him a snake?

Who come to you in sheep's clothing, but inwardly are ravening wolves

Foxes have holes, and birds of the air have nests.

I am sending you out as sheep in the midst of wolves.

So be wise as serpents and innocent as doves.

You strain out a gnat but swallow a camel.

You snakes, you brood of vipers!

As a hen gathers her chicks under her wings

As a shepherd separates the sheep from the goats

All ten of these paired animal images are in Matthew. None is in Mark or special Luke or John; three recur in Q passages in Luke.

How is this to be explained? Spurgeon argues that so striking a combination is likely to be the creation of a single mind. Jesus' perhaps, but then why have Mark and special Luke (and John) failed to pick up any of them? Ignoramus is a weak reply. Or it could be Matthew's mind, surely: all ten instances come in his Gospel (and other animals in pairs besides -- moth and rust, birds and lilies, sparrows and hairs, vultures and carcass). Mark would not have known his Gospel, and Luke could have selected a few appealing ones. Ah, but that would be, as Austin Farrer once said of another NT hypothesis, to touch the ark of the covenant, wouldn't it? It would mean halting the juggernaut.


1. D. Scholer helpfully collects lists of current work on Q in a series of SBL seminar papers: "Q Bibliography: 1981-1989," SBL 1989 Seminar Papers (Atlanta: Scholars Press, 1989) 23-37, with supplements in the 1990 Seminar Papers (pp. 11-13), 1991 Seminar Papers (pp. 1-7), 1992 Seminar Papers (pp. 1-4), 1993 Seminar Papers (pp. 1-5). See also A. Lindemann, "Literatur zu den Synoptischen Evangelien 1984-1991," TRu 59 (1994) 41-100, 113-85, 252-84.

2. A. Polag, Fragmenta Q (Neukirchen-Vluyn: Neukirchener Verlag, 1979); W. Schenk, Synopse zur Redenquelle der Evangelien (Düsseldorf: Patmos, 1981); see also J. S. Kloppenborg, Q Parallels: Synopsis, Critical Notes and Concordance (Sonoma, CA: Polebridge, 1987); and F. Neirynck, Q-Synopsis: The Double Tradition Passages in Greek (Studiorum Novi Testamenti Auxilia 13; Leuven: Leuven University Press, 1988).

3. The International Q Project has been under way since 1983. Its results are regularly reported in JBL: 109 (1990) 499-501; 110 (1991) 494-98; 111 (1992) 500-508; 112 (1993) 500-506; 113 (1994) 495-500; 114 (1995) 475-85.

4. R. A. Piper, ed., The Gospel behind the Gospels: Current Studies on Q (NovTSup 75; Leiden: Brill, 1995) 5-6: "[In Britain] the issue has remained the more basic one of whether Q existed at all."

5. Ibid., 19-47.

6. David Catchpole, The Quest for Q (Edinburgh: Clark, 1993).

7. A. M. Farrer, "On Dispensing with Q," in Studies in the Gospels: Essays in Memory of R. H. Lightfoot (ed. D. E. Nineham; Oxford: Blackwell, 1957) 55-88 [Note: Goulder is here referring to the reprint in 1957; the volume originally appeared in 1955, MSG]

8. Goulder, Luke: A New Paradigm (JSNTSup 20; Sheffield: Sheffield Academic Press, 1989). The first two, lengthy chapters offer a critique of Q as "the grandfather of all synoptic errors" (p. 27).

9. Piper, in Piper, ed., Gospel, 5.

10. "Michael Goulder [is] one of the most persistent and formidable of contemporary critics of the Q hypothesis" (Catchpole, Quest, 1); "[Goulder's] discussion is by far the most wide-ranging and comprehensive in the contemporary debate" (Tuckett, "Evidence," 32); "Two important works on the gospel of Luke were published in 1989: M. D. Goulder's Luke . . ." (F. Neirynck, L'Evangile de Luc: The Gospel of Luke [BETL 32; 2d ed., Leuven: Leuven University Press, 1989] 332).

11. "I consider [the Q theory] to be assured, despite all criticism. Neither the priority of Matthew [sc. the Griesbach theory] nor Proto-Luke hypotheses can resolve the Synoptic Problem" ("Q: Sayings of Jesus or Logia?" in Piper, ed., Gospel, 96). In Britain, Proto-Luke died with B. H. Streeter, and neo-Griesbach has never taken life.

12. I should add a further factor that has been significant in America: Prof. W. R. Farmer has for many years been an urgent advocate of the hypothesis that Mark was the last Gospel to be written and that there was no Q. Where the first-named part of the theory has seemed implausible, the second has naturally been dismissed also.

13. One such paradigm shift has been achieved by E. P. Sanders in his Paul and Palestinian Judaism (Philadephia: Fortress, 1977), which introduced a more realistic and respectful view of Judaism, and so a withdrawal from a Lutheran justification-by-grace exegesis of Paul. Sanders was able to achieve this because (a) he could show that none of the leading scholars had read the Jewish texts in their original settings, (b) earlier exegetes, on whose work more modern interpretation was based, had been religiously (and racially) biased, (c) the Holocaust had induced a change of climate, with a determination to be more fair to Jews, and (d) Sanders was both master of his subject and a writer of force.

14. See my Luke, 28-34: the association with Papias goes back to F. Schleiermacher in 1832.

15. Tuckett, "Evidence," 21.

16. Minor Agreements: Symposium Göttingen 1991 (ed. G. Strecker; GTA 50; G&ooml;ttingen: Vandenhoeck & Ruprecht, 1994), and literature therein referred to. The Minor Agreements are the problematic end of a spectrum for Q theorists. On my theory Luke sometimes copies Matthew word for word, sometimes with changes; sometimes he follows Mark with considerable, sometimes with minor, reminiscences of Matthew. With the Q hypothesis these are sorted into Q, Q?, Mark/Q overlaps, and MAs.

17. F. Neirynck, "TIS ESTIN O PAISAS SE: Mt 26,68 / Lk 22,64 (diff. Mk 14,65)," ETL 63 (1987) 5-47.

18. Neirynck, in Piper, ed., Gospel, 53.

19. Tuckett, "Evidence," 21.

20. Tuckett's instance is an unfortunate one. What is distinctive about Matthew's use of u(pokrith/j is not so much the number of uses (13/1/3) as the predominance of the vocative, u9pokrita//-ai/ (Matt 7:5QC; 15.7R; 22:18R; 23:13QD; 23:15M; 23:23QD; 23:25QD; 23:29QD; Luke 6:42QC; 12:56QD; 13:15L). [QC = Q words common to Matt and Luke; QD = Q words in which they differ; R = Redaction]. Mark 7:6 has peri\ u9mw~n tw~n u9pokritw~n. I am grateful to Dr. Mark Goodacre for this point.

21. C. M. Tuckett, 'The Beatitudes: A Source-critical Study," with a reply by M. D. Goulder, NovT 25 (1983) 193-216. 1 have argued the priority of the Matthean Lord's Prayer in "The Composition of the Lord's Prayer," JTS n.s. 14 (1963) 32-45, and, with a slight restatement, in Luke, 495-502.

22. Tuckett is arguing against my "hard-line" position that Luke took all the "Q material" from Matthew. Luke might, of course, also have had access to earlier forms preserved orally; and these would then be part of a more extended written-and-oral Q.

23. Neirynck, in Piper, ed., Gospel, 57-58.

24. 24. J. Schlosser, "Lk 17,2 und die Logienquelle," Studien zum Neuen Testament und seiner Umwelt 8 (1983) 77.

25. Tuckett, "Evidence," 39.

26. Goulder, Luke, 6-11. I used the expression "hard-liners" for those who helpfully stuck to their guns -- in this case the Two-Source (Mark, Q) Theory -- while "soft-liners" fell back on further hypothetical lost documents / traditions.

27. Tuckett, "Evidence," 40.

28. A practical demonstration that my hypotheses are falsifiable is offered by a recent D.Phil. thesis, Mark Goodacre's "Goulder and the Gospels" (Oxford, 1994). Goodacre has devised tests for my hypotheses on Matthean vocabulary, Minor Agreements, Lukan creativity, and Gospel lectionaries: three of these pass the test and one fails. Tuckett was examiner of the thesis.

29. Tuckett, "Evidence," 40.

30. It is desirable, when suggesting solutions to the Synoptic Problem, to use one's practical imagination so far as is possible. I suggested that Luke may have marked his texts with a pen and put one of his scrolls/codices on the floor. Tuckett gives the impression that such speculations are out of court by putting them in joke quotation marks: but I think realia are all part of the discussion.

31. Tuckett, "Evidence," 42.

32. Ibid.

33. Vincent Taylor, "The Order of Q," JTS n.s. 4 (1953) 27-31; expanded in "The Original Order of Q," in New Testament Essays (Festschrift T. W. Manson; ed. A. J. B. Higgins; Manchester: Manchester University Press, 1959) 95-118.

34. C. F. E. Spurgeon, Shakespeare's Imagery and what it tells us (Cambridge: Cambridge University Press, 1935).

Return to Top

Return to Main Page -O- Order The Case Against Q (US) / UK

This article is © 1996 Society of Biblical Literature, reproduced with permission.
This file was first published on 9 December 1999; this file was last updated on 5 March 2002
Web Site © 1999, 2001, 2002 Mark Goodacre