¹ See, for example, Davies, W. D. and Allison, D. C., A Critical and Exegetical Commentary on the Gospel According to Matthew, vol. 1 (ICC; Edinburgh: T. & T. Clark, 1988) 98.Google Scholar

² Thus for many the Strongest argument for Marcan priority remains the implausibility of the Griesbach hypothesis, for example Sanders, E. P. and Davies, M., Studying the Synoptic Gospels (London: SCM, 1989) 92 and 117;Google Scholar and in more detail, Tuckett, C. M., The Revival of the Griesbach Hypothesis (SNTSMS 44; Cambridge: CUP, 1983).CrossRefGoogle Scholar

³ The term ‘fatigue’ is coined by Goulder, Michael in Midrash and Lection in Matthew (London: SPCK, 1974) 35Google Scholar; and then used again in Luke: A New Paradigm (JSNTSup 20; Sheffield: JSOT, 1989) 109–10Google Scholar, with references on 274, 291, 334, 335, 368, 381, 423 (twice), 481, 512, 640 and 667. The term is also used by Drury, John, Tradition and Design in Luke's Gospel (London: Darton, Longman and Todd, 1976) 130 and 137;Google Scholar cf. 91, 142 and 162; and by Gundry, Robert, Matthew: A Commentary on His Literary and Theological Art (Grand Rapids: Eerdmans, 1982) 10.Google Scholar

⁴ This term is also used in passing by Goulder, Midrash, 35.

⁵ Styler, G. M., Excursus 4 in C. F. D. Moule, The Birth of the New Testament (3rd ed.; London: Black, 1981) 285–316Google Scholar; this example 294. The same example is used to make the same point elsewhere, for example Kümmel, W. G., Introduction to the New Testament (ET, London: SCM, 1966) 48.Google Scholar

⁶ Likewise, Luke correctly calls him ότετραάρϰηϚ (3.19, 9.7; Acts 13.1; cf. Luke 3.1).

⁷ The wording is slightly different between Matthew and Mark, καÌ Éλυπήθη ό βασιλεύς(Matt 14.9) and кαί περίλυπος γενόμενος ό βασιλεύς (Mark 6.26). It is not necessary, however, that Matthew should retain the exact wording of Mark: the thesis is one of direct dependence rather than direct copying in minute detail. In this and the forthcoming examples, the fatigue occurs where changes at the beginning of a pericope require changes in the overall plot or story-line which do not get made, so revealing dependence on the source.

⁸ This is Goulder's explanation of the fatigue here: Midrash, 35 and 376–7. On this example see also Davies and Allison, Matthew, 107, ‘Matthew unwittingly disturbs the coherence of Mark 6.17–29’. Styler (‘Excursus’, 294) adds that the story is introduced by Matthew in ‘flashback’ (14.3, ’For Herod …’) which Matthew has forgotten by 14.13 (‘When Jesus heard this…’).

⁹ Riley, H., in Orchard, B. and Riley, H., The Order of the Synoptics. Why Three Synoptic Gospels? (Macon, Georgia: Mercer, 1987) 100,Google Scholar attempts to answer by offering a counter-example, Mark's change from ‘Baptizer’ in 6.24 where there is no parallel in Matthew to ‘Baptist’ in 6.25 where there is a parallel (Matt 14.8). This is not strong: ‘Baptist’ and ‘Baptizer’ are equally correct. Mark's variation is not unusual or surprising in the way that Matthew's ‘tetrarch’ and ‘king’ would be on the assumption of Matthean priority, to say nothing of the king's ‘grief’.

¹⁰ See n. 3 above. Goulder takes for granted his solution to the synoptic problem in his discussion of‘fatigue’ but he does not use the phenomenon as an argument for that solution.

¹¹ This example is taken from Goulder, Midrash, 35 and 319.

¹² It is worth adding that Mark 3.20 contains reference also to Jesu's family, thus forming a famous Marcan sandwich: house and family (3.20–1) – Beelzebub (3.22–30) – house and family (3.31–5), a sandwich disturbed by Matthew's rearrangements.

¹³ I am unaware of any study or commentary which has noticed this feature in Matthew.

¹⁴ One can see an example of the phenomenon in Matthew here: Mark 4.7 has the thorns choking (συμπνίγω) the seed, ‘and it yielded no grain’ (кαì ĸαρπòν ούĸ ἒδωкεν). In Matt 13.7 they only choke (πνίγω) the seed. In the Interpretation (Mark 4.1911 Matt 13.22), however, anxieties and love of riches choke (συμίγω) the word, ‘and it proves unfruitful’ (кαì ĸαρποςγίνεται).

¹⁵ Evans, C. F., Saint Luke (London: SCM, 1990) 374,Google Scholar notes that Luke is revising Mark's Interpretation rather than commenting on his own version of the parable.

¹⁶ Luke does, however, omit the element of immediacy in both the Parable (Mark 4.5∥Luke 8.6) and the Interpretation (Mark 4.16 ║ Luke 8.13).

¹⁷ Luke, 331, though Goulder does not refer to this as an example of‘fatigue’.

¹⁸ Fitzmyer, J., The Gospel According to Luke I–IX (Anchor Bible 28; New York: Doubleday, 1981) 582,Google Scholar notices that Luke has omitted to mention the house and suggests that ‘they sought to bring him in’ (5.18b) is his substitute for it.

¹⁹ Plummer, A., Gospel According to St Luke (ICC; Edinburgh: T. & T. Clark, 1896) 155,Google Scholar sees this but feels that ‘λέγοντες may be used of thought’s, comparing Luke 12.17 and Matt 21.25. Fitzmyer, , Luke, 584Google Scholar, likewise appeals to 12.17, but this is one of Luke's soliloquies. A closer parallel is Luke 20.14 (a redactional addition to Mark on the assumption of Marcan priority) where the tenants are clearly debating aloud, διελογίζοντοπρòςάλλήλουςλέγοντες …

²⁰ There may be even a third example of the phenomenon in the same pericope: Conzelmann, H. and Lindemann, A., Arbeitsbuch zum Neuen Testament (5th ed.; Tübingen: Mohr, 1980) 56–7Google Scholar, suggest that Luke's switch from νομοδιδάσĸαλοι in 5.17 to the Marcan γραμματεīς 5.21 (Mark 2.6) is evidence of Luke's use of Mark.

²¹ It is seen by Goulder, , Luke, 433Google Scholar, though he treats this as an example of ‘muddle’, on which see my Goulder and the Gospels (JSNTSup 133; Sheffield: JSOT, 1996) chapter 11.Google Scholar

²² Some witnesses read ĸώμην, and others τόπον ἒ ρημονbut this simply shows that some scribes also felt our difficulty. The readings are discussed by Streeter, B. H., The Four Gospels: A Study of Origins (London: Macmillan, 1924) 569Google Scholar and Creed, J. M., The Gospel According to St Luke (London: Macmillan, 1930) 128.Google Scholar

²³ On the lodging element, which is unique to Luke, see Cadbury, H. J., The Making of Luk-Acts (London: SPCK, 1958) 250,Google Scholar and ‘Luke's Interest in Lodgin’, JBL 45 (1926) 305–22.Google Scholar It is actually another possible example of careless editing on Luke's part: having added a reference to lodging, he is unable to develop this idea in the remainder of the pericope which naturally focuses on the feeding element (cf. the reference to ‘prayers’ as well as ‘fasting’ in Luke 5.33, an element that has dropped out in Luke 5.34 ∥ Mark 2.19; cf. Goulder, , Luke, 109).Google Scholar

²⁴ Explanations for the incongruity are lacking in the commentaries. The mystery is why Luke would want to relocate to Bethsaida. Nolland, J., Luke 1–9.20 (WBC 35A; Dallas: Word, 1989) 440,Google Scholar says: ‘No really satisfactory explanation has yet been offered.’. Marshall, I. H., The Gospel of Luke, A Commentary on the Greek Text (Exeter: Paternoster, 1978) 360,Google Scholar attempts one: ‘The indications are that they were further round the E side of the lake than Bethsaida, possibly in gentile territory (the Decapolis) where Jews might not be sure of a welcome.’

²⁵ One of the most famous instances is Mark 4.10–12; see also 1.2–3 and the question of the names of the women in Chapters 15 and 16, among others.

²⁶ I am of course aware that one often defines what is characteristic of Matthew and Luke by how they differ from Mark and thus one is partly bound into a circle. It is not, however, the only means of establishing what is characteristic of each evangelist and in each case one has to ask whether the ‘fatigue’ explanation is more or less plausible than the alternatives.

²⁷ The problem with the king ‘grieving’ in 14.9 is similar. Matthew has characteristically depicted Herod as weak but evil (14.5). Herod the Great is much the same – frightened (2.3) but a murderer (2.16).

²⁸ These verses are redactional additions to Mark on the assumption of Marcan priority. The phrase is also at Luke 5.15 (at the end of this pericope – a minor agreement) and Luke 14.25. Altogether, the phrase comes five times in Matthew, never in Mark and twice in Luke.

²⁹ The construction comes eleven times in Matthew, never in Mark and once in Luke. At least four of the instances are redactional additions on the assumption of Marcan priority (9.10, 9.18,17.5 and here).

³⁰ Cf. Bultmann, R., The History of the Synoptic Tradition (ET, 2nd ed., Oxford: Blackwell, 1968) 360–1,Google Scholar which gives several examples. Further, έν (μι + genitive occurs in 5.12 and ένμι in 4.31, both redactional additions to Mark on the assumption of Marcan priority; cf. also 6.12.

³¹ On Lucan introductions to pericopae in relation to Matthew and Mark, see further my Goulder and the Gospels, chapter 5.

³² The exception is Mark 6.3044∥ Luke 9.10–17 (Five Thousand), on which see n. 24 above.

³³ De Wette did use arguments from Marcan incoherence in favour of Mark's use of Matthew and Luke, citing, for example, Mark 5.15 ∥ Luke 8.35 in comparison with Luke 8.27: see Bo, Reicke, ‘Griesbach's Answer to the Synoptic Question’, in Orchard, B. and Longstaff, T. R. W., eds., J. J. Griesbach: Synoptic and Text–Critical Studies 1776–1976 (SNTSMS 34; Cambridge: CUP, 1978) 50–67Google Scholar (64). But it never seems to be the case that the Marcan incoherence involves Mark specifically writing characteristically at the beginning of a pericope.

³⁴ It ought to be added that two of the examples from Luke above would make sense also on the theory of Matthean priority. The parable of the Sower (Matt 13.1–9; 18–23) and the Feeding of the Five Thousand (Matt 14.13–21) both share the elements in Mark with which Luke is apparently becoming fatigued. With the Paralytic, however, this is not the case since Matthew (9.1–8) does not have any reference to the house that is mentioned in Mark and assumed in Luke.

³⁵ Farrer, Austin, ‘On Dispensing with Q’ in Nineham, D. E., ed., Studies in the Gospels, Essays in Memory ofR. H. Lightfoot (Oxford: Blackwell, 1955) 55–88Google Scholar; for Michael Goulder, seen. 3, above. Griesbachian scholars also explain the double tradition by Luke's use of Matthew: Farmer, W. R., The Synoptic Problem. A Critical Review of the Problem of the Literary Relationships Between Matthew, Mark and Luke (New York: Macmillan, 1964).Google Scholar

³⁶ This example is taken from Sanders and Davies, Studying, 95, though they do not use the term ‘fatigue’.

³⁷ The pericope is, of course, usually taken as a ‘Mark–Q overlap’, but the feature discussed here is in Matthew and Luke alone.

³⁸ Luke could, logically, be fatigued with Q and not Matthew. Sanders and Davies (ibid.) note the possibility but think that it is unlikely. I will comment on the general issue below. In this instance, the International Q Project Reconstruction is unsure about ν δ ιδάσĸωντςπόλεως έĸεíνης(Q10.10) – it has a probability of only (C) (’a hesitant possibility’) on the descending scale of (A) to (D), no because of the difficulty that Luke’s wording differs between 9.4 and 10.10, JBL 114 (1995) 475–85 (480).Google Scholar

³⁹ This example is taken from Goulder, Midrash, 289 and 441– 2 andLuke, 681.

⁴⁰ The International Q Project reconstructs the wording of Q 19.20 with ό [[ ἓ ἔ ερος]], the double square brackets indicating a ‘hesitant possibility’. The translation given is ‘And [[another]] came …’, JBL 114 (1995) 475–85 (484).Google Scholar however, if Luke had meant ‘another’, he would, no doubt, have avoided the article: cf. 14.15–24 (Great Banquet), ό πρ τος… ἓ τερος… ἕτερος (14.18,19, 20).

⁴¹ This feature is often noticed but the significance for the question of Luke’s sources is less often realised. Jeremias, J., The Parables of Jesus (6th ed.; ET, London: SCM, 1963) 61,Google Scholar writes: ‘In 19.13 Luke would seem to have ten servants, but the continuation (cf. especially ό τερος in 19.20) betrays that three was the original number.’

⁴² Because of the closeness in wording, the International Q Project reconstructs Q 19.24 just as it stands here in Matthew, and Luke, , JBL 114 (1995) 475–85 (484).Google Scholar

⁴³ Goulder, , Luke, 681,Google Scholar ‘a mna is an absurd term, a tip’.

⁴⁴ Further, the naming of the servant as ‘the one who has ten pounds’ is in line with Matthew’s naming throughout, ‘the one who received the five talents’ etc. (25.20, 22, 24, 28), and not with Luke’s ‘the first’, ‘the second’, etc.

⁴⁵ Cf. Evans, , Saint Luke, 665–6,Google Scholar ‘Luke’s “ten of his servants” (v. 13) is very odd, and points to slovenliness at some stage.’

⁴⁶ Cf. Goulder, , Luke, 110 and 378.Google Scholar

⁴⁷ Cf. Goulder, , Luke, 481.Google Scholar

⁴⁸ For Luke’s fondness for fives and tens, see my Goulder and the Gospels, chapter 15.

⁴⁹ The critical text is reproduced in JBL 109 (1990) 499–501Google Scholar; 110 (1991) 494–8; 111 (1992) 500–8; 112 (1993) 500–6; 113 (1994) 495–500,114 (1995) 475–85. I have drawn attention to the relevant parts in notes 38, 40 and 42 above.