Published online by Cambridge University Press: 05 February 2009
The origins and the theological significance of the fourfold Gospel raise a set of teasing questions. Why did the early Church eventually accept four partly parallel foundation documents? There is no precedent for this either in the OT Scriptures or elsewhere in earliest Christianity. Did retention of four gospels assist or hinder the early Church in the presentation of its claims concerning Jesus? No doubt to some, insistence that there were four gospels implied that there were basic flaws in the single gospels. Was the second century church's decision to bring together four separate gospels wise? What were, and what are, the theological implications of the fourfold Gospel? A critical theology cannot avoid asking these questions.
1 von Harnack, A., The Origin of the New Testament and the Most Important Consequences of the New Creation (London and New York, 1925) 69–72;Google ScholarZahn, Th., Grundriss der Geschichte des neutestamentlichen Kanons (Leipzig: Deichert, 2nd ed. 1904) 35–41.Google ScholarGoodspeed, E. J. dated the origin of the fourfold Gospel to c. 125: see his The Formation of the New Testament (Chicago: University of Chicago, 1937) 33–41.Google Scholar Five years later John Knox rejected Good-speed's arguments and opted for the West between 150 and 175 as the time when ‘we get our first glimpse of the existence of the fourfold Gospel’, Marcion and the New Testament (Chicago: University of Chicago, 1942) 140–67.Google ScholarCarroll, K. L., ‘The Creation of the Fourfold Gospel’, BJRL 37 (1954–5) 68–77Google Scholar echoed Knox's claim that the fourfold Gospel was an answer to Marcion.
3 See especially Hahneman, G. M., The Muratorian Fragment and the Development of the Canon (Oxford: Clarendon, OUP, 1992) 101.CrossRefGoogle Scholar Note also, ‘It is difficult there fore to acknowledge that the fourfold Gospel was “firmly established” in the last quarter of the second century’ (100; and cf. 108). Hahneman develops considerably the arguments for a fourth century date first advanced by Sundberg, A. C. in two articles: ‘Towards a Revised History of the New Testament Canon’, Studia Evangelica 4/1 (1968) 452–61;Google Scholar‘Canon Muratori: A Fourth Century List’, HTR 66 (1973) 1–41.Google Scholar
4 See below, 326–9.
5 Cullmann, O., ‘Die Pluralität der Evangelien als theologisches Problem im Altertum’, TZ 1 (1945) 23–42Google Scholar; ET, in Cullmann's, The Early Church (ed. Higgins, A. J. B.; London: SCM, 1956) 37–54.Google Scholar See also Morgan, R. C., ‘The Hermeneutical Significance of Four Gospels’, Interpretation 33 (1979) 376–88;Google ScholarBurridge, R. A., Four Gospels, One Jesus? (London: SPCK, 1994) 163–79.Google Scholar
6 See Benoit, A., Saint Irénée. Introduction à l'Étude de sa Théologie (Paris: Presses Universitaires de France, 1960).Google Scholar Benoit notes that in Book III, ‘Gospel’ is used in the singular 41 times, 12 times for a particular Gospel, only 6 times in the plural. See also Blanchard, Yves-Marie, Aux Sources du Canon, Le Témoignage d'Irénée (Paris: Cerf, 1993) 157,Google Scholar who counts 75 occurrences of ‘Gospel’ in Book III, only five of which are in the plural.
7 See especially the Preface to Book III, and III.1.1. For a helpful discussion of the verbs used to refer to the ‘reporting’ or ‘recording’ activity of the evangelists, see Blanchard (n. 6) 161.
8 O. Cullmann (n. 5) ET 50–2 claims that Irenaeus's justification of the fourfold Gospel ‘is based on the same fundamental error as the Gnostics’ “docetic” arguments against it’: his appeal to four as a ‘divinely ordained number’ left out of account the purely human circumstances of the formation of the fourfold Gospel.
9 Zahn, Theodor, Geschichte des neutestamentlichen Kanons (Erlangen: Andreas Deichert, 1888) 1, 153,Google Scholar scornfully rejects attempts to write off Irenaeus's arguments as ‘dogmatic assertions and theosophical trifles’.
10 ‘… in accusationem convertuntur ipsarum Scripturarum, quasi non recte habeant, neque sint ex auctoritate et quia varie sint dictae …’ I have used the edition of the Latin text in Irénée de Lyon. Contre les Hérésies, III.II (eds. A. Rousseau and L. Doutreleau; Sources Chrétiennes 211; Paris: Cerf, 1974)Google Scholar which includes the Greek fragments and a retroversion of the Latin into Greek.
11 See Th. Zahn (n. 9) II, 43 and Merkel, H., Die Widersprüche zwischen den Evangelien; Ihre polemische und apologetische Behandlung in der alten Kirche bis zu Augustin (Tübingen: Mohr, 1971) 42–3.Google Scholar
12 Matt 11.27 is cited in three different ways in IV.6.1, 3, 7.
13 Cf. A. Benoit (n. 6,117): ‘La justification irénéenne ne veut pas être une démonstration, elle ne fait qu'augmenter la crédibilité du fait accepté par ailleurs.’ Skeat, T. C., ‘Irenaeus and the Four-Gospel Canon’, NovT 34 (1992) 193–9,Google Scholar claims that in his celebrated identification of the four evangelists with the four living creatures of the Apocalypse, Irenaeus has used an earlier source. His case is strong but not conclusive, so I have not drawn on it in this paper.
14 See n. 3 above.
15 In his article ‘Muratorian Fragment’ in The Anchor Bible Dictionary (ed. D. N. Freedman; New York: Doubleday, 1992),Google ScholarRobbins, G. A. claims that Sundberg's thesis ‘has won considerable acceptance and further confirmation’ (4, 929).Google Scholar In his article ‘Can on, New Testament’ in the same Dictionary, H. Y. Gamble accepts Sundberg's thesis cautiously (I, 856), as does Koester, H., Ancient Christian Gospels: their History and Development (London: SCM; Philadelphia: Trinity, 1990) 243.Google Scholar
16 I have used the critical edition edited (with a facsimile reproduction) by Tregelles, S. P., Canon Muratorianus (Oxford: Clarendon, 1867).Google Scholar Tregelles's learned notes are still worth consulting. See also Lietzmann's, H. edition, Das Muratorische Fragment und die monar-chianischen Prologe zu den Evangelien (Kleine Texte i; Bonn, 1902).Google Scholar
17 Similarly, E. Ferguson in his critical review of Hahneman's monograph in JTS 44 (1993) 691–7:Google Scholar ‘The pseudonymity would seem to be of doubtful value in a polemic against the Shepherd in the fourth century.’ See also Ferguson's, E. discussion of Sundberg's theory in ‘Can on Muratori: Date and Provenance’, Studia Patristica 18 (1982) 677–83.Google Scholar
18 Cf. J.-D. Kaestli: ‘Par son contenu et par sa forme, le CM (Canon de Muratori) est plus proche du genre des “prologues” que de celui des “listes canoniques”’, ‘La place du Fragment de Muratori dans l'histoire du canon. À propos de la thèse de Sundberg et Hahneman’, Cristianesimo nella Storia 15 (1994) 609–34,Google Scholar here 616. Similarly E. Ferguson (n. 17) 696.
20 A. T. Ehrhardt suggests that the comment about Luke in lines 6–7, ‘dominum tamen nee ipse vidit in came’, is sufficient evidence ‘to assume that Papias was responsible for the fragmentary remark about St. Mark in the Fragment', Muratorian: ‘The Gospels in the Muratorian Fragment’ in his The Framework of the New Testament Stories (Manchester: Manchester University, 1964) 11–36,Google Scholar here 13; this article was first published in German in Ostkirchen Studien 2 (1953) 121–38.Google Scholar
21 See especially A. T. Ehrhardt (n. 20) 18–25.
23 See also Th. Zahn (n. 9) II, 43.
24 Contra Celsum II.27 (ed. and transl. by H. Chadwick; Cambridge: CUP, 1953) 90.Google Scholar See also V.56 where Origen responds to Celsus's jibes concerning the number of angels at the tomb of Jesus. For other evidence of the problems caused by differences in the gospels, see Merkel, H., Die Widersprüche (n. 11) and Die Pluralität der Evangelien als theologisches und exegetisches Problem in der alten Kirche (Bern: Peter Lang, 1978).Google Scholar
25 So too Merkel, H., Widersprüche (n. 11) 11.Google Scholar See also Baarda, T., ‘δ0IAωNIA–σYMΦωNIA: Factors in the Harmonization of the Gospels, Especially in the Diatessaron of Tatian’, in Gospel Traditions in the Second Century (ed. Petersen, W. L.; Notre Dame and London: University of Notre Dame, 1989) 133–5,Google Scholar reprinted in his Essays on the Diatessaron (Kampen: Kok Pharos, 1994) 29–48;Google Scholar R. M. Grant (n. 22) 59–60.
26 Stanton, G. N., ‘The Two Parousias of Christ: Justin Martyr and Matthew’, in From Jesus to John, FS de Jonge, M. (ed. de Boer, M. C.; Sheffield: JSOT, 1993) 183–96.Google Scholar
27 A. T. Ehrhardt (n. 20) 11 suggests that the Fragment was produced at Rome, probably under Zephyrinus, 197–217.
28 I am not convinced by A. T. Ehrhardt's claim (n. 20,14–15) that the reference to John as ‘ex discipulis’ in line 16 betrays its origin from Irenaeus who refers to John the evangelist as ‘the disciple of the Lord’, but never refers to him as the son of Zebedee; this similarity could have arisen from independent use of the same tradition.
29 H. von Campenhausen (n. 2) ET 173–4.
30 The Chester Beatty Biblical Papyri, Fasciculi I and II, The Gospels and Acts, General Introduction and Text (London: Emery Walker, 1933)Google Scholar here I, 13. þ45 is made up of quires of two leaves – a single sheet of papyrus folded in two. The letters are small; the scribe was no calligrapher.
32 þ53, third century, with fragments of Matthew and Acts in the same hand, is an interesting possible partial exception, though the fragments may not be from the same codex. If they are, and if the codex included all four gospels, it would have run to a highly improbable 300 to 350 leaves. So a codex with one or two gospels, plus Acts, is possible. Cf. Aland, K.Repertorium der griechischen christlichen Papyri 1 (Berlin and New York: De Gruyter, 1976) 53 and 283.Google Scholar See also van Haelst, J., Catalogue des Papyrus littéraires juifs et chrétiens, Papyrologie 1 (Paris: Publications de la Sorbonne, 1976) Nr 380, 381.Google Scholar
33 T. C. Skeat (n. 31) 264. In his most recent article, ‘The Oldest Manuscript of the Four Gospels?’, NTS 43 (1997) 31,Google Scholar Skeat mentions in passing the possibility that perhaps there were two volumes, i.e. two single-quire codices bound separately.
34 T. C. Skeat (1997, n. 33) 1–34.
35 See Stanton, G. N., Gospel Truth? New Light on Jesus and the Gospels (London: HarperCollins, 1995) 1–19.Google Scholar For Thiede's theory, see his ‘Papyrus Magdalen Greek 17 (Gregory–Aland þ64): A Reappraisal’, ZPE 105 (1995) 13–20;Google Scholar reprinted in Tyndale Bulletin 46 (1995) 29–42.Google Scholar The theory has been rejected by numerous scholars, all of whom accept Roberts's original date for þ64, late second century. See especially, Birdsall, J. Neville, ‘The Dating of the Magdalen Papyrus’, The Church Times, 6 01 1995;Google ScholarWachtel, Klaus, ‘þ64/67: Fragmente des Matthäusevangeliums aus dem 1. Jahrhundert?’, ZPE 107 (1995) 73–80;Google ScholarHead, Peter M., ‘The Date of the Magdalen Papyrus of Matthew (P. Magd. Gr. 17 = P64): A Response to C. P. Thiede’, Tyndale Bulletin 46 (1995) 251–85;Google ScholarParker, D. C., ‘Was Matthew Written before 50 CE? The Magdalen Papyrus of Matthew’, ET 107 (1995) 40–3;Google ScholarPickering, S. R. in NT Textual Research Update 2 (1994) 94–8 and 3 (1995) 22–5;Google ScholarGrelot, P., ‘Remarques sur un Manuscript de l'Évangile de Matthieu’, RSR 83 (1995) 403–5;Google ScholarElliott, J. K. in NovT 38 (1996) 393–9.Google Scholar
36 After I had completed the research summarized in this paragraph, T. C. Skeat wrote to me (10 July 1996) as follows: The two-column format became the standard form throughout the whole of the Middle Ages, and has survived almost down to the present day in printed bibles and prayer-books. Why? I have never seen this considered, but I suppose the answer must be that this is the easiest to read. Certainly reading the lessons from þ45 with its very long lines and small script would have been quite difficult, and no doubt it wasn't intended for liturgical use.’
38 In 1970 the NEB was published with one single wide column on each page. In 1989 the REB reverted to the more traditional two-column format, partly in order to facilitate reading aloud in the context of worship.
39 Cf. E. G. Turner (n. 37) 36–7. Van Haelst (n. 32) Nr 336, 336, even says that this is the oldest example of a codex with two columns; presumably he means the oldest Biblical codex, though in the light of Turner's list, this is a doubtful claim.
40 This is Skeat's phrase (n. 33) 26.
41 Roberts, C. H. noted that ‘in its handsome script as well as in its organization … it is a thoroughgoing literary production’. Manuscript, Society and Belief in Early Christian Egypt (London: The British Academy, 1979) 23.Google Scholar Although Skeat (n. 33, 2 and 6–7) has been unable to find any trace of two of the examples given by Roberts – three different positions for punctuation as well as omission and quotation signs – he concurs with Roberts's general conclusions.
42 þ45, the latest of the three, is the least impressive hand. Whereas þ45 is made up of quires of two leaves – a single sheet of papyrus folded in two, þ75 probably contained two single-quire codices bound together. þ64 + þ67 + þ4 is a two-column, single-quire codex.
43 C. H. Roberts (n. 41) 53; Roberts notes some similarities with þ64 +þ67 +þ4, 23.
44 See above, n. 35.
45 ‘Die “Erinnerungen der Apostel” bei Justin’, in Das Evangelium und die Euangelien (ed. P. Stuhlmacher; Tübingen: Mohr, 1983) 341–54, esp. 341.Google Scholar
46 It was noted already by S. Tregelles in 1867 (n. 16, 71): ‘no smaller number (than four) could be implied by the two groups’.
47 E. J. Goodspeed (n. 1, 38) is much less cautious: ‘Justin became a Christian at Ephesus as early as 135 AD, and probably there became attached to the fourfold gospel.’
50 Koester, H., ‘The Text of the Synoptic Gospels in the Second Century’, in Gospel Traditions in the Second Century (ed. Petersen, W. L.; Notre Dame and London: University of Notre Dame, 1989) 30, and cf. 32.Google Scholar
51 W. L. Petersen mentioned the latter possibility to me in a letter dated 1.9.1996.
52 Similarly A. J. Bellinzoni (n. 48); Osborn, E., Justin Martyr (Tübingen: Mohr, 1973) 132;Google Scholar and L. Abramowski (n. 45) 352–3.
53 Hengel, M., ‘The Titles of the Gospels and the Gospel of Mark’ in his Studies in the Gospel of Mark (London: SCM, 1985) 64–84.Google Scholar
54 Hengel also notes that a page from þ64 + þ67 +þ4, ‘which belong together’ has the inscriptio εþαγγέλιον κτὰ Mαθθαîον (n. 53, 66). However, this example should not be set alongside the evidence of þ66 and þ75, for the inscriptio (which is now located with the þ4 fragments of Luke) is not in the same hand; it probably comes from a fly-leaf added to the codex at a later point.
56 Cf. H. Koester (n. 55, 381 n. 1) and M. Hengel (n. 53, 74–81).
57 See R. M. Grant (n. 22) 119–20.
58 For fuller discussion, see Stanton, G. N., ‘Matthew: BIBΛOσ, EYAГГEΛION, or BIOσ?’ in The Four Gospels 1992, FS Frans, Neirynck (ed. van Segbroeck, F. et al. ; Leuven: University, 1992) 1187–1202.Google Scholar For a different view, see Gundry, R. H., ‘EYAГГEΛION: How Soon a Book?’, JBL 115 (1996) 321–5.Google Scholar
59 For discussion of the relationship of Luke and Acts, see Parsons, M. C. and Pervo, R. I., Rethinking the Unity of Luke and Acts (Minneapolis: Augsburg Fortress, 1993).Google Scholar
60 þ53 is apossible exception; see n. 32 above. Metzger, B. M., The Canon of the New Testament: its Origin, Development, and Significance (Oxford: Clarendon, 1987) 296,Google Scholar lists a few late examples of Luke coming fourth in a sequence of the four gospels, perhaps from a desire to bring the two books by Luke side by side.
62 See A. Benoit (n. 6) 122–7 for a fuller discussion.
63 Similarly, Farmer, W. R. and Farkasfalvy, D. M., The Formation of the New Testament Canon (New York, 1983) 73,Google Scholar though they date the emergence of the fourfold Gospel somewhat later.
64 See Knox, John, Marcion and the New Testament (n. 1);Google Scholar so also K. L. Carroll (n. 1).
65 H. von Campenhausen (n. 2) ET 171 n. 113. Elsewhere von Campenhausen accepts that Marcion was very influential on the emergence of the NT canon; for discussion, see Bovon, F., ‘La structure canonique de l'Évangile et de l'Apôtre’, Cristianesimo nella Storia 15 (1994) 559–76.Google Scholar
66 W. R. Farmer and D. M. Farkasfalvy (n. 63) emphasize the importance of the decision taken by Irenaeus and Anicetus to agree to disagree over the date of Easter: ‘there was no other moment in Church history when it is more likely that the fourfold Gospel canon was, in principle, implicitly agreed upon’, 72. D. Trobisch also emphasizes the importance of this decision, reported by Eusebius, H.E. V.24.14: Die Endredaktion des Neuen Testaments. Eine Untersuchung zur Entstehung der christlichen Bibel (Freiburg: Universitätsverlag; Göttingen: Vandenhoeck & Ruprecht, 1996) 158–9.Google Scholar
67 So too H. von Campenhausen (n. 2) ET 123: ‘It is highly questionable whether the idea is correct, that originally each individual Gospel had its own territorial domain.’
68 See especially Epp, E. J., ‘New Testament Papyrus Manuscripts and Letter Carrying in Greco-Roman Times’, in The Future of Early Christianity, FS Koester, H. (ed. Pearson, B. A.; Minneapolis: Fortress, 1991) 35–56;Google ScholarEpp, E. J., ‘The Papyrus Manuscripts of the New Testament’, in The Text of the New Testament in Contemporary Research, FS Metzger, B. M. (ed. Ehrman, B. D. and Holmes, M. W.; Grand Rapids: Eerdmans, 1995) 3–21, here 8–10.Google Scholar
69 F. G. Kenyon (n. 30, Fasciculus I, 12f.).
70 Roberts, C. H. and Skeat, T. C., The Birth of the Codex (Oxford: The British Academy, 1983).Google Scholar
71 The possible exception is þ22 (= POxy, 1228) third century fragments of John 15 now in Glasgow with their recto inexplicably blank. It is worth noting that two of the three fragments of the Gospel of Thomas, POxy 654 and POxy 655 are from rolls; POxy 1 is from a papyrus codex. SeeNagHammadi Codex II,2–7 (ed. B. Layton; Leiden: Brill, 1989) 96–7.Google Scholar
72 T. C. Skeat (n. 31) 263–8, here 263.
73 Skeat (n. 31) 263, explains that he no longer accepts either of the two theories which he and C. H. Roberts had advanced in The Birth of the Codex. For discussion of the various explanations advanced for Christian adoption of the codex, see Haelst, J. van, ‘Les origines du codex’, in Les débuts du codex (ed. Blanchard, A.; Turnhout: Brepols, 1989) 13–36;Google Scholar and Llewelyn, S. R., New Documents Illustrating Early Christianity Vol. 7 (Macquarie University Ancient History Documentary Research Centre, 1994) 249–56.Google Scholar
Skeat's reference to the crisis caused by the publication of the Fourth Gospel recalls E. J. Goodspeed's suggestion that the fourfold Gospel emerged a few years after the appearance of the Fourth Gospel; it was intended to win a wider hearing for the Gospel of John than it would otherwise have received (n. 1, 35–6).
74 D. Trobisch (n. 66).
75 D. Trobisch (n. 66), especially 12 and 124.
76 I am not convinced by Trobisch's attempt (n. 66, 67–8) to overturn this generally held view of early second century Christianity.
78 Colette Sirat rejects the claim made by C. H. Roberts and T. C. Skeat (n. 70) that there was a Jewish origin for Christian use of the codex. She notes that in the first centuries of our era, traditional Jewish texts do not make any allusion to the codex: ‘Le livre hébreu dans les premiers siècles de nôtre ère: le témoignage des textes’, in Les débuts du codex (ed. A. Blanchard; Turnhout: Brepols, 1989) 115–24.Google Scholar
79 Cf. Gamble, H. W., Books and Readers in the Early Church (New Haven and London: Yale, 1995, 65):Google Scholar ‘To claim the most primitive edition of the Pauline letter collection was put out in a codex and that it was the religious authority of Paul's collected letters that set the standard for the transcription of subsequent Christian literature in codices is not to claim that this marked the first use of the codex in Christian circles. It is possible, perhaps likely, that the codex was first employed in primitive Christianity for collections of texts (testimonia) from Jewish scripture.’ I accept the latter point, but I am not convinced that it was a collection of the Pauline letters (which Gamble dates in the early second century, ‘and probably earlier’ ) which ‘set the standard’ for use of the codex. We have far more early codices of individual gospels and codices of the four gospels than we do of collections of Paul's letters; the gospels are quoted much more frequently in the second century than are the Pauline epistles. As I have noted above, early Christian use of the LXX in codices must also have encouraged Christians to adopt the codex as a standard format.
80 See Gamble (n. 79) 28–30; Stanton, G. N., ‘Form Criticism Revisited’, in What about the New Testament? FS Evans, C. F. (ed. Hooker, M. and Hickling, C.; London: SCM, 1975) 13–27.Google Scholar See also Alexander, Loveday, ‘The Living Voice: Scepticism towards the Written Word in Early Christian and in Graeco-Roman Texts’, in The Bible in Three Dimensions (ed. Clines, D. J. A., Fowl, S. E., Porter, S. E.; Sheffield: JSOT, 1990) 221–47.Google Scholar
83 See further, G. N. Stanton (n. 35) 84–93.
84 See further, G. N. Stanton (n. 35) 78–82.
85 T. Baarda (n. 25); W. L. Petersen (n. 49), and his Tatian's Diatessaron: Its Creation, Dissemination, Significance, and History in Scholarship (Leiden: Brill, 1994).Google Scholar
87 On this point, see especially R. C. Morgan (n. 5) 380–6.
88 Schnackenburg, R., Jesus in the Gospels. A Biblical Christology (Louisville: Westminster John Knox, ET 1995) 324–5.Google Scholar Schnackenburg refers to Gen 2.10–14, but does not note that this image also appealed to the patristic writers listed above. See H. Merkel (n. 11) 7. n. 1, who gives full references.
89 I have quoted the Greek text from fragment 11 (Anastasius Sinaita), as edited by A. Rousseau and L. Doutreleau (n. 10) 160–2.