The critical study of ancient documents means nothing else than the careful sifting of their origin and meaning in the light of history. The first principle of criticism is that every book bears the stamp of the time and circumstances in which it was produced. An ancient book is, so to speak, a fragment of ancient life; and to understand it aright we must treat it as a living thing, as a bit of the life of the author and his time, which we shall not fully understand without putting ourselves back into the age in which it was written. People talk much of destructive criticism, as if the critic's one delight were to prove that things which men have long believed are not true, and that books were not written by the authors whose names they bear. But the true critic has for his business, not to destroy, but to build up. He must review doubtful titles, purge out interpolations, expose forgeries; but he does so only to manifest the truth, and to exhibit the genuine remains of antiquity in their real character. . . . In a word, it is the business of the critic to trace back the steps by which any ancient book has been transmitted to us, to find where it came from and who wrote it, to examine the occasion of its composition, and to search out every link that connects it with the history of the ancient world and with the personal life of the author.-W. Robertson Smith.

APPENDIX

ON THE HYPOTHESES OF INTERPOLATION, COMPILATION, AND PSEUDONYMITY, IN RELATION TO THE NT LITERATURE

ANY discussion of the NT writings, especially with reference to their date, must include some attempt to appreciate the literary customs and conditions among which these writings took their earliest or final shape. The first and most obvious question is that of translation. It is introductory to the others, and scarcely as vital; but it cannot be passed over without some notice at least of its existence.

The question is, are all the NT writings extant in the language in which their authors originally wrote them? Or have any been subsequently translated from Hebrew and Aramaic (that "most concrete and unmetaphysical of languages," M. Arnold) into Greek? The bearing of this upon the problem of a book's date may be illustrated by the case of Ecclesiasticus. The Greek version of this book is a translation of the Hebrew original, prepared some fifty years later by the grandson of the author. Here translation implies a notable gap between the earlier and the final form of the book. When the translator is identical with the author of the original, the matter is comparatively speaking of less importance as in the case of Josephus, who composed his history of the Jewish war in the Aramaic vernacular, intending it for the Jews resident across the Euphrates (rois ava Bapßápois), but afterwards, to gain access for his work to the wider circles of the Roman Empire, rendered it into Greek. Still, apart altogether from the personality of the translator, the question of translation affects to some degree the date of a writing. The earlier instances in all departments of Jewish literature (e.g. 1 Maccabees, Judith, Seirach, Psalt. Sol., Enoch, Book of Jubilees, etc.) prove the abstract possibility of translation in regard to a NT writing, while the bi-lingual nature of the Roman Empire and the use of Aramaic and Greek in Palestine indicate that such a practice must have been necessary for the extensive circulation of literature. The hypothesis therefore has a legitimate claim to be at least tested. Whether it explains in a satisfactory fashion any or all of the NT records in connection with which it has been raised, is a question that depends upon the further examination of the particular case and its evidence.

1 In the case of this writing it is still a moot point whether the original was Aramaic or Hebrew.

2 Dalman (Die Worte Jesu, 1899, i. pp. 10–13) extends the practice further among the pseudepigrapha; he even conjectures an Aramaic original for the Hebrew of Dan 1-6, as Marshall (DB, i. p. 253) does for Baruch, 39-4. Similarly for 4th Esdras, Wellhausen (Skizzen u. Vorarbeiten, VI. p. 235 f.).

Only three books in the NT1 can be seriously supposed to admit even of the discussion of this hypothesis with regard to their contents. Chief among these is (a) the gospel of Matthew. The Papias - tradition of Matthew's Aramaic Logia (Eus. HE, III. 39, éßpaidi diaλékтo tà λóyia συνεγράψατο (ν. 1. συνετάξατο), ήρμήνευσε δ ̓ αὐτὰ ὡς ἦν δυνατὸς ἕκαστος) naturally led scholars of later ages to naïvely confuse or to deliberately identify this Aramaic compilation with the canonical gospel. The motives for such a belief were obvious. It secured apostolicity for the gospel, and it seemed to explain satisfactorily some of its linguistic and theological features. But if one result is certain in synoptic criticisın, it is that the extant Matthew is no translation. The Logia to which Papias refers formed one of its sources, but even this already existed in a Greek translation as it lay before the final author of the gospel, along with his other main source, the Greek Mark. Even apart from its dependence upon these documents, the linguistic phenomena of the gospel afford evidence that is practically decisive, e.g. the comparatively smooth Greek, the number of OT quotations that necessarily imply a use of the LXX, phrases of peculiarly Greek assonance and rhythm (like 67, BATTаλογήσητε πολυλογία; 616, ἀφανίζουσιν . . φανῶσιν ; 2141 247. 30), etc. The relation between this Aramaic (Dalman) or Hebrew (Resch, TU, x. 1. p. 90 f.) compilation of Logia and the canonical Matthew is extremely intricate (on the whole question, cp. Holtzmann, Theol. Jahresbericht, 1889, pp. 99-103, and Harnack's frank note, Chron. pp. 692–694). But at any rate the identification of the two is precarious in the extreme. It is even doubtful whether λóyia at that time could have been applied to a NT writing; and from what we know of the Hebrew "Matthew," the scanty traces of its nature and contents (in Origen and Jerome) indicate that the writing was very different from our extant gospel. Some editors, however, like Schanz (Comm. über das Evglm. d. heiligen Matthäus, pp. 8-23) and Carr (CGT, pp. xx-xxiii), still adhere to the translation hypothesis,* while Blass among others (including Nestle) goes back even to an Aramaic original not only for the first part of Acts, but even for Mark's Gospel

5

1 Leaving out of account the possibly Jewish (Hebrew) sources which, it has been conjectured, underlie the earlier chapters of Matthew and Luke, and portions of the Apocalypse. It is curious that an attempt has recently been made by Dr. H. P. Chajes to reconstruct the Hebrew original of Mark (Markus-Studien, 1899)

2 Cp. the statement and discussion in Weiss, Matthäus-Evangelium, pp. 35-48; Roberts' Greek the Language of Christ and his Apostles (1888), chaps. x.-xiii.; and Conybeare, DB, ii. p. 262. This position, reached by critics of the synoptic problem, has been recently corroborated, from the standpoint of an expert in Aramaic, by Dalman (Die Worte Jesu, pp. 47-57), and is conclusively stated by Weiss (-Meyer, 1898, Matthäus, pp. 4-13). Cp. also W. C. Allen (Exp. Ti. xi. pp. 135-137), who, however, seems impressed by the Aramaic phraseology of Mark (Exp. i. pp. 436-443). See further in Addenda.

3 He gives a lucid summary of his position in Th.St, pp. 95-128, a study of 729 y. For Zahn's theory, see his Einl. ii. pp. 295–322, a collection of good material and less acceptable inferences.

4 Cp. also Gla (Die Originalsprache des Mt.-Erglms. 1887), who holds to an Aramaic original for the gospel, but will not identify it with the Hebrew gospel. For the cognate theory that an original Aramaic gospel once existed, see Marshall's acute and elaborate papers (Exp. iii.-iv., résumé in Exp. Ti. iv. pp. 260-267). This and the theories of A. Meyer and Wellhausen upon the primitive form and dialect of the Urevangelium are discussed by Dalman, op. cit. Einleitung, VI.

5 One good feature of Swete's recent edition of Mark (also of Salmond's article, DB. iii. pp. 251, 252) is the reasoned opposition offered to such a hypothesis, which contradicts the earliest tradition (Papias), is not absolutely necessary for the textual phenomena, and must be pronounced a tissue of improbabilities.

(PG, pp. 190-218); but the most reasonable conclusion, with reference not merely to one but to all those gospels, is that, while the matrix of their original tradition was the Aramaic vernacular of Palestine, the extant gospels as well as their immediate sources-so far as these can be traced and felt-were composed with practical entirety in Greek.1

(b) The idea of an Aramaic original for the epistle of James has also been unconvincingly revived by Wordsworth (Studia Biblica, i. p. 144 f.),2 who conjectures that our present text forms one of two translations. The real impulse to this theory is the desire to do justice to the excellent Greek style of the epistle and at the same time to preserve its apostolic origin. But the hypothesis will not hold water. No NT writer moves with such vigour and freshness in Hellenistic Greek as the author of James. His book has assonances and idioms that preclude any idea of a translation, and ally him to the wisdom literature of Alexandria as well as to the Greek classics. The style of James embraces Hebraisms, as was to be expected; but it is as distinctively and independently Greek as a page of Marcus Aurelius.

3

(c) At an early period Hebrews was widely imagined to be a translation by Luke of Paul's originally Hebrew composition (Euseb. HE, VI. 14. 2, γεγράφθαι δὲ ̔Εβραίοις ἑβραϊκῇ φωνῇ, Λουκᾶν δὲ φιλοτίμως αὐτὴν μεθερμηνεύσαντα ἐκδοῦναι τοῖς Ἕλλησιν). Paul wrote it, says Jerome, “ uit Hebraeus, Hebraeis, Hebraice." The motive of this theory (from Clement to Thomas Aquinas) falls of course with the abandonment of the Pauline authorship. But it never had any real countenance from the internal evidence of the writing, with its verbal and close use of the LXX (especially 105. 10, also 17 1037 1221, etc.), its assonances (éμadev... erabev, 58 ; καλοῦ Kakoυ, 514 11 87, etc.), and the fine rhetorical periods that mark its structure. In face of the excellent Greek of the epistle, the translation-hypothesis is nothing better than a curiosity of criticism (cp. Westcott, Hebrews, pp. xxxii-xxxv).

These and every other application of the translation-hypothesis to the NT literature may be therefore set aside with almost absolute confidence. It is wrecked repeatedly and fatally upon the evidence of style. It is in conflict with the fact that long before any NT writing-and especially a gospel-was composed, the church was becoming full of members who knew Greek (Ac 61 929) and no Aramaic. Their needs became more and more paramount, and it was to the situation created by their presence and requirements that the NT writings later than Paul were mainly addressed. The translation-hypothesis in fact is bound up with a conception of the early Christian development which places several of the

1 Extremely fair summaries of the critical position in Bovon, NTTh, i. pp. 72-84; Zahn, Einl. i. pp. 1-51; T. K. Abbott, Essays on Original Texts of Old and New Test., 1891, pp. 154-158; and Wernle, Synoptische Frage, pp. 117-121 (Matthew), 221-223 (Mark and Luke), with Schmiedel, EBi, ii. 1870-1872.

2 His arguments (based mainly on the Latin translation in Codex Corbey, which seems to presuppose a Greek original differing from the extant text) are taken very seriously and refuted with complete success by Mayor, Epistle of St. James, pp. xli-xlii, ccv-ccxiii. So Zahn (Einl. i. pp. 84, 85), who remarks that if the extant James were not the original, it would imply a mastery of the art of translation such as could not be paralleled in antiquity.

3 The view still seems to linger. Panek (Commentarius in epistolam beati Pauli apostoli ad Hebraeos, 1882) considers the epistle was originally written by Paul in Aramaic, and Dr. Schiller-Szinessy (Exp. iv. p. 326) is "fully convinced from internal evidence that the epistle to the Hebrews, in the original, belonged to St. Paul, and that only the Greek now in our hands is a somewhat inexact translation of it."

NT writings at a period too early for their real nature and characteristics. It has never won serious or extensive support, and seems less likely to do so than ever. On the contrary, one of the postulates of modern criticism is that the NT writings, whatever processes they may have gone through in the course of their composition, are extant in the language in which they passed from their final authors to those who first received the autographs.2

The following collateral topics, however, need more serious discussion, as they closely touch the problem of the NT. documents and their dates. (1) The possibility of interpolation. Substantially a writing may bear evidence that it has originated in a certain period, while nevertheless it contains sections or verses which obviously belong to a different age, earlier or later. In such a case the hypothesis of interpolation becomes legitimate. Its relevance is always disputable: hardly ever, its possibility. During the pre-canonical age, and indeed for some centuries afterwards, the NT texts were exposed, in the course of things, to the possibility of such additions and incorporations. Habent sua fata libelli. No less than the gospels, the other writings of the NT": arose not in the sort of world where depositions are taken, nor in the sort of world where manuscripts are guarded. They arose, and they passed many years," as Matthew Arnold proceeds to point out, " in the immense, underground, obscure, fluctuating world of the common people. Probably even neighbours and contemporaries never knew, or cared to know, quite accurately, the literary history of a document like one of our gospels; and beyond question the knowledge, if it ever existed, was soon lost irrecoverably." Within the second century especially MSS had their vicissitudes. These were due partly to their occasional obscurity, partly to their very popularity and wide circulation. They were exposed to alteration, omission, addition at the hands alike of copyist and of reader. Traces of this still exist in our oldest MSS., e.g. the Marcan appendix (Ro 1624), and the Johannine pericopê (753-811). But there is no reason to deny the abstract probability of such interpolations even where the extant text no longer suggests any break. The history of the text does not in most cases reach back so accurately and so far, that room is not left between the autographs and the earliest known text for changes to have taken place. These changes

1 Nestle (SK, 1896, p. 102 f.; Phil. Sacra, passim) and Blass (Evangelium sec. Luc. pp. vi, xxi) have recently conjectured an Aramaic basis for the early part of Acts; but the linguistic evidence is not conclusive.

2 The case of the apocalyptic Jewish writings is scarcely analogous. As Gunkel points out (KAP, ii. p. 333), the number of "barbarian" languages in which so many of them are extant, is due to their historical fortunes. Jewish apocalyptic literature, he argues, has experienced two great catastrophes. One was the loss of the Hebrew originals, when the Jewish synagogue, on re-assembling after the Roman crisis, repudiated apocalyptic and Greek literature. That the whole Jewish-Greek literature did not perish at that time, was due to the fact that it had already flowed into the Christian churches, where the various translations made in languages so widely different show the popularity and enormous diffusion of the apocalyptic literature in this its second home. Yet it was there that a fresh catastrophe overtook it. The spirit of Greek philosophy, which possessed the Greek theologians, scented heresy; it would have nothing to do with the Oriental mythology in the apocalyptic literature. Hence the disappearance of Jewish apocalyptic from the Greek church. Hence, too, its preservation in "barbarian" tongues (Syrian, Armenian, Latin, Ethiopic, etc.). 3 God and the Bible, chap. vi. Cp. Schmiedel, IIC, II. i. pp. 80, 81; Blass, PG, p. 77 f.; and Reuss, pp. 367-380. In 3rd Maccabees, for example, the opening leaf or the introduction has been evidently lost at a very early date.