Text Of The Gospels

Text Of The Gospels

Text Of The Gospels

1. The problem. —All true criticism must begin by taking cognizance of, and as far as possible accounting for, existing facts. The leading facts in regard to the text of the Gospels may be briefly stated as follows:

(i.) A Greek text substantially the same as the text underlying the Authorized Version has been almost universally accepted by Christendom as the authentic Greek text from about the year a.d. 350 till the development in modern times of the critical study of the text of the Nt. This text is found in the great mass of existing Greek Manuscripts, and was used by almost all ecclesiastical writers from Chrysostom onwards. Translated into Syriac, under the name of the Peshitta version, it was used by most of the Syriac-speaking Churches from at least the 4th cent. onwards. It was the only Greek text printed on the revival of learning in the West, and received the name of Textus Receptus (Textus Receptus) from an expression used in the preface to the second Elzevir edition, 1633: ‘textum ergo habes nunc ab omnibus receptum, in quo nihil immutatum aut corruptum damus.’

(ii.) Against this general unanimity in regard to the Greek text must be set the fact that the Churches of the West read the Gospels in the Latin translation of Jerome (a.d. 384), according to a text substantially different from the Textus Receptus. Moreover, existing Manuscripts and Patristic quotations of the earlier Latin versions differed from the Textus Receptus even more fundamentally, and similar types of text are found to have been very widely spread, speaking in a geographical sense, and occur in some important Manuscripts, in many ancient Versions, and in the quotations of many Christian writers, especially in the earliest times. This text (or, more correctly speaking, texts of this type) has been named ‘Western’; and, although it has long been well known that the term is not exclusively applicable in a geographical sense (indeed, it is quite possible that at least some members of this family may have had their rise in the East), yet for the sake of convenience it must for the present be employed.

(iii.) But a few of our earliest Greek Manuscripts, supported by the quotations of the most scholarly Fathers of the earlier centuries, and by a few Versions, present a different text, which has commended itself on its intrinsic merits, as well as on account of its proved antiquity, to most modern critical scholars: it forms the base of practically all the modern critical editions, and of our English Revised Version Nt 1881, Ot 1885.

2. The Received Text. —A text substantially the same as the Textus Receptus has been called by Dean Burgon and his school the ‘Traditional Text’; by Dr. Hort (in the Introduction* [Note: This Introduction was written by Dr. Hort, and will in this article be cited under his name, though the two editors accept joint responsibility for it.] to Westcott and Hort’s The New Testament in the Original Greek ) the ‘Syrian’ Text. Hort also suggests the name ‘Antiochian,’ which is preferable, because it avoids any chance of confusion with the totally distinct Syriac versions. For reasons that will be explained later on in this article, Hort considers that the Antiochian text affords practically no evidence for the reconstruction of the original Greek of the Nt, and he may therefore be considered as the most extreme opponent of the Textus Receptus. In his opinion ( Introduction , § 185) the Antiochian text ‘must be the result of a recension in the proper sense of the word, a work of attempted criticism, performed deliberately by editors and not merely by scribes.’ He further distinguishes two stages in the revision, and thinks (§ 190) that the final process was completed by 350 or thereabouts, and that the first process took place at some date between 250 and 350. According to Burgon and his close follower Miller, these recensions are purely imaginary creations; they believe the Church of Antioch (in company, no doubt, with practically all the Greek-speaking Churches) to have preserved the pure text from the first. It is at any rate certain that Chrysostom used this text: he was born at Antioch about the middle of the 4th cent., and lived in that city till 398, when he became bishop of Constantinople. We have seen above that even the main opponents of this text allow that it took its final shape probably about the time of Chrysostom’s birth. From that time onwards it held practically undisputed sway, and the main mass of later Manuscripts contain it. When at length, some time after the introduction of printing, the first New Testaments in Greek were published, they naturally rested on the Manuscripts in ordinary ecclesiastical use, and thus the Antiochian text became the ‘Received’ Greek text of modern Christendom, from which our own Authorized Version was made.

As has been shown above, the history of the printed text in the 16th cent. is part of the history of the Antiochian text; although of no critical importance, it is a subject very full of interest. [[[A]] good short account of the early printed editions will be found in Scrivener’s Plain Introduction (ed. Miller, 1894), vol. ii. ch. 7. Cf. also Tregelles, Account of the Printed Text of the Greek Nt , 1854]. The Nt was first printed in Greek as vol. v. of the Complutensian Polyglott Bible. This magnificent work was prepared at the cost of Francis Ximenes de Cisneros, Cardinal-Archbishop of Toledo, and was printed at Alcalá ( Complutum ), where he had founded a university. The Ot was given in Hebrew, Latin, and Greek; the Apocrypha and Nt in Greek and Latin. The volume containing the Nt (which was the first to be printed) was completed on 10th Jan. 1514; but owing to the death of the truly great Cardinal, the publication of the whole work was delayed, the Pope’s license not being granted till 22nd March 1520. Meanwhile, in order to forestall the Spanish edition, John Froben, the celebrated publisher at Basle, employed Erasmus to prepare an edition of the Nt in Greek, accompanied by a revised Latin version: this was hurried through the press, and published in 1516. Erasmus published other editions in 1519, 1522, 1527, and 1535. Other important editions are those of Robert Stephen (especially the folio of 1550, which is regarded by many as the standard text), Theodore de Bèze (Beza), and the brothers Elzevir. All printed editions, even those prepared by the great founders of textual criticism, were based upon the Textus Receptus until 1831, when Lachmann published a text constructed directly from the ancient documents.

Whatever may be the ultimate verdict of textual criticism, the Textus Receptus must always remain a monument worthy of deep veneration and of close study. It is an essential factor in the history of the development of Christianity. Through it the Spirit of God has, during the greater part of the existence of the Church of Christ, spoken to the greater number of her members. It has controlled the doctrine and the life of Christians, and by its means we have been freed, in part at least, from the heavy yoke of mediaeval sacerdotalism and superstition. Those who translated it into modern languages have left us in their work something of their own life and spirit. If extent of influence for good is to be our criterion, then surely, whatever its origin, the Textus Receptus and the translations made from it bear the impress of the seal of God’s Spirit, and have an unsurpassed and almost unsurpassable claim to the veneration and gratitude of mankind.

This much every thinking Christian will surely grant. But it is a different thing to go on to say: ‘therefore this text must be the original authentic text.’ It would be as logical to argue that because the gospel was given to the world in the Greek language, therefore Jesus must have spoken in the same language. It is quite in accordance with our experience of God’s methods of working that He should employ an instrument fashioned and conditioned not only by the circumstances under which it took its rise, but also by those through which it has passed in the course of its history.

It is an unfortunate thing that Burgon and Miller’s writings seem to imply (we believe, indeed, that the Dean stated it in so many words) that of necessity God must have provided for the accurate preservation of the text of the book which He had given to man. It appears to have been inconceivable to Burgon that the true text should be any other than that commonly accepted by the Church: to him the Church was the guardian of Holy Writ in the same sense as some people believe her to be the guardian of doctrine. If this view, even though not expressly stated, is felt to underlie the student’s conclusions, then those conclusions are removed from the domain of matters with which the critic can deal. They may, as in the case of views as to the authority of the Church in matters of faith, or of theories as to the inspiration of the Bible, conceivably rest on a true spiritual perception, but they do not rest on evidence, with which alone the critic is competent to deal. We have pointed out above that a large, and the most enlightened, portion of the Christian Church read the Scriptures in the Vulgate, or Latin translation of Jerome, and regarded it as the only authoritative exponent of the true text and sense of the original. There never has been a unanimous tradition as to the text of Scripture: only for the three centuries that followed the first printing of the Greek Nt has there been even an appearance of such unanimity. But though the writings of Burgon and Miller force one to the conclusion that for them personally their theory rested on a priori grounds, yet they have with great labour, assiduity, and learning collected a vast amount of evidence in support of the ‘Traditional Text.’ Unfortunately, Burgon wrote in such a contemptuous manner of the leading textual critics and of the most ancient Manuscripts of the Nt that most of his work has the appearance of an ex parte statement rather than of a solid contribution to the investigation of a difficult problem. Miller, who edited and completed many of Burgon’s papers after his death, adopted a more temperate tone; but so much of Burgon’s language is incorporated, that the subject is still treated rather after the fashion of a polemical controversy than of a critical investigation, Moreover, Burgon’s contention was that the ‘Traditional Text’ is the only one that has any claim to be regarded as the true text; all documents that differ from it are treated as of practically no value. Hort, on the other hand, considered the ‘Traditional’ or ‘Antiochian’ text to be valueless as evidence. Thus the subject has been treated at its extreme points, and neither side has taken sufficient trouble to discover how much truth is contained in the views of the other side. We lay a good deal of stress on this matter, because we think there has been a strong disposition to regard the ‘Traditional Text’ as a hobby of Burgon’s, and to treat his defence of it with the same contempt that he poured so freely on others.

3. Hort’s ‘Syrian’ or ‘Antiochian’ Text. —In part iii. of Hort’s Introduction , chapter ii. bears the heading, ‘Results of Genealogical Evidence proper.’ Section i. (§§ 130–168) is devoted to proving the posteriority of Antiochian to other known types of readings. We hope to show later on that the evidence here adduced is not entitled to be called ‘genealogical’ in a strict sense, but with this we are not for the moment concerned. Hort begins (§ 130) by stating the incontrovertible fact that all great variations of text were prior to the 5th cent., since the text of Chrysostom and other Syrian Fathers of the 4th cent, is substantially identical with the common late text; and (§ 131) the text of every other considerable group of documents is shown by analogous evidence of Fathers and Versions to be of equal or greater antiquity. If we were living in the age of Chrysostom, the problem to be solved would in all essential points be the same as it is now. Hort then adduces three lines of evidence to prove the posteriority of Antiochian readings: (i.) by analysis of conflate readings (§§ 132–151), (ii.) by Ante-Nicene Patristic evidence (§§ 152–162), (iii.) by internal evidence of Syrian ( i.e. Antiochian) readings (§§ 163–168). We must deal with each of these divisions separately.

(i.) When one reading is found in one group of documents, another in a second group, and the two different readings are found combined in a third group, this reading is said to be ‘conflate.’ Of course it has to be assumed that the first two readings are prior to the conflate reading, or else it is not a conflate reading at all. Thus the argument goes in a circle, unless either it can be proved that the two separate readings existed at a time when it can be shown that the conflate reading did not, or the conflate reading is so obviously wrong that it cannot conceivably be the original reading. If neither of these conditions is fulfilled, then conclusions based on the so-called conflate readings are matters of judgment, not of evidence. Hort adduces and examines eight eases of readings which he believes to be conflate: in each case, according to his view, the Antiochian text has combined two separate readings found in earlier texts. Obviously eight examples, taken four from Mark and four from Luke, afford but a slender foundation on which to build: it may be, and has been, urged that these eight examples are only specimens taken from a large number available, but until further examples are collected and published the case must be judged by the eight given.

For the sake of illustration, we give here the main readings in the instance selected for special discussion by Hort. In  Mark 6:33 (following and the people saw them going, and many knew them, and they ran there together on foot from all the cities ) we find the following readings:

καὶ προῆλθον αὐτούς ( and outwent them ), אB lect 49 Lat. vg Boh Arm and (with προσῆλθον for προῆλθον) Lδ 13 lect 39; Syr. [Note: Syriac.] vg has καὶ προῆλθον αὐτὸν ἐχεῖ.

καὶ συνῆλθον αὐτοῦ ( and came together there ), D [Note: Deuteronomist.] gr 28, 604 b (2pe d ff r have καὶ ἦλθον αὐτοῦ, a simply et venerunt , Syr. [Note: Syriac.] sin and when they came  : these documents might be taken to support either of the shorter readings).

καὶ προῆλθον αὐτοὺς καὶ συνῆλθον πρὸς αὐτόν ( and outwent them, and came together unto him ), all known uncials, except the five named above, all cursives except eight, f q Syr. [Note: Syriac.] hcl aeth.

In this case it will be noticed that there is no evidence to show that καὶ συνῆλθον πρὸς αὐτόν alone was ever read; moreover, the evidence for καὶ συνῆλθον αὐτοῦ is very slender, and quite possibly later than the supposed conflation. Mill suggested with much probability that D [Note: Deuteronomist.] omitted the words and outwent them because they contradicted  Matthew 14:13 and  Luke 9:11 ‘the crowds followed him.’ Swete, ad loc. , quotes 33 as reading συνέδραμον πρὸς αὐτοὺς καὶ συνῆλθον πρὸς αὐτον: this appears to have been another way of getting rid of the words objected to. The reading of the mass of Manuscripts gives such good sense that Hort himself says (§ 136), ‘There is nothing in the sense that would tempt to alteration: all runs easily and smoothly, and there is neither contradiction nor manifest tautology’; and again (§ 138), ‘Had it been the only extant reading, it would have roused no suspicion.’ He does, indeed, argue that the fresh point made by and came together unto him ‘simply spoils the point of ἐξελθών in  Luke 5:34; the multitude “followed” (Mt., Luke) the Lord to the desert region (ἐκεῖ), but the actual arrival at His presence was due to His act, not theirs, for He “came out” of His retirement in some sequestered nook to meet them.’ But Swete, ad loc. , far more naturally takes the ἐξελθών to mean ‘having landed,’ and thus the only objection that Hort could find to the language of the fuller reading falls to the ground: the crowd were the first to reach the spot whither Jesus and His disciples were going, they ran together on the beach to meet Him; and as He landed He saw them, and realized that He could not secure the quiet He sought. It is therefore quite possible that the reading of אBl [Note: L Bampton Lecture.] Δ is due to the accidental omission of a clause.

In none of the eight cases can it be proved that the two parts of the longer reading both existed separately at a time when the combined reading did not exist, and it is a matter of opinion whether the readings in which the two separate ones are combined are likely to be right or not.

Dr. Salmon ( Some Thoughts on the Textual Criticism of the Nt , p. 68) says that ‘Canon Cook elaborately discussed Hort’s eight cases, contending that in every one of them the conflation hypothesis gives the less probable account of the facts.’ He adds: ‘In each of these cases I did not myself follow Hort altogether without misgivings.’ Miller also discusses the supposed conflations in Appendix ii. of his ‘Causes of Corruption,’ and makes out a fairly good case for the originality of the supposed conflate readings.

(ii.) Hort’s next argument to prove the posteriority of Antiochian readings is founded on Ante-Nicene Patristic evidence.

It will be convenient to follow Hort’s example in giving at this point some general considerations in regard to the character and the use of Patristic evidence. We will speak first of the disadvantages and difficulties experienced in using it. To begin with, the material is necessarily very fragmentary in more senses than one. Each writer quotes but a limited number of passages, so that it is only in the case of a few specially prominent passages that we can get together a really representative collection of Patristic quotations. It follows that any kind of Patristic apparatus is more or less deceptive. It may be, for instance, that Origen has a reading which agrees with Manuscripts most approved by critical writers, but that the passage in which it occurs is not quoted by Clement of Alexandria. Here we are placed in a difficulty, because Clement and Origen did not by any means always agree, and, if a quotation had been preserved in which Clement used a different reading, it would be probable that Origen’s reading did not belong to the text traditionally current at Alexandria, but that he had obtained it from some other source; his evidence, therefore, would be simply of a personal character. It is necessary, therefore, in weighing Patristic evidence to deal with the author’s quotations as a whole, in order to form a judgment of the character of the text he used. When Clement’s and Origen’s quotations are thus dealt with, it is found that Origen in part agrees with the text most favoured by critical editors, but that his predecessor Clement used a substantially different text of a ‘Western’ type; Origen too, in part, followed ‘Western’ texts: the conclusions to which these phenomena lead will be discussed later on. The important point to note at this stage is that the whole mass of a writer’s quotations must be treated as one whole, and that, while we can discover the type of text he used, our knowledge of it is only fragmentary, and necessarily confined as far as details are concerned to the passages explicitly quoted.

A moment’s reflexion on the way in which the Bible is quoted in extempore sermons or in conversation will be sufficient to show that a writer’s quotations may not always reproduce the text that he considered the best, supposing him to have formed a critical judgment on the subject. Natural looseness of quotation from memory, familiarity with more than one text, and confusion between parallel passages in the Gospels, will account for many deviations that cannot be considered genuine variant readings. A knowledge of the proneness of the human brain to repeat a mistake once made, will render us cautious even when a writer quotes a passage more than once in the same unusual form. Even with great care and wide experience it is difficult for a student to feel sure that a quotation gives the reading which the writer, in answer to a direct question, would have deliberately stated to be the right one.

Moreover, we often feel great doubt whether the quotation stands in our printed editions in its original form. The works of many Greek Fathers have been notoriously badly edited, and it is only when we have had personal experience of the editor’s methods that we can feel any security that full advantage has been taken of the Manuscripts and other evidence available. Dr. Nestle (in his Introduction to the Textual Criticism of the Greek Nt , English translation 1901, p. 145) refers to an extreme instance of supineness and ignorance on the part of even a fairly recent editor: he gave in his Ms the first and last words of quotations, and left the printer to fill them up from a printed copy of the Nt.

And when we go behind the editions, we often find that only comparatively late Manuscripts are now extant, and we have to allow for the natural tendency of scribes to substitute, both consciously and unconsciously, familiar for unfamiliar readings. Sometimes the comments that follow the quotation enable the student to detect the substitution, but such alterations must have been made by scribes in numberless passages in which there are no means of discovering them.

The case of Fathers writing in a language other than Greek presents further difficulties, because it is often impossible to say how far the form of the quotation is due to a knowledge of the original Greek, and how far to familiarity with the version in their own language. Analogous difficulties arise in the case of works which are preserved only in translations, because the translator was likely to introduce readings familiar to him in the vernacular.

We have enlarged somewhat on this matter in order to show how much care is needed in forming a judgment on the Patristic evidence in regard to individual readings. But, on the other hand, we desire to emphasize as strongly as possible the immense importance of Patristic evidence when employed with due precautions for its proper purpose, namely, the dating and localizing of special types of text.

But, again, we must remember that the remains of Ante-Nicene Christian literature that have come down to us are very fragmentary. ‘The only period for which we have anything like a sufficiency of representative knowledge consists roughly of three-quarters of a century, from about 175 to 250’ (Hort, § 158). Besides Clement and Origen, Hort names Irenaeus, Hippolytus, Tertullian, Cyprian, and Novatian, belonging to the period named; Methodius towards the close of the 3rd cent.; and Eusebius of Caesarea in the first third of the 4th century. ‘The text used,’ writes Hort (§ 159), ‘by all those Ante-Nicene Greek writers, not being connected with Alexandria, who have left considerable remains, is substantially Western.’

We are now in a position to consider the value of the argument for the posteriority of Antiochian readings which Hort bases on Ante-Nicene Patristic evidence: it is an e silentio argument—that no extant writer before Chrysostom used the Antiochian text. The force of this argument is considerably lessened if we reflect that, had the writings of Origen perished, we should have had practically no Ante-Nicene Patristic evidence for the type of text contained in the Revised Version Nt 1881, Ot 1885.

Miller ( The Traditional Text , p. 94 ff.) has attempted to prove the antiquity of the Traditional or Antiochian text by a wide appeal to Patristic evidence. In a sense he fails, because if a reading is shown to be older than the supposed revision which produced the Antiochian text, it is said by the school of Hort to be not distinctively Antiochian, but a ‘Western’ reading adopted by the revisers. To one who does not adopt an extreme view on either side, this will probably appear very like a fight over empty names. The Antiochian text confessedly contained an ancient element, and the real question is whether critical editors have paid sufficient attention to the evidence afforded by it. Call the text by what name you will, but let it be judged on the intrinsic value of its readings, not in accordance with uncertain theories. Its very existence forms evidence in favour of certain types of the Western text, which must go back to the 2nd cent., as is shown by Miller; and the real question at issue is, What weight is to be attached to the evidence of these texts?

(iii.) The judgment of such a scholar as Dr. Hort on the intrinsic value of the Antiochian readings must carry the greatest weight. It will be most satisfactory to quote his own words. ‘Another step is gained by a close examination of all readings distinctively Syrian (Antiochian) in the sense explained above, comparing them on grounds of Internal Evidence, Transcriptional and Intrinsic, with the other readings of the same passages. The result is entirely unfavourable to the hypothesis which was mentioned as not excluded by the phenomena of the conflate readings, namely that in other cases, where the Syrian text differs from all other extant ancient texts, its authors may have copied some other equally ancient and perhaps purer text now otherwise lost’ (§ 163). This decision may be regarded either as an expression of subjective judgment, in which case its value will vary according to the estimate formed of its author’s ability as a critic; or else it can be regarded as the result of certain lines of argument, in which case it is the business of other critics to examine those arguments.

The conclusions which Hort reached in regard to the conflate readings discussed above rest on, and indeed may be fairly considered to assume the truth of, his views as to the genealogical relations of the different families into which he divides all extant Nt documents. His whole text is indeed based on those views; and therefore, if we are to discuss the problem before us intelligently, it is essential to have correct knowledge of the exact nature of genealogical evidence, and of how far it is available for the criticism of the Nt text.

It is an obvious truth that, if the original of a document exists, no number of copies will possess any value for settling its text, which can be ascertained by reference to the document itself. This is the simple ground on which all genealogical evidence rests. If three independent copies have been made of a document which has itself perished, it may fairly be assumed that where all three agree they correctly represent the original; and further, in cases where two of the copies agree against the third, we shall confidently judge that these two preserve the right text, and that the third is in error. Now suppose that fifty copies have been made of this third original copy, and that it has itself perished, then it is clear that the evidence of the two extant primary copies outweighs the evidence of the fifty secondary ones. In this example it is assumed that the exact parentage of every copy is known. This is, of course, seldom the case with the Manuscripts of ancient authors; but. when the parentage of every Ms concerned can be ascertained, then genealogical evidence gives results from which there can be no appeal.

This matter is of such importance that it is worth while to illustrate further what we have said, by reference to an actual instance. A fair number of Manuscripts exist of the Paedagogue of Clement of Alexandria. In one family of these, consisting of eight or more members, a passage of considerable length is left out. Now two leaves have been lost from a Ms preserved at Florence (called F), which contained exactly this passage; it is therefore beyond doubt that the Manuscripts referred to were copied from F after the loss of these leaves, and they are therefore of no value as evidence. There exists also at Paris another Ms (P [Note: Priestly Narrative.] ), considerably older than F. At one time there was some little doubt about the relation existing between these two Manuscripts; but after a time it was pointed out by a German scholar, Dr. Stählin, that certain notes that were written in P [Note: Priestly Narrative.] by different people and at different times, are written in F in the hand of the original scribe; this makes it certain that F was copied, directly or indirectly, from P [Note: Priestly Narrative.] , and it can therefore also be put aside. Further researches showed that every known Ms of the work was derived from P [Note: Priestly Narrative.] , which consequently forms our only authority for the text. It is very seldom that such certain results as these, can be reached in actual practice. It is generally possible to group Manuscripts to some extent by observing their agreement in obvious errors, because it is not likely that different scribes would make the same mistakes independently in several different places. It is obvious that the confidence with which we can employ genealogical evidence is proportionate to the certainty with which the relations of the Manuscripts have been ascertained. In the case of certain cursive Manuscripts of the Gospels strictly genealogical evidence is forthcoming, and it has been shown that the cursives 13, 69, 124, 346, and certain others, are derived from one common ancestor; but, except for this one important and interesting case, the genealogical relations of Gospel Manuscripts are matters of deduction, if not of guesswork.

It appears, then, that it is impossible to acquiesce in Hort’s unqualified condemnation of the Antiochian text, so far as that condemnation rests on (i.) the analysis of conflate readings, which presupposes certain genealogical relations to exist between certain groups of Manuscripts, and involves an argument in a vicious circle, because those relations cannot be independently shown to exist; and (ii.) so far as it rests on Patristic evidence, this being precarious from its fragmentary character, while at the same time it does prove that the Antiochian text contains a very ancient element. It remains, therefore, to judge this text on its intrinsic merits.

4. The generally accepted Critical Text. —Once again, it is with Hort’s views that we must principally concern ourselves, because Wh [Note: H Westcott and Hort’s text.] ’s text is the only one published which can be regarded as in any way self-consistent. No textual student would place much confidence in Tischendorf’s judgment, which is embodied in his editio octava critica major  ; the Greek text underlying the Revised Version Nt 1881, Ot 1885 does not appear to have been formed in accordance with any ascertainable principles; and Weymouth’s ‘Resultant Text,’ and similar editions, founded on the consensus of critical editors, from their nature have no independent critical value. We have, therefore, to consider the principles on which Wh [Note: H Westcott and Hort’s text.] founded their text. We have already shown how the great mass of documents, containing an Antiochian text, were set on one side. The pre-Antiochian texts Hort divided into three families, and, on what appear to many students insufficient grounds, assumed that they stood in certain genealogical relations to one another. One of these families consists of the group of texts commonly called ‘Western’; after setting these aside as obvious corruptions of the original text, only a small body of Manuscripts, Versions, and Fathers remains. This small residuum, however, Hort proceeds to again divide into ‘Neutral’ and ‘Alexandrian’ documents. It is now, we think we may say, generally acknowledged that this distinction cannot be maintained (cf. Salmon, Some Thoughts on the Textual Criticism of the Nt , p. 50 ff.). Practically, he classes as ‘Alexandrian’ the readings of documents which usually agree with Codex B, when they differ from B and are not supported by much Western evidence. We shall therefore treat these documents as forming one group, and distinguish the readings, as Salmon suggests, as early and later Alexandrian. Hort frankly admitted the close relation existing between his Neutral and Alexandrian readings, since he conceived both sets of readings to be derived from a common non-Western ancestor; this led him, in the case of an important set of readings, which he called ‘Western non-interpolations,’ to prefer the testimony of a small group of Western documents to the practically unanimous evidence of all other documents.

It will be convenient here to give a list of the main documents with which criticism has to deal. We begin with those which more or less regularly support the Alexandrian readings. See also art. Manuscripts.

B, the famous Codex Vaticanus, assigned to the 4th cent., is by far the most interesting; according to Hort, it contains a purely ‘Neutral’ text in the Gospels.

א, Codex Sinaiticus, discovered by Tischendorf on Mt. Sinai, and probably to be assigned to the 4th century. This Ms is thought by Hort to be free from Antiochian readings, but to contain a ‘mixed’ text, that is, one in which Western, Neutral, and Alexandrian elements are all found, though in the Gospels he looks on it as largely Neutral; this is equivalent to saying that its agreements with B are very numerous.

C, Codex Ephraemi Rescriptus, a palimpsest preserved at Paris, and belonging probably to the 5th century. The text of this Ms is undoubtedly of great importance. Miller ( Plain Intr. 4 [Note: designates the particular edition of the work referred] vol. i. p. 123) well describes its text as ‘standing nearly midway between A and B, somewhat inclining to the latter.’ Hort considers C to contain an Antiochian and also a Western element.

L, Codex Regius, preserved at Paris, belonging to the 8th century. This Ms is especially remarkable for the number of readings it has in common with B. According to Hort (§ 209), ‘The foundation of the text is Non-Western Pre-Syrian.’ But he adds: ‘The fundamental text has been largely mixed with late Western and with Syrian elements.’

T, Under this symbol are placed several fragments of Manuscripts containing a Greek text and a translation in the dialect of Upper Egypt (Sahidic or Thebaic). They range in date from the 5th to the 7th century.

X, Codex Monacensis, preserved at Munich, of the 9th or 10th cent., has a fundamentally Antiochian text, but is of interest because it often joins with Cl in giving readings which may be regarded as late Alexandrian.

Z, Codex Dublinensis Rescriptus, perhaps to be assigned to the 4th cent., contains 295 verses of Mt. in 22 fragments. The text is apparently pre-Antiochian, and agrees more closely with א than with B.

Δ, Codex Sangallensis, of the 9th or 10th cent., has an ordinary Antiochian text, except in Mk., in which Gospel it has many readings in common with Cl.

S, Codex Xacynthius, a palimpsest, probably of the 8th cent., belonging to the British and Foreign Bible Society in London. This Ms contains 342 verses of Lk., giving an apparently pre-Antiochian text, in which both Western and Alexandrian elements are found.

1, A minuscule, preserved at Basle, assigned to the 10th, 12th, or 13th cent., often agrees with אB and Bl [Note: L Bampton Lecture.] .

33, A minuscule of the 10th cent., preserved at Paris, has been called ‘the queen of cursives.’ It has a very interesting text, containing many ancient elements, but agreeing now with one, now with another type of readings.

The ancient Egyptian Versions, as might be expected, to some extent support the Alexandrian text; but there is so much uncertainty in regard to these Versions that it is not easy to reckon with them as an element in the critical problem presented to us. Forbes Robinson, in his art. ‘Egyptian Versions’ in Hasting's Dictionary of the Bible, declines to follow Lightfoot and Hort in assigning one, if not both, of the principal Egyptian Versions ( i.e. the Bohairic and the Sahidic), or at least parts of them, to the close of the 2nd century. He gives good reasons for thinking that the Sahidic Version, which was current in Upper Egypt, was the earlier of the two; and it must be regarded as fundamentally Western rather than Alexandrian. The Bohairic (misleadingly called Coptic, and also Memphitic) Version, current in Lower Egypt, confessedly agrees in general with B, and perhaps even more closely with the text used by Cyril of Alexandria. If it has to be assigned to a date as late as the middle of the 3rd cent., it is evident that it may be the result of the type of text then current in Alexandria, and cannot be used as evidence for the greater antiquity of that text. The remains of the Bashmuric Versions—those current in Middle Egypt—are so scanty that they offer little help at present.

It would be easy to extend this list by including documents which occasionally support the Alexandrian text, but it will be found that the nucleus of the attestation for most of Hort’s readings lies practically in the group אBclx 33, often supported by the Egyptian versions.

At the same time, it is most necessary to bear in mind that the greater part of the attestation for Hort’s readings is often afforded by documents which he classes as Western, and whose evidence he would put on one side were it not supported by some member or members of the Alexandrian group. We proceed, therefore, to give a list of the main Western documents, which have not already been mentioned as containing an Alexandrian element.

D [Note: Deuteronomist.] . Codex Bezae Cantabrigiensis, of the 6th century. This is in many ways the most interesting Ms of the Gospels extant: its text is, to a great extent, unique, and gains in interest and importance from the support which it often receives from the most ancient versions known, the Old Latin and the Old Syriac. All evidence tends to show that it preserves for us a text which was widely read in the 2nd cent., and the questions connected with this text are likely to increase rather than to decrease both in importance and in practical interest.

P [Note: Priestly Narrative.] and Q. Two palimpsests preserved at Wolfenbüttel, assigned respectively to the 6th and 5th centuries. P [Note: Priestly Narrative.] contains 31 fragments, consisting of 518 verses from all four Gospels; Q 12 fragments of 247 verses from Lk. and John. The ancient element in these Manuscripts is partly Western and partly Alexandrian.

R. Codex Nitriensis, a palimpsest of the 6th cent., in the British Museum, contains 25 fragments of Lk., consisting of about 516 verses. The pre-Antiochian readings are mostly Western.

Two groups of minuscules are of importance. 1-118-131-209 are fairly closely related, and offer some interesting readings: but far more important are the minuscules of the Ferrar group mentioned above, 13-69-124-346-543-(788)-826. This group preserves the readings of a lost Ms containing a peculiar Western, text, different from that of D [Note: Deuteronomist.] , but in a manner parallel to it. Another important minuscule, of the 9th or 10th cent., is preserved at St. Petersburg, and is named by Miller-Scrivener 473 (565 of Gregory, 81 of Hort, 2pe of Tischendorf).

The evidence of the Versions is of great importance in regard to the Western text, for it shows how widespread this text was in the earliest times, and teaches us that the name ‘Western’ cannot properly be applied to it in a geographical sense. From East and West and from the south of Egypt we get evidence of the prevalence of distinctively Western types of readings.

The Old Latin ( i.e. the pre-Vulgate Latin) is found in different forms, which have been distinguished as African, European, and Italic; the last of these, however, approaches so nearly to the Vulgate text, that we shall now leave it on one side. The most important Ms of the African Latin is k (Codex Bobbiensis), of the 5th or 6th cent., preserved at Turin. Unfortunately, it contains only portions of Mt. and Mark. The close agreement of its readings with the quotations of Cyprian proves that it contains a text used in Africa in early times; e (Codex Palatinus), of the 4th or 5th cent., preserved at Vienna, contains a version of a similar type, though by no means so homogeneous as that of k . Of the European Latin there are several Manuscripts: a, b, ff (Mt. only), h (part of Mt.), i (part of Mk. and Lk.), m (not a Ms, but a collection of passages [ testimonia ] from the Ot and Nt, known as the ‘Speculum’); this type is also found to some extent in c, f , and q , and in many fragments of Manuscripts.

The text of the Latin Vulgate is preserved in very numerous Manuscripts. It is fundamentally Western in character, as being a descendant of the Old Latin, but has been much modified, especially in the Gospels, by the influence of Greek Manuscripts of the Antiochian type.

In Syriac, the Peshitta Version holds a place analogous to St. Jerome’s Vulgate in Latin, and supports the Antiochian text. Another Version, called by the followers of Hort the ‘Old Syriac,’ is preserved in two Manuscripts; one in the British Museum, the text of which was published by Cureton, is called after him the ‘Curetonian Syriac’ (Syr [Note: yr Syriac.] cur); the other was discovered by Mrs. Lewis and Mrs. Gibson in the Library of the Convent on Mount Sinai, and is known as the ‘Sinaitic Syriac’ (Syr [Note: yr Syriac.] sin). These Versions, allied, but by no means identical, have an essentially Western text. Another factor in the Syriac problem is the Diatessaron of Tatian (flourished a.d. 160), the text of which has to a great extent been recovered from an Arabic translation, from an Armenian translation of the Syriac commentary of Ephraem Syrus, and from the quotations of the Syrian writer Aphraates. The Diatessaron was a harmony of the four Gospels, which was widely used in Syriac-speaking countries in preference to the separate Gospels; and in compiling it Tatian used a Western text, similar in character to the Old Syriac. The mutual relations of these documents are still in dispute, but the most probable view is that the Old Syriac stands to the Peshitta as the Old Latin does to Jerome’s Vulgate. Two later versions must be mentioned; one is the Harkleian revision of the Philoxenian Syriac, made by Thomas of Harkel about the year 616, the text of which is based on the Peshitta, but important readings from Greek Manuscripts of a Western type are given in the margin; the other is an Evangelistarium, or Church-lesson book, of the 11th cent., known as the ‘Jerusalem Syriac,’ which sometimes offers very interesting readings of the Alexandrian type. It has already been stated that the Sahidic version of Upper Egypt is fundamentally Western.

In order to complete this brief survey of the most important documents, we must here mention A—the important Codex Alexandrinus of the 5th cent, preserved in the British Museum; it contains a pure form of the Antiochian text, and it is quite possible that critics will learn to allow more weight to its evidence than is at present the case. The main mass of uncials that have not been here mentioned, and of the minuscules, may be regarded as simply supplementing the evidence of A, because the importance to be attached to them depends upon the estimate formed of the value of the text of A.

We have now to consider in more detail the use which Hort makes of the Alexandrian group of documents. We have already tried to show how precarious any argument is which rests on genealogical considerations, owing to the lack of sufficiently full evidence; at the best, genealogical evidence affords us no help in judging between the Western and the Alexandrian texts, because they are confessedly parallel to each other, and have equal claims to consideration on genealogical grounds. But if it can be shown that the Alexandrian group consistently supports readings intrinsically better than those of the Western documents, this will afford good reason for following it. In other words, the question comes to this: Is the text of Wh [Note: H Westcott and Hort’s text.] , which all critics admit to be substantially a text used at Alexandria early in the 3rd cent., on the whole preferable to the Textus Receptus, and to such a text as would be formed by following exclusively Western documents? The answer of critics at the present time to this question would undoubtedly be in the affirmative. But, in the great majority of cases in which it differs from the Textus Receptus, Wh [Note: H Westcott and Hort’s text.] ’s text has the support of the best Western as well as of the Alexandrian documents; it is possible, therefore, to argue that its general excellence is due to the pure form of the Western element which it contains, and to look upon the distinctively Alexandrian readings as blemishes. On what grounds does Hort prefer these distinctively Alexandrian readings? His main argument is the internal evidence of groups; all the readings supported by a group such as אB or אD [Note: Deuteronomist.] are examined, and judgment is passed on them collectively, and also on the text common to the Manuscripts forming the group. Now, the text common to א and D [Note: Deuteronomist.] is, according to Hort’s classification, Western, and in his opinion gives inferior readings (of course, when unsupported by other primary documents); whereas the agreement of א and B almost invariably gives readings which he considers intrinsically excellent. This method of forming a judgment on a wide consideration of the general readings of a group, to a great extent does away with the personal element which is so great a danger when individual readings are considered each on its intrinsic merits, but it still leaves plenty of room for the personal equation, since a general judgment is based on a special individual judgment in a number of separate cases; thus Hort’s system is far less impersonal than it appears to be at first sight. It is obviously impossible to enter into all this minute research unless one is able to devote many years of close work to the subject; yet, without doing so, it appears presumptuous to dispute Hort’s conclusions.

But judgment in this matter really rests on a wider question. If it can be shown to be probable that the Alexandrian text is the result of a revision, then the greater part of Hort’s work has been expended in restoring the original text of that revision, and is only a step, though an important one, in getting back to the readings of the original autographs. Now, recent investigations seem to tend to render two facts probable: (1) that all documents giving an Alexandrian text are connected with Egypt, and (2) that the text current in Egypt prior to the time of Origen was fundamentally Western, not Alexandrian. If a strong probability can be made out for these two views, then it will be a reasonable conclusion that the Alexandrian text had its rise in Egypt during the early part of the 3rd cent., and it will have to be treated as parallel to, though earlier and more important than, the Antiochian text. Egypt was the home of scholars, and if such a recension was made there, it is natural that the conclusions of early scholars should commend themselves on their intrinsic merits to men of similar training even at a much later date; we have also to remember that it is quite probable that those early scholars, with more evidence before them than we now have, did select the best readings, and may have preserved to us many true readings which would otherwise have perished. The dislike with which the later students of Antioch regarded the opinions of the earlier Alexandrian Fathers, and the taint of heresy which attached to them, easily account for the text they preferred not having continued in general use, if indeed it was ever widely current. Hort has declared that there are no grounds at all for believing in this Alexandrian revision, but we are not aware that he has gone beyond assertion on this point. In the same way, Burgon and Miller declared that Hort’s Antiochian revisions were the creations of Hort’s imagination. But the fact remains that the Alexandrian text cannot be traced earlier than the first quarter of the 3rd century. Clement of Alexandria used a distinctively Western text; it is true that he sometimes has what are commonly regarded as Alexandrian readings, but it is manifestly impossible to prove that these may not have been part of the Western text, current in Alexandria, and naturally taken up by the revisers. If it is the case that the Sahidic version is earlier than the Bohairic, again we find the Western type preceding the Alexandrian; and if Robinson is further right in assigning the Bohairic to the 3rd, and not the 2nd cent., then it may very possibly have been made from Manuscripts with the revised Alexandrian text, and its character is thus accounted for.

The great importance which Hort assigns to the agreement of א and B depends on his contention that the two Manuscripts are independent of each other; but there are really strong reasons for doubting this. Hort (§ 288) admits the truth of the fact pointed out by Tischendorf, ‘that six leaves of the Nt in א, together with the opening verses of the Apocalypse, besides corrections, headings, and in two cases subscriptions, to other parts, are from the hand of the same scribe that wrote the Nt in B.’ He adopts the obvious conclusion that the scribe of B was the corrector of א, and adds that it shows that the two Manuscripts were written in the same generation, probably in the same place. He argues, however, that the evidence of the text, supported by differences in the order of the books and other externals, creates a strong presumption that they were copied from independent exemplars. But where so much depends on the absolute independence of two witnesses, this close local connexion must cause the most serious doubts. Have we any means of saying where it is likely that the two Manuscripts were written? Both Manuscripts contain a peculiar system of chapter numbers in the Acts, in each case in a very early hand, and with such differences that in neither case can the numeration have been copied from the other Ms, but must have come from a common original. Dean Armitage Robinson, in his ‘Euthaliana’ ( Ts [Note: S Texts and Studies.] iii. 3), gives reasons for believing that this chapter-numeration is the same as that connected with the name of Euthalius, and points out (p. 35) that a Euthalian codex claims to have been collated with the accurate copies in the library at Caesarea of Eusebius Pamphyli. The connexion of Origen with this great library is well known, and suggests (though it can hardly be called more than a suggestion) that the same library may have been the birthplace of these two great Manuscripts which, when in agreement, support the text which Origen mostly used, and with the rise of which he may well have been connected. It is impossible to speak with any confidence until a great deal more work has been done, but it does seem as if the evidence in favour of an Alexandrian revision is growing (cf., further, Burkitt, Ts [Note: S Texts and Studies.] v. 5).

We are able to judge of Hort’s work only by the results, and to some extent our judgment must be based on a consideration of extreme instances, that is, we must judge the theory by cases in which he has pushed it to its furthest limits. No one denies that the greater part of his text, right or wrong, is of extreme antiquity, being based on the agreement of Alexandrian and Western documents; the question is whether his theory has led to the inclusion of readings that cannot be shown to be earlier than Origen, and may therefore be due to an Alexandrian revision, or may be errors that had crept into the Western text current at Alexandria on which that revision was based. We propose to examine a few examples which throw light on the methods he employed.

One of the most important, Instructive, and truly typical examples occurs in  John 1:18. The passage has been exhaustively discussed by Hort in the first of Two Dissertations (1876). The verse runs in the Alexandrian Manuscripts: θεὸν οὐδεὶς ἑώρακεν τώτοτε· μονογενὴς θεὸς ὁ ὢν εἰς τὸν κόλτον τοῦ τατρὸς ἐκεῖνος ἐξηγήσατο. For μονογενὴς θεός the vast majority of documents give ὁ μονογενὴς υἱός: Hort’s reading is supported by a small, and nearly homogeneous, group of documents, אBc*L 33, the Pesh. Syriac, the margin of the Hark. Syriac, and the Bohairic. The Sahidic and Gothic Versions and the Sinaitic Syriac are not extant h