have generally been given up as purely visionary, and the very word families has come into disrepute by reason of the exploded fancies it recalls, we can discern not the less clearly that certain groups of them have in common not only a general resemblance in regard to the readings they exhibit, but characteristic peculiarities attaching themselves to each group. Systematic or wilful corruption of the sacred text, at least on a scale worth taking into account, there would seem to have been almost none; yet the tendency to licentious paraphrase and unwarranted additions distinguished one set of our witnesses from the second century downwards; a bias towards grammatical and critical purism and needless omissions appertained to another; while a third was only too apt to soften what might seem harsh, to smooth over difficulties, and to bring passages, especially of the Synoptic Gospels, into unnatural harmony with each other. All these changes appear to have been going on without notice. during the whole of the third and fourth centuries, and except that the great name of Origen is associated (not always happily) with one class of them, were rather the work of transcribers than of scholars. Eusebius and Jerome, in their judgments. about Scripture texts, are more the echoes of Origen than independent investigators.

Now, as a first approximation to the actual state of the case, the several classes of changes which we have enumerated admit of a certain rude geographical distribution, one of them appertaining to Western Christendom and the earliest Fathers of the African and Gallic Churches (including North Italy under the latter appellation); a second to Egypt and its neighbourhood; the third originally to Syria and Christian Antioch, in later times to the Patriarchate of Constantinople. We have here, no doubt, much to remind us of Griesbach and his scheme of triple recensions on which we enlarged in Chapter VI. (pp. 470-3,536-7), but with this broad distinction between his conclusions and those of modern critics, that whereas he regarded the existence of his families as a patent fact, and grounded upon it precise and mechanical rules for the arrangement of the text, we are now content to perceive no more than unconscious tendencies, liable to be modified or diverted by a thousand occult influences, of which in each single case it is impossible to form an estimate beforehand. Even that marked bias in the direc

tion of adding to the record, which is the reproach of Codex Bezæ and some of its compeers, and renders the text of the Acts as exhibited by DE, by the cursive 137, and the margin of the Philoxenian Syriac, as unlike that commonly read as can well be imagined', is mixed up with a proneness to omissions which we should look for rather from another class of documents (e.g. the rejection of eudóμevoι Matth. v. 11), and whịch in the latter part of S. Luke's Gospel almost suggests the idea of representing an earlier edition than that now in ordinary use, yet proceeding from the Evangelist's own hand (see p. 18)2. Again, the process whereby the rough places are made plain and abrupt constructions rounded, is abundantly exemplified in the readings of the great uncial A, supported as it is by the mass of later manuscripts (e.g. Mark i. 27; Acts xv. 17, 18; xx. 24); yet in innumerable instances (as we have sufficiently proved in pp. 543-52), these self-same codices retain the genuine text of the sacred writers which their more illustrious compeers have lost or impaired.

Hence it follows that in judging of the character of a various reading proposed for our acceptance, we must carefully mark whether it comes to us from many directions or from one. And herein the native country of the several documents, even when we can make sure of it, is only a precarious guide. If the Æthiopic or the Armenian versions have really been corrected by

1 So that we may be sure what we should have found in Cod. D, and with high probability in Cod. E, were they not defective, when in Acts xxvii. 5 we observe δι' ἡμερῶν δεκαπέντε inserted after διαπλεύσαντες in 137. 184, and the Philoxenian margin with an asterisk; as also when we note in Acts xxviii. 16 ¿§w TĤs #apeμßolîs before σvv in the last two and in demid. See also p. 552 note 1. 2 e. g. Luke xxiv. 3 TOû Kuρlov iŋσoû omitted by D. a. b. e. ff.2 l.; ver. 6 Ouk čσTIV ŵde àììà ǹyéρọn (comp. Mark xvi. 6) omitted by the same; ver. 9 åлò тоû μvnμciov by the same, by c. and the Armenian; the whole of ver. 12, by the same (except 2.) with fuld., but surely not by the Jerusalem Syriac, even according to Tischendorf's shewing, or by Eusebius' canon, for he knew the verse well (comp. John xx. 5); ver. 36 kaì λéyeɩ avtoîs, eipývŋ vuîv omitted by D. a. b. e. ff2. l. as before (comp. John xx. 19, 26); the whole of ver. 40, omitted by the same and by Cureton's Syriac (comp. John xx. 20); ver. 51 xai ȧvepéρeTO eis Tòv оúpavov and ver. 52 πроoκυvýσavres aúròv omitted by the same and by Augustine, the important clause in ver. 51 by * also, and consequently by Tischendorf. Yet, as if to shew how mixed the evidence is, D deserts a. b. 2. l. when, in company with a host of authorities, both manuscripts and versions (f. q. Vulgate, Memphitic, Syriac, and others), they annex кai άяò μeλioσlov кnplov to the end of ver. 42. See also Luke x. 41, 42; xxii. 19, 20, discussed in Chap. ix.

the Latin Vulgate (see p. 408, 410, &c.), the geographical remoteness of their origin must go for nothing where they agree with the latter version. The relation in which Cod. L and the Memphitic version stand to Cod. B is too close to allow them their full value as independent witnesses unless when they are at variance with that great uncial, wheresoever it may have been written the same might be said of the beautiful Latin fragment k from Bobbio (see p. 344). To whatever nations they belong, their resemblances are too strong and perpetual not to compel us to withhold from them a part of the consideration their concord would otherwise lay claim to. The same is incontestably the case with the Curetonian and margin of the Philoxenian Syriac in connexion with Cod. D. Wide as is the region which separates Syria from Gaul, there must have been in very early times some remote communication by which the stream of Eastern testimony or tradition, like another Alpheus, rose up again with fresh strength to irrigate the regions of the distant West. The Peshito Syriac leans at times in the same direction, although both in nation and character it most assimilates to the same class as Cod. A.

With these, and it may be with some further reservations which experience and study shall hereafter suggest, the principle of grouping must be acknowledged to be a sound one, and those lines of evidence to be least likely to lead us astray which converge from the most varied quarters to the same point. It is strange, but not more strange than needful, that we are compelled in the cause of truth to make one stipulation more: namely, that this rule be henceforth applied impartially in all cases, as well when it will tell in favour of the Received text, as when it shall help to set it aside. To assign a high value to cursive manuscripts of the best description (such as 1. 33. 69. 157. Evst. 259, or 61 of the Acts), and to such uncials as LRA, or even as & or C, whensoever they happen to agree with Cod. B, and to treat their refined silver as though it had been suddenly transmuted into dross when they come to contradict it, is a practice too plainly unreasonable to admit of serious defence, and can only lead to results which those who uphold it would be the first to deplore1.

So of certain of the chief versions we sometimes hear it said that they are

17. It is hoped that the general issue of the foregoing discussion may now be embodied in these four practical rules1:

(1) That the true readings of the Greek New Testament cannot safely be derived from any one set of authorities, whether manuscripts, versions, or Fathers, but ought to be the result of a patient comparison and careful estimate of the evidence supplied by them all.

(2) That where there is a real agreement between all documents containing the Gospels up to the sixth century, and in other parts of the New Testament up to the ninth, the testimony of later manuscripts and versions, though not to be rejected unheard, must be regarded with great suspicion, and, UNLESS UPHELD BY STRONG INTERNAL EVIDENCE, can hardly be adopted.

(3) That where the more ancient documents are at variance with each other, the later uncial and cursive copies, especially those of approved merit, are of real importance, as being the surviving representatives of other codices, very probably as early, perhaps even earlier, than any now extant.

(4) That in weighing conflicting evidence we must assign the highest value not to those readings which are attested by the greatest number of witnesses, but to those which come to us from several remote and independent sources, and which bear the least likeness to each other in respect to genius and general character.

less important in the rest of the N. T. than in the Gospels; which means that in the former they side less with NB.

1 Canon Kennedy, whose Ely Lectures exhibit, to say the least, no prejudice against the principles enunciated in Dr Hort's Introduction, is good enough to commend the four rules here set forth to the attention of his readers (p. 159, note). The first three were stated in my first edition (1861), the fourth added in the second edition (1874), and, while they will not satisfy the advocates of extreme views on either side, suffice to intimate the terms on which the respective claims of the uncial and cursive manuscripts, of the earlier and the more recent authorities, may, in my deliberate judgment, be equitably adjusted.

CHAPTER VIII.

CONSIDERATIONS DERIVED FROM THE PECULIAR CHARACTER AND GRAMMATICAL FORM OF THE DIALECT OF THE GREEK TESTAMENT.

1.

IT

T will not be expected of us to enter in this place upon the wide subject of the origin, genius, and peculiarities, whether in respect to grammar or orthography, of that dialect of the Greek in which the N. T. was written, except so far as it bears directly upon the criticism of the sacred volume. Questions, however, are perpetually arising, when we come to examine the oldest manuscripts of Scripture, which cannot be resolved unless we bear in mind the leading particulars wherein the diction of the Evangelists and Apostles differs not only from that of pure classical models, but also of their own contemporaries who composed in the Greek language, or used it as their ordinary tongue.

2. The Greek style of the N. T., then, is the result of blending two independent elements, the debased vernacular speech of the age, and that strange modification of the Alexandrian dialect which first appeared in the Septuagint version of the Old Testament, and which, from their habitual use of that version, had become familiar to the Jews in all nations under heaven; and was the more readily adopted by those whose native language was Aramæan, from its profuse employment of Hebrew idioms and forms of expression. It is to this latter, the Greek of the Septuagint, of the Apocalypse, and of the foreign Jews, that the name of Hellenistic (Acts vi. 1) strictly applies. S. Paul, who was born in a pure Greek city (Juvenal, III. 114-118); perhaps even S. Luke, whose original writings'

1 viz. Luke i. 1-4, some portion of the Gospel and most of the Acts: excluding such cases as S. Stephen's speech, Act. vii, and the parts of his Gospel which resemble in style, and were derived from the same sources as, those of SS. Matthew and Mark.