THE BIRTH OF THE NEW TESTAMENT by C. F. D. Moule, Lady Margaret's Professor of Divinity in the University of Cambridge. First published A & C Black Ltd 1962. This edition prepared for katapi by Paul Ingram 2013.




AFTER a century or more of discussion, it has come to be accepted by scholars almost as axiomatic that Mark is the oldest of the three synoptic gospels, and that it was used by Matthew and Luke as a source. This has come to be regarded as ' the one absolutely assured result' of the study of the synoptic problem.

It has also been usually agreed that, besides Mark, Matt, and Lk. shared another source of material, denoted by the symbol' Q'. Many have explored and accepted the hypothesis that it was a single clearly defined document, which can to a great extent be reconstructed. Others, however, have postulated a number of documents or traditions, known to both Matt, and Lk., sometimes in language closely similar, at other times less so. It may therefore be better to employ the symbol Q to denote the material common to Matt, and Lk. (but absent from Mk) rather than to denote a document, and thus prejudge the question of its unity. For the purpose or re-examining the priority of Mk, the unity of the Q-document or Q-material is irrelevant. But the validity of the Q hypothesis in some form or other is not wholly irrelevant, as will be seen.

The priority of Mk and the hypothesis of Q have been widely accepted in the present century, and are conveniently denoted by the name 'The Two-document Hypothesis', although it should be noted that the documents may well have been many more than two. The classical statement and defence was made by B. H. Streeter, [The Four Gospels (1924).] who attempted to reconstruct Q as a unitary document, but restricted it more narrowly than previous scholars. He gave the labels ' M' and 'L' to the material peculiar to Matt, and Lk., or (to be more precise) to the sources from which he took most of their peculiar material to be derived. Here again it may be noted that some scholars have been cautious in accepting the unity of the M or L material, and that since this material appears in only one gospel any reconstruction of its alleged source is even more speculative than the reconstruction of a Q.

It was not necessary to maintain that Mk's version must at every point be older than Matt.'s parallel version, since it was possible to say that anything in Matt, which in fact seemed more original than Mk could have been derived from Q. Further, there had been lingering doubts about the existence of Q. But it came as a shock when in 1951 Dom B. C. Butler published his book The Originality of St Matthew, attacking the Q-hypothesis and the priority of Mk at the same time. In a minutely detailed study he subjected both hypotheses to a severe criticism, and argued strongly for the priority of Matt. Mk, he argued, was dependent on Matt.; Lk. was dependent on Mk for the material which the two had in common, and on Matt, for the Q-material. Once the Q-hypothesis is abandoned, the priority of Matt., he claimed, quickly follows [Butler agreed with the general assumption that there is no case for maintaining the priority of Lk. ] from the existence of those passages in which Matt.'s text seems clearly more original than Mk's, or in some other way superior to it.

In spite of much close and careful reasoning, and the existence of at any rate some passages which tell in favour of Butler's conclusion, scholars have not abandoned the usual belief in the priority of Mk. In this Excursus it will not be possible to examine all Butler's arguments and instances one by one. [Nor to consider the various articles that have appeared subsequently.] But an attempt will be made to show that the belief in the priority of Mk is in fact securely grounded, and to make clear the principal arguments on either side, on which the decision must turn.

First of all, we must admit that Butler has exposed a serious logical error in many expositions of the priority of Mk. Many of its advocates have begun by stating certain formal relationships that hold between the three Synoptic Gospels, viz:

(i) The bulk of Mk is contained in Matt., and much of it in Lk.; there is very little that is not contained also in one or the other.

(ii) The Marcan material usually occurs in the same order in all three gospels; where Matt.'s order diverges from Mk's, Lk. supports Mk's order, and where (but this is very rare) Lk.'s order differs from Mk's, Mk's order is supported by Matt. In other words, Matt, and Lk. never agree with one another against Mk in respect of the order in which material common to all three Gospels is arranged.

(iii) The same relationship holds good for the most part in respect of wording. Matt, and Lk. are often closely similar in wording in Q-passages (i.e. where Mk has no close parallel), but in Marcan passages it is exceptional [There are a number of exceptions. Some of these may be coincidental; others are probably due to corruptions and assimilations in the course of the transmission of the text. (See Streeter, op. cit. 179 ff.) Perhaps there are still more than can easily be accounted for. But they remain few compared with the large number of passages when the generalization given in the text holds true. Butler agrees, and does not attempt to use them to establish Lk.'s knowledge of Matt.] for Matt, and Lk. to have significant words in common unless Mk has them also; frequently Mk and Matt, will share the same phrasing, while Lk. diverges, or Mk and Lk. will do so, while Matt, diverges.

From these facts it is clear that, as far as material common to all three is concerned, Mk is the middle term or link between Matt, and Lk.; Matt, and Lk. are not directly related here, but only through Mk.

Now it is obvious that the priority of Mk will satisfactorily explain these phenomena. But its advocates have made a serious mistake in arguing (or assuming) that no other hypothesis will explain them. [A variant theory is that the common source of Matt, and Lk. was not the Mk that we possess but an 'Ur-Markus', an earlier edition of Mk, similar but not identical. The only objection to this theory is that it seems to be unnecessary.] Butler is correct in claiming that they are guilty of a fallacy in reasoning. If Matt, were the original, followed by Mk with variations, and if in turn Lk. followed Mk, again with variations, these phenomena could well be the result. Alternatively, (as far as the phenomena go) Lk. might be the earliest of the three, Mk the second, and Matt, the last; Butler agrees that on other grounds that suggestion may be set aside. But he rightly insists that the phenomena are satisfied equally well by the standard view, that Mk is the common source of Matt, and Lk., or by the view which he supports, sc. that Matt, is the oldest Gospel, that Mk used Matt., and that Lk. in turn used Mk.

Butler is correct, therefore, in saying that the formal relationships do not by themselves compel one solution to the synoptic problem. The texts of the Gospels must be carefully studied side by side before we can decide on the question of priority. With this we agree. But we part company when he goes on to claim that this comparison points to the priority of Matt. Although the advocates of the priority of Mk have been wrong in claiming to establish it through the statistics etc., yet there are cogent arguments which retain their force. In the present writer's judgment Butler has not succeeded in destroying them, and they outweigh the arguments that Butler has adduced on his side.

Before turning to them, however, let us examine the place of the Q-hypothesis in the enquiry. It is relevant to the priority of Mk to this extent: there are passages, especially in sections of teaching, where Matt.'s version may well be judged more original than Mk's parallel. If the existence of a non-Marcan source is denied, it will be difficult to maintain that in such passages Mk is prior to Matt.; rather, they support the priority of Matt, to Mk. But if Matt, had access to a non-Marcan source, then there is no problem for advocates of Marcan priority. Q is therefore relevant, since it is just what is required – sc. a non-Marcan source. The evidence of Lk. therefore will be relevant to the enquiry if it supports the existence of Q. If, however the Q-hypothesis is rejected, the position of 'Marcan priorists' is weakened; they must still postulate a non-Marcan source for Matt, to have used, but they can no longer point to Q as constituting that source.

Butler's attack on Q, therefore, is an important preliminary to his attack on the priority of Mk. If the former succeeds it increases his chance of succeeding in the latter also. But only slightly. Marcan priorists will be driven on to more uncertain ground; they must now postulate an unknown source, instead of being able to point to one which is partially known. But in this there is no intrinsic improbability. [Butler rightly asserts the principle that sources should not be multiplied needlessly. But there is no objection to postulating such a source; and it may be necessitated by other reasons, viz. the arguments for the priority of Mk, if they prove to be sound.] In other words, it is perfectly possible to believe that Lk. obtained his so-called Q-material directly from Matt., and at the same time that Matt, obtained it from an earlier source, possibly known to Mk. [E.g. Dr A. M. Farrer holds (like Butler) that Lk. knew and used Matt, (as well as Mk), but continues to accept the standard view that Matt, used Mk. Cf. his essay 'On Dispensing with Q' in Studies in the Gospels, ed. D. E. Nineham (1955), 55-88, and St Matthew and St Mark (1954).]

It is agreed that Matt, has not borrowed from Lk. The two possible explanations of the fact that they have some closely similar passages are (a) that Lk. borrowed from Matt, and (b) that both depend on a common source (or sources), i.e. the Q-hypothesis. Any argument against (a) is therefore ipso facto an argument in favour of Q. There are three principal ones: first, that in at least some of the parallel passages Lk.'s version seems more original than Matt.'s; [Cf., e.g., Lk. vi. 20 ('Blessed are you poor') with Matt. v. 3 ('Blessed are the poor in spirit'). Although it is true that Lk's. version fits his own special infterests, it still seems to the present writer that it is nearer to the probable original than Matt.'s 'poor in spirit'. Cf. also Lk. iii. 8 with Matt. iii. 9; Lk.'s μὴ ἄρξησθε may well be nearer to the original than Matt.'s μὴ δόξητε.] secondly, that it is hard to see why, if Lk. has borrowed material from Matt, he has so violently and frequently disturbed Matt's. order; and thirdly, it seems inexplicable why he has consistently ignored [Or 'almost entirely ignored'; cf. p. 224, n. 3.] Matt, in any passage where he follows Mk, and has made no use of Matt.'s narrative wherever Mk has no parallel account – e.g. of the Nativity stories and Resurrection appearances.

Butler faces these arguments, and dismisses them. He denies that there are any cogent examples of Q-passages in which Lk.'s version is more original than Matt.'s; and he claims that Lk. makes no attempt to put material derived from Matt, into the Marcan context in which Matt, had it. He claims that it would be a tricky task to find out what that context was, since already there are large variations in order in the first half of Matt, and Mk. [Dr A. M. Farrer (in Studies in the Gospel, 67 ff.) argues that Lk.'s order is typological, and was never intended to reflect the order of Matt. But although his argument is persuasively written, the present writer finds his thesis incredible.]

But in spite of Butler's arguments, which deserve careful study, the present writer continues to find it hard to believe that Lk. used Matt.

After disposing of Q, Butler turns to the question of the relative priority of Matt, and Mk, and rightly asks that it should be discussed on the basis of a direct comparison of the parallel passages without prejudice. This is a fair challenge, which the 'Marcan priorist' need not evade. Butler himself examines a large number of passages, and claims that here on a straight comparison of Matt, and Mk the preference will nearly always go to Matt.'s version as the more original. The most convincing of his examples are cases [Cf., e.g., the preaching of John the Baptist.] where Matt.'s version is a coherent whole, and Mk's seems to be an excerpt, in which knowledge is betrayed of some phrase or fact which Mk does not reproduce. Unless we are allowed to appeal to Q, and say that it is some knowledge of Q [i.e., of Matt.'s source for this passage.] (not Matt.) which Mk betrays, Butler's conclusion does indeed seem to be forced on us. Rather less convincing are passages where Mk refers to Jesus' teaching, or to parables (in the plural), and goes on to produce only one example. Certainly Mk here betrays knowledge of more than he reproduces, and certainly Matt, shows more knowledge than Mk; but it is gratuitous to suppose that it must be from Matt, that Mk's knowledge comes.

Less convincing still are passages where Matt.'s version is more smooth than Mk's; rather, they tell the other way. In all his arguments of this type, Butler defeats himself; the better his defence of Matt, the harder it becomes to see why Mk should have altered something smooth into something less smooth. In fact, the relative roughness of Mk is one of the strong arguments on the other side. In textual criticism it is an accepted canon that, other things being equal, the harder reading is to be preferred, since it is more probable that the harder should have been altered to the easier than vice versa. Numerous examples can be produced in which, in one way or another, Matt.'s version looks easier than Mk's. They may be grouped under several heads:

(a) Grammatical variants, where Mk is wrong and Matt, correct; e.g. Mk x. 20 ἐφυλαξάμην (wrong), = Matt. xix. 20 ἐφύλαξα (right).

(b) Stylistic variants, where Mk is sprawling, and Matt, tidy; e.g. Mk x. 27 = Matt. xix. 26. Butler argues regularly for the superiority and greater originality of Matt, in such cases, on the ground that his version is closer to an authentic Semitic parallelism and even an original poetic strain in the teaching of Jesus. To the present writer it seems far more probable that Mk represents the earlier version, and that Matt, by careful rewriting has achieved a greater polish.

(c) There are the well-known examples [E.g., Mk iv. 38 (cf. Matt. viii. 25); Mk vi. 5-6 (Matt. xiii. 58); Mk. x. 17-18 (Matt. xix. 16-17), where Matt.'s question τί με ἐρωτᾶς περὶ τοῦ ἀγαθοῦ, and the appended comment εἷς ἐστιν ὁ ἀγαθός are intelligible as a rewriting of Mk, but most odd otherwise.] where Mk's version appears lacking in respect for the apostles or even in its estimate of the person of Christ, and where Matt.'sversionavoids all such implications. Without attempting to establish any rigid law of development, we must surely say that all such passages tell strongly in favour of the priority of Mk, and that Butler's attempts to evade this inference are unconvincing.

(d) In some passages Mk is suggestive but obscure, and Matt.'s parallel looks like an attempt to leave the reader with an edifying message; but we are left with the suspicion that Matt, has not penetrated to the real sense. Compare, e.g., Mk viii. 14-21 with Matt, xvi. 5-12 where Matt, interprets the 'leaven' against which Jesus warns his disciples as ' the teaching of the Pharisees and Sadducees'.

But the best instance is the difficult passage about the purpose (or effect) of parables. [Mk iv. 10-12 = Matt. xiii. 10-15.] Butler's treatment [Op. cit. 90-92.] of this leaves me quite unconvinced. Matt, seems here to be trying hard to extract a tolerable sense from the intolerable statement that Mk appears to be making, sc, that Jesus taught in parables to prevent the outsiders from having a chance of understanding and being converted. He assumes that Mk's ' all things are (done) in parables' means ' I speak in parables'. But recent commentators have suggested a line of interpretation of Mk's text which the present writer finds wholly satisfying; viz. that the same teaching is put before all by Jesus, but whereas some by God's grace penetrate to its inner meaning, for others it remains external, a parable and nothing more; [Or even 'a riddle'; the same word in Hebrew or Aramaic can mean 'parable' or 'riddle'. Cf. J. Jeremias, The Parables of Jesus (Eng. trans. 1954), 12-14. ] and herein the dark purpose of God, as predicted in Isaiah, is fulfilled. Mk may have partly misunderstood what he recorded; but it seems certain to the present writer that his words are closer to the original, and that Matt.'s version is an unsuccessful attempt to simplify what he found intolerable.

Another example of misunderstanding by Matt, may be claimed at Mk ii. 18 (= Matt. ix. 14). Mk's first sentence ('the disciples of John and the Pharisees were fasting') sets the scene; then ' they' ask Jesus why his disciples, unlike those of John and of the Pharisees, do not fast. The persons who ask are surely not the ones who were the subject of the previous sentence; they are persons unspecified. [For another clear example of Mk's 'impersonal' plural cf. v. 35. iii. 21 and 32 are other probable examples.] Matt.'s sentence is much shorter and neater. But he obviously assumes that ' they' are the disciples of John and the Pharisees.

We have passed on to an argument which to the present writer puts the priority of Mk beyond serious doubt, viz. that there are passages where Matt, goes astray through misunderstanding, yet betrays a knowledge of the authentic version – the version which is given by Mk. The two accounts of the death of the Baptist (Mk vi. 17-29, Matt. xiv. 3-12) contain clear examples of this. Mk states fully the attitude of Herod to John; he respected him, but was perplexed; and it was Herodias who was keen to kill him. And the story that follows explains how in spite of the king's reluctance she obtained her desire. Matt., whose version is much briefer, states that Herod wanted to kill John. But this must be an error; the story, which perfectly fits Mk's setting, does not fit Matt's introduction; and at xiv. 9 Matt, betrays the fact that he really knows the full version by slipping in the statement that 'the king' [Mk calls Herod 'king'; Matt, correctly calls him 'tetrarch' in xiv. 1, but lapses into calling him 'king* at xiv. 9. Butler attempts to base an argument for the priority of Matt, on his superior knowledge at this point; and also on the knowledge he displays from time to time of Jewish and Palestinian customs. Such arguments are tenuous, and are more than counterbalanced by Mk's superior knowledge of the story he is relating.] was sorry. It is surely clear that Matt., in a desire to abbreviate, has oversimplified his introduction.

Further, both Mk and Matt, relate this story as a 'retrospect' or 'flashback', to explain Herod's remark that Jesus was John risen from the dead. Mk quite properly finishes the story, and then resumes his main narrative with a jump; Matt., failing to remember that it was a ' retrospect', makes a smooth transition to the narrative which follows: John's disciples inform Jesus; and 'when Jesus heard ... etc.' (Matt. xiv. 12-13).

Butler rightly asks that the comparison should be made without prejudice. But, in the course of it, impressions are necessarily received. One impression which is received of Matt, is that he regularly aims at giving a smooth version, without any kind of roughness; and also that he is somewhat pedantic. [Cf. xxi. 2 ff., where he apparently takes Zech. ix. 9 to mean two animals. Cf. also xiv. 21, where he speaks of 5,000 men 'apart from women and children'; surely this is a pedantic gloss on Mk's plain ἄνδρες.] It is at least in line with this impression if, as the Marcan priorists maintain, Matt, regularly conflates Mk and his other source. Butler pours scorn on the suggestion, regarding it as cumbrous and pointless; but if Matt, really was pedantic he may not have thought it so.

Of all the arguments for the priority of Mk, the strongest is that based on the freshness and circumstantial character of his narrative. [Including touches that might well come almost directly from an eyewitness, e.g. the cushion in the boat, Mk iv. 38; and the Aramaic words and phrases, of which Mk preserves more than Matt.] Tradition connects his Gospel with St Peter, and this trait has strengthened the belief that the tradition may be sound. But it should be noticed that the same character is to be found even in narratives of events at which St Peter was not present. Butler concedes this quality to Mk, and explains it by a daring suggestion: he accepts the tradition that St Peter is often Mk's source – hence a vividness and wealth of detail greater than in Matt. – and saves the priority of Matt, by suggesting that St Peter himself [Or, presumably, it might have been that Mk, while using Matt, as his written source, called to mind the fuller account he had heard from St Peter. In any case, Matt, is acquainted with the same account as Mk.] had access to a copy of Matt, while speaking.

In effect, this suggestion amounts to the view that Mk had direct access to what was in fact Matt.'s ultimate source; to the authentic version of the story which Matt, has often abbreviated or modified. Clearly it makes dependence on Matt, unnecessary. These are the cases on which the priority of Mk strongly rests, and they counterbalance the rival group of passages, mostly teaching and not narrative, on which Butler relies for his hypothesis. In those, it will be remembered, Marcan priorists are sometimes on the defensive, and appeal to Q. [Cf. supra p. 223; 'Q' need not here mean a source common to Matt, and Lk.; all that is needed is a source used by Matt, which Mk knew, and occasionally used.]

For those passages,


In the narrative passages, the situation is exactly reversed; here the simple view is that Matt, depends on Mk; and it is now Butler who requires a three-term diagram, viz.


But whereas it seems perfectly credible to the present writer that (i) a non-Marcan source existed, known to both Mk and Matt., and (ii) Matt, conflated it with Mk, Butler's suggestion on the other hand seems to be pure fantasy. Our explanation of his favourable cases may be cumbersome; but his explanation of our favourable cases is incredible.

In the next place, Mk, if he is using Matt., has used only about 50 per cent of his subject-matter, but has expanded it in the telling. [But has expanded it in a natural way; the extra sentences seldom look like extraneous insertions.] But it is hard to see why he should have omitted so much of value if  he was using Matt.: not only the Sermon on the Mount and much teaching besides, but also the narratives of the Infancy of Jesus. Mk does include teaching; and so it cannot be replied that he was only interested in narrative. [True, he has in any case made his selection from the material available to him. Butler urges this point, and claims that if there is a difficulty here it is almost as great for us as for him. But surely not. It is harder to see why Mk should have omitted these if they were already incorporated in a full-scale document accessible to him.]

The point may be put like this: given Mk, it is easy to see why Matt, was written; given Matt., it is hard to see why Mk was needed. [i.e. by the early Christians. Mk is of course valued most highly by the modern scholar.]

If Matt, was using Mk, and incorporating other material, it is easy to understand why he should regularly have abbreviated Mk whenever he could safely do so. [The story of Jairus' Daughter is a good example (Mk v. 21-43 = Matt. ix. 18-26). Matt, omits one whole episode in the story. His version is more compact, but as a result it is historically far less credible; according to him, the father tells Jesus that his daughter has just died, and begs him to come and restore her to life. In Mk, the girl is near to death, and the father appeals to Jesus for help.] In spite of the postulates of the Form Critics, [The Form Critics postulate that conformity to a regular pattern is an indication that a passage goes back to an earlier date than one which does not conform to the pattern.] it is likely that Mk's sprawling and circumstantial stories are more original than Matt.'s shorter and more formal ones.

Lastly, an examination of Matt.'s additions tells heavily against his priority. Under this head there are two classes of passage,

(i) First, pieces of teaching included by Matt, but absent from the parallel section of Mk. Butler claims that Matt.'s whole context hangs together; and that if he has really inserted them into a framework provided by Mk he has done so with a felicity that is beyond belief. But, with some exceptions, [Viz. the passages where an overlap of Mk and Q is postulated by Marcan priorists.] this judgment will be challenged. Thus, in spite of Butler's claim that the famous Tu es Petrus passage has parallels or antitheses with both the preceding and the following verses, few will find him convincing. On the contrary, the passage will still seem to many to be an insertion into Mk's account of Peter's confession of faith [Mk viii. 27-33, cf. Matt. xvi. 13-23. Cf. also Matt.'s parable of the Labourers in the Vineyard (xx. 1-16), which is placed after πολλοὶ δὲ ἔσονται πρῶτοι ἔσχατοι καὶ ἔσχατοι πρῶτοι, and rounded off with a variant of the same logion. But the real point of the parable is different; in inserting it after this logion, Matt, has accentuated a secondary feature.] – although not, of course, necessarily an invention.

(ii) There are also some narrative additions in Matt, which seem to stem from later apologetic, or even from the stock of legendary accretions which are evident in the apocryphal Gospels. Butler argues strongly that any such judgment is premature and unwarranted; that if the detailed comparison of Matt, and Mk proves Matt, to be older, then that verdict must be accepted, and any suspicion that Matt.'s special narratives are ' late' must be mistaken. But since we hold that the detailed comparison of Matt, and Mk tells in the other direction, in favour of Mk's priority, then the judgment that Matt.'s narratives are late, and sometimes close to the legendary, must be given full weight.

In conclusion it should be said that, although Butler naturally gives most space to the passages which seem to tell in his favour, he does not attempt to conceal the fact that there are strong arguments on the other side. To some extent, however, he weakens their force by admitting them quickly, and attempting to explain them all away in a few words by his suggestion that St Peter had access to a copy of Matt. Until some less incredible explanation is forthcoming, the natural conclusion that Mk is prior to Matt, will continue to hold the field. [For a review of the way in which the priority of Mk came to be accepted, and the dubious arguments often used in its defence, cf. W. R. Farmer, ' A " Skeleton in the Closet" of Gospel Research', Biblical Research vi (1961), 3 ff., which the present writer had not seen at the time of writing. But it seems that, however insecure the arguments used in the past, the reasons for accepting the priority of Mk are in fact strong.]