THE FOUR GOSPELS - A Study of Origins, The Manuscript Tradition, Sources, Authorship, & Dates. By B.H. Streeter, Hon.D.D.Edin. Fellow of Queen's College, Oxford; Canon of Hereford. First published by MacMillan & Co Limited 1924. - Prepared for katapi by Paul Ingram 2004.

Chapter III

THE TEXTS OF THE GREAT SEES

Home | List of MSS | synopsis | the versions as a clue | the text of Alexandria | corrected Alexandrian texts | the Western text | the text of Ephesus | the text of Antioch | (pages 51-76)

SYNOPSIS

the versions as a clue

The antecedent probability that the oldest forms of the Coptic, Latin and Syriac versions were derived respectively from the Greek texts current in Alexandria, Rome and Antioch is confirmed in the first two cases by quotations of early Fathers, in the third by less cogent evidence.

the text of alexandria

Summary of evidence that B א represent, not what Hort called a "Neutral" text, but the purest type of Alexandrian text.

The text found in Clement, which is largely Western, probably not really Alexandrian. The B א text used in Origen's Commentary on John begun at Alexandria before AD 230.

Doubt whether the "partially degenerate form of the B text" (found especially in C L and the Bohairic), to which Hort gave the name "Alexandrian," ever existed as a definite recension.

The distinction between degeneration of a text caused (a) by scribal blunders or stylistic emendation, which are necessarily wrong readings, (b) through infiltration of occasional readings from other ancient local texts, which, in certain cases, may preserve a true reading.

corrected alexandrian texts

The survival of B א side by side with certain MSS. that represent the Alexandrian text partially corrected to the Byzantine standard, enables us to study the actual process of standardisation. It appears (a) that the revision was often very irregular, (b) that the text of Mark has frequently escaped with much less revision than that of the other Gospels. Hence emerges the canon of criticism—

"research into the pedigree of a MS. should begin with the study of its text of Mark."

the western text

The Old Latin version survives in some few MSS. in a very pure form, in others with a greater or larger amount of correction to the standard of Jerome's Vulgate. There are two main families:
(1) the African, best preserved in k (shown by Hort and Sanday to preserve the text used by Cyprian of Carthage, c. 250);
(2) the European, of which b is the type MS. It is possible that a may represent a third local type. The probability is that the African Latin (k supported by e) was translated from a very old form of the Roman text. The Codex Bezae D for all four Gospels, and the recently discovered W for Mark only, give, roughly speaking, the Greek equivalent of the type of text found in the Old Latin, and in the quotations of Irenaeus of Lyons, c. 185.

the text of ephesus

Our evidence far too scanty to justify a definite conclusion; but, such as it is, it suggests that the old text of Ephesus may have been allied to that of D.

the text of antioch

The new evidence discovered since Hort wrote makes it possible to make a clear distinction between an Eastern and a Western text. The Old Syriac and the mixed cursives can no longer be treated as authorities for the "Western" text.

Summary of reasons for supposing that the Old Syriac represents approximately the ancient text of Antioch. Relation of this to the later Syriac and to the Armenian versions.
top

CHAPTER III

THE TEXTS OF THE GREAT SEES

the versions as a clue

OUR jumping-off point, so to speak, for a scientific study of the text of the New Testament is the consideration that the churches of Rome, Alexandria and Antioch were the frontier stations of Greek-speaking Christianity. After the fall of Jerusalem, these naturally became the "home base" of missions to the peoples whose native speech was Latin, Coptic or Syriac. This fact facilitates our quest for the early local texts of the Gospels; for there is obviously a presumption that the Latin, Egyptian and Syriac versions were derived from the Greek texts current respectively in Rome, Alexandria and Antioch. In the case of Rome and Alexandria this is more than a presumption. Marcion c. 140, Justin c. 150, and Hippolytus c. 190-236 wrote in Rome, and Tatian about AD 172 compiled the Harmony of the Four Gospels, known as the Diatessaron, either at, or immediately after leaving, Rome. All these wrote theological works in Greek and so presumably read the Gospels in Greek, especially as this was the language of the liturgy of the Roman Church. But their quotations show that the text they used was similar to that which appears in the surviving MSS. of the Old Latin. A similar inference may be drawn from the general coincidence between quotations of the Gospels by Origen, Athanasius and Cyril of Alexandria with the type of text found in the Coptic (i.e. Egyptian) versions.

Unfortunately we cannot test the early text of Antioch in the same way. Theophilus, AD 180, is said to have composed a commentary on the Four Gospels, but in his one surviving work his quotations of them are too rare to be of use to the textual critic; and the next writer of this church is Chrysostom, 360. Nevertheless reasons, less than demonstrative but still cogent, can be produced in support of the view, in itself antecedently probable, that the Old Syriac approximately represents an early text of Antioch. The identification of the old text of Caesarea, which is the main contribution that I personally have to make to the subject, will be discussed in the following chapter. But in the course of this chapter I venture in regard to the text of Ephesus to put forward a suggestion that avowedly is no more than mere conjecture.
top

the text of alexandria

Our first step is to scrutinise the Greek MSS., especially those of early date, to see if the text of any of them exhibits any close connection with that of one or other of the three types that the early versions attest. At once our search is rewarded by the discovery that the text of the two oldest MSS. B א and their VIIIcent. Ally, L is closely connected with that of the Coptic version which exists complete in two dialects, the Sahidic and Bohairic and to that implied in quotations of the New Testament by Origen and Cyril of Alexandria. Moreover, a text identical with that found in א B L is found in the fifth century fragments of Luke and John, known as T. T is bilingual, Graeco-Sahidic, so that this alone would, as it were, anchor this type of text in Egypt. Besides this, papyrus fragments of the third and fourth centuries have been found at Oxyrhynchus agreeing closely with B א,
[Of. esp. the fragments of John in Oxyrhynchus Papyri 208 and 1781, probable date AD 250-300. Though printed in different volumes (ii. and xv.) these are part of the same MS. This is the oldest known MS. of any part of the Gospels and is in book (not roll) form.]
while the text found in later papyri is predominantly, though not exclusively, of this type. Additional evidence may be found in the fact that readings found in the B א text are sometimes spoken of as "Alexandrian" in the scholia in certain MSS.
[E.g. in c, to Mt. xxv. 1, there is the note, "sponsa non in omnibus exemplariis invenitur, nominatim in Alexandrino"; to Lk. x. 43-44, in Syr. Hiermg, "This section is not found in the Gospels among the Alexandrians." Cf. Tischendorf, ad loc.]

Lastly, the readings specially characteristic of this text are not found in the quotations of any early Fathers outside Egypt.

Bousset has compiled a series of half a dozen tables of various readings to illustrate the relation between the text of א B L and the various Graeco-Sahidic fragments T. As these tables also serve to illustrate the relations of א L to one another and to B, I reproduce the first of them, [W. Bousset, Textkritische Studien zum Neuen Testament (Leipzig, 1894), p. 77.] in which he analyses the 104 variants occurring in the fragment containing Lk. x. 20-xi. 20. The left-hand column shows the number of variants supported by each of the four MSS.; the others show the number of times that T is supported by B, א and L respectively.

    B. א. L.

B א L T

64

64

64

64

BאT

7

7

7

 

BLT

11

11

-

11

BT

15

15

-

-

אT

4

-

4

-

א L T

1

-

1

1

T

2

-

-

-

Total

104

97

76

76

This table is fairly typical of the series, and it shows the close relation of this group of MSS., but the central position occupied by B.  When we find that in 97 out of 104 variants the reading of B has the support of one or more of the other MSS., while there are only five cases in which the others combine against B, the inference that in at least 90% of its readings B preserves the text of the common ancester of the group can hardly be resisted.

In 1913 a IVcent.  MS. of John, representing an older form of the Sahidic, was found in Egypt. Sir E. Maunde Thompson courteously informed me that, so far as he could judge from a first rapid examination, its text is akin to B א.

The Sahidic, the older of the two complete Egyptian versions, has only recently been made known to the world [The Coptic Version of the New Testament in the Southern Dialect, otherwise called Sahidic and Thebaic (no name), Oxford, Clarendon Press, 1921.] in a reliable form in the magnificent edition of Dr. Horner, along with the fullest Apparatus Criticus of the Greek text at present available in English. The Sahidic, it is now clear, goes in the main with the B א text; but in an important minority of readings it goes over to the side of the text represented by D and the Old Latin version especially in its African form. From the figures given in Dr. Horner's analysis of readings it would appear (op. cit. iii. p. 387) that for all four Gospels the Sahidic has 505 readings characteristic of D (with or without Latin support) and 157 distinctively Old Latin. The meaning of this "Western" element in the Sahidic cannot be appreciated if considered in isolation. It must be studied in connection with the appearance of Western readings in א, in L, and in the other MSS. that have a text akin to B. Of these the most important are C 33, and, for Mark, Δ Ψ.
[Actually the closest supporters of א B L T are the two fragments Z Ξ, which contain respectively about one-third of the Gospels of Matthew and Luke, Z being nearer akin to א and Ξ to LC and 33 have a considerable mixture of other texts. With a very much greater amount of mixture the Alexandrian text is preserved in four cursives, 167, 579, 892 (esp. in Mk.), and 1241 ( = Sod. ε371). A few notable readings of the later Alexandrian type are found in X. Soden, somewhat perversely I think, classes 157 as a supporter of the "I text." He does the same (here I have not checked him) with the uncial fragments P Q R, except that for John he regards Q as Alexandrian. For 157 see pp. 49 and 76 note. 892 was collated by J. Bendel Harris, Journal of Biblical Literature, 1890.]
Again, Dionysius, bishop of Alexandria c. 260, seems also to have used a form of the B text which had an infusion of Western readings. The notable fact, however, is that whenever one or more of these authorities desert B to give a Western reading, almost always there are others of them found ranged in support of B. The natural conclusion to draw from this is that B represents approximately the oldest text of Alexandria, but that at a very early date MSS. with a Western text were in circulation in Egypt. Considering the close connection between Alexandria and Rome, which was mainly dependent on Egypt for its corn supply, there must frequently have been Christians from Rome coming to Egypt on business and bringing with them copies of their Gospels. Odd variants from these would naturally be entered on the margin of local MSS., and would thus creep into the text. But, since this happened in a haphazard way, one set of Western readings would get into one Egyptian MS., a quite different set into another.

To the view that B represents, not merely the Alexandrian text but also the earliest form of it, an objection, at first serious, arises from the Gospel quotations of Clement of Alexandria, 190-200, since these are found to have a specially large infusion of Western readings. [P. M. Barnard in Texts and Studies, v. 5 (Cambridge, 1899).]  
It ought not, however, to be taken for granted that these quotations represent an average Alexandrian text of that date.

  1. It is thought by some that his extant writings were composed after he had left Alexandria.
  2. Clement was not a native of Alexandria, but came there fairly late in life.

He had lived for many years in S. Italy. Is it likely that, when he migrated from thence, he left his copy of the Gospels behind? Clement usually quoted from memory; now in regard to the vagaries of the human memory an appeal to personal experience is valid. I am myself in the habit of reading the New Testament in the Revised Version, but I was "brought up" on the Authorised, and it is still the version commonly read publicly in the Church of England. As a result of this I find, when revising MS. for the press, that, when I have quoted from memory, the resultant is nearly always a mixture of the old and the new versions. Now, if Clement's own copy of the Gospels represented an early form of Western text but he commonly heard them read in church from a text akin to B, it would be inevitable seeing that he was apparently not interested in textual criticism that his quotations would represent now one, now the other, type of text. I would go further and suggest that, as Clement was head of the catechetical school, his pupils would be likely to note on the margins of their own copies notable variants from the master's; and as these pupils subsequently became leaders of the Church, the readings from their copies would tend to get into the texts used in some of the principal churches. In that case the text of Clement, so far from representing the earliest text of Alexandria, would be a main source of its decline.

But the determining piece of evidence (cf. p. 93 fi.) that the B text represents the early text of Alexandria is its use by Origen in the earlier books (the limitation is intentional) of the series of homiletical lectures known as his Commentary on John. Of course the MS. used by Origen was not absolutely identical in text with B.   No two MSS. are exactly identical.    Sometimes Origen agrees with X against B; more rarely he agrees with one of the other manuscripts belonging to the same family; occasionally he has a reading characteristic of D.    The few readings in which he appears to support the Byzantine text may be suspected as probably the result of corruption in the text of our only MS. of this work of his.    But, all said and done, it would be safe to say that the manuscript used by Origen for the first ten books of his Commentary on John differed from B. less than B and א differ from one another.   This evidence is highly important for three reasons.

  1. The Fathers, including Origen himself, frequently quote from memory; but in this work, which contains a long series of quotations from John with a running commentary upon them, we have absolute security that, in regard at any rate to the longer quotations, Origen is not quoting from memory but reproducing a written MS.
  2. Textual criticism belonged to the tradition of Alexandrian scholarship and it was a subject in which Origen himself was supremely interested. He had already spent some years on his famous Hexapla a critical edition of the Greek Old Testament. There is no reason to suppose that at this time he contemplated a critical edition of the New Testament. But it would be incredible that he should not have provided himself with the best text of the Gospels available, before starting his Commentary on them. To an Alexandrian critic the best text meant one based on the oldest MS. procurable. But the oldest MS. that Origen could procure would have gone behind the time of Clement, his immediate predecessor in the Catechetical School. Indeed, its date may well have been nearer the middle than the end of the second century.
  3. We can exactly date the evidence.
    Origen himself tells us that the first five books of the Commentary on John were written before he left Alexandria for good and migrated to Caesarea. This took place in the year 231. We have therefore a fixed point for the textual criticism of the Gospels. The text of the Gospels preserved in B א (practically, that is, the text printed by Westcott and Hort) is to all intents and purposes the text on which Origen lectured in Alexandria in the year 230.

In the quotations of the Alexandrian fathers, especially in Cyril (d. 444), and in the Bohairic version which on the whole is even closer to B than is the Sahidic occur a number of readings, which look like attempts at grammatical and stylistic improvements of the B text. Readings of this class crop up in all the Alexandrian MSS., except B. Some are found even in א; but they occur most thickly in L, and next to that in C 33 Ξ ΔMk. ΨMk.  Hort attributed an importance greater than either their number or their character deserves to them.  Hort declined to recognise any connection of B א with Alexandria; the B א text he named the "Neutral text," and assigned it to no definite locality. And he gave the name "Alexandrian" to a text conceived of as the B text modified by the minor stylistic improvements found in the readings of C L 33, etc., wherever these differ from both B and the " Western" texts in any readings not also found in the "Syrian" (=Byzantine) text. This so-called "Alexandrian" text he described as "a partially degenerate form of the B text." While admitting that no single MS. preserved this text entire, he ranged it alongside the "Neutral" and "Western" as one of three great families of pre-Byzantine, or, to use his own title, "pre-Syrian," texts. It is generally recognised that this was a mistake. MSS. in which readings of this type are found may well be described in Hort's words as exhibiting "a partially degenerate form of the B text." But all of them include a number of Western readings; and there is no evidence that any MS. ever existed which contained what Hort calls the "Alexandrian" readings but did not also include many Western readings. Nor is it certain that the whole number of the "Alexandrian" readings ever coexisted in any single MS. It is quite as likely that "Alexandrian" stylistic correction and infiltration of Western readings were two concurrent processes gradual in character acting upon individual MSS. in the natural course of textual corruption, and thus affecting different MSS. in different degrees.

If, however, we use a word like "corruption" or "degenera­tion" in this connection, we must be on our guard against an easy fallacy. Any departures in these MSS. from the B type, which are of the nature of grammatical and stylistic correction, are "corruptions" in the strictest sense of the term; that is to say, they are alterations, intentional or accidental, of what the original authors wrote. But departures from the B text, which consist in the substitution of a reading found in ancient authorities belonging to the Western family, are corruptions in a quite different sense. In so far as they are departures from the oldest form of the Egyptian text they are a degeneration of that particular textual tradition. But the Western authorities represent a textual tradition of great antiquity belonging to a different locality, and it may well happen that they sometimes preserve a true reading which has been lost in the B text. This is a point on which clear thinking is extremely necessary. To say that the text of B is a purer representative of its type than א is by no means the same thing as saying that it is a purer representative of the original text of the Gospels as the authors wrote it. That is quite a different question. For example, the incident of the Bloody Sweat (Lk. xxi. 43 f.) was, I am inclined to think, in the original copy of Luke's Gospel. If so, א, which contains the passage, preserves the true text; B, which omits it, does not. But a comparison of the MSS. and Versions which contain or reject the passage shows that its absence is characteristic of the particular line of textual transmission which א B on the whole represent. If so, its presence in א is due to "contamination" with a MS. of the "Western" type. But what follows? Whenever א has a "Western" reading, it ceases to be an authority for the Alexandrian text; but it becomes for the time being the oldest Greek MS. with a "Western" text. To put it metaphorically, B is a thoroughbred;
א
is a cross, but a cross between two thoroughbreds of different stocks. Hence, as evidence for what the authors wrote, the "Western" readings of א are a most valuable authority; but, if we mistake them for evidence of the primitive text of Alexandria, we fall into hopeless confusion. The reader may ejaculate that he is not interested in the primitive text of Alexandria, but only in what the original authors wrote. In textual criticism there are no short cuts; and, since local texts of the Gospels came into existence in the second century, it will not be till we have got back to these in their most primitive form that we have all the materials on which to base our judgement as to what the original authors wrote.
top

corrected alexandrian texts

The fortunate preservation of representatives of the Alex­andrian text so ancient and relatively pure as B א presents us with an exceptionally favourable opportunity of studying the phenomenon of correction of MSS. of an ancient local text to the Byzantine standard. It will be instructive to consider some examples of this.

The first point to notice is that correction was often ex­tremely "patchy" and unsystematic. Thus the corrector of א, whom Tischendorf styles אb, corrected the earlier chapters of Matthew with scrupulous care; but his interest in the work seems to have flagged, and he makes few corrections in the latter part and hardly any in the other Gospels. This actual instance of a corrector's work explains at once a peculiarity of L, an ancestor of which must have suffered a similar fate; for the text of L is almost the Byzantine in Mt. i.-xvii., but it has only a thin sprinkling of Byzantine readings in the latter part of Matthew and in the other Gospels.
[Interesting figures as to Byzantine correction in L, C and Δ may be found in E. A. Button, An Atlas of Textual Criticism, p. 13 (Cambridge, 1911).]
The V or VIcent. correctors of א, however, whom Tischendorf calls אc and אcb, were more systematic than אb. Indeed, if א had been copied after these various correctors had finished, the result would have been a mixed text of components and general character very like C, which has a large proportion of Byzantine (and a few miscellaneous) readings in all four Gospels, but more so in Matthew and Luke than in Mark and John.

The Icent. cursive 579 (cf. p. 50) has in Matthew an ordinary Byzantine text. In the other Gospels, especially in Luke, it has a considerable number of Alexandrian readings. In the Introduction to his edition of this MS., Schmidtke gives a list of readings in which it differs from the Byzantine text in order to support one or other member of the Alexandrian group. On the basis of these lists I have compiled the following figures:

agreements with B א,

31;

with B against א,

132;

with א against B,

111;

with one or other of the group C L 33 Δ Ψ 892 against both B and א,

134.

We must, however, recollect that the Byzantine text is much more closely allied to the Alexandrian than to the Western. Hence the great majority of the agreements of 579 with א B C L 33 are necessarily excluded from a list of readings, which purports to include only differences between 579 and the Byzantine text. Again, many of the readings in which B א agree against that text are of a conspicuous nature, and therefore would be particularly likely to attract the notice of the corrector. Bearing this in mind, the interest of the above figures is the demonstrative proof they afford that, although the majority of readings in 579 are in agreement with the Byzantine text, an ancestor must have been a MS. the text of which stood right in the middle of the Egyptian group of MSS. The value of a MS. of this kind appears where it supports a reading of B, א or L, which otherwise is unsupported. Every MS. has a larger or smaller number of errors due to mistakes by the scribe who wrote it or one of its immediate ancestors. Such errors in no sense constitute readings characteristic of the local text, which the MS. as a whole represents. If, however, a "singular" reading of any MS. is supported by another MS., which on other grounds we can connect with the same family, we have sufficient proof that the reading in question is not an accident or idiosyncrasy of the particular MS. in which it occurs. If that MS. happens to be B or א, such support for "singular" readings is of special interest.

Another point about mixed texts is illustrated from the Egyptian group.
Except in Mark, Δ and Ψ have the ordinary Byzantine text with a few scattered readings of the later Alexandrian type; but in Mark this state of things is exactly reversed. The fundamental text is the later Alexandrian with a few scattered Byzantine readings the proportion of Alex­andrian readings in the other Gospels being rather larger in Ψ than in Δ. If these two MSS. stood alone we should infer that Mark was copied from an exemplar belonging to a different family from that used for the other Gospels. But these two MSS. are only an extreme example of a regularly recurrent phenomenon. A study of mixed texts belonging to other families than the Alexandrian shows that it is not the exception but the rule for the Gospel of Mark to have a much smaller proportion of Byzantine readings than the other Gospels. Nearly every one of the MSS. of which the text is discussed in our next chapter exhibits this more especially the Codex Theodorae 565. The phenomenon must, therefore, be explained on the hypothesis that these MSS. were copied from exemplars in which the text in all four Gospels had originally been pre-Byzantine, but which had been more thoroughly corrected to the standard text in the first three Gospels than in Mark. Mark provided very few lessons for the selection read in the public services of the Church. It was much less used and much less commented on than the other Gospels; of this an interesting illustration is the Xcent. MS. X, which has a marginal commentary on the three longer Gospels, but merely gives the bare text of Mark. Hence the comparative carelessness shown in correcting Mark to the fashionable type of text is easily accounted for.

There emerges a principle of some importance, but one that heretofore has been insufficiently emphasised. Seeing that the Gospel of Mark has escaped Byzantine revision in more copies and to a greater extent than the other Gospels, it follows that our materials for reconstructing the old local texts are far more abundant and more trustworthy in this Gospel. From this we deduce the following canon of textual criticism.
Research into the pedigree of a MS. should begin with a study of its text of Mark.
top

the western text

Jerome's Vulgate, as has been already indicated, played in the Latin Church the same part as the Byzantine text in the Greek. The process of haphazard correction of older MSS. to the standard text resulted in the production of a number of copies having a mixed text, partly Vulgate, partly Old Latin. Fortun­ately, besides these, a few MSS. with a text entirely Old Latin, or with only a small admixture of Vulgate readings, still survive. These differ from one another very considerably; and they differ even more in the Latin words chosen to represent the Greek than in the underlying Greek text they presuppose. Hence many scholars think there must have originally been two independent translations from the Greek, which subsequently have become somewhat confused by mixture with one another as a result of sporadic correction of MSS. of the one translation by MSS. of the other. However this may be, it is the fact that the Old Latin MSS. which it is customary to cite by small italic letters sort themselves roughly into two main groups, of which the most typical representatives are k (Bobiensis) and b (Veronensis).

One of the most important contributions to the criticism of the Gospels made in recent years was the demonstration by Dr. Sanday that the text of k is to all intents and purposes identical with that used by Cyprian of Carthage c. 250. This gives us another fixed point towards the determination of early local texts. Accordingly the type of text found in k is commonly spoken of as the "African Latin." Unfortunately k is only extant for Mk. viii.-xvi. and Matt, i.-xv. But another MS. e (Palatinus), also incomplete, while overlapping with k to some small extent, contains those parts of the Four Gospels, which in k are missing. In e we have a somewhat later form of the same text as k, with a slight mixture of European Latin readings. In fact the type of text in e has much the same relation to that of k as L bears to B. The African text, so far as Mark and Luke are concerned, is also supported by c, a twelfth-century MS. which has a text, roughly speaking, half Vulgate and half Old Latin, though in the other Gospels the Old Latin approximates nearer to the type of b. [Cf. F. C. Burkitt, J.T.S., Jan. 1908, p. 307 B. The Old Latin and the Itala (Texts and Studies, iv. 3, 1896) by the same author must be read by all students of the Latin versions.] The Speculum, a collection of proof-texts, cited as m, perhaps of Spanish origin, helps to eke out our scanty authorities for this African Latin text.

Tertullian, the predecessor of Cyprian at Carthage, speaks of the Apostolic Sees, with special reference to Rome, as the "wombs of the Catholic Church." [De Praescr. Haer. 21.]
From this and from the general probabilities of the case we may tentatively infer that African Christianity came from Rome, and that the African Latin was ultimately derived from an early form of Roman text. Further evidence, slight but pointing in the same direction, may be seen in the noticeable points of contact between the African Latin MSS. and Fathers and what little we know of the text of Marcion, who was in Rome 140-144. The African Latin, it is important to notice, in many of its readings agrees with B א against the type of Old Latin of which I will now proceed to speak.

Some scholars call the other type of the Old Latin the "European," others the "Italic." The MS. that occupies the same sort of central position among the European Latin MSS. as B does among the Alexandrian is, curiously enough, denoted by the letter b (Veronensis). As the Dean of Christ Church puts it, "b indeed seems to be almost a typical European MS.; as the other MSS. of European and of Italian origin, such as a f h i q r, all resemble b more closely than they resemble each other."
[H. J. White, Old Latin Biblical Texts, iii. p. x., Oxford, 1888. This vol. contains an edition of
q, the Introduction to which is invaluable to the student of Old Latin texts.] The most constant supporter of b is ff 2 (Corbiensis II.). Of the MS. mentioned above the oldest [Burkitt argues that k also is IVcent. (J.T.S., Oct. 1903, p. 107); b and e. are uniformly dated Vcent., ff 2 V or VIcent.] is a (Vercellensis) IVcent.
This MS., supported by the fragment n, stands a little apart from the others. The difference between a and b is at its maximum in Mark, so that the critical canon just enunciated justifies the suspicion that it may possibly represent a third local type, intermediate between b and k, which in the other Gospels has been partially conformed to the b text.
[The VI cent. MS. f (Brixianus) has a large number of readings that occur in the Byzantine text but not in other Old Latin MSS.; many of these occur in the Vulgate. There is a difference of opinion among experts as to whether this MS. represents an attempt earlier than Jerome's to revise the Latin by comparison with the Greek, or whether it is a Vulgate MS. corrupted by the influence of the text of the Gothic version, as it seems to have been copied from the Latin side of a bilingual Gothic-Latin MS. Cf. Wordsworth and White, Nov. Test. ... Hieronymi, p. 656, and F. C. Burkitt, J.T.S., Oct. 1899.]

Of the European Latin as a whole it may be said that it repre­sents a type of text at the furthest remove from that of B. And even the African Latin, which in small points frequently deserts the European to support B, is conspicuous for the number of striking additions—"interpolations" Hort calls them—to the B text.

The Roman theologians, up to and including Hippolytus, d. 236, wrote in Greek; and the liturgy of the Roman Church was in Greek possibly till an even later date. The number of quotations in the fragments of Marcion or in the writings of Justin and Hippolytus sufficiently definite to be used for critical purposes is not very large, but such as they are they imply the use of a Greek text, roughly speaking, corresponding to the Old Latin. We know also that the Diatessaron or Harmony of the Four Gospels compiled by Tatian was produced about 170, either during, or im­mediately after, his long residence in Rome; and this had a text of the same character. But the use of the Greek language slowly died out in the West. Hence it is not surprising that we have few MSS. which preserve the type of Greek text used in the West during the period when Greek was still spoken there. For the Epistles of St. Paul four such survive; but for the Gospels until recently there was only one, the Codex Bezae D.

Being practically the sole representative of the Greek text used in the West towards the end of the second century, D has a quite unique importance, and a large literature has come into existence about it. Its text stands fairly well in the middle of the various MSS. of the Old Latin. Where these differ from one another, D sometimes supports one type, sometimes another; but on the whole it is nearer to the European than to the African type. In a certain number of readings it supports B against the Old Latin; in a much larger number it agrees with B against the Byzantine text. Its date, according to the latest authority, is fifth century.

D is a bilingual MS., having the Greek on the left, the Latin on the right-hand page.
[It is customary to cite the Latin half of D, which not infrequently differs from the Greek, as d. Similarly the (far less important) interlinear Latin of Δ is cited as δ.]
Theodore Beza, who presented the MS. to the University of Cambridge, states that it was found in the monastery of St. Irenaeus at Lyons; and Dr. E. A. Lowe [J.T.S., April 1924, p. 270 ff.] produces good reasons for the belief that it was already in Lyons in the ninth century. Where it was originally written is a question on which there is at present no agreement among experts. Southern Italy, Sardinia, or the Rhone valley are the favourite guesses. In favour of the last-mentioned locality is the close relation of the text of D to that used by Irenaeus, Bishop of Lyons c. 177-195. This relation was noted long ago, but is re-affirmed with further confirmation in Sanday and Turner's recent edition of the New Testament Text of Irenaeus. [Novum Testamentum S. Irenaei (Oxford, 1923).] We may, however, infer that it was written in some rather out-of-the-way church or monastery, for two reasons. First, the corruptions in the text imply an ignorant scribe. Secondly, the story of the man working on the Sabbath inserted by D after Lk. vi. 4, and the attempt to assimilate the genealogy of our Lord in Luke to that found in Matthew, are readings so remarkable that they almost demand comment. Yet, to the best of my knowledge, no allusion is made to either by any ecclesiastical writer. This is easily explicable if these readings were current only in some out-of-the-way church.

The solitary position of D as the Greek representative of the Old Latin text has been partially relieved, so far as the Gospel of Mark is concerned, by the discovery in 1906 of W (the Freer MS. Vcent.), the possession of which gives the library of Washington the distinction of containing one of the six most ancient copies of the Gospels. The text of W presents a unique problem. Its editor, Prof. A. T. Sanders, [Cf. the elaborate introduction to his collation of W, The New Testament MSS. in the Freer Collection, Part I. (The Macmillan Co., New York, 1912).] thinks it is descended from an ancestor made up of fragments from different rolls of the Gospels pieced together in some district where the attempt of Diocletian to crush Christianity by destroying the sacred books had been more than usually successful. Some portions seem to have been drastically revised to conform to the B א, others to the Byzantine, text. In the whole of Matthew and in Lk. vii. 13 to the end it presents a text mainly Byzantine; since most of Matthew is wanting in Δ (the only other MS. as early as the fifth century which gives a text closely allied to the Byzantine) we have in W a welcome accession to the early evidence for this type of text. For the first seven chapters of Luke and in John (v. 12 to the end) the text of W is mainly of the Alexandrian type. But the most notable feature of W is its text of Mk. i.-v. 30. This is almost word for word identical with the Greek text underlying the African Latin e. Unfortunately hardly anything of Mark after vi. 9 is preserved me; so no comparison of W and e is possible in the later chapters; but in the rest of Mark W is still found to agree, though less closely, with the Old Latin and especially with k. So far, then, as the Gospel of Mark is concerned we have in W a valuable addition to the evidence afforded by D for the ancient Greek text current in the West to which text alone the title "Western" will in these chapters be applied.
top

the text of ephesus

Constantinople, as we shall see shortly, appears to have adopted its text as well as its theology from Antioch. It was inevitable, therefore, that at Ephesus, situated as it was between these two dominant patriarchates, the old local text should succumb at an early date to the standard text used by both these Sees. There is reason to believe that some time in the fourth century Ephesus was compelled to surrender its ancient liturgy in favour of the Byzantine. [W. Palmer, Origines Liturgicae, i. 106.] But there is a certain amount of evidence that in the second century the text used at Ephesus was akin to that found in D and the European Latin.

(1) The recently discovered second-century document known as the Epistula Apostolorum is supposed by its editor Carl Schmidt [In Texte und Untersuchungen, 1919.]  to be of Ephesian origin. The evidence adduced is not conclusive, but this view is more probable than any other. The author of this work is clearly familiar with all four Gospels. But for our present purpose the most interesting point is that he seems to have read them in a text like that we call Western in the strictest geographical sense.

(a) In ch. 2 he has a very remarkable list of the apostles in which the name of John stands first; and one of the names is Judas Zelotes. It might have been imagined that this was due to a conjectural combination, made by the author himself, between the names of the two apostles mentioned in Lk. vi. 16, Judas of James and Simon Zelotes; but this same combination, Judas Zelotes, occurs (Mt. x. 3) as a substitute for the ordinary Thadaeus =Lebbaeus in the Old Latin a b h q etc. It also occurs in the fifth-century mosaics in the Baptistry of the Orthodox at Ravenna, which are in another respect connected with the Western text, in so much as they arrange the evangelists in the order Matthew, John, Luke, Mark, found in D, W, several Old Latin MSS., and in the Monarchian Prologues (P. 11).

(b) In ch. 3 ad fin. we read

"He is the word made flesh, born in the sacred Virgin's womb, conceived by the Holy Ghost, not by carnal lust, but by the will of God."

This seems to imply the famous Western reading of Jn. i. 13, which substitutes ὃς ... ἐγενήθη for οἳ ... ἐγενήθησαν, and thereby makes the fourth Gospel also assert the Virgin Birth of Christ. This reading is found in b, in three quotations of Irenaeus, two of Tertullian, and was also known to Ambrose, Augustine, and probably to Justin Martyr.

(c) Possibly his text also included the Longer Conclusion of Mark. [Cf. C. Schmidt, op. cit. pp. 219, 224, also below, p. 348.] This is a characteristic Gallic and Italian reading.

It is absent from the African Latin, from the oldest Alexandrian MSS., from the majority and the oldest MSS. known to Eusebius of Caesarea, from Syr. S., which seems to represent the old text of Antioch, and from the Eastern authorities mentioned p. 88. But it is found in D, in all Old Latin MSS. except k, and in the text used by Irenaeus and Tatian.

These striking agreements between the text quoted in a document of the second century, probably Ephesian, and the text used in Italy and Gaul, compel us to review the nature of the patristic evidence for the Western text. We notice at once a special connection between most of our earliest authorities for the Western text and the Roman province of Asia of which Ephesus was the capital.

(a) Justin Martyr was converted to Christianity in Ephesus; and according to the evidence given by himself at his trial, though he had lived and taught in Rome, he had done so without any very close affiliation to the local Church. [Cf. the Martyrdom of Justin and the discussion by K. Lake, Landmarks of Early Christianity, p. 127 (Macmillan, 1920).] The text of the Gospels he used is therefore more likely than not to have been the one he brought from Ephesus. Tatian was a pupil of Justin, and may well have used his master's text.

(b) Irenaeus as a boy sat at the feet of Polycarp of Smyrna, and never tires of emphasising the value of the apostolic tradition of the Churches of Asia and Rome. But the connection between Asia and the Church of Lyons, of which Irenaeus was a member and ultimately bishop, was in no sense personal to Irenaeus himself. Eusebius preserves the letter written by the Churches of Lyons and Vienne to the Church of Ephesus to tell them the story of the martyrdoms in the persecution of 177. This implies a special affiliation of these Gallic Churches to the Church of Asia. The Greek-speaking com­munities of the Rhone valley seem always to have kept up a connection, mainly no doubt for trade purposes, with the cities of Ionia, of which they had originally been colonies. [Three inscriptions have been found in Lyons set up by persons described as natione Graeca, and one by a lady, natione Asiana. Cf. Vasile Parvan, Die Nationalität der Kaufleute im römischen Kaiserreiche, pp. 90, 107 (Breslau, 1919).]
It is, therefore, exceedingly probable that the Christianity of the Rhone valley was derived from Ephesus. In that case the text of the Gospels used there would naturally be the Ephesian text.

(c) Since Justin, Tatian, and Irenaeus all resided and taught in Rome, some readings from the text they used would get into the local text. The text of Irenaeus, we have seen, is closely related to that of D and the Old Latin. This suggests the possibility that the earliest Latin translation used in Gaul was derived, not from the Greek text used in Rome, but from that used in the Rhone valley. This translation might have spread thence into Gallia Cisalpina, the consanguineous district of N. Italy.

The evidence available is quite insufficient to justify any definite conclusion, but it at least suggests the tentative hypothesis that while regarding the African Latin as a descendant of the older Roman text, we should look on D and the European Latin as representing a mixture, varying with individual MSS., between the Roman and the Ephesian text.
top

the old text of antioch

We pass on to consider a field of inquiry which more than any other has been illuminated by recent discovery the old local texts of the Asiatic provinces of the Roman Empire. When Hort wrote, the materials at the disposal of critics were insufficient to justify any definite conclusion, and we are still hampered by the lack of ecclesiastical writers from whose quotations of the New Testament the type of text current in these provinces can be ascertained. In fact the only early writer native to these provinces of whom enough survives to be of any practical use for this purpose is the historian, Eusebius of Caesarea, c. 325. It had long been recognised that Eusebius used a "Western" text, but one of a peculiar kind, that is to say, a text which, although more closely allied to D than to א B L, is markedly distinct from D. But the only MS. giving a continuous text of early date which could be certainly assigned to the Eastern provinces was the fragmentary Cureton MS. of the Old Syriac version (Syr. C.), which contains less than half the total contents of the Gospels, and of Mark only four verses. Hence Hort was justified in including the Old Syriac and the sporadic non-Byzantine readings of a similar character found in cursives like 565 or the Ferrar group under the general designation of the "Western text" a title which he inherited from Griesbach, although he quite recognised the geographical inappropriateness of this extended use of the adjective.

Since Hort wrote, the situation has been completely changed by a series of discoveries. Of these the one which has opened up the prospect of our obtaining at least a general idea of the ancient text of Antioch was the discovery in 1892 of the Sinaitic Syriac (which I shall cite as Syr. S.), a fourth-century palimpsest containing, with some lacunae, a fairly complete text of the Four Gospels in the Old Syriac version. The Syriac text of Syr. S. and Syr. C., along with an English translation, purposely so literal that even the order of words in the original can often be followed, was published in 1904 by F. C. Burkitt under the title Evangelion Da-Mepharreshe = "Gospel of the Separate," as distinguished from the Diatessaron or "Gospel of the Mixed." The Introduction and Notes to this edition form a contribution to textual criticism the value of which to the advanced student cannot be over-estimated.

It appears that Syr. S. and Syr. C. represent fundamentally the same version, but that one or the other must have been revised partially from a Greek MS. having a slightly different type of text. Burkitt thinks that Syr. S. gives the version most nearly in its original form, while Syr. C. has been revised here and there by a Greek MS. more or less similar to the Codex Bezae. However this may be, the fact remains that the version as a whole is more closely allied to the Western text than to B, although Syr. S., especially in the matter of omission, frequently supports B against both Syr. C. and D. For instance, it omits the last twelve verses of Mark and the two notable passages Lk. x. 43 (the Bloody Sweat) and Lk. xi. 34 ("Father, forgive them"). But, though on the whole it ranges itself on the side of D and the Old Latin against B and its allies, the Old Syriac has a sufficiently large number of distinctive readings found neither in D nor B to justify our regarding it as a third type of text.

Burkitt was, I believe, the first to work out in any detail the suggestion that the Greek text underlying the version of the Old Syriac preserved in Syr. S. was derived from the older text of Antioch. [Evangelion Da-Mepharreshe, ii. p. 254.]   His argument, briefly, is as follows.   Tatian, who seems to have been the first effectively to plant Christianity in Mesopotamia, introduced there, not the Four Gospels, but the Syriac Diatessaron which for centuries was spoken of as "the Gospel."    The Four Gospels, known by contrast to the Diatessaron as "the Gospel of the Separate," were a later introduction.
Syr. S.
seems to be an earlier form than Syr. C. of the Syriac version of the Separate Gospels.   Its translator was familiar with the Diatessaron, and its readings, as well as its renderings, may sometimes have been affected by that fact; hence the original Greek text from which Syr. S. was translated will have differed from the Diatessaron even more than does the translation.    Now the text of the Diatessaron is closely akin to D and the Old Latin.   This resemblance, coupled with the fact that Tatian came from Rome about AD 172, makes it highly probable that the text used for the Diatessaron was the Roman text. Where, then, did the text of Syr. S. come from?    Geographically, the province of which Antioch is capital marched with the Syriac-speaking district.    More than this, there is evidence that, after the disorganisation caused by a period of persecution, Serapion the Patriarch of Antioch, c. AD 200, re-established the Syriac Church by consecrating Palut, the bishop from whom in after years that Church reckoned its episcopal succession. Thus the revived Syriac-speaking Church was in a special way a daughter church of Antioch, and would naturally obtain therefrom the text of the "separate" Gospels which hitherto it did not possess. The presumption, then, that the Old Syriac represents the second-century text of Antioch is decidedly high. Moreover, I would observe, to any one who prefers the view of certain scholars that the old Syriac version of the Gospels is earlier than the Diatessaron, the presumption that its text came from Antioch is considerably enhanced; for the only reasonable ground for doubting the Antiochene origin of the Syriac text arises from the known connection of Tatian with Rome.

Burkitt points out that a number of readings of the Old Syriac, which are not found in any other Greek MS., occur in one or more of the cursives 1 &c., 13 &c., 28, 565, 700. Those MSS., I shall argue in the next chapter, preserve (with much Byzantine admixture) the old text of Caesarea. Seeing that Caesarea and Antioch were the capitals of adjoining provinces, the discovery that those MSS. represent the text of Caesarea cannot but add weight to the view that the cognate text implied in the Old Syriac has some special connection with the neighbouring Church of Antioch.

Any evidence is welcome which throws further light on the true text of a version that survives only in two MSS., both imperfect. For this purpose some use can be made of the Peshitta the Syriac version used at the present day in all branches of the Syriac-speaking Church. Of this we have MSS. as early as the fifth century. Burkitt ["S. Ephraim's Quotations from the Gospels," in Texts and Studies, vii. 2 (Cambridge, 1901).] has shown that the Peshitta represents a revision of the Old Syriac made by Rabbula, Bishop of Edessa, about 425. The MSS. used by Rabbula evidently represented the Byzantine text, and his revision was fairly thorough. Nevertheless the number of readings of the Old Syriac which survive in the Peshitta is considerable, though the proportion which these bear to the whole is less than that borne by the Old Latin readings surviving in Jerome's Vulgate. Wherever, therefore, the reading implied in the Peshitta differs from that of the Byzantine text, there is a fair presumption that it represents the reading of an Old Syriac MS.

The other Syriac versions are of less value as evidence for the text of the Old Syriac. We have a number of MSS. of a revision made for the Jacobite sect by Thomas of Harkel in 616; but this revision went still further in the direction of assimilating the Syriac to the Byzantine Greek text. He noted in the margin the readings of three old Greek MSS. in the Gospels and of one in the Acts, which differed from the Byzantine text. These readings (cited as Syr. Hclmg.) are of considerable interest; but their importance lies in the evidence they afford for pre-Byzantine Greek texts, not for the light they throw on the Old Syriac.

Much the same may be said of the "Palestinian" or "Jerusalem" Syriac (cited as Syr. Hier.). Burkitt has shown that this was not a native Palestinian product. [J.T.S., Jan. 1901, pp. 174 ff.] It was produced in a monastery near Antioch as part of an effort of Justinian to combat Nestorianism in Palestine by providing orthodox litera­ture in the vernacular. It is probable that the translators made some use of previous Syriac versions, but Syr. Hier. cannot be safely quoted as an authority for the Old Syriac still less, as von Soden thought, for the Greek text used at Jerusalem. [Hoskier, J.T.S., Jan. 1913, p. 242, notes points of contact between the text of this version and the mixed cursive 157.]

The Armenian version is held by Dean Armitage Robinson [Cf. "Euthaliana" in Texts and Studies, iii. 3 (1895).] to have been originally made, wholly or in part, from the Old Syriac. In that case it may be used as supplementary evidence for the original form of that version. This view, however, has lately been disputed (cf. p. 104); I shall venture later on (p. 104 f.) to put forward a suggestion of my own in regard to the most debateable point.
top