79-89) of Matthew. 1 This gospel he reckons the earliest of the synoptics, written even before the fall of Jerusalem, in all probability about 66 A.D. But this extremely early date requires some sacrifices. In the extant gospel, Keim admits, there lie several passages which point unequivocally to a later period, and could hardly have been written previous to 70. Such passages are in addition to slighter accretions (26. 18 415 215, prophetic comments, and illustrations from the Hebrew text)-the incident of the ass and the colt (212-7), the account of Judas and the thirty pieces of silver (2615 273-10), the narratives of the virgin-birth, etc. (2), the parables of the wedding feast (221-14) and of the ten virgins (251-12), the incidents of the watch at the grave (2762-66), of Pilate's wife (2719), of the resurrection of the saints (2752. 53), and sayings like those of the Baptizer (314. 15), and of Jesus (811. 12). These Keim is forced to cut out of the original gospel; they must be "slight and yet essentially consistent additions made to the gospel after the destruction of Jerusalem by a zealous Jewish-Christian contributor, dating from about the time of Mark and Luke, in the sense of a more liberal Christianity." On the other hand, it is clear that when Matthew's gospel is relegated to a date later than this period of exaggerated antiquity, recourse to the interpolation-theory is unnecessary. The whole writing then can be taken as a practical unity. Or, to turn the problem round, if it can be established that such passages are an integral part of the gospel, then they preclude any date for the whole work earlier than that at which they themselves can be reasonably fixed. The applicability of the interpolation-hypothesis_to Paul's writings has been totally denied by several critics, even by Renan (Saint Paul, chap. ix.); it is sparingly adopted by the more sensible among those who, like Schmiedel and Lipsius, admit its relevance (cp. Heinrici, Die Forschungen über die paul. Briefe, 1887, p. 67 f.). In itself the method is legitimate enough. The trouble is that it has been frequently discredited. As the evidence is almost uniformly internal, a passage is often rejected or retained upon absurdly a priori notions of what is Pauline or un-Pauline, or of what the writer should and would have said at such a point in the argument. Two considerations also have to be borne in mind. One is the subtlety of Paul's mental processes; these do not work always in a very obvious fashion, but imply fluctuations of his temper, his habit of going off on a word, his repetitions and allusions, and what Irenaeus once called the velocitas sermonum suorum. Consequently, when the question is one of purely internal difficulties, it is only fair to remember that "in a writer so subtle and abrupt as St. Paul, obscurity is not a strong ground of objection" (Jowett). Otherwise one is apt to do injustice to the writer's arguments and illustrations by too hasty recourse to a method which tends to become easy rather than accurate, courageous rather than sympathetic.2 Besides, there is the mechanical condition. Paul merely 1 Cp. too, Mr. Badham's very similar theory developed in his Formation of the Gospels, and St. Mark's Indebtedness to St. Matthew. A similar instance of the bearing of the interpolation-theory upon a book's date is afforded by Dr. Salmon's verdict upon Mk 169-20 (cp. p. 553). Bartlet also (DB, iii. p. 304) has to take Mt 226a. 7b. as post-70 additions, to secure 68-69 as a date for the whole work. E.g., occasionally in its clever application to the pastoral epistles by H. Bois, JpTh (1888), pp. 145-160. 3 For the conditions of ancient authorship and literary composition in their bearing on the structure of a writing, cp. Birt, Das antike Buchwesen in seinem Verhältniss zur Litteratur (1882). The possibility of deviation or misapprehension through the employment of amanuenses is usually admitted, even in conservative wrote the postscripts or occasionally the benedictions to his epistles, when he wrote anything. (The note to Philemon is the exception which proves the rule.) What appear to the modern eye, then, to be gaps, edges, and corners, abrupt transitions, indistinct arguments, left upon the surface of the writings, may quite well have been due to the fact that the speaker outran the amanuensis (notarius, actuarius) to whom he was dictating,1 or that the latter now and then missed, or perhaps caught up wrongly, words and sentences (Laurent, pp. 3–38). It is useful to think of his style as a "stenographed conversation." But that is a feature which explains its occasional obscurity 2 as well as its vivacity. Further, the digressions and pauses which appear disjointed to a modern reader, lose something of their strangeness when it is recollected that an ancient writer, even when he wrote himself and no less when he dictated, lacked many of those aids which a modern author possesses, in the form of notes and parentheses. In ancient MSS the whole is fused together. There is no accessible means of correcting or amplifying what is once written. Consequently, the argument has occasionally an appearance of being interrupted by pieces of foreign matter which really have only to be interpreted as asides, or read apart, in order to let their secondary connection with the central idea become visible. There is shrewd sense in Coleridge's remark upon the obscurity of Romans: "Some of the difficulty," he urged, "is accidental, arising from the form in which the epistle appears. If we could now arrange this work in the way in which we may be sure St. Paul would himself do, were he now alive, and preparing it for the press, his reasoning would stand out clearer. His accumulated parentheses would be thrown into notes, or extruded to the margin." The spirit of this paragraph is historically true. It is sounder to criticise the Pauline epistles with a recollection like this of their natural structure, than with a detective's suspicion or with a cleverness which often fails to understand the naïveté of an age less acute and literary than itself. That way lie subjectivity, prejudice, and doctrinaire circles (cp. Romans, ICC, p. lxv). It is increased if the Pauline letters are held to have been originally taken down in shorthand, like the speeches in the Roman Senate (Plut. Cato Min. 23), or the later lectures of Origen (Eus. HE, vI. 23. 2). The Pitman of the first century B.C. seems to have been Tiro, Cicero's freedman. 5 3 1 Cp. Norden, Die antike Kunstprosa (1898), p. 954 f. Prof. Rendel Harris adds (Exp. viii. pp. 402, 403): "When Paul replied to a letter, he held the letter that he was replying to in his hand and followed closely the points in it that needed attention. He is not responsible for all that is printed under his name, for there may be whole sentences that belong to the earlier and antecedent factors of the correspondence, and there may be cases where the language is not his own, but is either that of his secretary or that which is common to all secretaries." 2 There is a similar element of obscurity in Arrian's diarpiai of Epiktetus, due to the constant interchange of question and answer. 3 The Pauline letters are not entirely unpremeditated effusions, at least not the major epistles. A study of their structure proves that in many passages the ordinary methods of ancient rhetoric have been somewhat elaborately followed. How far this practice was conscious or unconscious, it is hard to say. (See J. Weiss, Beiträge zur paulinischen Rhetorik, pp. 5, 6.) Instances of words or sentences which are conjectured to have been either scholia or added in the margin after the whole composition had been written out, are possibly to be found in 2 Co 614-71, Philemon 19a. (Zahn, Einl. i. p. 322), 1 Co 112, iya de XpIoT (Heinrici), Ro 214.15 (Wilke, Die neutest. Rhetorik, pp. 216-228 f.; Laurent, 17-19; Blass; cp. J. Weiss, 56), Mark 73.4 (Blass, PG, pp. 212, 213), Ro 1621-23 (Hoc ad marginem adscripserat Tertius," Grotius), etc. Fortunately, as Semler once put it, "Sanctis doctrinis ipsis innititur religio Christiana, non vero hinc decreto, Pauli epistolas omnes uno tenore atque ordine totas perscriptas inde ab initio fuisse." 4 The rigorous application of the interpolation-hypothesis is widely discarded at the opinions. At the same time it is seldom safe to ignore the possibility of later editorial changes or marginal glosses even in so compact and straightforward a piece of writing as 1 Corinthians, where (as Heinrici admits) one may still detect the work of a later hand in the occasional improvement of detailed sentences and in the "attempts made" by subsequent revisers "to smooth over or to supplement harsh or defective passages." (2) To the amateur nothing appears quite so absurd and unreal as the way in which modern critics analyse various historical writings in the OT or in the NT, producing in each case several documents which have been partly copied, partly edited, with parallel or slightly discrepant accounts of the same affairs, which have not been fused together into a perfectly smooth literary unity. The unreality, however, lies not so much in the theories as in the difference between modern and Oriental conceptions of a historian's task. The latter allowed a writer not merely to epitomise and omit, but to reproduce his sources literatim or alter them in so far as his special purpose demanded, instead of completely rewriting the whole in the style of a modern historian. Thus an ancient history like 1-2 Kings or Acts preserves extracts from documents far older than itself. Fragments of these survive in its pages, slightly edited but generally distinguishable from one another and from the final editorial matter. To reject such a method as slavish, and its results as a mere patchwork unworthy of a good historian, is simply to betray ignorance of the literary conditions under which these ancient books were written. Compilation, in fact, is a providential habit. In some cases its prevalence secured the life of several early documents which would otherwise have perished, since it delivered ancient historians from the need of producing free compositions in which it would have been impossible for later ages to distinguish between the substratum of good tradition and the overlying structures of the editor. The Matthaean logia preserved and reset in Matthew and Luke, the small apocalypse of the synoptists, the sources framed in Acts, the apocalyptic fragments transcribed and recoloured in the Apocalypse, the Pauline notes embedded in the pastoral epistles, are all NT instances of our debt to a habit of compilation whose very roughness (from the standpoint of modern literature) is its chief recommendation to historical research. present day by most schools of criticism, from Weiss and Zahn to Hilgenfeld and Lipsius. Jülicher finishes his prolegomena (Einl. p. 18) with a contemptuous sentence upon the modern school, chiefly Dutch, represented by Straatman, Rovers, and Baljon, who find Paul's letters strewn with interpolations: "Hätten Jene Recht, so müsste der liebe Gott im 1 und 2 Jhdt. 90 bis 120 Hände in Bewegung gesetzt haben zu einer beispiellosen Verkrüppelung aller NTlichen Texte mit dem Zweck, dem Scharfsinn der Theologen des 19 Jhdts., der sonst keine Aufgaben mehr kennt, ein Feld zu glänzendster Bethätigung zu schaffen." 1 The reliance upon sources of different value (earlier histories, state papers, evidence of eye-witnesses), and the obvious combination of these, can be easily seen, e.g., in Josephus' 'Ioudaïz 'Apxaioλoyiz, of which chaps. xii.-xviii. form an excellent instance of literary method in the first Christian century. For Eastern historians cp. Robertson Smith (OTJC,2 pp. 113f., 129 f., 328 f.) and Bennett (BI, pp. 17-19). The extent to which the habit is employed varies with the writer, e.g. 3 Esdras is little more than a cento of older passages, chiefly taken from 2 Chron and Neh; while Tacitus, Strabo, and as a rule Josephus, use their sources in quite a free and masterly fashion. 1 Macc incorporates various letters, which in some cases may be of first-rate authority; but 2 Macc is a clearer instance of pure compilation, consisting in its present form of two letters, followed by an abridgment of the earlier history of Jason. The book of Baruch, again, includes three or four pieces, of which one (115-38) may be nearly two centuries earlier than the whole volume. 1 Along the lines of this method 1 the author of Acts would proceed to work, upon a plan not essentially different, we may presume, from that followed by Layamon. He would work up his sources into the extant history very much as the mediaeval chronicler is reported to have compiled his "Brut" or "Chronicle of Britain" from the three "noble books" in which he found his chief materials, namely: the English book of Bede, Albin, and the fair Austin's Latin treatise, but especially Wace's Norman-French version of Geoffrey's history. Then Layamon before him laid these books, And turned their leaves and lovingly beheld them; And so he wrote on book-skin, and true words Three books.2 At the same time this illustration must be admitted to be deficient in representing the persistent and free process of editorial treatment in the NT histories. These books are not mosaics pieced together from early sources. The interstices are filled up by explanations, alterations, expansions, and corrections. The purpose which the author set before himself dominates his materials throughout, and determines where they are to be admitted, re-arranged, or supplemented. As a result, the better the historian, the more intricate the history-intricate, that is, from the point of view of the modern analyst, whose business it is to recover, if possible, the more primitive traditions which have been embellished and overlaid by subsequent deposits. In certain forms this hypothesis is cognate to that of interpolation. A writing interpolated on a large scale practically comes to be indistinguishable from a composite work, especially if this process is due to one hand. When an author had the opportunity for doing so, it was open to him either to revise previous work of his own, or to re-cast the work 1 Harnack's outline of a single Jewish document and its use in later writings affords an interesting illustration (Die Apostellehre u. die Jüdischen beiden Wege, 1886, pp. 31, 32) :— Jewish catechism of the "Two Ways" 2 Abbey, Relig. Thought in Old Eng. Verse, p. 30. The use of documents in the synoptic gospels has also been paralleled from chroniclers-" conspicuous among mediaeval writers for intelligence and trustworthiness "-like Benedict of Peterborough and Roger of Hovenden (cp. C. Plummer, Exp.3 x. pp. 23-35). Proof is led that not merely are later usages transferred to an earlier period by such historians in all good faith, but also "even in documents given textually, changes-unimportant, no doubt, but still changes-may be made more or less unconsciously by a perfectly honest and scrupulous writer." The monastic chronicles and charters furnish similar evidence, but the extent of their periods and the more deliberate motives at work in the process of their accretion bring them nearer to a book like Chronicles than to any of the NT histories. of others. On this hypothesis, a book may be found to contain one or more pieces, each issuing from a different period, so that what is to all appearance a literary unity, really consists of several smaller unities. Two alternatives are possible. Either, as in the case of John's apocalypse, the whole was substantially re-cast at the time of the final editing, or, as in the closing chapter of Mark's gospel (?), and (as Destinon and Wellhausen conjecture) 1 Mac 14-16, and Jn 21 (?), one part was simply added to another, both remaining as they originally stood. In the sphere of apocalyptic literature especially, the use of older fragments was the more natural owing to the traditional character of the writings. The apocalyptic writers were not distinctively creative. They inherited a set of ideas, forms of phrases, terms of speech. These were their stock-in-trade, and it was to be expected that such writers should frequently adopt and adapt earlier pieces which had come down to them upon the stream of an honoured antiquity. 1 A particular phase of compilation is to be seen in the use of florilegia by the NT writers and the early fathers in their quotations from the OT. The deliberate and composite character of these quotations suggests that they are secondary, taken not from the originals but from collections of texts upon different subjects which were arranged in a certain order. Hence it has been conjectured that such a florilegium of Messianic passages-drawn up to illustrate topics like "the Fore-runner," "the sufferings of Messiah," "the call of the Gentiles," etc.--was employed by Paul and the other authors of the NT. The needs of controversy would lead to the production of such an anthology among the Jews, who constantly fell back upon the proof from scripture, yet could not depend upon their hearers possessing a convenient copy to which appeal might be successfully made. "The hypothesis is hardly to be rejected. The Jewish catechetical and missionary instruction in the Diaspora needed such collections, and their existence seems to be proved by the Christian apologies and the Sibylline books" (Harnack, HD, i. p. 175). This would help to account for the grouping of different quotations under an inexact title (e.g. Mk 12. 3), as well as for some of the more common divergences from the LXX text. Questions of structure, however, do not rank so prominently in the NT as in the OT literature. 2 The three certain instances are, Luke, Acts, and the Apocalypse. The third gospel is far from being a mere compilation, but it has used sources freely, including the gospel of 3 1 On the contemporary habit of drawing up such manuals, and the proofs for it in the early Christian writers, see especially Hatch (Essays in Bibl. Greek, p. 203 f.). So Sanday and Headlam (Romans, ICC, pp. 204, 282), the author of The Logia of Papias (1894), pp. v-vii, and Vollmer (Die alttestamentliche Citate bei Paulus, 1895). The last-named argues that while Paul quoted usually from the LXX, some hypothesis like this is needed to explain deviations such as 1 Co 29. It has been similarly suggested that a number of the Latin, French, and Provençal books quoted by Dante were known to him merely through the indirect form of collected extracts. 2 On the composite nature of Semitic literature, cp. Sayce, Higher Crit. Monum. (1894), chap. ii.; for apocalyptic literature, cp. R. H. Charles' editions of Enoch, the Book of Jubilees, and Baruch, besides the respective introductions in KAP. Most valuable of all, perhaps, is Professor Carpenter's scholarly discussion in The Hexateuch (1900), vol. i. chap. i., where both evidence and analogies for the stratification of literature are given with much clearness and precision. 3 On these see Resch, TU, x. 3. pp. 333-347, and below, pp. 652, 653. If the distinction be allowable, in the process of evangelic composition the |