THE TEXT OF THE GREEK BIBLE - A STUDENT'S HANDBOOK by Frederick G Kenyon, late Director & Principal Librarian of the British Museum. First published 1937 New edition 1949 Reprinted 1953. (Prepared for katapi by Paul Ingram 2015).
HOME | contents | << | introduction | new mss discoveries | textual classification : streeter, von soden | further discoveries : egypt, chester beatty mss | textual families | bibliography | >> |


The publication in 1881 of Westcott and Hort's Greek Text, and of the Revised Version of the New Testament which largely (but not by any means wholly) reflected it, let loose a flood of controversy. The battle was at first between the advocates of the Received Text and its opponents. Dean J. W. Burgon, the protagonist of the former, was a doughty controversialist, and could deal a swashing blow with gusto. He had also behind him the sympathy of those who resented the changes in their beloved Authorized Version and did not understand the reasons for them. Among scholars, however, he found little support, and though his disciple and assistant, E. W. Miller (who, besides directly controversial books, edited the fourth edition of Scrivener's useful Introduction ), maintained the fight after his death, this conflict soon died down. The relative lateness of the Byzantine or Received Text was regarded by almost all scholars not only as proved by the evidence of the Fathers, but as in accordance (though on a much larger scale) with the textual history of all ancient literature. Subsequent discoveries have done nothing to disturb this verdict, and this particular issue may be regarded as closed. ← It is not to be understood that this equally closes the discussion as to the comparative merits of the Authorized and Revised Versions. This involves not only the text translated but the methods of translation. So far as the text is concerned, there is no doubt that the translators of the Revised Version had a much better text before them than those of the Authorized Version ; but as to their command of the English language and the wisdom of some of their verbal changes a different opinion may be held. That however does not concern the present handbook.

It was far otherwise with regard to the question of the comparative claims of the Neutral and the Western types of text. Here, if all readings for which there is early evidence, but which were not in the Neutral text, were to be reckoned as Western, the evidence of the Fathers, which Westcott and Hort used so decisively against the Syrian text, was rather in favour of the Western, and not a few of the leading scholars were disposed to challenge Westcott and Hort's almost exclusive reliance on B, if not definitely to prefer the Western text. This, therefore, became and still remains the outstanding issue in the criticism of the New Testament, and it is for their bearing upon this problem that the discoveries of the ensuing fifty years were diligently scrutinized. It is with these discoveries, and with the theories to which they give rise, that this chapter is concerned.

The two decades following 1881 saw the appearance, at different times and from different places, of four manuscripts (or rather portions of manuscripts) which are nevertheless closely related in appearance and character. These are the four manuscripts of the Gospels on purple vellum described above as N, Ο, Σ, and Φ. N and Σ are so closely related in text as to be practically sister manuscripts, and Σ and O are connected by the fact that both contain illustrations. Σ and Φ contain Matthew and Mark, O portions of Matthew, N portions of all Gospels, but chiefly of Luke and John. All are approximately of the same date, probably sixth century, and must have emanated from the same centre.

There is nothing to show what that centre was, but since N appears to have been broken up by Crusaders in the twelfth century, Constantinople or some other large town in the Eastern Empire seems probable. In text, all represent the Byzantine type in an early stage, but O and Φ are not so closely related to the others as N and Σ are to one another. Φ in particular is remarkable for including the long passage after Mt. xx.28, for which the only other Greek witness is D, but which is found in most MSS. of the Old Latin and a few of the Vulgate, and also in the Curetonian Syriac. Before the discovery of W, Σ and Φ (N and O being defective here) were the earliest authorities for the doxology to the Lord's Prayer in Mt. vi.13. On the whole, however, the discovery of this group of MSS. did not materially affect the textual position. The same may be said of the few other uncial MSS. that have come to light, though Ψ is noteworthy as falling into the same class as L, like it agreeing in general with B, and containing both the shorter and the longer endings to Mk.

Far more important was the discovery of the Sinaitic palimpsest of the Old Syriac version made by Mrs. Lewis and Mrs. Gibson in 1892, which has been described above (p. 117). This, with its considerable proportion of non-Neutral, and sometimes definitely Western, readings (markedly greater than in the Curetonian copy of the same version) was a distinct encouragement to the Westerners. It will be clear, however, from the readings quoted in the description of the version, that its adhesion to the Western group is only very partial, and that its Western readings may be due to the influence of the undoubtedly " Western " Diatessaron. For the moment, however, the extent to which it supported the Neutral text was ignored, and it was regarded as a proof that the Western text was prevalent in early times not only in the West but in the East.

Washington Codex W

In 1906 a fresh discovery of interest was made, that of the Freer or Washington MS. of the Gospels (W, see p. 101 above). Here again attention was naturally attracted to the features which differentiate it from either the Syrian or the Neutral type; and while in Mt., Lk. and Jn. it belongs to one or other of these families, in Mk. it not only has an apocryphal addition after xvi.14 (quoted above), but is throughout plainly neither Syrian nor Neutral, and in the earlier chapters shows strong affinity to the Old Latin. Another Western characteristic is the order of the Gospels, Mt., Jn., Lk., Mk. It was natural at first to reckon the whole of this Gospel to the Western side of the account. It was only later research and discovery that threw an altogether fresh light on this estimate.

This research began at a remote and apparently quite unconnected point, in the identification by W. H. Ferrar, of Trinity College, Dublin, of four minuscule manuscripts (those known as 13, 69, 124 and 346) as forming a single group, closely allied in text and presenting many remarkable readings. The discovery was made by Ferrar as far back as 1868, and after his death the investigation wg.s carried on by T. K. Abbott and published in 1877.

Three of the MSS., 13,124 and 346, were written in Calabria in the twelfth or thirteenth century; 69, written in England by Emmanuel of Constantinople in the fifteenth century, must have been copied from an exemplar of the same provenance. Subsequent research has added to this group (known as the Ferrar group or fam. 13) several other MSS. (230, 543, 788, 826, 828, 983,1689,1709), as descended from the same source and retaining many of its characteristic readings, though all have been more or less brought into conformity with the Received Text. The most remarkable variant characteristic of this group is its transference of the pericope adulterae (Jn. vii.53— viii.11) to a position after Lk. xxi.38. Other notable readings are its agreement with the Curetonian Syriac in Mt. i.16, with both OS. MSS. in omitting Mt. xvi.2, 3, and with Sin. in omitting Lk. xxvi.43, 44; and in many other passages it separates itself from the Received Text. In Mt. xiv.24 it reads σταδίους πολλοὺς ἀπὸ τῆς γῆς ἀπεῖχεν, with B and OS. ; xviii.11 it omits ἦλθε ... ἀπολωλός, with א B and the Sinaitic (but not the Curetonian) O S. ; Mk. vi.3 ὁ τοῦ τέκτονος υἱός, with O L. ; Lk. vi.1 it omits δευτεροπρώτῳ with א B W; xi.2 it omits γενηθήτω ... γῆς, with B L and OS.; xv.16 χορτασθῆναι, with א B D and Curetonian (but not Sinaitic) OS.: Jn. i.28 Βηθεβαρᾷ with the Sinaitic OS. At first the significance of this group was far from clear. It was merely a rather peculiar text, emanating, so far as could be seen, in Southern Italy at a relatively late date. Subsequent developments were to add to its importance.

Another group with somewhat similar characteristics was identified by Prof. Kirsopp Lake in 1902 in his study of Codex 1 of the Gospels and its Allies. This consists of the four minuscules 1, 118, 131, 209, of which 1 is the most important. Here, as in W and some other MSS., it is in Mk. that the divergence from the Received Text is most marked, but the characteristic readings of the group are not confined to this Gospel. It agrees with א B and OS. in the omissions in Mt. i.25, v.44, and xviii.11, but not in those in xii.47, xvi.2, 3; in xix.17 it agrees with א B D and O S.; in xix.29 with B D and the Sinaitic (not Curetonian) OS.; xx.22, 23, it omits καὶ τὸ βάπτισμα ... βαπτισθῆναι and καὶ τὸ βάπτισμα ... βαπτισθήσεσθε, with א B D and OS.; xxiv.36 it omits οὐδὲ ὁ υἱός with the Sinaitic O S.; xxvii.16, 17 it reads Ἰησοῦν Βαραββᾶν with the same (Cur. is defective); Mk. i.27 it agrees with א B against almost all others; ix.44, 46, 49 it has the same omissions as B W and OS.; it gives Mk. xvi.9-20 with a note that the passage is omitted in some copies; Lk. i.28 it omits εὐλογημενη οὺ ἐν γυναιξὶν with א B; iv.44 Ἰουδαίας with א B and O S.; vi.1 it omits δευτεροπρώτῳ with א B W; ix.35 ἐκλεκτός against έκλελεγμένος of א B and ἀγαπητός of most authorities; ix.55, 56 it has καὶ εἶπεν ... σῶσαι with D and most authorities against א A B W; xi.2-4 (the Lord's Prayer) it has the shorter version with B, O S., and generally א ; xv.16 χορτασθῆναι with א B D and the Curetonian; in xxiv it has none of the omissions which characterize D and O L.; Jn. i.28 Βηθαβαρᾷ with the Curetonian; iii.31 it omits ἐπάνω πάντων ἐστίν with א D and O S.; v.3, 4 it includes the angel and the stirring of the water; vi.69 Χριστὸς ὁ υἱὸς τοῦ Θεοῦ with O S.; at vii.53 1 omits the pericope, but adds it at the end of the Gospel (so also 1582) while 209, after writing the first words of viii.12, deletes them and inserts the pericope. The importance of this group, known as fam. 1, lies not only in its own readings which seem to have an early origin, but still more in the fact, as shown by Lake, that at any rate in Mk. it forms part of a family which includes fam. 13 and the minuscules 28, 565, and 700, and in which some affinity with the Old Syriac is discernible. ← Some other minuscules (22, 872 in Mk., 1278, 1582, 2193) are included by von Soden in this group ; according to Streeter only 1582 is comparable with 1 in importance.

A third stage in a structure which was already assuming some importance was reached when the Koridethi Gospels (Θ) came to light. Attention was first called to it in 1906 by von Soden, who associated it closely with D; but after its publication by Beerman and Gregory in 1913 it was made the subject of a careful study by Lake and R. P. Blake in 1923, who showed that (at any rate in Mk., to which the study is confined), while it had no special association with D, it was closely connected with famm. 1 and 13 and the other minuscules (28, 565, 700) which have some relation to them. All these might in fact be linked together into a single family under the title fam. Θ. Moreover through Θ, which Blake believed to have originated in the Georgian colony in Sinai in the ninth century, the Georgian version was brought within the ambit of the group.

The coping stone was placed on this structure, the foundations of which were laid so far back as 1868, by Dr. B. H. Streeter in his work, The Four Gospels, published in 1924; and it will be convenient to desert chronological sequence in order to complete the story. In the first place he reinforced Lake's demonstration of the existence of fam. Θ by showing that it was not confined to Mk. but applied also to the other Gospels. He found also that a number of other MSS. occasionally preserve readings of this type, which suggests that they descend from an ancestor of this character, though they have been even more drastically revised into accordance with the Received Text than the more recognizable members of the group. The most important of these is a group headed by the minuscule 1424. The text of fam. Θ stands about midway between Neutral and Western, and has a large number of readings which do not belong to either. It may therefore claim to be a distinct family, to take its place alongside the three main families which Hort called Neutral, Western and Syrian. It stands nearer to the Old Syriac than any other Greek MSS., but not by any means so close as to represent the same text. It is about equally connected with the Armenian, and more decidedly with the Georgian. Now on examining the quotations by Origen in his commentaries on Jn. and Mt. and, his Exhortation to Martyrdom (all written between 231 and 240) the surprising result emerged that while in the first ten books of the Commentary on Jn. Origen used a copy of Mk. of the same type as א B, in the remaining books of this Commentary and in the subsequent works he used one of the same type as fam. Θ. Now it is known that Origen began his Commentary on John at Alexandria and finished it at Caesarea; hence Streeter felt justified in arguing that while at Alexandria he naturally used MSS. of the type particularly associated with that city (the א B type), when he came to Caesarea he found MSS. of a different type there (theΘ type) and thereafter used it. If this is so, the text of fam. Θ may be rightly named the Caesarean text.

This discovery of Streeter's, summing up so much earlier work, and proving that the texts of Θ, fam. 1, fam. 13 and their satellites can be joined together into a single family, and that this family, in spite of the lateness of its extant members, can be associated with the great names of Origen and Caesarea, is an epoch-making event in the history of Biblical criticism, which deserves the fullest recognition. In one respect, however, it was speedily shown to require qualification. Prof. Kirsopp Lake, who had already made such an important contribution to the subject by his study of Codex 1, in 1924 published a most valuable article on the Caesarean text of Mark, ← Harvard Theological Review xxi, 207-404. in which, after pointing out that the text of W in the latter part of Mk., which had hitherto been vaguely classed as Western because it was neither Neutral nor Syrian, is in fact Caesarean, he proceeded to examine more closely the connection between this text and Caesarea. Streeter's argument rested on the fact that in the first ten books of the Commentary on John, Origen quoted from a text differing from that which he used subsequently. But the fact (as quoted by Eusebius from Origen's own statement) is that only the first five books were written at Alexandria; and, according to Lake, in these five books the evidence as to the text used is very slight, and is quite as consistent with the use of a Caesarean text as of an Alexandrian, while in books 6-10 (the first written after his arrival at Caesarea) he certainly used an Alexandrian text. In the remaining books he seems to have used an Alexandrian text when quoting from the greater part of Mk., but a Caesarean text after about the middle of chapter xii; and in all his subsequent writings his text is definitely Caesarean.

It is therefore necessary (if Lake's arguments remain unshaken) so far to modify Streeter's conclusions as to say that Origen may have used the Caesarean text before he left Alexandria; that he certainly used the Alexandrian text on his first arrival at Caesarea; and that for the rest of his life at Caesarea he certainly used the Caesarean text. From its being thus domiciled in Caesarea (whence its connection with the Oriental versions would not be difficult) it may rightly retain the name Caesarean; but the possibility remains open that it existed already in Egypt before Origen's Hejira, and even that it was Origen himself who brought it to Palestine. On this point some further evidence may be adduced when we come to the Chester Beatty papyri.

Meanwhile a fresh classification of the textual evidence had been put forward by H. von Soden in the second part of the prolegomena to his new edition of the New Testament, published in 1906, before the identification of the Caesarean text. The familiar threefold division is retained, but with some differences, and with some peculiar developments. The division is as follows:

1. K (Κοινή): The Textus Receptus, found in the great mass of MSS. This is divided into about 17 groups, with a somewhat bewildering nomenclature, which represent a progressive modification of the text in verbal and stylistic details in accordance with the taste of the times. The three main classes are K1, the earliest form of this family, Kr the latest, and Kx, which covers all that lies between them. The earliest complete K1 MS. is ε 61 [Gregory's Ω of the eighth century, and 75 [V], 92 [0136], and 95 [047] of the ninth; but traces of K readings occur plentifully in earlier MSS. mainly belonging to other families, notably δ 4 [A], δ 3 [C], δ 2 [א] and δ 6 [Ψ ], Ki and Kik are sub-groups in which the K1 text has been influenced by other groups indicated as J and Jk respectively. Kx is the dominant text of the Middle Ages, at least from the tenth or eleventh century. Its subgroups are distinguished by the different forms in which they have (or omit) the pericope adulterae, Kμν, Kμ2, Kμ3, Kμ4, Kμ5, Kμ6, with minor varieties in which these have reciprocally influenced one another. Kr is the final revision made about the twelfth century with special reference to lectionary requirements, and this from the thirteenth century onwards became the accepted ecclesiastical text. The MSS. of this class are generally furnished either with tables of lections or with clearly marked beginnings and endings of lections in the text. They amount in number to about 200, and fall into eight divisions according to the fullness of their lectionary apparatus. Other classes are Ak, consisting of commentaries with a K text, and Ka, a variant of K1, to which over 100 MSS. are assigned, notably δ 4 [A], though this has also been influenced by other groups. This is more fully dealt with below as a member of the third family. ← For a study of Ka see S. Lake, Family π and the Codex Alexandrinus, who decides that π is the archetype of this family, which is akin to A but not directly descended from it. In an appendix Mrs. Lake gives the readings of Kr in Mark. This analysis of the great mass of K MSS., the result of a vast amount of detailed labour, is a contribution to the history of the New Testament text in the Middle Ages, but in view of the relative lateness of the K text as a whole, and its entirely secondary character, it has little bearing on the recovery of the primitive text, which is the prime object of textual criticism.

2. H (Hesychius): roughly equivalent to Hort's Neutral text. This family includes 50 MSS. headed by δ 1 [B], δ 2 [א], δ 3 [C], δ 6 [Ψ ], δ 48 [33]. δ 1 and δ 2 have a common ancestor, δ1-2, which is by far the best representative of the H text, though not free from corruption. It has been influenced both by the Coptic versions and by Origen. δ 1 has been influenced by K1 (not by Kx) and by I, and the Sahidic. δ 2 departs more from H, both by scribal errors and by alterations due to carelessness or intention. It has a considerable number of readings adopted from parallel passages (which δ 1 has not), also readings from K1, I, and Sah. It also has some elements in common with δ 5 [D], especially in Jn. δ 3 is more correctly written than δ 1 or δ 2, but is more influenced by parallel passages and by K, less so by I, and hardly at all by Sah. Other MSS. of some importance in the H group are those known in Gregory's list as L, Z, Δ, 579, 892, 1241. There are no subdivisions of the H family, and its readings can in the very large majority of cases be determined with certainty.

3. I (Jerusalem): a family not preserved in substantial integrity in any outstanding MSS., as H is in B and א, and never acquiring a dominating position, as did K, but to be elicited from a number of authorities of mixed characters. This is claimed by von Soden as his special discovery and his main contribution to textual history. It rests upon the observation of the existence of a number of readings common to the Old Latin, the Sinaitic Syriac, famm. 1 and 13, and D, sometimes agreeing with H, sometimes with K, sometimes with neither. It has fourteen subgroups, the nomenclature of which is bewilderingly arbitrary: (1) Hr = fam. 1; (2) J = fam. 13; (3) Φ 2, headed by δ 30; (4) Φ b, headed by δ 250 [2191]; (5) Φ c, headed by ε 1091 [1223]; (6) Φ r, an offshoot of Φ, of which there are three branches, in none of which is much of Φ left; (7) B, a group much influenced by K, of which there are two branches, the better one headed by ε 1043 [1216], the inferior one by δ 152 [491]; (8) Ka, which has already appeared among the K family, and which is a mixture of K1 and I, the latter furnishing the basic text, which has been drastically revised from the former, with some slight influence from H; (9) Ir, the weakest member of the I family, being 90 per cent. K; (10) Ak, like Ka, a mixture of K and I; and the same may be said of (11) Ki, and indeed of several other branches of the K family and of various individual members of it. More important is (12) Π, composed of the four purple MSS. described above as N, O, Σ, Φ which are all fundamentally I texts largely altered to K; and (13) O, a small group headed by ε 90 [U]. Finally (14) Ia is the most important group of all, headed by δ 5 [D], δ 050 [Φ ], δ 93 [565], δ 168 [28], etc. This is the best representative of the I family. It was the discovery of ε 050 that first gave it clear definition. δ 5 is a mixture of I and K1, with many scribal errors, but also with many deliberate alterations, attributable to various ancestors. It is strongly affected by parallel passages, also by the Old Latin; but this is due not to common descent from the original, but to editorial revision of an ancestor of δ 5.

The constitution of the I family is the most original feature of von Soden's work, but it is to be feared it is also the least sound. It will be observed that it combines, as closely related, the principal representatives of the Western and Caesarean families, and this does not appear to be valid. At the same time the immense labour devoted to the analysis of all these various groups and the tabulation of their characteristic readings deserves grateful recognition.

Behind these three great families lies the fundamental text I-H-K. It is possible, but not probable, that some readings of this fundamental text have disappeared;, but in some cases they may be very slightly supported, so that their identification remains insecure. By far the greater part of the variants are due to the influence (conscious or unconscious) of parallel passages; this is especially strong in K, which originated in Syria, where it was no doubt affected by the Diatessaron. Of the three families, K is much the farthest removed from the original. Besides being full of verbal and stylistic alterations, it shows some influence from the Old Syriac. It has also in its turn affected the later Syriac versions, the Peshitta being a revision of the Old Syriac by means of K. The text of Chrysostom is K1, with some traces of I (derived from Origen), but none of H; in Jn. his text is Ka. The Cappadocian Fathers also exhibit a mixture of I and K, much resembling the group Π. The K text originated in Syria, and is perhaps to be attributed to Lucian, the well-known editor of the Septuagint.

H seldom departs from I-H-K except in matters of verbal style, notably the omission of unnecessary words. It is substantially the text of Athanasius, Didymus and Cyril of Alexandria; also of the Sahidic and Bohairic versions. It is obviously Egyptian in origin, and may be assigned to Hesychius, to whom the Alexandrian edition of the Septuagint is due.

I, when recoverable, in general preserves the common original most faithfully. It must have been widely known for a long period, since its influence is far-reaching; but it was eventually submerged by K. It is the text of Cyril of Jerusalem and Eusebius of Caesarea, and is the basis of the Palestinian Syriac. It is therefore Palestinian in origin, and may presumably be assigned to Eusebius.

The relation of Origen to the I-H-K text would appear to be that he used it, but did not make it, since he often departs from it. His evidence can be used along with that of I, H, and K for the recovery of I-H-K. The later Egyptian Fathers (Dionysius, Alexander, Methodius) use I-H-K more accurately than Origen; and this was also the text used by Jerome.

Going back behind Origen, we have evidence of the early Fathers and the Old Latin and Old Syriac versions. These show many variations from I-H-K, often palpably inferior, though they may be supported by mofe than one witness. The cause of this disturbance in the text is to be found, according to von Soden, in Tatian. Such material (as distinct from stylistic) divergences as are found in K are to be attributed to Tatian, and to him is also due the strong influence of parallel passages, in which the preference is generally given to Mt. Of the versions, the African and European Old Latin are independent translations, almost always differing where difference is possible. When either or both differ from I-H-K, the difference is generally attributable to Tatian. The same is largely the case with the Old Syriac; the translator would appear to have had an I-H-K text before him, but allowed himself to be much influenced by Tatian. Of the Fathers, Origen keeps clear of the influence of Tatian, and there is little of it in Tertullian or Hip-polytus. The main text of the two last named is I-H-K, as also is that of Clement of Alexandria, who however has many readings peculiar to himself, in part due to quoting from memory. Irenaeus is mainly I-H-K, with more Tatianic influence. Justin's departures from I-H-K are personal liberties; he sets no store by accuracy in quotation. Among the principal MSS. the influence of Tatian is by far most evident in δ 5 [D]; it is noticeable in δ 2 [א], but hardly at all in δ 1 [B] or in δ1-2. The very various and sporadic appearances of Tatianic readings is a proof that the Diatessaron was the originating influence, not vice versa. The appearance of such readings both in East and West thus becomes intelligible. It is by the elimination of Tatianic corruption that the text of I-H-K can be recovered, i.e., the text of about A.D. 140, with the strong probability that it had not been seriously modified before that date.

The above analysis applies only to the Gospels. In Acts and the Catholic Epistles the H and K texts are easily distinguishable, and H is nearer to the original. A third text, which may be identified as I, is a mixture of these two, with some readings of its own. H is to be found best in δ 1 and δ 2, but also in δ 3, δ 4, δ 6, δ 48, and in the Egyptian Fathers and Versions. With regard to the latter, von Soden remarks that since δ 2 contains characteristic Boh. readings which are plainly not original H readings, Boh. must be earlier in origin than δ 2; and the same argument shows that Sah. is earlier than δ 5. The principal representative of I is δ 5; it is also found in the scanty quotations of Eusebius, Cyril of Jerusalem and Epi-phanius. The special I readings may perhaps be attributable to Tatian, if, as Eusebius records, he made a revision of the book. Origen, Clement, and Jerome used I-H-K; also Tertullian and Irenaeus, but with more intermixture of Tatian, especially in the latter.

In the Pauline Epistles the three families are all discernible, but the differences are less important. The main representatives of the several families are the same. Marcionite influence is observable, especially on the Latin versions, and to this a great part of the variants in K are probably due. In the Apocalypse, H is represented by δ 2, δ 3, δ 4, and α 3 [א C A P], K by most MSS., with greater uniformity than elsewhere.
Such is, in summary, the textual theory expounded by von Soden. In view of the great amount of labour expended on it, it has seemed right to set it out in some detail; but it must be added that its most original features, the constitution of the I family and the wide-reaching influence ascribed to Tatian, have met with little acceptance, though so far as the Old Syriac is concerned, Burkitt agrees with him.

It will naturally be asked what is the effect of all this research and theory on the text as finally constituted. Here the results may be regarded as disappointing or reassuring according to the view taken. The principles laid down by von Soden are, first, that when the readings of the three main families can be established, the reading supported by two of them is generally adopted; which must mean that where I and H differ, K has the casting vote. Where, however, two families give a text agreeing with a parallel passage, while the third gives a different one, the latter is preferred; which would rule out many K variants. Readings supported by Tatian are generally suspect; so that even if all three families agree with Tatian, an alternative that, has the support of early versions or Fathers is preferred. The result is a text not materially different from that of most other modern editors. In Sanday's 168 select passages in the Gospels, von Soden agrees with Westcott and Hort in 117, and with their second reading in 11, while he differs in 40; among the latter are included the " Western non-interpolations " at the end of Luke, which Westcott and Hort, though printing them in their text, believe to be non-authentic and enclose between double brackets, while von Soden accepts them.

Meanwhile, during all the years from 1881 onwards, a constant stream of fresh evidence was coming from the discoveries of papyri in Egypt. Unfortunately the fragments of New Testament MSS. thus discovered were until recently almost all too small to be individually of much importance. In 1932 the list of them compiled by the Rev. P. L. Hedley (still unfortunately unpublished) contained 157 items (as against 174 of the Old Testament), including vellum fragments from Egypt and ostraka (inscribed potsherds) as well as papyri. The few that have some individual importance have been described above (pp. 71-2). For the most part they play much the same role as Greek and Latin inscriptions do for ancient history, in that from a multitude of small details a general impression may be derived. The general deduction in this case is that, while a few of the fragments seem to support the Neutral text, the majority of them show no such close adherence to it as would justify the belief that that text was universally dominant in Egypt. It was at bne time attractive to believe that א B and their allies represented a pure type of text which (under the influence of Alexandria, the headquarters of Hellenistic scholarship) had maintained itself in Egypt, while elsewhere corruption had been freely admitted. It is now evident that this is not the case. In Egypt, as elsewhere, variants from this type were freely current; and though these seldom gave support to the characteristic readings of the rival Western text, they sufficed to show that the history of the so-called " Neutral" text, whatever it was, was different from what had been supposed. What explanation of it is to be substituted must be considered later.

In another direction the contribution of the papyri has been more material; for to them we owe in the main the recovery of the Sahidic version. Many fragments of this have been found, including some of bilingual Graeco-Coptic MSS., showing that such bilingual copies did exist, though the extent of their use and their influence on the Greek text has sometimes been exaggerated. But there were also a number of substantial volumes, some of which have been mentioned above (p. 131), and from all of these resources it has been possible for Horner to construct an almost complete text of the Sahidic New Testament, and to enable scholars at last to form an estimate of its real character.

But the great papyrus discoveries were reserved for the most recent years. In the winter of 1930-31 Mr. A. Chester Beatty, whose remarkable fortune as a collector enabled him to acquire (in addition to his renowned collection of Eastern and Western illuminated MSS.) at about the same time a number of valuable ancient Egyptian texts and a group of Coptic versions of lost Manichean works, obtained from dealers the papyri of the Greek Bible which have been individually described above (pp. 42-6,72-4). Of these the three New Testament MSS. may all be assigned with confidence to the third century: the Pauline MS. to the very beginning of it, the Gospels and Acts to the first half of it, the Apocalypse perhaps to the second half. They are very far from perfect, the Gospels and Acts papyrus containing not more than one-seventh (perhaps less) of the text, the Pauline papyrus nearly the whole (excluding the Pastorals and with the loss of the beginning of Romans), and the Revelation papyrus just one-third of the book; but between them they cover nearly the whole New Testament, and give us a substantial insight into the condition of the text a century before the Vatican and Sinaitic codices with which our manuscript evidence previously began. The discovery was announced in November, 1931; the Gospels and Acts text was published in July, 1933 (with a complete photographic facsimile of the papyrus in 1934), the ten leaves of the original Pauline acquisition and the Revelation in March, 1934, and the complete Pauline papyrus, with all the leaves subsequently acquired, in June, 1936. The Chester Beatty Biblical Papyri: Descriptions and Texts of Twelve Manuscripts on Papyrus of the Greek Bible, edited by F. G. Kenyon, London, Emery Walker, 1932-6. A facsimile of the Revelation papyrus was published in 1936; that of the Pauline Epistles is in preparation.

Naturally the greatest amount of interest was taken in the Gospels and Acts MS., since here the textual problems are most numerous and most important. Here it was at once apparent that the new MS. could not be assigned wholly to any of the three main families, whether they be called Neutral, Western and Syrian, or I, H, and K. On the contrary, it appeared that in Mk. at any rate (the book in which textual peculiarities are often most prominent), if it was to be assigned to any one family, it would be to the newly discovered Caesarean. It may be instructive to repeat from the introduction to the published text the table which shows its agreements and disagreements with the principal MSS.:

  with papyrus against   with papyrus against

א 42 108 W 107 52
A 54 94 Θ 65 91
B 44 106 Fam. 1 72 80
C 31 67 Fam. 13 79 73
D 49 100 Cod. 565 68 74
ς (Textus Receptus) 55 94 Cod. 700 57 87

The contrast between the two columns is striking, and is still more so when it is remembered that many of the differences in the second column (especially in Θ and the minuscules) may be attributed to the revision in the direction of the Byzantine text which all of them have more or less undergone. Even without this, the preponderance towards the Caesarean text is unmistakable, while as between the Neutral, Western and Byzantine texts it stands almost equidistant from all, and that at a considerable distance.

In Luke and John the position is rather different. Here the Caesarean text has not yet been established, so that it is impossible to say whether the papyrus represents it; but it corresponds with Streeter's definition of that text as standing midway between Neutral and Western, while it is nearer to both than it is in Mark. Thus in Lk., of which more is preserved than of any other Gospel, it has 158 agreements with B as against 130 disagreements, while for D the figures are 136 and 137 respectively. In Jn. the figures are B 40 and 40, D 43 and 37. Of Mt. too little has been preserved to allow of any conclusion. On the whole, however, the papyrus adds a conclusive proof that the text of א B was not exclusively dominant in Egypt in the third century, and also increases the probability that the Caesarean text did not originate at Caesarea, but was extant in Egypt before Origen left Alexandria, and may have been taken by him to Caesarea.

In Acts the papyrus distinctly agrees more with the Neutral group, א A B C, than with the Western. It has a number of agreements with D in minor variants, but none of the major variations so characteristic of D and its colleagues in this book. Of 77 readings printed by Clark in heavier type as especially characteristic of the Western text in those portions of the text that are preserved in the papyrus, not one is supported by the papyrus. It is therefore quite definitely not a " Western " text, and it is a further proof that not by any means all readings which have Western attestation are Western in the sense that the greater variants are.


In the Pauline Epistles the textual variations are less important, but the distinction between the Neutral and Western texts is clearly marked, and here the papyrus (P46) ranges itself definitely on the Neutral side. In passages where the two families plainly take different sides, P46 agrees in Romans with the Neutral reading in 89, and with the Western in 51. In the other Epistles the difference is even more strongly marked, the agreements with Neutral being 412 and with Western 79.

In Revelation the textual position is rather different. Here there are on the one hand the four uncials,א A C P2, then a group of minuscules headed by the late uncial B2 or 046, and finally the remaining minuscules representing the Byzantine text. As between these P47 is definitely to be classed with the four uncials rather than with either of the other groups, but the four differ considerably among themselves, and P47 differs as widely from all of them, though slightly less from א and C than from the other two. The textual variants in Revelation are not of great importance, but the papyrus must take its place as one of the principal, as it is the oldest, authority for the book.

Von Soden's edition having proved unsatisfactory, for the reasons given above, and fresh discoveries having accumulated since, the need for a new critical edition to take the place of that of Tischendorf, now antiquated by the lapse of some sixty fruitful years, was very generally felt; and eventually the task of preparing one was undertaken by an English committee, under the chairmanship of Dr. A. C. Headlam, Bishop of Gloucester, with the promise of collaboration from textual scholars in Germany and America. The editorship was entrusted to the Rev. S. C. E. Legg, who had assisted Dr. White in the great Oxford edition of the Vulgate. It was decided, after much deliberation, (1) to print the text of Westcott and Hort, rather than the Received Text or a text newly prepared, as the basis of the edition, and to adapt the apparatus of variant readings to it, (2) to make a beginning with Mark, as the book on which most textual work had been done, and in which the textual phenomena are of special interest. The fasciculus containing this Gospel appeared in 1935; and Matthew is now far advanced in preparation. The apparatus includes the variants of 29 uncials and 29 uncial fragments, of the Chester Beatty papyrus P45, of the families 1 and 13, and of a large number of other minuscules which contain notable readings; also of the Latin, Syriac, Coptic, Georgian, Aethiopic, and Armenian versions, and of 39 early Fathers. The general arrangement follows that of Tischendorf; that is, the various authorities are quoted individually, without attempting (as was done by von Soden) to prejudge results by grouping them in families. If this enterprise can be carried through (for which larger financial assistance is urgently needed), Biblical students will have an indispensable tool for their work in an edition recording all the evidence available up to the present date.

It may be useful to refer also to some smaller critical editions which will answer the purpose of those who do not require the larger work. One is the Oxford Greek Testament edited by Prof. A. Souter (Oxford, 1910), in which the text represented by the Revised Version is printed with a select apparatus recording the more important variants of the principal MSS. (uncial and minuscule), versions and Fathers—an exceedingly serviceable manual edition. The other is the edition prepared by the late Dr. Eberhard Nestle for the Wiirtembergische Bibelanstalt (13th edition by Dr. Erwin Nestle, 1927), based upon the texts of Tischendorf, Westcott and Hort, and Weiss (following as a rule the verdict of the majority), with an apparatus containing the readings of these editors when not accepted for the text, with most of the variants of von Soden, and a selection of readings from the principal groups of MSS. (von Soden's H and K, famm. 1 and 13), individual MSS. (chiefly א B D W Θ ), versions and Fathers. When once the rather elaborate system of symbols used in the apparatus has been mastered, this makes a convenient pocket edition. An edition on similar lines, with a text revised according to the editor's judgment and a select apparatus criticus, is that of H. J. Vogels (1920).

The Chester Beatty papyri are the latest of the major discoveries of Biblical MSS., and indeed the greatest since the Sinaiticus. But to complete the story up to date it is necessary to record two very recent discoveries of papyrus fragments which have a special interest. The first of these, published so recently as April, 1935, consisted of two leaves (with a tiny portion apparently of a third) of a papyrus codex ← Now in the British Museum (Egerton pap. 2), Edited by H. I. Bell and T. C. Skeat, Fragments of an unknown Gospel and other early Christian Papyri, London, 1935. containing a hitherto unknown narrative of our Lord's life. Apocryphal Gospels are by no means uncommon, and many are known to have existed, notably the Gospel of Peter, of which a considerable portion (including the Crucifixion and Resurrection narrative) was recovered (with portions of the Apocalypse of Peter and of the Book of Enoch) from a small vellum book discovered at Akhmim in Egypt in 1886, but not published till 1892. This Gospel, which was found in circulation near Antioch about the end of the second century, has features which at once show that it has no claims to authenticity; and the same is true in greater measure of the other apocryphal Gospels so far as they are known to us. But the new narrative differs from all others by reason of its early date and its sober and straightforward character. Palaeographers are agreed that the manuscript must be assigned to the first half of the second century, which makes a first century date for the narrative itself at least possible, not to say probable. The fragments contain records of four incidents in our Lord's life. Three of these are incidents which also occur in the canonical Gospels: (1) a controversy with the Jews, with phrases that recur in Jn. v and ix; (2) the healing of a leper (Mt. viii.2-3, Mk. i.40-42, Lk, v.12-13); (3) the tribute-money (Mt. xxii.16, Mk. xii.14, Lk. xx.21). The fourth is new, but unfortunately so much mutilated that its exact character cannot be determined. The style is simple and unexaggerated, very similar to that of the Synoptists; and what is remarkable is that it contains close verbal similarities with all the Gospels. Therefore, either we have here one of the early narratives to which Luke refers in his preface, using the materials out of which the canonical Gospels grew, or we have a very early use of the canonical Gospels by a writer engaged on a work of edification. The opinion of scholars, so far as it has been yet expressed, is in favour of the latter alternative; in which case the papyrus becomes an additional proof of the first century date of all Gospels, including the fourth, which at one time was hotly contested.

An even more decisive proof of the early date of the fourth Gospel, which would have been invaluable for controversial purposes sixty years ago, came to light when Mr. C. H. Roberts, working among the papyri in the John Rylands Library at Manchester, found a tiny scrap of a codex containing a few words of Jn. xviii.31-33, 37, 38, in a hand which can be referred with some confidence to the first half of the second century. ← Published in November, 1935: An Unpublished Fragment of the Fourth Gospel in the John Rylands Library : edited by C. H. Roberts (Manchester, 1935). Small as the fragment is, and therefore textually unimportant, it is of great value as the earliest extant manuscript of the New Testament, and a conclusive refutation of those who would bring the fourth Gospel far down into the second century.

These recent discoveries of exceptionally early MSS. (to which should be added the Tatian fragment described on p. 115) justify the hope of other discoveries which may clear up the many obscurities that still beset the early history of the Bible text, with the present position of which we have now to deal.


Westcott and Hort, op. cit.; G. Salmon, Introduction to the New Testament (7th ed. 1894), Some Thoughts on the Textual Criticism of the New Testament (1897); von Soden, Die Schriften des Neuen Testaments, vol. I, pts. 2-4 (1906-10); B. H. Streeter,The Four Gospels (1924); F. G. Kenyon, Recent Developments in the Textual Criticism of the Greek Bible (1933); The Chester Beatty Biblical Papyri, fasc. i-v (1933-6); S. C. E. Legg, Novum Testamentum Graece ; Evangelium Secundum Marcum (1935).