¹ The New Testament in the Original Greek [ii] Introduction, Appendix (2nd edn.London, 1896), 91.Google Scholar

² ibid., 107–15. Kenyon, F. G. in fact called this ‘the cornerstone of his theory’ (Recent Developments in the Textual Criticism of the Greek Bible [London, 1933], 8)Google Scholar, although he later modified this to ‘a cardinal point in Hort's theory’ (The Text of the Greek Bible [new ednLondon, 1949], 167).Google Scholar

³ See Dicks, C. D., ‘The Matthean text of Chrysostom in his Homilies on Matthew’, J.B.L. 67 (1948), 365–76Google Scholar. He concludes (p. 375): ‘We may have in Chrysostom the first step toward the more lucid K [Byzantine] text-type. His text was not yet K but was on the way to becoming K.’ In other words, it is ‘early’ Byzantine.

⁴ See Gignac, F. T., ‘The text of Acts in Chrysostom's Homilies’, Traditio 26 (1970), 308–15CrossRefGoogle Scholar. He concludes (p. 314): ‘Chrysostom's text resembled those of the Antiochene-Byzantine group out of which grew the medieval text which came to be the T.R.’

⁵ Cf. Gifford, C. K., Pauli epistolas qua forma legerit Johannes Chrysostomus (Halle, 1902)Google Scholar, and Förster, J., ‘Gerechtigkeit für Lucian und die antiochenischer Text’, Monatschrift für Pastoraltheologie 45 (1956), 267–72Google Scholar. Both of these scholars were interested in the deviant element in Chrysostom's text, which they demonstrated to have a clearly Western strain. But they incidentally also demonstrated the fundamentally Byzantine character of his text.

⁶ For this judgement as to the ‘growth’ character of the Byzantine text-type, see Lake, K., Blake, R. P., and New, Silva, ‘The Caesarean text of the Gospel of Mark’, H.T.R. 21 (1928), 208–404Google Scholar, esp. Excursus 1 by Lake, 338–57. Although he did not use the language ‘early’ and ‘late’ for this growth, he demonstrated beyond serious question (1) that von Soden's K² ( = K Π et al.) and the Ecclesiastical text ( = the ninth- to fifteenth-century Majority text) are related texts, (2) that K² represents an early stage of this text, and (3) that in all likelihood Chrysostom also belonged to this early stage.

⁷ ‘The text of John in Origen and Cyril of Alexandria: a Contribution to Methodology in the recovery and analysis of Patristic Citations’, Biblica 52 (1971), 357–94, esp. pp. 371 f., n. 1.Google Scholar

⁸ ‘Chrysostom's text of the Gospel of Mark’, H.T.R. 24 (1931), 133.Google Scholar

⁹ See, e.g., Metzger, B. M., ‘The Lucianic Recension of the Greek Bible’ in Chapters in the History of New Testament Textual Criticism (Leiden, 1963), pp. 21 f. CfGoogle Scholar. some modern disciples of Dean Burgon who consider this conclusion as ‘canon’ for Chrysostom's biblical text to the neglect of all other studies. See Pickering, W. N., The Identity of the New Testament Text (Nashville, 1977), 62–3Google Scholar, and Hodges, Z. C., ‘Modern textual criticism and the majority text: a response’, Journal of the Evangelical Theological Society 21 (1978), 149Google Scholar; repeated in his ‘Surrejoinder’, 161.

¹⁰ Op. cit., 141.

¹¹ Hodges, op. cit.

¹² In my previous study (op. cit., 358–61) I urged that fully reconstructed texts be the basis of such studies as this, not merely lists of variants. I have made such reconstructions for this paper, but space simply prohibits their inclusion. Hopefully, several such reconstructed texts will eventually be published.

¹³ Cf. Fee, G. D., ‘The text of John in The Jerusalem Bible: a critique of the use of patristic citations in New Testament textual criticism’, J.B.L. 90 (1971), p. 169, n. 26Google Scholar. My attention was first called to this phenomenon by Sister Thomas Aquinas Goggin in her translation of homilies 48–88, vol. 41 of The Fathers of the Church (New York, 1960), vii–viii.Google Scholar

¹⁴ ‘The text of John in Origen and Cyril’, 362–4.

¹⁵ Extensive work on the Johannine text of each of these Fathers made certain priorities for the reconstructions of their texts become very clear. For Origen: (1) the Commentary ad loc., (2) the lemmata of the Commentary, which in his case are remarkably free from later contaminations, (3) citations elsewhere in the Commentary, and (4) citations from other works. The material in Latin and from the catena fragments is nearly worthless in the recovery of Origen's actual text.

For Chrysostom: (1) the homilies ad loc., although the homiletical style requires a considerable amount of caution, (2) citations from the controversial or dogmatic treatises, and (3) citations from other homilies, both from John and elsewhere. The lemmata of the homilies have suffered extensive corruption; for example, the lemma of homily 48 (on John 7. 1) reads ούκ ηἤελεν, but the extended comment in the text ad loc. makes it abundantly clear that ούκ εΙχεν έξουοσΙαν is the only text of this passage Chrysostom knew.

For Cyril: much the same as with Origen. In his case the lemmata, especially as preserved in codex B, are of primary value.

¹⁶ Chrysostom's one citation actually reads οτός έστιυ, which is also the reading of G ψ 1093. This is probably a ‘homiletical’ variation on Chrysostom's part.

¹⁷ Both Origen and Chrysostom have Βηθανία in their own Bibles; however, both prefer the Βηθαβαρά they acknowledge to be in other known copies. Our reconstructed text of both Fathers is concerned with the text they actually used, not with texts they may have known.

¹⁸ This is the reading of the lemma, which in this case is the only ‘citation’ available.

¹⁹ Chrysostom cites the text twice, once with ὄτι and once without. A decision as to his actual text is nearly impossible in such cases.

²⁰ Chrysostom reads ποιῶ once and ποιήσω twice. Which may have been his actual text is impossible to discern.

²¹ Chrysostom reads ἀληθής with f¹ 1071 pc, but has the article, which the others do not.

²² This factor must be kept in mind for all of Chrysostom's biblical citations. See, e.g., the judgement of Dumortier, J., ‘Les citations bibliques des Lettres de S. Jean Chrysostome à Théodore (PG 47, 277–316)’, Studia Patristica 4 (TU 79; Berlin, 1961), 78–83Google Scholar. After noting this as the recurring phenomenon in Chrysostom's Old Testament citations, Dumortier says (p. 80): ‘Le texte du Nouveau Testament subit un traitement analogue à celui de l'Ancien. Pour obtenir plus d'euphonie ou de rythme, Jean modifiera l'ordre des mots ou supprimera telle expression jugée par lui superflue.’

²³ Codex ∏ is lacking John 8. 5–39. In its place I have used codex 1219 which Geerlings prints for this section. See Family ∏ in John (SD 23; Salt Lake City, 1963), 1.

²⁴ Origen cites the text this way (έστε…ποιεῑτε) no less than 10 times in the commentary (Book x); however, elsewhere he cites it τε…;έποίειτε (J0. 6. 4; mart. 38; hom. 4. 5 in Jer). Since this, too, is a reading that circulated in Egypt, it may be that Origen knew both readings. But the commentary makes it clear that έστε…ποιειτε is the text of his own Bible.

²⁵ In the one citation in his homilies the editions are split (Savile has ἁν; Montfaucon = MPG omits). In three other citations, two include ἄν and one omits.

²⁶ Cyril's evidence is especially difficult to decipher. In the commentary ad loc. there is not a hint that Cyril knew this addition. The lemma cites through Ιερουῦ and stops. In Book XII in his comment on John 19. 10 he says, ‘according to the Evangelist, they wished to seize him, but ἑξελθὠν διἀ μέσουαυύτωῆν έπορεύετο, καἰ παρηύγεν ουῆτως’. This is the language of the Byzantine text of John 8. 59, and therefore must have been known to Cyril, although one cannot tell whether ευύαγγελιστής here refers to the passage in John or to Luke 4. 30.

²⁷ Since B⁷⁵ is lacking this chapter, I have included the evidence from C, which is extant for vv. 8–38.

²⁸ In the first citation ad loc. Chrysostom reads τουύς πόδας μου; in two later citations he reads the possessive first.

²⁹ Origen's text is very difficult to assess. The lemma reads εἰ μὴ τουύς πόδας, which is also the text presupposed in the Latin translation of hom. 1. 4 in Lev. However, this latter could also reflect a conformation of the Old Latin. On the other hand, in the commentary itself he three times cites the text without these words, and makes no suggestion in his comments that he knew them. Cf. ℵ it^{aur, c} vg Tertullian. Hautsch, E. (Die Evangelienzitate des Origenes [TU 34², Berlin, 1909], pp. 149Google Scholar f.) is probably correct in judging them as unauthentic. None the less, the ‘addition’ in the lemma is made from an Egyptian text, not the Byzantine.

³⁰ Origen's evidence is quite ambiguous here (two or three citations with μετ’ ἐμουῦ [if one accepts the reading of V in the lemma]; one citation and three adaptations with μετ’ ἐμοū). Hautsch is probably correct in leaning toward μετ’ ἐμοū (op. cit., 150), since the two citations with μου could reflect Origen's own assimilation toward the LXX.

³¹ The split evidence (2/4) makes a firm choice impossible here.

³² Chrysostom's evidence here is limited to the lemma.

³³ It is perhaps of interest to note how many fewer of the total number of variants Chrysostom has here in contrast to Origen and Cyril, although his homilies cover the whole chapter. The reason for this is that so many of the Byz/UBS³ variants in this chapter are found in the introductory formulae to the discourses. This is the kind of material which simply gets lost in the homilies.

³⁴ Tasker, R. V. G. (‘The Text of the Fourth Gospel used by Origen in his Commentary on John’, J. T. S. 37 (1936), 146–55)Google Scholar concluded, to the contrary, that in Bock XXVIII, covering John 11. 39–57, ‘the text used by Origen is here less Neutral in complexion and more Caesarean’, thus suggesting that there at least ‘he made his quotations from the text which he found at Caesarea’ (p. 153). But this is not even partially correct. Tasker's errors are the result of several methodological failures. For example: (1) although the commentary here is on John 11, he makes no distinction between these citations and others; (2) at least 5 of his variants (including 3 of the 4 distinctively ‘Caesarean’ readings) are simply incorrectly listed; while another (om. α*ν 18. 28) is the kind of random variant that can never be used, since particles and conjunctions – especially their omission in a Father's citation - are the least certain elements in patristic evidence; (3) he has 12 variants from the TR listed for 11. 39–57, only 8 of which are correct, but he has neglected 7 others; (4) of the 15 variants from the TR in this section of John Origen agrees with the Egyptian text in every case, 14 of which are in common with B, while only 6 of them are also supported by the ‘Caesarean’ MSS; (5) Codex Θ joins other ‘Caesareans’ in 9 distinctive variants for this section, not one of which is read by Origen. As elsewhere Origen's text of John in Book XXVIII is thoroughly Egyptian, without a shred of influence from the so-called Caesarean MSS.

³⁵ Op. cit. 133–5.

³⁶ For the sake of completeness the authors listed all the references in categories (b) and (c).

³⁷ The homily De poenitentia (MPG 49. 323–36) belongs to Severian. See Clavis Patrum Graecorum 2, no. 4186.

³⁸ See e.g. the citations of Mark 10. 32 in hom. 65 in Mt. (Field 2, 253 c) and of Mark 11. 13 in hom. 67 in Mt. (Field 2, 279b).

³⁹ This singular reading is a clear illustration of Chrysostom's faulty memory. The citation begins, σιὰ τουῦτο καì ἔτερος εύαγγελιής φησιν.

⁴⁰ This is another clear illustration of the adaptive nature of Chrysostom's ‘citations’. In this case he specifies ὁ δὲ Μᾶρκος φησίν.

⁴¹ The parallel in Matthew that Chrysostom is commenting on suggests that Mark 4. 10 is what he intended to refer to, but the language is much closer to 4. 34.

⁴² All items from the Eclogae ex diversis homiliis (MPG 63.567–902) have been excluded because of their patently secondary character.

⁴³ Asterius the Sophist (d. post 341) appears to have been the first to use such a text.