¹ Luce, T. J., Livy: The Composition of His History (Princeton, NJ: Princeton University, 1977) 185.Google Scholar

² Downing, F. G., ‘Compositional Conventions and the Synoptic Problem’, JBL 107 (1988) 69–85;Google Scholar‘A Paradigm Perplex: Luke, Matthew and Mark’, NTS 38 (1992) 15–36.Google Scholar

³ Because of the virtual absence of any parallels to John in the Gospel of the Ebionites, Bertrand, D. A. (‘L'Évangile des Ébionites: Une harmonie évangélique antérieure au Diatessaron’, NTS 26 [1980] 551)CrossRefGoogle Scholar suggests a terminus ad quem of around 150 John, CE.Kloppenborg, S. (‘Introduction to “The Gospel of the Ebionites”’ [unpublished paper printed at the University of St Michael's College, Toronto, 03 15, 1992] 8–9)Google Scholar agrees that ‘the early to mid-second century fits well’, on account of Irenaeus' testimony, the presence of the phrase μαρτύριον τοῦ Ὶσραηλ in one of the fragments (paralleled in Barn. 8.3), and similarities in wording between the Gospel of the Ebionites and passages of other early texts. George Howard (‘The Gospel of the Ebionites’, ANRW II.25.5 [1988] 4035–6) is less confident in assigning a terminus ad quem, but allows that if it was in fact the Gospel of the Ebionites to which Irenaeus made reference, then ‘its date can be assigned somewhere in the first three quarters of the second century’.

⁴ Neirynck, Frans, ‘The Apocryphal Gospels and the Gospel of Mark’, in The New Testament in Early Christianity (BETL 86; ed. J.-M., Sevrin; Leuven: Leuven University, 1989) 158.Google Scholar

⁵ Bertrand, ‘L'Évangile des Ébionites’; Howard, ‘Gospel of the Ebionites’, 4037. Kloppenborg has personally expressed to me some doubts as to whether the Gospel of the Ebionites represents an early gospel harmony produced by a close literary conflation of the synoptic texts.

⁶ Neirynck, ‘Apocryphal Gospels’, 160. I have numbered the lines, and printed some of the parallels in heavy type.

⁷ Bertrand, ‘L'Évangile des Ébionites’, 553. Although Bertrand lists Markan parallels, he does not believe that these were directly used by the author: ‘On peut ainsi isoler plusieurs emprunts à Matthieu et à Luc, mais aucun à Marc …’ (554).

⁸ Kloppenborg, ‘Testimonia and Fragments, “Gospel of the Ebionites“’ (printed April 13, 1992)5.

⁹ Howard, ‘Gospel of the Ebionites’, 4039.

¹⁰ Ibid.

¹¹ Some have argued that codices were used as early as the time of the composition of the synoptic gospels. See Roberts, C. H., ‘Books in the Graeco-Roman World and in the New Testament’, Cambridge History of the Bible 1 (ed. Ackroyd, P. R. and Evans, C. F.; Cambridge: CUP, 1970) 53–6.Google ScholarThrockmorton, B. H. (Gospel Parallels: A Synopsis of the First Three Gospels [3rd ed.; Nashville and New York: Thomas Nelson, 1967]Google Scholar v) remarks that, although the papyrus codex probably goes back to the first century CE, the earliest extant example of the gospels in codex form comes from the third century, and even then the use of codices was not widespread. Of the 304 extant pagan manuscripts of the third century, 275 are papyrus rolls (90%), 26 are papyrus codices, and 3 are vellum codices.

¹² For Bertrand (‘L'Évangile des Ébionites’, 554, 559) only the two scrolls of Matthew and Luke would have been used (see n. 7).

¹³ Kloppenborg, ‘Testimonia and Fragments’, 6; Howard, ‘The Gospel of the Ebionites’, 4039; Bertrand, ‘L'Évangile des Ébionites’, 553.

¹⁴ Kloppenborg (‘Testimonia and Fragments’, 5) remarks that the ‘text of Epiphanius appears to be defective’. Bertrand (‘L'Évangile des Ébionites’, 553) offers another possibility: ‘… la coupure s'explique par un saut d’ Ανδρέαν à Άλφαίου'.

¹⁵ P. G. Walsh as repr. by Luce, Livy: Composition, xv.

¹⁶ Pelling, C. B. R., ‘Plutarch's Method of Work in the Roman Lives’, JHS 99 (1979) 96.CrossRefGoogle Scholar

¹⁷ Petersen, William L., ‘Textual Evidence of Tatian's Dependence upon Justin's Apomnemoneumata’, NTS 36 (1990) 520–1.CrossRefGoogle Scholar Both Bellinzoni, A. J. (The Sayings of Jesus in the Writings of Justin Martyr [Leiden: E. J. Brill, 1967])CrossRefGoogle Scholar and Koester, Helmut (Ancient Christian Gospels: Their History and Development [London: SCM/Philadelphia: TPI, 1990] 360–402)Google Scholar hold that Justin relied on more than one written source. According to Koester (Ancient Christian Gospels, 374, 378), Justin's sources are ‘several gospel writings among which Matthew and Luke predominate’ together with a ‘gospel harmony’ for the narrative material; but for the sayings material, ‘catechetical collections’ or ‘clusters’ of ‘sayings which were composed on the basis of harmonized gospel texts and which incorporated additional sayings from the non-canonical tradition … grouped together thematically before their use by Justin in his writings’.

¹⁸ Luce, Livy: Composition, 143–4.

¹⁹ E.g., Petersen, ‘Textual Evidence’, 513, 533–4, and Baarda, Tjitze, ‘ΔΙΑΦΩΝΙΑ – ΣΥΜΦΩΝΙΑ: Factors in the Harmonization of the Gospels, Especially in the Diatessaron of Tatian’, in Gospel Traditions in the Second Century: Origins, Recensions, Text, and Transmission (ed. Petersen, William L.; Notre Dame: University of Notre Dame, 1989) 142.Google Scholar

²⁰ A survey of the quotations presented in both Bellinzoni (Sayings) and Koester (Ancient Christian Gospels, 360–402) amply demonstrates this fact. One striking exception is Dial. 49.3 = Matthew 3.3_legacy1–12 = Luke 3.3_legacy6b–17, the passage on John the Baptist's proclamation of the one who is coming (see Koester, Ancient Christian Gospels, 391–2). Significantly, this is also one of the sayings displaying the highest degree of verbatim agreement to be found in the synoptic tradition. Another quotation in almost exact agreement with a synoptic parallel is 1 Apol. 34.1 = Matt 2.6 (from the Matthean special tradition). This quotation, based on Mic 5.1, does not follow the text of the LXX, but follows the form as presented in Matt 2.6 (see Koester, Ancient Christian Gospels, 382).

²¹ On the odd occasion, the harmonizer inserts explanatory phrases. The Diatessaron also contains Western readings and some non-canonical material.

²² Downing, ‘A Paradigm Perplex’, 19–20.

²³ See the majority of the examples presented by Tigay, Jeffrey H., ‘Conflation as a Redactional Technique’, in Empirical Models for Biblical Criticism (ed. Tigay, Jeffrey H.; Philadelphia: University of Pennsylvania, 1985) 61–83.Google Scholar A perusal of Roger of Hovedon's conflation of Benedict of Peterborough and the Passio Sancti Thomae, as presented by Longstaff, Thomas R. W. (‘The Minor Agreements: An Examination of the Basic Argument’, CBQ 37 [1975] 184–92),Google Scholar reveals that even a thousand years later block-by-block conflation in which the combined material is transcribed almost perfectly verbatim continued to be practised. It is curious that Longstaff would find in the conflational patterns of Roger of Hovedon evidence to support the Griesbach hypothesis (also see Evidence of Conflation in Mark? A Study in the Synoptic Problem [SBLDS 28; Missoula: Scholars, 1977]). Throckmorton's, B. H. (‘Mark and Roger of Hovedon’, CBQ 39 [1977] 103–6) critique of Longstaff is similar to my own.Google Scholar

²⁴ According to Frend, W. H. C. (Rise of Christianity [Philadelphia: Fortress, 1984] 217),Google Scholar it is only in Syria and towards the Euphrates frontier that Marcionism took permanent roots, where it ‘may have been the faith of the majority of Christians … throughout the third century’. It can scarcely be considered coincidental that it was specifically in Syria in the third and fourth centuries that the Diatessaron gained preeminence over the canonical gospels.

²⁵ See Downing, ‘Compositional Conventions’, and ‘A Paradigm Perplex’.

²⁶ Downing, ‘Compositional Conventions’, 70.

²⁷ Ibid., 75.

²⁸ Downing, ‘A Paradigm Perplex’, 20.

²⁹ E.g., see Downing, , ‘Redaction Criticism: Josephus' Antiquities and the Synoptic Gospels (II)’, JSNT 9 (1980) 29–48;Google Scholar and Talbert, C. H., ‘Oral and Independent or Literary and Interdependent? A Response to Albert B. Lord’, in The Relationships Among the Gospels: An Interdisciplinary Dialogue(ed. Walker, William O. Jr; San Antonio: Trinity University, 1978) 95–9.Google Scholar

³⁰ Downing, ‘Redaction Criticism (II)’, 29.

³¹ See Feldman's, Louis H. survey of this debate in ‘A Selective Critical Bibliography of Josephus’, in Josephus, the Bible, and History (ed. Louis, H. Feldman and Gohei Hata; Detroit: Wayne State University, 1989) 352–3.Google Scholar

³² Downing, , ‘Redaction Criticism (II)’. Also see Part I of this article in JSNT 8 (1980) 46–65.Google Scholar

³³ On the odd occasion, official documents were transcribed word-for-word. For example, Walsh, P. G. (Livy [Oxford: Clarendon, 1974] 19)Google Scholar notes that where Livy systematically recorded diplomatic exchanges between Rome and other states, formal agreements and treaties were apparently set down verbatim.

³⁴ It has been suggested by T. D. Barnes that ‘rewriting’ more accurately describes how classical authors adapted their sources than ‘paraphrasing’, as the latter term implies a closer rendering of the wording of sources than actually most often took place.

³⁵ Thackeray, H. St J., Josephus: The Man and the Historian (New York: Ktav, 1967) 107.Google Scholar

³⁶ Fornara, Charles W., The Nature of History in Ancient Greece and Rome (Berkeley: University of California, 1983) 48.Google Scholar Not all sources for contemporary historians were oral. Military notes, archives, official reports, memoirs of participants, letters, and other written accounts of the same events, were frequently used as supplementary sources.

³⁷ Schepens, G., ‘Arrian's View of His Task as Alexander-Historian’, Ancient Society 2 (1971) 256–7.Google Scholar

³⁸ Jane, Hornblower, Hieronymus ofCardia (Oxford: Oxford University, 1981) 28.Google Scholar

³⁹ Pelletier, André, Flavius Josèphe, Adapteur de la Lettre d'Aristée (Paris: Klingksieck, 1962) 221.Google Scholar

⁴⁰ Sid Leiman, ‘Josephus and the Canon of the Bible’, in Josephus, the Bible, and History, 55.Google Scholar

⁴¹ While the synoptic discourses are set up to be speeches of Jesus, in their form they find much closer analogues in hellenistic gnomologia and chriae collections. See Kloppenborg's discussion of genre for the sayings designated to ‘Q’ in The Formation of Q: Trajectories in Ancient Wisdom Collections (Philadelphia: Fortress, 1987) 289–316.Google Scholar The discourses may stem from early catechetical collections of sayings that were later interpolated into the biographical frameworks of the gospels. Koester (Ancient Christian Gospels, 49–162) suggests that such collections were common in the early church.

⁴² For discussions of Tacitus' adaptation of the Claudian Tablet, see Miller, N. P., ‘The Claudian Tablet and Tacitus: A Reconsideration’, in Rheinisches Museum für Philologie (Frankfurt am Main: J. D. Sauerländers, 1956) 304–15;Google Scholar and Griffin, M. T., ‘The Lyons Tablet and Tacitean Hindsight’, CIQ 32 (1982) 404–18.CrossRefGoogle Scholar

⁴³ See Rosché, Theodore ‘The Words of Jesus and the Future of the “Q” Hypothesis’, JBL 79 (1960) 210–20.Google Scholar This does not mean that all sayings exhibit high verbatim agreement. The ‘Q’ sayings, for instance, range from very high to quite low agreement, the majority (68%) having a level of correspondence below their mean (Rosché, ‘Words’, 215–18). But the overall agreement is higher for sayings in both the triple and double traditions than for the narrative material.

⁴⁴ Perhaps the closest match to the synoptic pattern of agreement is displayed between rabbinic anecdotal parallels (see Alexander, Philip, ‘Rabbinic Biography and the Biography of Jesus: A Survey of the Evidence’, in Synoptic Studies: The Ampleforth Conference of 1982 and 1983 [ed. Tuckett, C. M.; Sheffield: JSOT, 1984] 29–33Google Scholar). Jeremias, Joachim (‘Zur Hypothese einer schriftlichen Logienquelle Q’, ZNW [1930] 147 n. 2)Google Scholar comments: ‘Wieviel aus diesem synoptischen Problem für das synoptische Problem der Evangelien und umgekehrt zu lernen ist, das ist leider nur wenig bekannt.’

⁴⁵ Downing, ‘Redaction Criticism [II]’, 33.

⁴⁶ Honoré, A. M. has attempted to subject this variability to statistical analysis. See his coefficients of variance and coefficients of variance of variance (‘A Statistical Study of the Synoptic Problem’, NovT 10 [1968] 127, 131, 144–7).Google Scholar Honoré strives to draw conclusions from these c.v.s regarding the ‘policies’ of the derivative authors towards their sources. But these c.v.s are so high and are spread over such a wide range (see ‘Appendix C’, 144–7) that it is difficult to know what to conclude statistically from them except what is already patently obvious – that the pattern of verbatim agreement over the synoptic data is highly variable and erratic. This is particularly the case for the ‘Q’ sections (the Matt-Luke double tradition gives the highest overall c.v. – see Table 13, 127). John, O'Rourke (‘Some Observations on the Synoptic Problem and the Use of Statistical Procedures’, NovT 16 [1974] 277)Google Scholar offers a similar critique of Honoré's c.v.s to my own.

⁴⁷ Luce, Livy: Composition, 205–6.

⁴⁸ The following analysis is based on Josephus (ed. and trans. Thackeray and Ralph Marcus; vol. 5; London and Cambridge, MA: Heinemann and Harvard University, 1934)Google Scholar and The Septuagint with Apocrypha: Greek and English (Lancelot C. L. Brenton; Grand Rapids, Mich.: Zondervan Pub. House, 1986?).Google Scholar

⁴⁹ Downing (‘Redaction Criticism [I]’, 61) argues that Josephus ‘renders almost every phrase’ of 1 Chron 10.1–10 and 1 Sam 31.1–10 because he finds the two accounts so largely in agreement and is therefore convinced of their accuracy. In Against Apion 1.1–46, Josephus stresses the importance of unanimity of sources. This is probably also the motivation behind his careful harmonization of the divergent accounts of Saul's death (he evidently did not perceive the Amalekite to be a deliberate deceiver), for Josephus had criticized Greek historiography for its numerous contradictions, setting against it the harmony of the Jewish records. Given the extent to which he rewrites, ‘renders almost every phrase’ is nonetheless a somewhat misleading description of Josephus’ adaptation of scripture here, and Downing implies that Josephus had the scrolls of both 1 Chronicles and 1 Samuel unrolled before him while composing. This is impossible to prove for the very reason that the two accounts are so nearly identical. Josephus' main source for this section was most likely 1 Samuel.

⁵⁰ For instance, at Ant. 4.196–301, Josephus both condenses and rearranges the Mosaic code (see Downing, ‘Redaction Criticism [I]’, 57–60). He also omits embarrassing episodes, and elaborates and expands on others with romanticizing, dramatic additions. For a survey of these, see Talbert, ‘A Response to Albert B. Lord’, 97–8; and Feldman, ‘Josephus’ Jewish Antiquities and Pseudo-Philo's Biblical Antiquities’, in Josephus, the Bible, and History, 74.

⁵¹ The details fabricated include Saul and his sons ‘laying many of the Philistines low’ before their demise (Ant. 6.369), the massacre of the fleeing Hebrew host (Ant. 6.369), and the bodyguard of ‘the ablest men’ who accompany Saul in his flight and who perish forthwith (Ant. 6.370 and Ant. 6.372).

⁵² For instance, David Dungan (‘Mark – The Abridgement of Matthew and Luke’, in Jesus and Man's Hope: Pittsburgh Festival on the Gospels [ed. David G. Buttrick; Pittsburgh: Pittsburgh Theological Seminary, 1970] 63) quotes it as corroboration of the Griesbach hypothesis, while Talbert (‘A Response to Albert B. Lord’, 99) proposes it as a second ‘control’ that validates the two document hypothesis.

⁵³ Greek and English taken from: Arrian with an English Translation: Anabasis Alexandri, Books I-IV in Two Volumes (vol. 1; trans. E. IliffRobson; London: Heinemann/Cambridge MA: Harvard University, 1967)Google Scholar. The more literal translation in square brackets is my own.

⁵⁴ Hornblower, Hieronymus, 3. Hornblower refutes the complex ‘sceptical position’ (22) represented by A. Burton who is of the opinion that Diodorus spliced widely different sources together into a ‘framework of his own construction’ (21).

⁵⁵ Luce, Livy: Composition, xix.

⁵⁶ Pelling, ‘Plutarch's Method of Work’, 92. Walsh (Livy, 14) also corroborates Luce's model.

⁵⁷ Hornblower, Hieronymus, 49.

⁵⁸ This statement is based on a discussion of the topic with Barnes at the University of Toronto.

⁵⁹ Luce, Livy: Composition, xix, 74, 137–8, 179, 199, 204. For instances of conflation in Diodorus, see Hornblower, Hieronymus, 50, 94, 97.

⁶⁰ Stadter, Philip A., Arrian of Nicomedia (Chapel Hill: University of North Carolina, 1980) 70.Google Scholar

⁶¹ Ibid., 71.

⁶² Ibid.

⁶³ Tacitus does not bother to make note of their agreements: ‘My plan is to indicate such individual sources only when they differ. When they are unanimous, I shall follow them without citation’ (The Annals of Imperial Rome [trans. Michael Grant; Harmondsworth: Penguin Books, 1977] 293)Google Scholar.

⁶⁴ Pelling, ‘Plutarch's Method of Work’, 92.

⁶⁵ Pelling outlines some examples in ‘Plutarch's Method of Work’, 93–4. He provides a list of other scholars who argue for a large use of memory on the part of Plutarch when composing (74 n. 3). He seems to be of the opinion that Plutarch alternated between having the source before his eyes for the sections of his narrative displaying more careful reworking, and relying on memory for the sections reflecting freer composition (see ‘Plutarch's Adaptation of His Source-Material’, JHS 100 [1980] 130–1.Google Scholar

⁶⁶ Russell, D. A., ‘Plutarch's Life of Coriolanus’, JRS 53 (1963) 22.Google Scholar

⁶⁷ Luce, Livy: Composition, 212–15.

⁶⁸ Pelling, ‘Plutarch's Method of Work’, 92–3.

⁶⁹ Parássoglou, G. M., ‘A Roll upon His Knees’, Yale Classical Studies 28 (1985) 273.Google Scholar

⁷⁰ Downing, ‘A Paradigm Perplex’, 19.

⁷¹ Pelling, ‘Plutarch's Method of Work’, 95.

⁷² Wiseman, T. P., Roman Studies: Literary and Historical (Liverpool and Wolfeboro, NH: F. Cairns, 1987) 253–5.Google Scholar

⁷³ Alexander, Loveday, ‘Luke's Preface in the Context of Greek Preface-Writing’, NovT 28 (1986) 50.Google Scholar

⁷⁴ Ibid., 61.

⁷⁵ Ibid., 69.

⁷⁶ Ibid., 71.

⁷⁷ Luce, Livy: Composition, 185.