1 ‘Q’, with the addition of quotation marks, indicates any entity (other than Mark) that is shared by both Luke and Matthew. Q, without quotation marks, indicates the conception derived from the 2DH and reconstructed by the International Q Project (IQP).

2 J. S. Kloppenborg, Q the Earliest Gospel: An Introduction to the Original Stories and Sayings of Jesus (Louisville/London: Westminster John Knox, 2008) 21: ‘No hypothesis is without its difficulties, and for any of the existing Synoptic hypotheses there are sets of data which the hypothesis does not explain very well.’ See also similar comments in J. S. Kloppenborg, ‘Is There a New Paradigm?’, Christology, Controversy, and Community: Essays in Honour of David Catchpole (ed. D. G. Horrell and C. M. Tuckett; NovTSup 99; Leiden/Boston/Cologne: Brill, 2000) 37.

3 Garrow, A., ‘Streeter's “Other” Synoptic Solution: The Matthew Conflator Hypothesis’, NTS 62.2 (2016) 207–26CrossRefGoogle Scholar.

4 Garrow, ‘Streeter's “Other” Solution”, 212–13.

5 Two Low DT passages with credible examples of internal Alternating Primitivity are: On Retaliation and Love of Enemies (Matt 5.38–48 // Luke 6.27–36), and Woe to the Scribes and Pharisees (Matt 23.23–36 // Luke 11.39–51). Cf. D. R. Catchpole, The Quest for Q (Edinburgh: T&T Clark, 1993) 23–6, 55–6.

6 Garrow, ‘Streeter's “Other” Solution’, 213–15.

7 Garrow, ‘Streeter's “Other” Solution’, 212–13.

8 As, for example, proposed in J. S. Kloppenborg Verbin, Excavating Q: The History and Setting of the Sayings Gospel (Minneapolis: Fortress, 2000) 104–11, esp. 109.

9 The Didache was rediscovered in 1873 by Philotheos Bryennios, who published the first critical edition in 1883. For further details of the discovery see, K. Niederwimmer, The Didache (trans. L. M. Maloney; Hermeneia; Minneapolis: Fortress, 1998) 19–21.

10 In a personal communication in 2004 Helmut Koester generously admitted that, when writing his ground-breaking volume Synoptische Überlieferung bei den apostolischen Vätern (TU 65; Berlin: Akademie-Verlag, 1957), ‘I did not dare to . . . ask the question: Why could Matthew not be dependent upon the Didache – in whatever form it existed at the time?’. Another influential volume, A Committee of the Oxford Society for Historical Theology, The New Testament in the Apostolic Fathers (Oxford: Clarendon Press, 1905) 24–36, similarly fails to countenance the notion that the Didache might be a source for the Gospels, despite a willingness to consider every other option.

11 That the Didache has a complex compositional history is very widely accepted. See, for example, W. Rordorf, ‘Does the Didache Contain Jesus Tradition Independently of the Synoptic Gospels?’, Jesus and the Oral Gospel Tradition (ed. H. Wansborough; JSNTSupp 64; Sheffield: Sheffield Academic Press, 1991) 396: ‘The Didache cannot, of course, be considered a homogenous text. Even those who attempt to attribute it to a single author must unhesitatingly grant that older material is used in it. This is especially true in the first five chapters.’ Also, J. A. Draper, ‘The Jesus Tradition in the Didache’, The Didache in Modern Research (ed. J. A. Draper; AGJU 37; Leiden: Brill, 1996) 74–5: ‘. . . the text shows signs of considerable redactional activity, which defies any theory of unity of composition, even allowing for the activity of an interpolator. The Didache is a composite work, which has evolved over a considerable period.’ See also the works cited in n. 12 below.

12 Two recent and full-scale treatments of the Didache's compositional history, A. J. P. Garrow, The Gospel of Matthew's Dependence on the Didache (JSNTSupp 254; London: T&T Clark International, 2004) and N. Pardee, The Genre and Development of the Didache: A Text-Linguistic Analysis (WUNT2 339; Tübingen: Mohr Siebeck, 2012), both assign Did. 1.3–5a to a pre-Gospel stage of the Didache's development (Garrow, Matthew's Dependence, 216–37; Pardee, Genre and Development, 183, 191).

13 Did. 1.3b–2.1 is commonly regarded as a later addition to the Didache on the grounds that these verses do not appear in the Doctrina Apostolorum. Garrow, Matthew's Dependence, 68–75, notes, however, indications that the Doctrina was, after all, aware of Did. 1.3–6.

14 Tuckett, along with most other scholars, treats Did. 1.3–5a and Did. 1.2 separately.

15 C. M. Tuckett, ‘Synoptic Tradition in the Didache’, The New Testament in Early Christianity: La Réception des Écrits Néotestamentaires dans le christianisme primitif (ed. J.-M. Sevrin; BETL 86; Leuven: Peeters, 1989) 197–230.

16 Tuckett, ‘Synoptic Tradition’, 89. This method, in instances where it may be applied, continues to command respect. See, for example, A. F. Gregory and C. M. Tuckett, ‘Reflections on Method: What Constitutes the Use of the Writings That Later Formed the New Testament in the Apostolic Fathers’, The Reception of the New Testament in the Apostolic Fathers (ed. A. Gregory and C. Tuckett; Oxford: Oxford University Press, 2005) 61–82, esp. 71; and S. E. Young, Jesus Tradition in the Apostolic Fathers (WUNT2 311; Tübingen: Mohr Siebeck, 2011) 45–67.

17 Tuckett, ‘Synoptic Tradition’, 230.

18 If the MCH is correct, the difficulties of reconstructing ‘Q’ are exponentially increased.

19 A. Gregory, The Reception of Luke and Acts in the Period before Irenaeus (WUNT2 169; Tübingen: Mohr Siebeck, 2003) 124. Tuckett receives similar criticism in Rordorf, ‘Does the Didache?’, 406–7; Garrow, Matthew's Dependence, 224; and Young, Jesus Tradition, 206.

20 Gregory, Reception, 124.

21 A. J. Bellinzoni, ‘Luke in the Apostolic Fathers’, Trajectories through the New Testament and the Apostolic Fathers (ed. A. Gregory and C. Tuckett; Oxford: Oxford University Press, 2005) 57.

22 Draper, ‘Jesus Tradition’, 84–5. Earlier in his discussion, Draper concludes: ‘In this group of sayings [1.3b–c], the Didache thus represents an independent text which cannot realistically be viewed as a harmony of the Gospels. It seems to have independent access to the tradition on which the Gospels also draw.’ (p. 83)

23 D. A. Hagner, ‘The Sayings of Jesus in the Apostolic Fathers and in Justin Martyr’, The Jesus Tradition outside the Gospels (ed. D. Wenham; Gospel Perspectives 5; Sheffield: JSOT, 1985) 241–2.

24 For example, Glover, R., ‘The Didache's Quotations and the Synoptic Gospels’, NTS 5 (1958) 12–29CrossRefGoogle Scholar; Draper, ‘Jesus Tradition’, 79–85, 90–1; Young, Jesus Tradition, 203–13; Hagner, ‘Sayings of Jesus’, 241–2; Rordorf, ‘Does the Didache?’, 396–412; Milavec, A., ‘Synoptic Tradition in the Didache Revisited’, JECS 11 (2003) 443–80Google Scholar, esp. 449.

25 Given our almost complete ignorance about the shape of traditions circulating in the first century, it is also always possible that the feature original to text ‘A’ was taken up by text ‘C’ and thence to text ‘B’. For a helpful discussion of factors relevant to assessing the probability of direct or indirect relationship see, A. Bellinzoni, ‘Luke in the Apostolic’, 46–52.

26 Under the influence of the Doctrina Apostolorum most scholars use ‘1.3b–2.1’ to denote the section inserted into the Didache's Two Ways. However, as noted above (n. 13), the Doctrina does not offer a secure insight into the prehistory of the Didache's Two Ways. If its influence is removed, then the logical starting point for the insertion of this group of sayings is Did. 1.3. The group of sayings continues until at least Did. 1.5a, but Did. 1.5b–6 may be a latter insertion to combat abuse of Did. 1.5a. Consequently, the insertion commonly referred to as Did. 1.3b–2.1 is, in the following discussion, referred to as Did. 1.3–5a. Further, I use the label ‘Sayings Catena’ to denote this group of sayings, instead of the more common, but rather less neutral, ‘Evangelical Section’. These details do not materially affect the case for Luke's use of Did. 1.2–5a.

27 Epistle of Barnabas 18–20 and 1QS 3.13–4.26.

28 F. E. Vokes, The Riddle of the Didache: Fact or Fiction, Heresy or Catholicism? (London: SPCK, 1938) 92 suggests that the Didachist may have made this change to ‘conceal the borrowing’. The weakness of this suggestion only serves to emphasise the puzzle.

29 C. N. Jefford, The Sayings of Jesus in the Teaching of the Twelve Apostles (VigChrSup 11; Leiden: Brill, 1989) 33.

30 Koester, Synoptische Überlieferung, 168–9 notes the rarity of the positive form of the Golden Rule. Thus, it appears in ancient sources only in Matt 7.12 (which, under the MCH, depends on Luke), 1 Clem 13.2c and Justin's Dial. 93.1. Koester notes, on this basis, that the positive form appears to have been introduced by the Gospels.

31 Niederwimmer, The Didache, 76 tries to deal with the anomalous status of Did. 1.4a by identifying it as a later gloss. However, as Garrow, Matthew's Dependence, 78–9 notes, it is difficult to detect a likely motive for such an awkward insertion. See also Draper, ‘Jesus Tradition’, 83.

32 A. K. Kirk, The Composition of the Sayings Source Q: Genre, Synchrony, and Wisdom Redaction in Q (NovTSup 91; Leiden: Brill, 1998) 163 notes that a hermeneutically open central gnome is sometimes set within other sayings designed to interpret and expand it.

33 There are no earlier examples of ‘love your enemies’, despite the appearance of similar sayings in Romans 12.14,20–21.

34 J. S. Kloppenborg, ‘The Use of the Synoptics or Q in Did. 1:3b–2:1’, Matthew and the Didache: Two Documents from the Same Jewish-Christian Milieu? (ed. H. van de Sandt; Assen: Van Gorcum/Minneapolis: Fortress, 2005) 105–29 argues that the compiler of Did. 1.3b–2.1 knew Luke and Matthew/Q. His essay variously illustrates the complexities entailed by this arrangement. For example, with reference to Did. 1.4, Kloppenborg proposes that ‘Didache's rather odd formulation depends logically on Luke's reformulation of Q. What is awkward about this explanation is that it requires imagining that the Didache is following Q or Matthew in 1.4bc but then prefers Luke's robbery scene over Q/Matt's lawsuit. This probably implies that the compiler of Did. 1:3a–2:1 is not looking at the text of the gospels (or Q), but rather harmonizing from memory’ (p. 126, emphasis added). When it comes to the Didachist's treatment of the saying in Luke 6.30 (discussed above), however, Kloppenborg requires the Didachist to behave as the opposite of a harmoniser, succeeding instead in ‘reformulating it as a separate admonition’ (p. 127, emphasis original).

35 This conclusion raises, of course, the question of whether Luke made further use of the Didache. This is the subject of a forthcoming project.

36 After the completion of this article my attention was drawn to the reconstruction of the order of Q proposed by D. R. Burkett, Rethinking Gospel Sources, vol. ii:The Unity and Plurality of Q (Atlanta: SBL, 2009) 90. He proposes that Luke's source originally had ‘love your enemies’ (Luke 6.27–8) immediately followed by the justification of this command (Luke 6.32–3). Remarkably, this ‘original’ sequence is what occurs in Did. 1.3.

37 Bellinzoni, ‘Luke in the Apostolic’, 48–50 and Young, Jesus Tradition, 65–6 note that one text is ‘accessible’ to another if it was written at an earlier date and in a theoretically accessible location. The ‘chain of use’ Did. 1.2–5a -> Luke 6.27–36 -> Matt 5.38–48 establishes that Did. 1.2–5a was accessible, in this sense, to Matthew. Incidentally, this chain also eliminates the possibility that Matthew was accessible to Did. 1.2–5a.

38 Similar conflation happens, for example, in: Matt 27.55–6 // Mark 15.40–1 // Luke 23.49; Matt 12.22–30 // Mark 3.22–7 // Luke 11.14–15,17–23; and Matt 24.23–8 // Mark 13.21–3 // Luke 17.23–4, 37b.

39 Garrow, ‘Streeter's “Other” Solution’, 212–13, 219–22.

40 Other examples of redactional elements of Luke 6.27–36 that reappear in Matt 5.38–48 are: the call to act as υἱός of the Father/Most High; the inclusion of the idea that God is generous to the evil (πονηρούς) and the good (Luke 6.36 // Matt 5.45); and the call to be merciful/perfect [καθ]ως ὁ πατὴρ ὑμῶν . . . (merciful/perfect) ἐστιν (Luke 6.36 // Matt 5.48).

41 These synopses are designed to show where Matthew's deviations from Luke are matched by the Didache. Matt // Did. verbal parallels are not highlighted, therefore, when Matthew's text most credibly comes from Luke. To make clear where Matthew deviates from Luke, however, all Luke // Matt verbal parallels are rendered in bold.

42 This raises the question of whether more of the Didache was known to Matthew. Detailed arguments for Matthew's knowledge of Did. 1.1–6, and most other parts of the Didache, are presented in Garrow, Matthew's Dependence.

43 The full complexity of these relationships, as commonly understood, is obscured by scholars' (understandable) preference for treating the relationship between Matt 5.38–48 and Luke 6.27–36 separate from the relationship between the Didache and the Gospels. Strategies to explain the former include: the presence of different recensions of Q, U. Luz, ‘Sermon on the Mount/Plain: Reconstruction of Qmt and Qlk’, SBL 1983 Seminar Papers (ed. K. H. Richards; SBLASP 22; Chicago: Scholars Press, 1983) 473–9; the influence of oral tradition, Dunn, J. D. G., ‘Altering the Default Setting: Re-envisaging the Early Transmission of the Jesus Tradition’, NTS 49 (2003) 139–75CrossRefGoogle Scholar, esp. 163–5; and Luke's rearrangement and interpretation of selections taken from Matthew, F. Watson, Gospel Writing: A Canonical Perspective (Grand Rapids: Eerdmans, 2013) 165–7. Explanations for the latter include: the use of shared traditions, Glover, ‘The Didache's Quotations’, 12–16, 25–9; the influence of oral transmission within a shared milieu, S. E. Young, Jesus Tradition in the Apostolic Fathers (WUNT2 311; Tübingen: Mohr Siebeck, 2011) 210–29, 283; the Didachist's use of free allusion, Tuckett, ‘Synoptic Tradition’, 199; oral composition modified under the influence of Matthew, D. C. Allison, The Jesus Tradition in Q (Harrisburg: Trinity, 1997) 90–2; depends on Synoptic texts derived from Q, Jefford, The Sayings, 38–53; and the Didachist's capacity to harmonise Luke and Matt/Q from memory, Kloppenborg, ‘The Use’ (cf. note 35). Each of these strategies appeals either to an additional intermediary source or sources, and/or to a particular flexibility in the way sources are treated. These complicating factors are compounded when the three sides of the triangle are brought together.

44 Garrow, ‘Streeter's “Other” Solution’, 213–14.

45 Cf. Garrow, ‘Streeter's “Other” Solution’, 215. This phenomenon is illustrated in Synopses 2 and 3, above. As Matthew conflates Luke 6.27–36 with Did. 1.2–5a he preserves the (necessarily more primitive) wording of the Didache more closely than Luke on a number of occasions, for example: ‘if someone strikes you on your right cheek, turn your other to him also’; ‘if someone forces you to go one mile, go with him two’; ‘do not even the Gentiles do the same?’; and ‘pray for those . . . persecuting you’. In the last three instances Matthew is judged to be more primitive than Luke in J. M. Robinson, P. Hoffmann, J. S. Kloppenborg, The Critical Edition of Q (Leuven: Peeters, 2000). In the case of ‘turn the other cheek’ and ‘give your shirt also’ the Critical Edition does not agree with the Didache's wording. Catchpole's reconstruction of this verse of Q (The Quest, 23–6) does, however, match the Didache.

46 For discussion of the differing compositional practices exhibited by Luke and Matthew, see Garrow, ‘Streeter's “Other” Solution’, 215–19.

47 E. Mazza, The Origins of the Eucharistic Prayer (Collegeville: Liturgical Press, 1995) 12–41, treating the prayers independently of their wider context, dates them prior to 49 ce.

48 Scholars dispute whether references to ‘the Gospel’ are to known canonical Gospels. Garrow, Matthew's Dependence, 141 concludes that Matthew's Gospel is ‘very probably’ in view.

49 This view is reflected, for example, in Rordorf, ‘Does the Didache?’, 409: ‘it is commonly accepted that the Didache comes from a marginal community’.

50 The most recent and detailed treatments of the Didache's compositional history are Garrow, Matthew's Dependence and Pardee, Genre and Development. There are, however, fundamental points of disagreement between these two studies, and with the many other treatments that predate them.

51 This concluding statement alludes to Goulder, M. D., ‘Is Q a Juggernaut?’, JBL 115 (1996) 669Google Scholar, where he complains that, if Q existed, ‘it is ex hypothesi older than the canonical Gospels and must have enjoyed enormous (probably apostolic) prestige’. The wider project, of which the current pair of articles is a part, includes the pursuit of the possibility that the original Didache did indeed enjoy apostolic prestige.