¹ Professor Bonner has discussed this sentence in H. T. R., 1938, 175 ff. and again in the introduction of his edition, p. 30 ff. He has stated the problems involved and presented the reader with valuable material and interesting hints for their solution. I offer the following suggestions with full appreciation of the painstaking and helpful work which is evident in every page of his edition (references to which are given in the following notes by Ed.). I would also here express my gratitude to P. Maas for repeatedly discussing with me the present subject and helping me both by his criticism and by his suggestions.

² Philo combines the adjective Ἑβραϊκός with ἔθνος στρατιά γύναιον; Clement, with τρὁπος, φιλοσοφία, βασιλεῖς etc. But a nomen actionis would be a different proposition.

³ LXX Exod. XIX, 1; Test. XII Patr. Jos. 20^6c, Benj. 12³; Epiphanius, De mens. et pond., 3.

⁴ LXX, ps. 113¹.

⁵ Clem. Al. Strom. I, 21 (II, 84²⁴ Stählin).

⁶ Sophocles Trach. 51 τὴν Ἡράκλειον ἔξοδον might be quoted as a counterinstance. But this usage, peculiar to tragedy, has no bearing upon Christian rhetoric.

⁷ See Ed., p. 21.

⁸ Πῶς ὁ λαὸς σώζεται refers to the ‘angel of death’ who was averted from the Israelites (Exod. XII¹³).

⁹ Cf. p. 2, l. 23–45. — It is a possible, though not a necessary, inference from p. 3, l. 2–p.7, l.27, that the text read included also Ex.XII, 28–30, in a similar way as vv. 43–49 are added in Hippolytus. The main thesis would not be affected by this assumption. I incline however towards regarding the looser paraphrasis p. 3, l. 2–6 as Melito's addition to the text actually read. It serves as a basis for the long ekphrasis of (a) the sufferings of the Egyptians (p. 3, 15–p. 5, 20) and (b) the protection of Israel by the τύπος of the suffering Christ (5, 20–7, 27) which opens with the elaborate period p. 3, 6–14.

¹⁰ Cf. Büchler, A., Jew. Quart. Rev., V, 1893, 443 f., 449, 453Google Scholar; Elbogen, I., Der Jüdische Gottesdienst, 1913, 158 and 163Google Scholar; Thackeray, H. St. J., The Septuagint and Jewish Worship, 1921, 127Google Scholar; Moore, G. F., Judaism, I, 1927, 298Google Scholar.

¹¹ Prophetologio n, L. 41d.

¹² Morin, Dom G., Le plus ancien comes de l'église Romaine, Rev. Bened., 1910, 54.Google Scholar

¹³ Missale Mixtum Mozarabicum (Migne 85, p. 451).

¹⁴ Ranke, E., Das kirchliche Perikopensystem, 1847, 344Google Scholar.

¹⁵ Cabrol, Dict. Arch. Chrét., V, 293, nr. 70.

¹⁶ Conybeare, F. C., Rituale Armenorum, 1905, 522.Google Scholar

¹⁷ Burkitt, F. C. in Proceedings of the British Academy, 1923, p. 825Google Scholar.

¹⁸ There are two exceptions in the far West: the Liber Comicus Toletanus (indicating vv. 1–42) and the Luxeuil-lectionary (vv. 1–50).

¹⁹ Cf. also Philo, Migr. Abr. 3⁴ (II, 275⁵) τὴν … τῶν κατὰ φιλοσοφίαν δογμάτων γραφήν and Origen, Joh. XI, 4 (p. 185⁶) ἐν τῆ προκειμένη τοῦ εὐαγγελίου γραφῆ.

²⁰ In the New Testament and in the Apostolic Fathers γραφή in the singular as a rule means ‘Scriptural passage’; see (e.g.) Acts 8³² and ³⁵ with K. Lake's note, Mk. 12¹⁰, Luke 4², John 7⁴², 13¹⁸ etc.; II Clem. Cor. 2⁴; cf. Lightfoot, Galatians 7th ed., 147. For later instances cf. Eusebius, in psalm. LXVII 23 (P.G. XXIII, 706 D) ἡ παροῦσα γραφἠ (cf. ib., 761 C) and the preceding note.

²¹ See the passages collected by Preuschen, E., Antilegomena², 1905, p. 5 ffGoogle Scholar.

²² Where else was there, in Greek literature, an opportunity to quote a book-title with an adjective qualifying the language? I can only think of the critical work on the Old Testament by Origen and his successors. They, however, nearly always compare single passages with the corresponding Hebrew. This gives no occasion to indicate the book-title: τὸ Ἑβραϊκόν, ἡ Ἑβραϊκὴ γραφἠ or ἀνάγνωσις ἑβραϊστἰ are the natural forms of reference. Rarely a passage from another book is adduced, hardly ever with the addition of the Ἑβραϊκὴ ἀκρἰβεια. In the two cases of this kind which I have been able to trace (Maximus Confessor, Schol. in Dionys. Areop. De Myst. Theol. I (P. G. IV, 421 B) and Eusebius, De Eccl. Theol. III, 2 (P.G. XXIV, 977 D)) the Hebrew is in the usual way superadded to the Greek: there was then no reason to repeat the book-title.

²³ H.T.R., XXXI, 176, n. 2; Ed., p. 31.— Suidas διασαφῆσαι· ἑρμηνεῦσαι cf. Hesysychius s.v.; Acta Petr. et Andr., XV, p, 124 Bonnet(διασάφησις equalling ἑρμηνεία); Cyril of Alexandria, De Ador., I, 3 (P.G. LXVIII, 137 B): διασάφησις of the hidden typological sense of the law; Insidore of Pelusium, Epist, I, 73 (P.G. LXXVIII, 233 A): διασάφησις of a problem of theodicy. — I am indebted to the Librarian of Pusey House for allowing me to use the manuscript of the Dict. Patr. Greek, from which I have borrowed several instances quoted in this article.

²⁴ Ed., p. 36.

²⁵ Ed., p. 32 with note 2 and ib. 33; H.T.R., XXXI, 177; Melito 2²⁵, 7²⁷. The same applies to Hippolytus' Paschal homily.

²⁶ H.T.R., XXXI, 178–182; Ed., p. 32 f.; cf. C. Höeg, La Notation Ecphonétique, 1935, 152. As to the passage quoted from Origen on Numbers, I cannot but subscribe to Prof. Ch. Martin's interpretation: the dozens of parallel passages in Origen (suitably collected in Cabrol's Mon. Lit.) would seem to put it beyond doubt that ‘explicuit’ is a merely stylistic variation of the regular reference to the reader's preceding recitation.

²⁷ Ed., p. 35 f.

²⁸ Aristeas has a rather strange notion of the difficulties facing the translators of the Septuagint and of the methods with which these were met. The real translators apparently worked with word lists and stuck to some rough rules in equating Hebrew and Greek modes of inflection. But “Aristeas” — who apparently did not know Hebrew — does not consider this task as a matter mainly of grammar and lexicon. To him it means “transcription” (μεταγραφή), from strange parchment scrolls (3, 176 f.), of the mysterious Hebrew letters into Greek ones (10, 30, 38 f., 45 f., 307), just as Ezra is credited in the Talmud (L. Goldschmidt, VII, p. 81 f.) with transcribing the sacred letters of the Torah into ‘Assyrian’ ones. True, the Jews have also a language of their own (11); but Aristeas has not one word to say of the linguistic problem. His ‘interpreters’ are well-bred noblemen, versed in the Law and able to expound its letters (32, 39, 121; cf. 3); only once (121) it is added that they had given good heed also to the Greek κατασκευή (the noun means anything and nothing; to Philo, Vita Mosis 6, it is παιδεἰα). Their procedure accordingly is this: the letters on the sacred roll are read out by one of them (305); bit by bit their implications are considered by the rabbis — this is ἑρμηνεἰα or διασἀφησις — until by common agreement τὸ ἀκριβές is achieved (32, 39, 302, 305). This ἑρμηνεία somehow at the same time results in an agreed Greek wording to be set down by Demetrius acting as secretary (302). Strange though it may seem to us this conception does not lack an element of truth: rem tene verba sequentur; besides we must keep in mind the amazing novelty, even uniqueness, of this translation viewed as a Greek literary phenomenon. Josephus, then, was entirely justified in rendering Aristeas' ἀνάγνωσις καὶ διασάφησις, in his own condensed paraphrase (Antiq., XII, 2, 13, § 106), by ἑρμηνεἰα.— Little is gained by the reference to an unattested work of Origen, De LXX interpretibus, which Meecham took from Wendland, Wendland from Harnack, and Harnack from Boulenger. It is a slip of the last named, the passage in question being a verbatim quotation from Josephus l.c., and quite unrelated to Batiffol's Tractatus Origenis De libris SS. Scripturarum.

²⁹ Bab. Talm. Nedarim 37b (transl. London, 1936, p. 116 with note; Megilla 3a (L. Berliner, III, p. 536); cf. Jer. Talm., cap. IV.

³⁰ The decisive central passage of the sentence is in the LXX at variance with the Hebrew; ἐδίδασκεν there looks like an alternative for the following διέστελλεν rather than an equivalent for שדפמ.

³¹ In view of their synonymous character ἑρμηνεύω might be considered an equally good equivalent, resulting in a less complicated argument. Yet the fact is that this root is used for םגוח the Septuagint is normally described as ἑρμηνεία, the meturgeman (v. infra) is in Greek the ἐρμηνεύς; Esdr. B’ 4₇ the Sept. has ἡρμηνευμένην for Aramaic םגוחט; Gen. 42₂₃ ὲρμηνευτήα corresponds to ןטגוחט in Targum Onkelos.

³² Happily I need not here inquire into the real problems of this word and of its relation ןגשחפ B in Esther: in the present context only its interpretation by Hellenistic rabbis needs to be considered.

³³ Esdr. B’ 5₆ άντίγραφον, 7₁₁ ὑποκίμενον. In the different rendering of Ezr. 4₁₁ in Esdr. A’ 2₁₂ ὑπογεγραμμένην probably reflects the same interpretation.

³⁴ The admissibility of this interpretation (independent of the rabbinic etymology) is considered in Batten's commentary, pp. 135 and 356.

³⁵ Cf. Kahle, P., Theol. Stud. und Krit., 88, 1915, 412 ffGoogle Scholar. — About targum and Septuagint cf. the following section.

³⁶ See (e.g.) Bousset, W., Kyrios Christos, 1913, 367Google Scholar; Lietzmann, H., Geschichte der alten Kirche, I, 1982, 153Google Scholar; Höeg, C., La notation ecphonétique, 1935, 150 fGoogle Scholar.

³⁷ Cf. W. Susemihl, Literaturgeschichte der Alexandrinerzeit, I, 685; C. Höeg l.c., 152.

³⁸ The Talmudic passages referring to the meturgeman have been often quoted; most completely by Strack-Billerbeck, III, 1926, 465 ff. The persistence, among the Palestinian Jews, of the alternate reading of single verses in Hebrew and Aramaic appears to be indicated by the fact that the Genizah fragments of the Pentateuch targum are written in the same way (see P. Kahle, Masoreten des Westens, II, 1930, p. 2).

³⁹ Moore, G. F., Judaism, I, 1927, 174Google Scholar.

⁴⁰ Cf. Zunz, L., Die Gottesdienstlichen Vorträge der Juden, 1832, 75 ff.Google Scholar; Kahle, P., Masoreten des Westens, II, 1930, passimGoogle Scholar.

⁴¹ Cf. P. Kahle, l.c., p. 3 f.; Baumstark, A., Biblische Zeitschrift, 1931, 257 ffGoogle Scholar.

⁴² The analogy between Targum and Septuagint has often and in various connections been remarked upon (for older literature see Hodius, H., De Bibliorum text. orig., 1705, 240 ffGoogle Scholar.; more recently see E. Schürer, Geschichte des Jüdischen Volkes, III⁴, 140, n. 17 and H. Lietzmann, l.c., I, 84). Kahle, P. (Theol. Stud. und Krit., 88, 1915, 410 ffGoogle Scholar.) has shown new ways of approach. — In Melito the introduction of the angel, instead of Jahwe himself, slaying the Egyptians (cf. Num. 20¹⁶) is a haggadic feature of the kind typical both of targum and Septuagint; it is presupposed, and rejected, in the old midrash published by L. Finkelstein, H.T.R., XXXI, 1938, 296; cf. ib., 307, note 32, and H. Lietzmann, l.c., 85 f. on Greek Halacha and Haggada.

⁴³ Cf E. Schürer. l c., 141, note 22.

⁴⁴ It is perhaps not even necessary to suppose that the lector must always have been expert in Hebrew: he might memorize his text from a transcript. So do Coptic priests to-day.

⁴⁵ Cf. Frey, J. B., Corpus Inscriptionum Judaicarum, I, 1936Google Scholar, index, p. 663 f. s.v. Rouleau sacré, Etui, Aron and ib., CXLII f.; cf. the decorations of the aedicula in the Dura synagogue, where the repository of the torah is still discernible (Excavations at Dura, VI, 1936, 320, 343, 351, 371 ff.); for the portrait of one reciting the torah (Ezra?) next to the aedicula see ib., pl. XLVIII and Rostovtzeff, Dura-Europos and its Art, 1938, p. 100 ff., pl. XX and XXI.

⁴⁶ R. Simon, Hist. Crit. du V. T., 1685, 293: “Il y a même lieu de douter, qu'on n'ait autrefois lû dans les Synagogues des Juifs Hellénistes, que la seule Version des Septante, comme on le croit communément. Il y a au contraire bien plus d'apparence que parmi les Juifs Hellénistes … on ne lisoit cette Traduction Grecque, que comme une explication ou paraphrase; de la même manière que dans les Synagogues des Juifs de Babylone, de Jerusalem et des autres endroits où la Langue Caldaique était en usage, il y avoit un Interprète qui paraphrasoit le texte Hébreu en Caldéen…. Il est donc fort vraisemblable qu'on a lû la Loi en Hébreu dans les Synagogues des Juifs qu'on appelle Hellénistes aussi bien que dans les autres Synagogues” etc., with reference to Justinian Nov. 146; cf. ib., 181 ff. and 502. R. Simon repeated his assertion, with greater assurance, in Hist. Crit. du N. T., 1689, 316 and 324.

⁴⁷ Epiphanius, Panarion, 30¹¹, Ἀζανιτῶν τῶν παρ᾽ αὐτοῖς διακόνων ἑρμηνευομένων ἢ ὑπηρετῶν (quoted by J. B. Frey, l.c., XCIX, with literature).

⁴⁸ The greater part of the material here presented has been brought together already by the great scholars of the 17th and 18th centuries; see (e.g.) Hodius, H., De Bibliorum textibus originalibus, 1705, 334Google Scholar; cf. Zahn, Th., Geschichte des Neutestamentlichen Kanons, I, 1888, 43Google Scholar and Schürer, E., Geschichte des jüdischen Volkes, II ⁴, 1907, 85 and 534Google Scholar.

⁴⁹ Megillah 4¹; Elbogen, I., Der jüdische Gottesdienst, 1913, 187Google Scholar.

⁵⁰ An excerpt in the original Greek of the passage in question is preserved in the Byzantine Synaxarion; see Zahn, Th., Tatians Diatessaron, 1881, 19Google Scholar and Violet, P., T. U., XIV, 1896, 110Google Scholar. It is unfortunate that since the days of Valesius it is quoted time and again from the Latin Acta Martyrum — as if the Eusebian source had not been traced.

⁵¹ Cf. the “presbyter” of II Clem, reading and expounding the lesson from Isaiah 54.

⁵² Wobbermin, L., T. U., XVII, 3, 1899, 18Google Scholar; Brightman, J. T. S., 1900, 103, cf. ib., 93; F. X. Funk, Didasc. et Constit., II, 1905, 169.

⁵³ Cf. E. Schürer, l.c., 85; Clark, The Acts of the Apostles, 1933, p. LIX ff.; Byzantion, XIV, 1939, 568 f.

⁵⁴ Itin. Hieros., ed. Geyer, 1898, p. 99.

⁵⁵ G. Morin, Anecd. Mareds., II, 220 ff.

⁵⁶ Liturgies Eastern and Western, I, 1896, 501²⁶, 502³¹; cf. ib., 578.

⁵⁷ De Fide 21 (Panarion, p. 522²² Holl).

⁵⁸ According to Const. Apost. II 4⁴ the bishop is expected to be πολὺς ἐν ταῖς άναγνώσμασιν, ἲνα τὰς γραφὰς ἐπιμελῶς ἐρμηνεύῃ (cf. ib. 7 πάντα ἐπιμελῶς κατὰ λέξιν ἑρμηνεύσῃς). Here, as in the old Apostolic Church-order 16, this injunction apparently refers to exegesis proper (as e.g. Eusebius Vita Const., 4, 54² ἄλλοι δὲ ἑρμηνείας τῶν θείων ἀναγνωσμάτων ἐποιοῦντο, τὰς ἀπορρήτους ἀποκαλύπτοντες θεωρίας). The parallel passage in the Syriac Didaskalia is rendered “vertat et interpretetur” by Funk, and “interpret and expound” by Conolly. Is it just redundancy of expression? or is a reference to bilingual service to be acknowledged?

⁵⁹ I hope shortly to produce a theory of the “Western Text” of the New Testament based in part on the facts mentioned in this section.

⁶⁰ Cf. H. Lietzmann, Geschichte der alten Kirche, I, 204 ff., especially 209.

⁶¹ Cf. H. Lietzmann l.c., I, 150, II, 122. — With diffidence I would here mention one possible argument: Hippolytus in his paschal sermon (the affinities of which with Melito's were recognized by Professors Ch. Martin and C. Bonner) reads out the LXX version of the whole Exodus passage on which he is going to comment. An exegetic homily not preceded by a lesson is unbelievable; so why this duplication? Had the lesson been in Hebrew?

⁶² Cf. W. M. Ramsay, The Letters to the Seven Churches, ch. XII.

⁶³ The references, in Irenaeus and his sources, to the ancient Christian presbyters as κρείσσονες ἡμῶν are in Jewish style (Iren. haer., I, praef. 2; I, 13³; III, 17⁴ etc.; cf. Philo, Migr. Abr., 89).

⁶⁴ Eusebius, Hist. Eccl., IV, 26.

⁶⁵ Cf. (e.g.) Zahn, Th., Geschichte des Kanons, I, 1888, 117Google Scholar; Thomas, C., Melito, 1893, 107 fGoogle Scholar.; Swete, H. B., Introduction to the Old Testament in Greek, 1900, 221Google Scholar. But for the exclusion of Esther, Melito's canon is identical with the Jewish in Epiphanius, Panarion 8, 6 and the Nasorean ib., 29, 7⁴. — I have refrained from using the fourth Catena-fragment (Pitra Spic. Sol., II, LXIII; reprinted (e.g.) in E. Goodspeed, Die ältesten Apologeten, 1914, 313) in evidence of Melito's Jewish affinities. The style of its relevant second part is that of Caesarean and Antiochian textual criticism; ὁ Ἑβραῖος are ὁ Σύρος are typical Hexaplaric terms (Field, LXXV ff.); already Combefis showed that the same scholion is preserved under the name of Eusebius of Emesa (Montfaucon, Hexapla ad Genes., 22¹³). Piper then (Theol. Stud. und Krit., I, 1838, 65 ff., quoted by Harnack T. U., I₁, 254) was right in assigning it to this fourth century author; C. Thomas (l.c., 38 f. and 213) did not grasp the conclusiveness of his argument.

⁶⁶ Cf. Lietzmann l.c., I, 129.

⁶⁷ This possibility was first mentioned to me by P. Maas.

⁶⁸ Epiphanius Haer., 29, 7⁴ (p. 329, 14 ff. Holl).

⁶⁹ See Epiphanius (quoted above p. 311).

⁷⁰ See above p. 311.

⁷¹ The Syriac eclogues from Melito's oration are more likely, with the bulk of Syriac translations of Greek theological literature, to have originated in Upper Mesopotamia at some later date than to have been adapted from some old Palestinian rendering (the manuscript is dated 562 A.D.). And Melito's paraphrase (unlike Hippolytus' repetition) of the lesson can be regarded as reinforcing rather than duplicating the διασάφησις.

⁷² On “Hellenists” in Palestine see R. Simon, l.c. (above, note 46), p. 294, and the same in his Castigationes, 1685, p. 22, against Isaac Vossius (who had asserted that a Greek paraphrase was read also in Jerusalem): “Longe verius dixisset, his locis ubi lingua Syriaca apud Judaeos erat vernacula, accessisse lectioni Hebraicae sacri contextûs interpretationem Syriacam; in iis vero locis ubi apud Judaeos vigebat sermo Graecus, accessisse Graecam interpretationem. Sic etiam Hierosolymis in Synagoga Alexandrinorum, qui Graece loquebantur, non dubito quin sacer contextus primo Hebraice pro antiquo synagogarum usu lectus fuerit, deinde Graece … agnosco cum Vossio etiam in Judaea linguam Graecam” (for this we have more evidence to-day from archaeology, inscriptions, the study of Greek words in the Talmud, etc.).