Chapter III: Theory and Praxis of the Textual Criticism of the New Testament.129 (1)

There is no special theory of the textual criticism of the New Testament. The task and the method are the same for all literary productions. The task is to exhibit what the original writer intended to communicate to his readers, and the method is simply that of tracing the history of the document in question back to its beginning, if, and in so far as, we have the means to do so at our command. Diversity of treatment can only arise when the fortunes of one written work have been more chequered and complicated than those of another, or when we have more abundant means at our disposal to help us in the one case than in the other. The task is very simple when we have only one completely independent document to deal with, as in the case of several of the recently discovered papyri, but this occurs very seldom with literary texts. In this case all that we have to do is to see that we read the existing text correctly, and then by means of the so-called internal criticism to determine whether the text so received can be correct. [Sidenote: Internal Criticism.] Even when several witnesses are at our command, we cannot altogether dispense with this internal criticism in the matter of sifting and weighing their testimony, only it would be unfortunate were we left with such a subjective criterion alone. For not only in such a case would different scholars come to very different conclusions, but even one and the same scholar would not be able to avoid a certain amount of uncertainty and inconsistency in most cases. The principle laid down in the maxim, _lectio difficilior placet_, or, as Bengel more correctly and more cautiously puts it, _proclivi scriptioni praestat ardua_, is perfectly sound; that reading is correct, is the original reading, from which the origin of another or of several others can be most easily explained. But how seldom can this be established with certainty! Take an illustration:—

[Sidenote: Conclusion of the Apocalypse.]

How does the Apocalypse, and the New Testament with it, conclude? Leaving out of account additions like “Amen” or “Amen, Amen,” and variations like “The grace of the Lord Jesus,” and “our Lord Jesus,” and “the Lord Jesus Christ,” and “Christ” simply, we find that the following forms are given:—

(1) μετὰ πάντων ὑμῶν
(2) μετὰ πάντων ἡμῶν
(3) μετὰ πάντων τῶν ἁγίων
(4) μετὰ πάντων
(5) μετὰ τῶν ἁγίων

How are we to decide without external evidence which is the correct form? Even supposing we know that the first two are out of the question, and why they are so, it is very difficult on internal grounds alone to decide between the other three. Lachmann, who did not know of (5), decided in favour of (4). But so does Tischendorf, Weizsäcker, and Weiss, the latter giving as his reason for doing so that (5), τῶν ἁγίων, is explanatory of (4), πάντων, which is manifestly too general, and that (3) is the result of a combination of these two. On the other hand, Tregelles and Westcott and Hort favour (5), without so much as mentioning (4) in their margin; while Bousset, the latest expositor of the Apocalypse, regards (3) as the correct reading, and thinks that in all probability both (4) and (5) are due to a transcriptional error. Who is to decide when doctors disagree? Manifestly one might argue on quite as good if not better grounds than those of Weiss to the very opposite conclusion—viz. that a later writer who wished the Apocalypse, and with it the New Testament, to conclude with as comprehensive a benediction as possible, substituted the words “Grace be with all” in place of the restricted and somewhat strange expression “Grace be with the saints.” I did not observe that Bousset still defends the third form when I said in the first edition of this work that this reading does not fall to be considered at all. But my reason for saying so was not “because this form proves to be a combination of the other two,” or “because the authorities for it are later,” but because it could be shown that its supporters follow a corrected text in other places as well as this; and I concluded with observing that the decision between (4) and (5) could not be made to depend solely on internal criteria either, but depended on the decision come to regarding the general relationship between the witnesses that support each one, in this instance between A, as supporting (4), and א, as supporting (5).

(1) It may be stated here, merely by way of comment, that the first
form of the benediction, “with you all,” was clearly translated into
Greek by Erasmus from his Latin Bible, without the authority of a
single Greek manuscript. But in spite of this, it is still
propagated in the _textus receptus_ by the English Bible Society,
and even in the last revision of Luther’s German Bible it was
allowed to stand without demur. The English Authorised Version had
it in this form, but the Revised Version adopts the fifth form “with
the saints,” and puts (4) in the margin, with a note to the effect
that “two ancient authorities read ‘with all.’”

The second form, “with us all,” which was adopted by Melanchthon in
his Greek Bible of 1545, published by Herwag, is just as arbitrary
an alteration. The third form, “with all the saints,” is read by the
Complutensian with Q, with more than forty minuscules, and the
Syriac, Coptic, and Armenian versions. The fourth, “with all,” is
found in A and Codex Amiatinus, while the fifth, “with the saints,”
is given by א and the Old Latin g. In the Syriac version of the
Apocalypse, edited by Gwynn in 1897, a sixth form seems to have been
brought to light, which Baljon, who himself decides for (5), cites
as μετὰ πάντων τῶν ἁγίων αὐτοῦ: Syr^{gwynn}. But the pronoun, which
in Syriac is indicated by a suffix only, is employed now and again
merely to represent the Greek definite article, so that this new
Syriac manuscript does not give us a sixth form but only another
witness to the third. On the other hand, Gwynn mentions the omission
of the entire verse in Primasius, a fact that neither Tischendorf
nor Weiss takes the least notice of, and he adduces lastly that a
manuscript of the Vulgate reads “cum omnibus _hominibus_.” One sees
from an illustration like this what an amount of pains is required
seriously to apply, even in a single point, Bengel’s principle that
the smallest particle of gold is gold, but that nothing must be
passed as gold that has not been proved to be such (_Introductio in
Crisin Novi Testamenti_, § 1, p. 572).

(2) LITERATURE.—See especially Gebhardt (_Urt._, p. 16). Ed. Reuss,
_Geschichte der h. Schriften des N. T._, Braunschweig, 1887, § 351
ff. S. P. Tregelles, _An Introduction to the Textual Criticism of
the N.T._ (= vol. iv. of Horne’s _Introduction_, 1877). F. H. A.
Scrivener (see above, p. 6); also _Adversaria Critica Sacra_, edited
by Miller, Cambr. 1893. B. F. Westcott, _The New Testament_ in
Smith’s _Dictionary of the Bible_, vol. ii., London, 1863. C. E.
Hammond, _Outlines of Textual Criticism_, Oxford, 1890.
Westcott-Hort, vol. ii. (see p. 21). B. B. Warfield, _Introduction
to the Textual Criticism of the N. T._, New York, 1887; London,
1893. J. W. Burgon, _Last Twelve Verses of the Gospel according to
St. Mark_, Oxford and London, 1871; also _The Traditional Text of
the Holy Gospels vindicated and established_, edited by Miller,
London, 1896; also, _The Causes of the Corruption of the Traditional
Text of the Holy Gospels_, edited by Miller, London and Cambridge,
1896. _The Oxford Debate on the Textual Criticism of the N. T. held
at New College on May 6, 1897_; with a preface (by Miller)
explanatory of the Rival Systems, 1897, pp. xvi. 43. Ed. Miller,
_The Present State of the Textual Controversy respecting the Holy
Gospels_ (see above, p. 152).

Martin (Abbé J. P.), _Introduction à la Critique textuelle du N.T._,
in five volumes, with plates and facsimiles: vol. i. pp. xxxvi. 327,
Paris, 1884, 25 fr.; vol. ii. pp. ix. 554, 1884, 40 fr.; vol. iii.
pp. vi. 512, 1885, 40 fr.; vol. iv. pp. vi. 549, 1886, 40 fr.; vol.
v. pp. xi. 248 and 50 pp. of facsimiles, 1886, 20 fr. Also by the
same author, _Description technique des manuscrits grecs relatifs au
N.T. conservés dans les Bibliothèques de Paris_. Supplement to the
foregoing, Paris, 1884, pp. xix. 205, with facsimiles, 20 fr.;
_Quatre manuscrits importants du N.T. auxquels on peut en ajouter un
cinquième_, Paris, 1886, pp. 62, 3 fr.; _Les plus anciens mss. grecs
du N.T., leur origine, leur véritable caractère_, in the _Revue des
Quest. Hist._, 1884, No. 71, pp. 62-109; _Origène et la Critique
textuelle du N.T._, Paris. Reprinted from the _Rev. des Quest.
Hist._ for Jan. 1885, No. 73, pp. 5-62.

Th. Zahn, _Geschichte des N.T. Kanons_: vol. i., _Das N.T. vor
Origenes_, Part 1, 1888; Part 2, 1889. Vol. ii., _Urkunden und
Belege zum ersten und dritten Band_, Part 1, 1890; Part 2, 1892. The
third vol. has not yet appeared. The order of the books of the N.T.
is discussed in vol. ii. p. 343 ff., and the conclusion of Mark’s
Gospel in the same vol., p. 910 ff.

Salmon (Geo.), _Some thoughts on the Textual Criticism of the N.T._,
London, 1897, pp. xv. 162. Blass, _Philology of the Gospels_,
London, 1898, pp. viii. 250. Ada Bryson, _Recent Literature on the
text of the N.T._ in the _Expository Times_ for April 1899, pp.
294-300. M. Vincent, _History of the Textual Criticism of the N.T._,
1900. G. L. Cary, _The Synoptic Gospels, with a chapter on the
Textual Criticism of the N.T._, New York, 1900. See also Prof.
Jannaris in the _Expositor_, vol. viii. of Series V. There is an
article in the _American Journal of Theology_, 1897, iv. p. 927 ff.,
entitled _Alexandria and the N.T._, which I have not been able to
consult.

In attempting to restore the text of the New Testament as nearly as possible to its original form, it is essential to remember that the New Testament, as we have it to-day, is not all of one piece, but consists of twenty-seven separate documents now arranged in five groups, and that every several document and every several group has had its own peculiar history. Of these groups the most complicated, perhaps, is the one with which the New Testament opens—viz. the Gospels.

[Sidenote: Gospels.]

It is quite uncertain when our four Gospels were first written together in one volume and arranged in the order that is now common. The Muratorian Fragment on the Canon[130] is defective at the beginning, but seems to imply this arrangement. It was supposed that the Gospels were written in the following order—viz. Matthew first and John last. The order, Matthew, Mark, Luke, John, which is found in nearly all the Greek and Syriac manuscripts, was made popular by Eusebius and Jerome. The former followed it in his _Canons_, which were afterwards adopted by Jerome in his Latin Bible. According to Eusebius (_Eccl. Hist._, vi. 25),[131] Origen knew this order, though he very frequently cites the Gospels in the order Matthew, Luke, Mark.

The following arrangements are also found:—

[Transcriber’s Note: The following list begins with (2).]

(2) Matthew, Mark, John, Luke, in the earlier (Curetonian) Syriac and in the Canon Mommsenianus, a catalogue of the Books of the Bible and of the works of Cyprian, originating in Latin Africa about the year 360, and first published by Mommsen.[132]

(3) Matthew, Luke, Mark, John, in the so-called Ambrosiaster and in a Catalogue of the Sixty Canonical Books.

(4) Matthew, John, Mark, Luke—_i.e._ the two Apostles put before the two pupils of Apostles, in the Codex Claromontanus.[133] This order occurs also in the Arabic writer Masudi’s _Meadows of Gold_.[134]

(5) Matthew, John, Luke, Mark, in Codd. D and X, in the _Apostolic Constitutions_, in Ulfilas, and especially in the Old Latin Manuscripts; see Corssen, _Monarchianische Prologe_, p. 65, in _TU._ xv. 1.[135]

(6) John, Luke, Mark, Matthew, in Codex k.

(7) John, Matthew, Mark, Luke, in the Vocabularies of the Egyptian versions.

(8) John, Matthew, Luke, Mark, in Tertullian and cod. 19. See Arthur Wright, _Some New Testament Problems_, p. 196 ff.[136]

This very variety shows that for a long time, perhaps till the third century, at all events much longer than the Pauline Epistles, the Gospels were propagated singly, perhaps on rolls, and only afterwards incorporated in a codex. And this makes it probable that the text of our manuscripts was not taken from a single copy of the first Tetraevangelium. More than probable we cannot call it, seeing that a copyist may have had any sort of reasons of his own for disarranging the order of the books given in his exemplar, as may still be gathered luckily from the position occupied by Hebrews in Codex B. The probability is heightened, however, by the fact that our manuscripts display a considerably greater amount of textual variation in the Gospels than in the Pauline Epistles, though not in all to the same extent as in D which contains an entirely peculiar recension, especially in Luke. One of the most remarkable indications of this is afforded by the discovery made by E. Lippelt, a pupil of Professor Blass. The order of the books in D is Matthew, John, Luke, Mark, Acts, where it will be seen that the two portions of the book inscribed to Theophilus are separated by Mark. Now Lippelt observed that while the name Johannes is regularly spelt with two ν’s (Ἰωάννης) in Matthew, John, and Mark, it is just as regularly spelt with one (Ἰωάνης) in Luke and Acts, sundered though these two books are by Mark, where the other spelling prevails.[137] This shows an accuracy of tradition which is surprising, but till now it has only been traced in this one manuscript. The others write the name throughout with two _ν_’s and B as consistently with one. In this connection the question naturally arises whether certain liberties were not taken with the books on the occasion of their collection and arrangement. Resch, _e.g._, thinks that it was then that the second Gospel received the conclusion or appendix which is found in most of our manuscripts, and Rohrbach holds a similar opinion.[138] I have elsewhere expressed the idea that the peculiar opening of Mark is to be accounted for in this way.[139] Zahn, however, doubts whether the use of ἀρχή and τέλος for ἄρχεται and ἐτελέσθη, _incipit_ and _explicit_, can be established for early times.[140] I have found it in Greek Psalters, though not very early, I admit, where αρχη των ωδων occurs instead of ωδαι as the superscription of the Hymns at the end of the Psalter.[141] However, there is no need to dwell further on this point. Zahn (p. 174) is quite right in his contention that the usual titles κατὰ Μαθθαῖον, etc.,[142] imply a collection of the Gospels of which Εὐαγγέλιον is the general title.

If, then, for the sake of simplicity, we take as our goal the first manuscript of the Tetraevangelium, one would think it must be possible with the means at our command gradually to work back to it. Even the latest of our manuscripts is surely copied from an earlier one, and that one from another, and so on always further and further back, so that all we have to do is to establish their genealogy, pretty much as Reuss has done for the printed editions of the New Testament; and seeing we have manuscripts as old as the fourth and fifth century, that means that the entire period of a thousand years prior to the invention of printing is bridged over at once, so that the task would appear to be simply that of throwing a bridge over the first few centuries of the Christian era. And by going on comparing the witnesses and always eliminating those that prove unreliable, it must be possible, one would suppose, in this way to arrive at the original. But a little experience will shortly moderate our expectations.

At the outset it is very much against us that we have no really serviceable text for comparison. The text of our present critical editions is a patchwork of many colours, more wonderful than the cloak of Child Roland of old. In fact it is a text that never really existed at all. In the preparation of my _Supplement_, which I undertook with the object of making the text of Codex Bezae easily accessible to every one, I compared the text of that manuscript with that of Tischendorf-Gebhardt’s edition, and I saw clearly that my work would necessarily present a very confused appearance indeed. I also issued an interleaved edition of my Stuttgart New Testament with a similar object—viz. to furnish a convenient means of comparing the text of manuscripts and of Patristic quotations, but that, too, labours under the same disadvantage. Whoever intends really to further the textual criticism of the New Testament will have to issue a copy of a single manuscript printed in such a way as will make it practically convenient for the comparison of different texts, something like Tischendorf’s edition of Codex Sinaiticus (_Novum Testamentum Sinaiticum_, 1863), which, however, is of little use for other purposes, or like Schjøtt’s edition of the New Testament (see above, p. 24). But as these are in the hands of very few, there is nothing for it at present but to take one of our most common texts, always bearing in mind its composite character. This feature of the text appears at the very outset in the title. In א B (D) it is κατα Μαθθαιον. Codex D is defective at the beginning down to c. i. 20, but κατ Μαθθαιον is found regularly as the title at the top of the pages, a fact which Tischendorf has overlooked. Most other manuscripts, C E K M etc., have Ευαγγελιον κατα Ματθαιον. If this latter is held as incorrect, then all these manuscripts should for the future be dropped out of account and א B D alone be regarded as authoritative.

Again, in verse 2, א* has Ισακ twice, while the others have Ισαακ, so that א too would drop out, leaving B standing alone. But then in verse 3 our editors forthwith reject B, which reads Ζαρε, and decide in favour of the others which have Ζαρα. Whether this may not be a little premature, seeing that there are other places where ε is found for final ח,[143] and that one manuscript, 56, has deliberately corrected Ζαρα into Ζαρε in Gen. xxxviii. 30, where a third has Ζαρε, we do not pause to determine. The point is simply this, that in these first three verses there is no manuscript that is always right in the judgment of our editors. True, the cases we have been considering are trifling, the differences being of an orthographical nature merely, and one must not be too particular in such matters, though at the same time the oft-quoted maxim, _minima non curat praetor_, is nowhere less applicable than in textual criticism. But the same state of things reappears immediately where we have differences involving important matters of fact. What is the fact in verse 11? Did Josias beget Jechoniah, or did he beget Joachim and Joachim Jechoniah? Verse 16 has already been referred to: in this case our oldest Greek manuscripts would give no occasion to mention the verse. But in verse 25 we have again to ask which is correct, ἔτεκεν υἱόν or ἔτεκεν τὸν υἱὸν αὐτῆς τὸν πρωτότοκον? [Sidenote: Dogmatic influence.] And when we hear Jerome say—Ex hoc loco quidam perversissime suspicantur et alios filios habuisse Mariam, dicentes primogenitum non dici nisi qui habeat et fratres, we learn already how dogmatic motives may have some influence upon the form of the text. And, moreover, when we call to mind the words of Luke ii. 7, we are made aware of another thing that may exert a disturbing influence in the Gospels—viz. the tendency to alter the text in conformity with the parallel passage. [Sidenote: Parallel passages.] Apart from the stylistic peculiarities of Codex D, we meet with no materially important variants in our Greek manuscripts of Matthew till we come to the Sermon on the Mount. The only thing is in iii. 15, where two Latin witnesses have an addition which is evidently taken from a Greek source: et cum baptizaretur, lumen ingens circumfulsit de aqua ita ut timerent omnes qui advenerant (congregati erant). This interpolation, however, does not concern the criticism of the text of the New Testament, seeing that it is derived from some source outside the Canon.

On the other hand, there is a great question as to the order of the first three Beatitudes in Matthew v. 3-5, whether they are to be read in the order given in the common text, πτωχοί ... πενθοῦντες ... πραεῖς ..., or as our recent editors prefer πτωχοί ... πραεῖς ... πενθοῦντες.[144] The latter arrangement is attested by only two Greek manuscripts—D and 33. Now, if their evidence is accepted here in spite of its apparent weakness, how can we justify the refusal to acknowledge the authority of D in other similar cases? Verse 22, but a short way down, is a case in point. Here D, with most authorities, exhibits the sorely-contested εἰκῇ. But our modern critics will have nothing to do with it, going by א B, Origen, Jerome, and Athanasius. Merx (_Die vier kanonischen Evangelien_, pp. 231-237) has recently come forward as a strong supporter of it, on the ground that Syr^{sin} also has it,[145] but how is its omission, especially by Jerome, to be explained? The Vulgate itself shows that it was easier to insert it than to omit it, because out of twenty-four manuscripts collated in W.-W. three have it, though it certainly does not belong to the text of the Vulgate.[146]

In view of these illustrations, which serve to show the somewhat haphazard way in which the text of our editions hitherto has been arrived at, the question becomes very important how the original text is to be restored in disputed or doubtful cases.

[Sidenote: Conjectural emendation.]

The first case, or, if we like to call it, the last, but at all events the one most easy of settlement, is when the correct reading is no longer found in any of our witnesses, neither in Greek manuscript, version, nor patristic quotation. Here we must simply have recourse to =conjecture=. Not long ago philologists evinced such a fondness for conjectural emendation that the question might not unreasonably be asked why they did not rather themselves write the text that they took in hand to explain. At the same time, the aversion to this method of criticism which till recently prevailed and still to some extent prevails, especially in the matter of the New Testament text, is just as unreasonable. Tischendorf, _e.g._, did not admit a single emendation of this nature into his text, while Westcott and Hort consider it to be necessary in only a very few cases, such as Colossians ii. 18, though they also decline to adopt any conjectural readings in their text. For ΑΕΟΡΑΚΕΝΕΜΒΑΤΕΥΩΝ in this passage, which Weizsäcker renders “pluming himself upon his visions,” they would read ΑΕΡΑ ΚΕΝΕΜΒΑΤΕΥΩΝ, which is obtained by the omission of a single letter and a different division of the words. In Holland conjectural criticism is freely indulged in,[147] the example of Cobet and his school being followed by such critics as S. A. Naber, W. C. van Manen, W. H. van de Sande-Bakhuyzen, van de Becke Callenfels, D. Harting, S. S. de Koe, H. Franssen, J. M. S. Baljon, J. H. A. Michelsen, and J. Cramer.[148] Baljon has adopted a great number of such conjectural emendations in his edition of the text published in 1898 (see above, p. 24). In place of πολλοὶ διδάσκαλοι, _e.g._, in James iii. 1, Lachmann would read πῶλοι δύσκολοι, Naber πλανοδιδάσκαλοι, while Junius, de Hoop-Scheffer, and Bakhuyzen prefer πολυλάλοι on the ground that m^{64} has _nolite multiloqui esse_.[149] So far, therefore, this last is not pure conjecture. For κρινέτω in Col. ii. 16 Lagarde wished to read κιρνάτω, because the verb דוד found in the Peshitto at this place is elsewhere used to translate θροεῖν (Matt. xxiv. 6), ταράσσειν (John xiv. 1, 27), ἐγκόπτειν (Gal. v. 7), and also διαστρέφειν (Eccl. vii. 18; xii. 3). My proposal to read ἐπὶ πόντον in Apoc. xviii. 17, a reading adopted by Baljon in his text, instead of ἐπὶ τόπον or ἐπὶ πλοίων as given in our manuscripts, was a pure conjecture, but it has the support of _super mare_ in Primasius.[150] There is therefore no objection on principle to the method of conjecture, nor to the adoption of conjectural readings in the text, though it is only to be resorted to as the _ultima ratio regis_ and with due regard to all the considerations involved, transcriptional, linguistic, and otherwise.[151] There is no essential difficulty in supposing, _e.g._, that κιρνάτω in Col. ii. 16 was first corrupted into κρινάτω and then into κρινέτω. Such a transposition of the liquid is quite common in all languages.[152] But we must see if κιρνάτω has the sense required in the passage. There is no doubt a reference to drinking here, and so far, therefore, the word seems to suit the context better. It is also true that evidence is not wanting of the metaphorical use of the word proposed to be inserted. Passow, _e.g._, gives τὸ τῆς φύσεως σκληρὸν κιρνᾶν from Polybius iv. 21, 3, and τὴν πόλιν κιρνᾶν from Aristophanes i. 1. In spite of this, however, I have considerable misgivings whether this sense of the word is in harmony with Pauline usage and is suitable to the context of the passage. If it is sought to justify a conjectural reading on transcriptional grounds, then, as has been observed (p. 82), a strict account must be taken of the manner of writing prevalent at the time when the corruption is supposed to have originated. Luke’s handwriting must have been very bad indeed if we are to suppose that the scribe of D or the parent manuscript mistook ἠρνήσασθε for the enigmatical ἐβαρύνατε in Acts iii. 14, though it is quite conceivable how he came to write δόξῃ instead of δεξιᾷ in v. 31, or conversely wrote ἐδέξαντο instead of ἐδόξασαν in xiii. 48, or that ΚΑΙΤΟΥΤΩΣΥΝΦΩΝΟΥΣΙΝ was made into ΚΑΙΟΥΤΩΣΣΥΝΦΩΝΗΣΟΥΣΙΝ or _vice versa_.[153] A slight experience in the reading of ancient manuscripts shows how easy it is to make mistakes of this sort. And if we wish to see what mistakes of this sort actually do occur in Codices Vaticanus, Alexandrinus, Sinaiticus, and Ephraemi, we have only to look into Morrish’s _Handy Concordance of the Septuagint_,[154] though of course the examples there are all from the Old Testament. We have, _e.g._, ἀγαπάω for ἀπατάω; ἀγάπη for ἀπάτη; ἁγιάζω for ἀγοράζω; ἅγιος for αἴγειος, ἀγγεῖον, ἀγρός, γῆ; ἀδιάλυτος for διάλυτος; βάλλω and its compounds for λαμβάνω and its compounds; λαός for ναός, etc.

[Sidenote: Eclectic method.]

It is more difficult to answer the question how the text is to be restored in cases where there is no lack of external evidence. We have already seen that critics have hitherto adopted an =eclectic= mode of procedure. In general, whenever Vaticanus and Sinaiticus agree, editors, Tischendorf as well as Westcott and Hort, give the preference to their testimony. But if they do not agree, what is to be done? And what if a third reading seems on internal grounds to be better than either? In his _Thoughts on the Textual Criticism of the New Testament_, Salmon very pungently, but not altogether incorrectly, describes Westcott and Hort’s method on the lines of an anecdote told of Cato by Cicero: “To the question what authorities should be followed, Hort answers, Follow B א. But if B is not supported by א? Still follow B, if it has the support of any other manuscript. But suppose B stands quite alone? Even then it is not safe to reject B unless it is clearly a clerical error. But suppose B is defective? Then follow א. And what about D? What about killing a man!” Lagarde has said that the gag is the modern equivalent of the stake. Codex D has not been gagged outright, to be sure, but it has been shoved aside, and only now and then with remarkable inconsistency has its evidence been accepted as trustworthy. For one must surely call it inconsistent to follow one side as a rule and then all at once to take sides with that which is diametrically opposed to the first. In his _Introduction_, Hort, in the most brilliant manner one must admit, has established the principle that the restoration of the text must be grounded on the study of its history, and no one has studied that history as carefully as Hort has done. But the question remains whether he has not interpreted the history wrongly, whether what he calls the Neutral text is really the original, and whether that which he rejects as a Western Corruption is really to be regarded as such.

[Sidenote: History of transmission.]

I cannot presume to judge; but I have the feeling that the history of the transmission of our New Testament text must be studied in quite another way from that in which it has been done hitherto, and in a twofold direction:—

(1) The manuscripts and their relation to each other must be subjected to a still more searching investigation, and

(2) The works of the ecclesiastical writers, especially the Commentaries and the Catenæ, must be thoroughly explored for any information they may have to give regarding the history of the text of the New Testament, and these two results must then be set in relation with each other.

With regard to the former task, it might not be essential to make such a minute collation of the manuscripts as Ferrar, Hoskier, and other investigators deemed necessary, and as is certainly the right thing to do in the case of the oldest documents. With such a mode of procedure, the task could not be accomplished in any conceivable time. But suppose the work was organised the way that Reuss did with the printed editions, by selecting say a thousand passages for comparison, it would be possible, and in a very short time we should be much better informed than we are at present as to the state of the text in our manuscripts, and especially in the minuscules.

Such a task, moreover, must be preceded by a fresh scientific statement of the way in which the text was propagated previous to the invention of printing, on the lines laid down by Hort in the first fourteen paragraphs of his _Introduction_.[155] A necessary preliminary to this is the study of genealogy, in which we have an excellent guide in Ottokar Lorenz’s _Lehrbuch der gesamten wissenschaftlichen Genealogie_ (Berlin, 1898). See especially the first chapter of Part I. on the distinction between Genealogical Tree (Table of genealogy) and Table of Ancestors, and the third chapter of Part II. on the problem of Loss of Ancestors.

All the ideas pertaining to the genealogy of living creatures, such as crossing, heredity, and so forth, fall to be considered also in the genealogy of manuscripts, the only difference being that in the latter case new features make their appearance. It has been asserted somewhere that if an Englishman, a Dutchman, a German, a Frenchman, and an American meet in a company, the nationality of each is at once recognisable, but it is impossible to determine their exact genealogical relationships, and that the same impossibility exists in the case of the manuscripts of the New Testament. That is perhaps an exaggeration, but it is certainly a surprising fact that so few even of our latest manuscripts can be proved with certainty to be copies of manuscripts still in existence, or at least to be derived from a common original.

[Sidenote: Analogous works.]

It will be a very great help, particularly to those beginning work in this field, to compare the method and results of investigations pursued in similar and perhaps easier departments of study. Apart from the works of classical philologists, or works like the new edition of Luther’s writings, a great deal of most valuable research has been carried on of late years in the matter of textual criticism, some of it very extensive, some of it less so. Ed. Wölfflin, _e.g._, devoted his attention to the Rule of St. Benedict of Nursia, who died some time after the year 542. [Sidenote: Rule of Benedict.] His Rule, which extends only to eighty-five pages of the Teubner size, is extant in manuscripts dating as far back as the seventh and eighth centuries. By a comparison of these, Wölfflin was convinced that we still possess the Rule essentially in the identical wording of the original vulgar Latin, that Benedict himself had afterwards made certain alterations and additions, and that we have therefore to distinguish several (fortasse tres) editions.[156] Wölfflin purported to give the text of that recension which he took to be the earliest. But we had no more than time to congratulate ourselves on the satisfactory result arrived at by this experienced philologist, when behold, another totally different conclusion was announced by a younger worker in the same field. Wölfflin had done little more than compare the manuscripts, but Lud. Traube applied also the external evidence afforded by the history of the text, and discovered that certain manuscripts that Wölfflin had thrown aside possessed a greater claim to originality.[157]

[Sidenote: De Viris Illustribus.]

Similarly, E. C. Richardson gave several years to an examination of all the accessible manuscripts of the _De Viris Illustribus_ of Jerome and Gennadius, a work not much larger than the Rule of Benedict. These manuscripts, about 120 in number, he grouped, and then framed his text in accordance with them.[158] While his work was in the press, an edition was published by C. A. Bernoulli, based on some of the manuscripts that Richardson also had used.[159] But from the very first sentence onwards, the two editors follow contradictory authorities, so that while one gives _parvam_ as the correct reading, the other reads _non parvam_. But more than that, the same Part of _Texte und Untersuchungen_ that gave us Richardson’s laborious work contained a second piece of work on the same material—viz., O. v. Gebhardt’s edition of the so-called Greek Sophronius, which is an old version of Jerome’s book. And the last chapter of this version, which is autobiographical, contains indications, according to v. Gebhardt’s Introduction, that Jerome issued two editions of his book, so that, if this be so, an entirely new grouping of the manuscripts becomes necessary.

[Sidenote: Julian’s Letters.]

In the case of Holl’s researches on the _Sacra Parallela_ of John Damascene, published in the same Collection,[160] matters are too complicated for beginners in textual criticism, but of non-biblical texts mention may be made of the _Recherches sur la tradition manuscrite des lettres de l’Empereur Julien_ by S. Bidez and Fr. Cumont (Brussels, 1898), as showing how much can be attained by combining the internal and external history of the transmission of literary texts.[161] [Sidenote: Latin New Testament.] In the field of Biblical texts, and particularly of the New Testament, the study of Wordsworth and White’s _Epilogus_ to the first volume of their _Novum Testamentum Latine_ is to be specially recommended, particularly c. iv. _De Patria et Indole Codicum nostrorum_, c. v. _De Textus Historia_, and c. vi. _De Regulis a nobis in Textu constituendo adhibitis_. As was said before, Jerome undertook his revision of the Gospels in the year 383; his work was of an entirely uniform texture, apart from a few passages where a correction may have occurred even in the original manuscript; and there is a sufficiency of manuscripts extant, some of them going back to the sixth century. One of these, which formerly belonged to the Church of St. Willibrord at Echternach, contains a note dated 558, and copied into it from the parent manuscript, to the effect: proemendavi ut potui secundum codicem in bibliotheca Eugipi presbyteri, quem ferunt fuisse sancti Hieronymi. (See above, p. 122.) In these circumstances it must surely be possible, one would think, to arrive at some satisfactory conclusion. And yet, in spite of all this, and in spite of long years of labour, many a problem remains to perplex the editor of the Vulgate. To begin with, there is one striking circumstance. Jerome executed his work at Rome, in obedience to the commission of Pope Damasus. One would therefore expect to find the best manuscripts in Rome, or at least originating from Rome. But that is by no means the case: “praeter expectationem accidit ut pauci vel nulli ex codicibus optimis et antiquissimis originem Romanam clare ostendant.” The manuscript that editors consider the best—viz. Codex Amiatinus, now at Florence—was certainly at Rome for a long time, but it was sent there as a present from beyond the Alps; indeed, it came from England. And on the other hand it was not from Rome that the Latin Bible came to England, or to Canterbury in particular, although Augustine was sent thither by Pope Gregory the Great, but from the South of Italy; in fact, it was from Naples. Codex Fuldensis, which may have been brought to Fulda by Boniface, formerly belonged to Bishop Victor of Capua; the Echternach manuscript referred to above came from the Lucullan Monastery at Naples, while the manuscript from which Codex Amiatinus was copied was written by Cassiodorus of Vivarium, and therefore came from Calabria. The history of the transmission of the Latin Bible reveals many other facts as strange as these in connection with the locality to which manuscripts belong. We must be prepared, therefore, for similar surprises in the case of Greek manuscripts also, and need not be astonished to hear that the Greek-Latin Codex Bezae, so much decried as “Western,” takes us back to Smyrna or Ephesus by way of Lyons and by means of Irenæus.[162]

It is also very instructive to observe that after long years of the most thorough study of their manuscripts, Wordsworth and White refrain from constructing =stemmata= or genealogies of these. All they venture to do is to distinguish certain large classes or =families=, and within these again to bring certain manuscripts into somewhat closer relationship with each other. They distinguish two main classes. In the first they reckon five, or it may be four, Italian-Northumbrian manuscripts, A Δ H* S Y, two Canterbury O X, three Italian J M P, the two mentioned already from Capua-Fulda F and Lucullan-Echternach EP, and the Harleian Z, so called from a former possessor and now in the British Museum. To the second class belong five Celtic D E L Q R, three French B BF G, and two Spanish C T. After these come the witnesses to the history of the text in the stricter sense of the term, the recensions made by Theodulf (H^c Θ) and Alcuin (KV M̅), and, for the form which the text assumed in the later manuscripts and in the printed editions, the Salisbury Codex W. We should have great reason to congratulate ourselves, could we arrive at like certain results in the region of the far older and more diversified history of the transmission of the Greek text, but there too we shall encounter the same general features—viz., a form of the text in the printed editions, in the later manuscripts, in the Recensions, the dates of which are still to be determined (Lucian, Hesychius?, Pamphilus), and in the families, which are only to be classified in a general way.

It is also instructive to find that in the case of several manuscripts, Wordsworth and White are obliged to observe that they seem to have been corrected from the Greek, H* (p. 709), M (p. 711), EP* (p. 712), D R L (pp. 714, 716), which suggests the possibility of Greek manuscripts also having been influenced by one of the versions, be it Latin or Syriac.

Finally, when we enquire as to the relationship between Jerome and the Greek manuscript or manuscripts used by him, we find that that manuscript must have been most nearly related to the Sinaiticus, while it had no sort of connection with Codex D. Whether this result tells in favour of א or the reverse, we will not pronounce at present: Jerome certainly avers that he made use of a Graecorum codicum emendata conlatione ... sed veterum (p. 108), only _veterum_ is a comparative term, and it might quite well happen that to Jerome that form of text appeared to be the best which was most recent or most widely circulated in his neighbourhood, and that he would have nothing to do with such a singular form of text as D exhibits, even supposing he was acquainted with it, a point we cannot decide. The intimate connection in various passages between his text and that of the cursive 473 (2^{pe}) is remarkable, and especially the many points of resemblance between the Irish manuscripts (D L R) and the members of the Ferrar Group.

And this brings us back from our survey of the history of the Latin text to that of the Greek, where we seem to have got at least one fixed point to begin with in this perplexing chaos, for that is the first impression we gain on glancing at the mass of Greek manuscripts before us. But here again our too sanguine hopes are likely to be disappointed, and we shall learn only too soon that even this is no Archimedean point from which we are able to regulate this world of disorder.

[Sidenote: Ferrar Group.]

If we find that in a certain number of manuscripts the passage usually indicated as John vii. 53-viii. 11 occurs in Luke, and in exactly the same place in each one—viz. after Luke xxi. 38—we must needs conclude that the manuscripts exhibiting this common peculiarity are intimately connected with each other. This is the case with the cursives 13, 69, 124, 346, 624, 626. Whether the others that are reckoned in this group have this same peculiarity, I am not perfectly sure. Now one would naturally expect that these manuscripts would also coincide in the other peculiarities characteristic of this group. [Sidenote: Variation.] But on the contrary, they part company over the very first and most conspicuous of these—viz. Matthew i. 16. Here, unfortunately, 13 and 69 are defective, but we can compare 124 and 346. And we find that whereas the former has the usual text, the latter, with the support of 556, 624, 626, exhibits in place of τὸν ἄνδρα ... Χριστός, the reading already mentioned, ᾧ μνηστευθεῖσα παρθένος Μαριὰμ ἐγέννησεν Ἰησοῦν τὸν λεγόμενον Χριστόν. Which of these readings is right, or whether both of them may not be wrong, we need not enquire at present. It is sufficient to point out that one and the same mother may give birth to very different children. The specific difference will be inherited from the other of the two parents, in this case represented by the copyist, and will depend on whether he is painstaking, careless, violent, arbitrary, well-informed, or the reverse. But the mother herself has a great number of hereditary or acquired peculiarities which, the latter no less than the former, may be transmitted to the children in a variety of ways. There is perhaps no manuscript in existence which is entirely free from corrections, while, on the other hand, there are many so overlaid with corrections that the original writing is scarcely now recognisable. Codex B, on the whole, is in a very good state of preservation, but it was supposed lately (see _ThLz._, 1899, col. 176) that its first hand wrote in John viii. 57, “and Abraham saw thee” instead of “and hast thou seen Abraham,” as all our editors read in that passage.[163] The supposition proved to be incorrect, but if that could possibly happen with B, what must have happened in the case of manuscripts that are so full of corrections as א and D, if they came to be copied in later times? Suppose that one scribe took the trouble to copy the text of the first hand, while another thought it his duty to follow the corrections, the result would be two manuscripts whose common origin would be scarcely recognisable. And in the course of centuries how often may not this process have been repeated. No wonder, therefore, that so few manuscripts have as yet been clearly made out to be the descendants of our oldest codices, so as to permit of their being removed from the board with one sweep, as having no independent testimony to contribute, as is the case with E_{3} which is merely a bad copy of D_{2} (p. 77), or that scholars cannot agree as to the relationship existing between two manuscripts like F_{2} and G_{3} (p. 77).[164] Nor need we be surprised to find that some of our most peculiar witnesses seem to have remained absolutely childless, while a less valuable race is perpetuated in many copies. The words of Homer—οἵη περ φύλλων γενεή, τοίη δὲ καὶ ἀνδρῶν—may be applied conversely to the leaves to which we entrust our immortality: habent sua fata libelli. The only unfortunate thing is that we are so little able to follow the course of these fates by means of external testimony.

[Sidenote: External testimony.]

When the Emperor Constantine _e.g._ asks Eusebius to supply him with fifty copies of the Scriptures at once, we cannot but suppose that these became authoritative over a large area. But in which of the classes into which our manuscripts have hitherto been divided are we to look for these now?[165] And conversely if, in another locality, heathen persecution was directed specially against the Bibles of the Christians, this cannot have been without some effect. According to C. Hülsen (_Bilder aus der Geschichte des Kapitols_, Rome, 1899), even the Roman Bishop Marcellinus, with his deacons Strabo and Cassianus, in the year 304, burned, in front of the temple of Juno Moneta, those Gospels which ten years later were made the law of Christian Rome by the Emperor Constantine. These scattered notices must be much more carefully collected and considered than hitherto, and combined with the results obtained from the collation of the manuscripts.

[Sidenote: Recensions.]

Most of our information with respect to the =recensions= of the Bible comes from the Syrian Church, and is concerned rather with the Old Testament than with the New; but the recensions of the former may throw some light on those of the latter.

[Sidenote: Lucian.]

(_a_) As the result of their researches, Westcott and Hort have made it very probable, on internal grounds, that a recension of an official nature was undertaken in Syria, perhaps at Antioch, about the fourth century, and that to this recension are due the origin and propagation of that form of the text of the New Testament which was widely disseminated in the Byzantine Empire, now represented in our later minuscules, and made the _textus receptus_ by means of the first printed editions.

=Lucian=, the celebrated founder of the Antiochean school of exegesis, suffered martyrdom in the year 311 or 312, most probably the latter. Of all the names that we know, none has a better claim than his to be associated with such a recension, and the conjecture derives some support from the passage of Jerome cited above (p. 85).[166] It is, perhaps, even better supported by what we know of Lucian’s recension of the Old Testament. In his preface to the Chronicles, Jerome wrote: “Alexandria et Aegyptus in Septuaginta suis Hesychium laudat auctorem, Constantinopolis usque Antiochiam Luciani martyris exemplaria probat, mediae inter has provinciae Palaestinae (_v.l._ Palaestinos) codices legunt, quos ab Origene elaboratos Eusebius et Pamphilus vulgaverunt; totusque orbis hac inter se trifaria varietate compugnat.” Now it is true that the words of a man like Jerome must not be pressed too far, and what may have been true in his day might be quite different in a comparatively short time—think _e.g._ of the fifty copies of the Bible that Eusebius Pamphili had sent to Constantine, or the Bible or Bibles sent by Athanasius to Constans[167]—at the same time it has been established beyond all doubt by Field and Lagarde, that it is to Lucian we must refer a peculiar recension of the Greek Old Testament preserved in a good many manuscripts, the one found in the unfortunately small remnant of the Gothic Old Testament and especially in the numerous Biblical quotations of the famous theologian John of Antioch, better known as Chrysostom of Constantinople, who was a pupil of Lucian. The probability, therefore, is very great that the same thing will hold good of the New Testament portion of the Bible of Ulfilas and Chrysostom. As regards the former, Kauffmann has expressed a decided opinion to this effect in his work on the Gothic Bible cited above (p. 139). The supposition would be converted into something like certainty if it could be proved on palæographic grounds that this or that New Testament manuscript belongs to this or that Old Testament manuscript of Lucianic derivation as part of what was originally one and the same complete Bible. This is a point which I am not in a position personally to investigate, and I must therefore content myself with throwing out this suggestion, and with adding in support of it that we have the express testimony of the Menologies for saying that Lucian bequeathed to his pupils a copy of the Old _and New_ Testaments written in three columns with his own hand.

There is a statement in the Pseudo-Athanasian Synopsis,[168] which seems, however, to refer only to the Old Testament portion, to the effect that it was discovered in a concealed cupboard (πυργίσκος) in the house of a Jew at Nicomedia in the time of Constantine. Long ago Hug and Eichhorn attributed the “Asiatic” or “Byzantine” recension to Lucian, and no decided objection can be established to the view of Hort, which Gregory also is inclined to accept (_TiGr._, p. 814).[169]

[Sidenote: Hesychius.]