ON THE OBIGg^ OF THE FEEEAB-GEOUP, LECTURE ON THE GENEALOGICAL RELATIONS OF. NEW- TESTAMENT MSS. : Delivered at mansfleld college, oxford, . ON NOV. 6th, 1893,. BY I. RENDEL HARRIS, M.A., D.Litt. (Dubl.) *, ;,; EELIW OF QiARB COLLEGE, CAMBRIDGE. LONDON: 0. J.' CLAY and SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,1, . ,AVE MARIA LANE. ; 1893 '-Price One Shilling and Sixpence. ON THE ORIGIN OF THE FEEEAE-GEOUP. - SonHon: C. J. CLAY and SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AVE MARIA LANE. dnmltiSBC : DEIGHTON, BELL AND "CO. Eeipjtjj: F. A. BROCKHAUS. #ris -gorli: MACMILLAN AND CO. ON THE ORIGIN OF THE FEEEAE-GEOUP. A LECTURE ON THE GENEALOGICAL RELATIONS OF NEW TESTAMENT MSS. DELIVERED AT MANSFIELD COLLEGE, OXFORD, ON NOV. 6th, 1893, BY J. RENDEL HARRIS, M.A., D.Litt. (Dubl.) FELLOW OF CLARE COLLEGE, CAMBRIDGE. LONDON : C. J. CLAY and SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AVE MARIA LANE. 1893 [All Rights reserved."] ffiatnbrfoge: PRINTED BY C. J. CLAY, M.A. AND SONS, AT THE UNIVERSITY PRESS. The history of the text of the New Testament is the most perplexing of the unsolved problems of the universe, and has almost as many missing links as the chain of life itself ; indeed ERRATA. Page 9, line 4, after Syr. add 10 „ 11, „ 20, „ noted ,, above, and which we shall observe which all schools are agreed^ and one oT~tEese points is the derivation of the four manuscripts which form the so-called Ferrar-group from a common lost original. This lost original, which was edited from the manuscripts by Ferrar and Abbott, is held to be the equivalent of an uncial text of the first water, 1 Burgon : The Revision Revised, p. 256 : " Assemble an Englishman, an Irishman, a Scot ; a Frenchman, a German, a Spaniard ; a Russian, a Pole, an Hungarian ; an Italian, a Greek, a Turk. From Noah these twelve are all con - fessedly descended : but if they are silent, and you know nothing about their ante cedents — your remarks about their respective ' genealogies ' must needs prove as barren — as Dr Hort's about the ' genealogies ' of copies of Scripture." b ON THE ORIGIN OF on account of its frequent agreement with the oldest known codices. Burgon suggested that when the four manuscripts agreed it would be well to denote the reading in which they agree by a symbol or a single capital letter ; and von Gebhardt, in his New Testament of 1881, has actually assigned to the lost archetype of the four one of the few remaining letters of the Greek alphabet [<£]. Now the question forces itself upon us ; supposing that a number of genealogical relations, similar to that which has been demonstrated to hold for the four MSS. in question, could be established, what would be the meaning, historically, of the separate ramifications ? These divergences are the key to the life-history of the text, and they must be interpreted. The New Testament in its textual transmission is unlike that of any other known book, it is complicated by commentary and irregular revision, and con trolled by re- translations, and therefore the leading groups of the MSS. which are the carriers of the text have a significance which is linguistic, chronological, geographical, and historical. We ought not merely to try to determine the groupings of the MSS., but to ask at what time the groupings diverge from the main tradition, and if possible in what part of the world, and under what surroundings in the history of the Church. A formidable series of problems, in all conscience ; let us, how ever, begin to attack them by unravelling some of the threads of the textual tangle as regards the Ferrar-group. First of all, let us recall briefly the work which has already been done on this famous group : the classical work is the restoration of the text of the lost original made by Prof. Ferrar, and completed, after his death, by Prof. T. K. Abbott. We have here an editorial restoration which shews by spaced type all the deviations from the printed text of Stephen of 1550 accom panied by a critical apparatus which furnishes the individual variations of the manuscripts in question amongst themselves and a valuable, though all too brief, introduction dealing with the relations of the four texts to the leading uncial authorities and to one another. This was published in 1877. Next in importance was the tract of the late Abbe" Martin THE FERRAR-GROUP. 7 published in 1886, entitled Quatre Manuscrits importants du Nouveau Testament auxquels on petit en ajouter un cinquieme. The Abbe complains that the editors of the group of texts had made no effort to determine the geographical home of the lost original. " II semble qu'ils auiaient du s'attacher a. en ddter- miner 1' origine, si cela leur eut etd possible... A-t-on fait ces efforts ? A-t-on signale des faits qui trahissent la patrie, l'origine et la provenance de ces manuscrits ? — Pas le moins du moude ? " After which complaint, the Abbe- proceeded to shew that three out of the four texts (13, 124, and 346), were demonstrably written in Calabria. And he also pointed out that there were other MSS. (notably the MSS. Ev. 348 and Ev. 556), which belonged, in all probability, to the same group. This tract was a valuable contribution to the history of the text, for, in locating three out of the four texts, within a small radius of country, we may reasonably conclude that the lost common original was also located ; the hypothetical Codex is therefore a Calabrian MS. The next step was taken by myself, in 1887, in an attempt to trace the history of the Leicester Codex (Cod. 69), the only one of the four which the Abbe" Martin had not seen. I was able to throw a good deal of light on the migrations of the MS. though not to write a complete life-history of it. Some further notes were published by Batiffol, who was rather contemptuous of my work, on the ground that I made a long journey to results which he was able to reach much more rapidly. Since the year 1887 not much has been done on the matter, with the exception of some fresh identifications of MSS. more or less closely related to the famous group. My object in the present discourse is to throw a fresh ray of light on the Ferrar-text, and so, by implication, on the text of the Gospels at large. And as it frequently happens that the solution of a large problem depends upon the solution of a small one, I shall begin by trying to resolve an obscurity in the general paleography of New Testament MSS. In Scrivener's Introduction to the New Testament (ed. 3, p. 62) we are told that " besides the division of the text into tniyoi or lines, we find in the Gospels alone another division o ON THE ORIGIN OF into prf/Mara or prjcre^, 'sentences,' differing but little from the o-tiyoi in number. Of these last the precise numbers vary in different copies, though not considerably, etc." And on p. 66 we find the following statistical statement : Matthew has 2522 pii/Mtra Mark „ 1675 Luke „ 3803 John „ 1938 These figures are derived from MSS. of the Gospels, in which we frequently find the attestation given both of the pr\p.ara and the oTi%oi : e.g. Cod. Ev. 173 gives for Matthew ,/3^>k/3' prjfiara ,/3(/>f o-Tt'vot, while the corresponding figures for Mark and Luke are Mark ,a%oe'| aB(^ Luke /ywy' 1 No explanation, as far as I know, has ever been given of these curiously numbered ptjfiara. The word is, certainly, a peculiar one to use, if short sentences are intended, such as are commonly known by the terms 'cola and commata.' It has occurred to me that perhaps the explanation might lie in the fact that pfjfia was here a literal translation of the Syriac word r*i2a^o\&. Let us then see whether Klsa^Aia is the proper word to describe a verse, either a fixed verse, like a hexameter, or a sense-line. A reference to Payne Smith's Lexicon will shew that it may be used in either of these senses for example, we are told that it is not only used generally of the verses of Scripture, but that it may stand for " comma, mem- brum versus, sententia brevior quam versus, a-ri^of, Schol. ad Hex. Job ix. 33; Ssa n£sa\£i& Tit. ib. Ps. ix. h£so^\* .sj rdAiri?, ib. Ex. xxx. 22 marg.: insunt in Geneseos libro riisa^Ava mmmmdix, coloph. ad Gen., it. c. s. b. 2 et sic ad fin. cuiusque libri; in libris poeticis sententia est hemistichio minor, e.g. in Ps. i. insunt versus sex sed .T» A\&. ; in Ps. ii. versus duo decim, sed -ma o\&. THE FERRAR-GROUP. 9 It seems, therefore, to be used in Syriac much in the same way as o-Tt'^o? in Greek. Now there is in one of the Syriac MSS. on Mount Sinai (Cod. Sin. Syr.} a table of the Canonical books of the Old and New Testaments with their measured verses. We will give some extracts from this table ; but first, notice that the Gospels are numbered as follows : Matthew has 2522 rcisa^Ava Mark „ 1675 Luke „ 3083 John „ 1737 and the whole of the four Evangelists 9218, which differs slightly from the total formed by addition, which, as the figures stand, is 9017. On comparing the table with the numbers given by Scrivener from Greek MSS., viz. Matt. = 2522 p^ara Mark = 1675 Luke = 3803 John = 1938, we see at a glance that we are dealing with the same system ; Luke should evidently have 3083, the Greek number being evidently an excessive one ; and if we assume that John should be 1938 the total amounts exactly to the 9218 given for the four Gospels. This is very curious, and since the prifiara are now proved to be rightly equated to Klsa^ova, and this latter word is a proper word to describe a verse or cttiyo?, the pr^iaTa appear to be a translation of a Syriac table. Perhaps we may get some further idea about the character of the verses in question by turning to the Sinai list, which is not confined to the Gospels, but ranges through the whole of the Old and New Testaments. The Stichometry in question follows the list of the names of the seventy disciples, which list is here assigned to Irenaeus, bishop of Lugdunum. After which we have 1—5 10 ON THE ORIGIN OF rcia&i.l ri'-liVga.i n^iaJt-CU* -=o*» r^Av.'ira : m^_j\co3SO .t*» •"»*" : TQQi^.a\o\s-c\ i.e. followed by Genesis has 4516 verses Exodus 3378 7i Leviticus 2684 » Numbers 3481 JJ Deuteronomy 2982 J> Total for the Law 17041 J> Joshua 1953 JJ Judges 2088 ;j etc. When we come to the New Testament, it seems at first sight as if the verses which are there reckoned cannot be the Greek equivalent hexameters : for we are told that Philemon contains 53 verses, and the Epistle to Titus 116, numbers which are in excess of the Euthalian reckoning, 38 and 97 verses respectively, and similarly in other cases. The sugges tion arises that the lines here reckoned are sense lines, and this is therefore the meaning to be attached to the prmara of the MSS. But upon this point we must not speak too hastily. The interest of the Sinai stichometry is not limited to this single point : its list of New Testament books is peculiar in order and contents. There seem to be no Catholic Epistles, and amongst the Pauline Epistles, Galatians stands first ; note also the curious order Hebrews, Colossians, Ephesians, Philip pians. While I do not think there can be the slightest doubt that our explanation of the origin of the pij/j,ara is correct, it may be worth while to point out a curious case of the occurrence of THE FERRAR-GROUP. 11 the same mistake in English writers as we have attributed to an unknown Greek scribe. If we turn to the Letters of Dr Olinthus Gregory in defence ofthe Inspiration and Authenticity ofthe Scriptures1, we shall find the writer, in the course of his argument alluding to the evidence for the antiquity of the text afforded by the Syriac version. And he tells us that this Syriac version was preserved from very early times in the churches of Malabar. In proof of which he says that " Dr Buchanan who in 1806 visited the Syrian churches, amounting to 119 in Malayala, was informed by the inhabitants that no European had, to their knowledge, visited the place before," and that amongst the Syriac MSS. which Dr Buchanan lighted on in Malabar Churches " there is one volume found in a remote church of the mountains, which merits particular description; it contains the Old and New Testaments, engrossed on strong vellum, and written with beau tiful accuracy. The character is Estrangelo Syriac, and the words of every booh are numbered." It may readily be supposed that we have lighted here upon a precisely similar case to that which we notedflh the Ferrar- Group. So we read a little further in order to find out some more about this interesting MS. Gregory goes on to tell us that "this copy of the Scriptures has admitted as cano nical the epistle of Clement, in which respect it resembles the Alexandrine MS.: but it has omitted the Revelation — that book having been accounted apocryphal by some churches during a certain period in the early ages." If we search for a MS. which we suspect to have the numbers of the prj/iara, and find, in the quest, a copy of the Syriac Clement, we shall have supplied one more illustration of the Scripture story of King Saul, who went to seek his father's asses and found a kingdom. But Dr Gregory goes on to say that " this most valuable and interesting MS. is now in England. Mar Dionysius the resident bishop at Cadenette, presented it to Dr Buchanan, who again has presented to the University of Cambridge, in whose public library it is now lodged. It has been lately examined with ' Ed. Lond. 1812, p. 132. 12 ON THE ORIGIN OF great care and skill by Mr Yeates, who has published a more minute account of it than the above, in the Christian Observer1." So we must now turn our attention to Dr Buchanan, Mr Yeates, and the great MS. itself, which is in one ofthe show cases in the University Library. We shall soon see whether the " words are numbered," and whether the epistle of Clement is there. We must go to Dr Buchanan's account, in order to be sure that we have identified the MS. correctly. He tells us in his Christian Researches in Asia (published at Cambridge in 1811) that " some of the Malabar copies are certainly of ancient date. Though written on a strong thick paper, like that of some MSS. in the British Museum, commonly called Eastern paper, the ink has in several places eaten through the material in the exact form of the letter. In other copies, where the ink has less of a corroding quality it has fallen off and left a dark vestige of the letter, faint indeed, but not in general illegible. There is a volume, which was deposited in one of the remote Churches, near the mountains, which merits a particular description. It contains the Old and New Testa ments, engrossed in strong vellum, in large folio, having three columns in a page : and is written with beautiful accuracy. The character is Estrangelo Syriac and the words of every book are numbered. But the volume has suffered injury from time or neglect. In certain places the ink has been wholly obliterated from the page, and left the parchment in its state of natural whiteness : but the letters can, in general, be distinctly traced from the impress of the pen, or from the partial corrosion of the ink — I scarcely expected that the Syrian Church would have parted with this manuscript. But the Bishop was pleased to present it to me, saying, ' It will be safer in your hands than in our own,' alluding to the revolu tions in Hindostan — 'And yet,' said he, 'we have kept it, as some think, for near a thousand years.' ' I wish ' said I, ' that England may be able to keep it a thousand years.' " A comparison of Dr Buchanan's description with the features of the great Malabar MS. in the University Library 1 Christian Observer for May and June 1810. THE FERRAR-GROUP. 13 shews a complete agreement. The MS is a noble folio, written in three columns, in the Estrangelo hand, and far advanced in the stages of decay. Olinthus Gregory's account is taken, partly verbatim from Dr Buchanan's own story, and partly from Yeates' articles on the MS. in the Christian Observer. His reference to the epistle of Clement is a misunderstanding of the following sentence in Yeates' account : "The books of Clemens, called Clemens the Saint and Disciple of Peter, eight in number are subjoined to the books of the New Testament. They are not preserved entire in this copy : the four last being mere parts of leaves and fragments." These books of Clement are the same, in text, with those published by Lagarde in his Reliquiae Iuris Ecclesiastici, with whose text the Malabar MS. ought to be collated1. And now at last we come to the pr/fiara, of which Mr Yeates remarks that " it is to be noted that subjoined to the subscript of each book of the Old Testament is the number of words contained in each book respectively, which number is expressed in words at length and written in red characters." A reference to the MS. shews that all the subscriptions referred to are the numeration of the r«Lsa^o\A. For instance, the subscriptions to the four gospels run as follows : Math. Here endeth the holy Gospel of the blessed Matthew, which he spake and wrote in Hebrew in Palestine. [No verses legible.] Mark. Here endeth the holy Gospel, the preaching of Mark the Evangelist, which he spake and wrote in Roman in Rome. [No verses given.] Luke. Here endeth the preaching of the Evangelist Luke, which he spake and wrote in Greek in the city of Alexandria; verses 3803. John. Here endeth the holy Gospel, the preaching of John 1 It is strange that no one attempted to publish the text from this MS. Yeates simply says, as if to provoke one, that ' ' search has been made in the books of Clemens published in Greek and Latin, but no translation has been found corresponding with the Syriac text of the above." 1—7 14 ON THE ORIGIN OF the Apostle, which he preached in Greek in Ephesus, and its verses are 1938. And is ended, by the help of the Divine grace, all the holy book of the Gospel, of Matthew, and of Mark, and of Luke and of John, all of them ; its complete verses are 9938. Comparing the figures with those given above from the Sinai MS.: we see that the same error occurs in the Gospel of Luke in both MSS. and that our correction of the number for John is confirmed. The total for the four gospels is again corrupt: (apparently the 938 is a repetition from the correct number for John). We have, however, without going further into the matter at this point, the evidence which we were looking for, that the peculiar Syriac word for verses was misunderstood by the Cambridge scholars of the beginning of the nineteenth century, exactly as it was by an ancient scribe of the Greek Gospels. Having now cleared up the obscure subscription which occurs in the New Testament MSS. and, for obvious reasons, is peculiar to such manuscripts ; we proceed to ask the question as to what manuscripts have been translating or propagating such Syriac subscriptions, and with this we cannot avoid asking two other questions; the first is as to whether there are sug gestions of a common genealogy in manuscripts which have prjixara subscribed ; the second, as to whether such re-transla tion from Syriac is limited merely to the subscriptions. Now it is matter of no small interest that amongst the handful of MSS. which have the subscriptions in question, there are found the leading members of the Ferrar-group. The subscriptions are, indeed, wanting in the Leicester Codex which has no subscribed matter ; but from the other three MSS. we have a subscription as follows for the lost archetype, with a few variations for the individual MSS. of the group. etc tov Kara fiardaiov eiayyeXiov' iypd(f>rj i^paiaTi ev ¦jraXcuarlvT) p,era rj errj tjj9 dva\i]^rea><{ tov lev' eyei Se pr/fiaTa fi(pKj3'' ej(€t Se ctt('%ous ,/S^>^'. eiayyeXiov kcito. fidp/cov eypdipr) ptdfMilaTrj ev poofiy fieTa THE FERRAR-GROUP. 15 tp err) TJ79 dvaXtjyjreay; tov Kvpiov' e%et, Be prjfiaTa ,a%oe' eiayyeXiov Kara Xovicav eypdipr) eXXrjvio-fi et? dXet-avhpiav Trjv p.eydXrjv fiera te err) ti?? dvaXq-^reas tov Kvpiov, eY,et Se pijfiara flcoy' <7TtY,ov? fiyfrv . evaryyeXtov KaTa 'Icodvvov eypdxpr) eXXrjvto'Ti, eh e<^>eaov yttera errj X' t>/? dvaXrrfreco? tov Kvpiov' ej(ei Se prjfiaTa ptrXr) ctIxovs //SkS/. eirl Sop.eTiavov 0acnXe : iypdijyr) eXXr/vio-Ti et? etpecrov fieTa err) X Trjs dvaX-rjyfrew tov kv + e%ei Se prjfiaTa pirXrj + e%et Se o-rtYov? ]3icS + iiriSofieTi avoir tov /SacrtXeo)?. The last subscription shews the same numerical blunder (¦n- for ~*\) as was observed in the Milan MS., with which indeed it closely agrees. There is no doubt that the Burdett- Coutts MS. belongs to the family, and that the agreement in the subscriptions is here also a just index of genealogy1. Whether the same thing is true of all the MSS. which carry the p-rj/iaTa subscriptions, it is impossible to say, as the major part of them are imperfectly described in the catalogues, and hardly any have been collated. The principal texts to be tested are as follows : Codd. 9, 163, 168, 173, 174, 211, 230, 233, 345, 427, 507, 592, 709, 715, 716, 828, 873. It is worthy of note that Cod. 9 is supposed to have been written in Sicily, Cod. 174 in Calabria, and Cod. 873 probably in Calabria ; Cod. 211 is supposed by the Abbe- Martin to be a Sicilian MS. and to be closely related to the Milan MS. (Cod. 346). For want of further information we are obliged to leave the discussion of the possible relation of these Codices to the Ferrar- Group on one side. There are also some other copies which are supposed to be in textual relation with the Group, which do not (as far as we can judge from descriptions in catalogues) contain 1 This MS. was imported from Janina in Epirus. THE FERRAR-GROUP. 17 the subscribed prjfiaTa, as Codd. 348, 788, 826. Of these all three are Calabrian in origin1. As far then as our enquiry has gone we have proved that certain Greek MSS. of the Gospels contain Syriac subscriptions ; and that those Greek MSS. which exhibit the Syriac table of verses are suspiciously related; some of them are certainly derived from a Calabrian ancestor, and others appear to have been written in Calabria and in Sicily. We now pass on to our remaining question, which has regard to the possibility of Syriac influence in the Ferrar-Group, and those texts which stand with it. As we have already said, it would be an unlikely thing that the scribe who first translated the Syriac subscriptions should have made no use of the Syriac text to which the subscriptions were attached. So that if we were discussing the question for the first time we should certainly read the Ferrar-text for Syriasms. The text itself is easily seen to be composite in character, derived from two sources at least and perhaps from many. The possibility of a Syriac element in the text is favoured by such a consideration ; but we do not need to discuss the subject de novo, for the Ferrar-text has already been challenged for Syriasms, before any attempt had been made to solve the puzzle of the prjfiaTa. In a tract on the Diatessaron of Tatian published by me in 1890 (but written two years before that date) I argued for the possibility of a dependence of the Ferrar-Group upon the text of Tatian *, on the ground that there existed in the Ferrar-text (a) actual Syriasms, (/3) curious coincidences in text with the Curetonian Syriac and with the recovered readings of the Harmony, (7) harmonistic readings which agreed with conjunctions made in the Harmony. For the sake of convenience I reproduce a few of the leading points of the argument. In the account of the Transfiguration the Ferrar-Group 1 We have not succeeded in throwing any fresh light on the relation between the interesting MS. Cod. 561 and the Ferrar-Group. It is quite possible that the connexion is not so intimate as was imagined by me in an article published in the Journal of the Society for Biblical Literature and Exegesis. 3 See Diatessaron of Tatian, pp. 45 — 50. 18 ON THE ORIGIN OF gives us for vecpeXr) tpcoTeivrj, vecpeXi) TO<; (Matt. xvii. 5). The reading may fairly be described as a Syriasm, and is actually found in the Curetonian Syriac, and in Ephrem Syrus. It is only found in Greek in the Ferrar-Group and a few cursives. In the same account in Luke ix. 29, the Harmony reads, ' Et dum ipsi orarent,' and the Ferrar-Group in Mark ix. 3, gives Kal iv tg3 7rpoo-ev')(ecr0ai avTois fieTefiopTiov avTov has interrogantes eum, dixerunt illi; and accordingly the Ferrar text gives iirrfpaiTaiv avTov, XeyovTe<:. The addition might be due to Matthew (elirov), but is a natural expansion in Syriac. No other evidence for it exists, beyond one or two cursives. In Luke xxiv. 40 the Harmony has " ostendit eis manus et pedes et latus," a reading which is apparently compounded of Luke xxiv. 40, and John xx. 20. Now if we turn to John xx. 20 in the Ferrar text, we shall find the verse run eSeigev avTol<; to? %et/aa? Kal toi)? 7r6Sa? Kal Trjv TrXevpdv avTov. There is no other Greek support for this reading, and only one Latin Vulgate MS. It may be suspected to have originated in a combination made by Tatian. A similar case may be found in the Arabic Harmony at Matt. xvi. 8" sciens autem Jesus, dixit eis : Quid cogitatis intra vos, modicae fidei. Mark viii. 17b et solliciti estis quia etc. While the Ferrar-Group reads in Mark ti SiaXoyi£ea0e ev THE FERRAR-GROUP. 19 kavTOis 6Xty6trio-Toi though the last three words are not con sidered to form a part of the true text of Mark. Other instances may readily be found ; as I pointed out in my Tatian notes, the real difficulty in the determination of the source of the readings lies in the discrimination between Tatian readings proper and readings absorbed by Tatian or by old Syriac texts dependent on him which are classed as Western. This dis crimination required the complete solution ofthe genesis ofthe Western readings and cannot be discussed here. But it seemed reasonable to believe, that where Tatian or the old Syriac con spire with the Ferrar-Group in singular readings for which little or no Greek attestation can be found, the Ferrar-Group have derived their Greek text by the expedient of a translation. The result of this investigation accords so well with the evidence derived from the subscriptions, that I think we may take it to be demonstrated that there is a decided streak of Syriac in the Ferrar-text. And we may go a step further in the determination of the way in which the Syriac element was introduced. It did not come from the Harmony directly, but from a Syriac MS. of the Gospels which was coloured by the Har mony ; there are no calculated verses in a Harmony, but in the text which coloured the Ferrar archetype the verses were subscribed. We are entitled therefore to say that some of the intrusive elements in the archetype came from an old Syriac MS., which was, like all Old Syriac MSS. of the Gospels, affected by Tatianisms. Again, this Old Syriac text was of a late type, and comparable rather with the Curetonian text than with the earlier Sinai MS. Some of its coincidences with the Cureton text are very decided. For example, in Matt. i. 17 we have in the Curetonian text -.co rc'AvAoAv^D -o_,T-t3) cnA chocn r^i » VW.1 Geo and in Cod. 346 $ fivr)o-Tev0eiaa trapdivoi M.apidfi iyevvqerev 'Irjcrovv tov Xeyofievov Y,j0t<7Toi>. 20 ON THE ORIGIN OF This did not come from Tatian, for in his text the genealogies were absent ; there is no Greek authority for it, beyond the MS. quoted (though Latin authority is abundant) ; whatever may be the ultimate source of the gloss, it seems reasonable to say that the Ferrar text took it from an old Syriac text; which is as much as to say that the Greek evidence for the reading has now disappeared. It may be relegated to the category of Syro-Latin readings. But there is another reason for saying that the Syriac element in the Ferrar archetype is relatively late in origin. The numbering of the verses (whatever be the kind of verses counted) is not a feature of any Old Syriac text which has yet come to light. It is not in the Sinai MS. which is probably our oldest form, nor is it found in the Cureton text. We infer therefore that it belongs to some later period chronologically than that to which these two closely related MSS. are referred. In con clusion we may offer a few remarks as to the resultant impression which our enquiry has made upon us. It seems that the direction in which we ought to work for the solution of the origin of peculiar readings in Greek copies of the Gospels is the study of the separate versions. The existence of a reading attested by a copy and a version is of itself suspicious; for we see in such cases a tendency on the part of the copy to run back and hide under the wings of the version. Many Greek variants may turn out not to be Greek at all. It becomes, therefore, necessary that the text of the New Testament should pass into the hands of Latin, Syriac and Coptic scholars. They must tell us how much Latinism, Syriasm and Coptism they are able to detect. And we must wait the result of their investigations. The work is however one that requires some caution and self-restraint, in order that the explorer may not be betrayed by his enthusiasm so as to import into the texts upon which he is engaged the phenomenon of which he is in search. We may remark further that the general criticism of the text will be much affected if it should turn out that in zealously collating such cursive texts as shew singularities THE FERRAR-GROUP. 21 we are merely repeating the evidence of versions; and if it should be found that when we describe a manuscript or group of MSS. as striking or ' containing an original text' or the like, we are praising some translator for his misdeeds, and des pising the mass of transcribers who have committed no sin except that of want of originality. APPENDIX. I add some further notes on Syriac and quasi-Syriac Stichometry. The transmission of a stichometric record in Syriac is as might be anticipated, subject to errors of its own, which are totally different from those made in Greek MSS. In the first place the confusions which are common are found in different numerical signs, and in the next the count is made in a different manner. The count is safest when it is made at length, the thousands, hundreds, tens, and units being expressed in distinct words. In this case the only error, other than those arising from wandering eyes, occurs between the tens and the units, as when, for instance ^isnt. is confused with s -it.. But where the numbers are given by signs, or where the direct expression in words is limited to the thousands, we are in difficulty directly : one method of counting prevalent in Syriac MSS. represent the hundreds by the same sign as the tens, only using a diacritical point; (e.g. ^_^Ar^= 1893), and it need hardly be said that this method is peculiarly liable to errors of transmission, especially where the number for the APPENDIX. 23 tens or hundreds is absent, or where the number of tens is the same as the number of hundreds. We have had a case of this above in the enumeration of the prjfiaTa of Luke as 3803, where it should have been 3083. A second method represents the hundreds by direct signs as far as the limits of the alphabet, and then employs com binations : thus we have st = 100, \ = 200, a. = 300, h\ = 400, after which we proceed with jio>, iw, juo\, and a\a\, and write jscWeW for 900. The error to which this count is most liable is the dropping of a sign. Where the signs are taken over from a previously existing Greek rendering, we must allow for other directions of error; for, even where no signs are misunderstood by the transcriber, we must remember that when we pass tr and A, the two alphabets diverge ; st which stands for 100 in Syriac is there fore not the equivalent of the Greek Koppa, which stands for ninety : neither is the Greek p to be taken over into a Syriac i, which represents 200: but how easy the mistake, on the part of a rapid or ill-informed transcriber, to increase the number of hundreds by one in the very act of transference of the record, without ever intending to do so. It will be seen from the foregoing remarks that the Syriac stichometries are very liable to error; but when we realise the forms of error which recur most frequently, we ought to be able to edit, by comparison of MSS. inter se, the primitive calculated extent of the book, and we may, perhaps, be able to determine whether the count was originally made in Greek or Syriac. We will shew, for example, that the Sinai and Malabar stichometries at all events in regard to the Pauline epistles, are both derived from the Greek, and that their count, in spite of all confusions and misplacements is demonstrably that of Euthalius. The comparison is a little difficult because the Malabar MS. is in many places illegible from age, and my copy of the Sinaitic list is imperfect, while the books are arranged in an unusual order. 24 APPENDIX. The comparison may be made by the following table : Euthalius. Sinai: Syr. 10. in words. Malabar. lomans. 920 — ca*.2k = 826 1 Cor. 870 — \=*J«-. = 949 2 Cor. 590 — .^co = 653 Gal. 293 — ca!k.i = 275 Eph. 312 — — Phil. 208 235 238 in words. Col. 208 — — 1 Thess. 193 417 V»& = 417 2 Thess. 106 118 ~MLU5 = 118 Heb. 703 — A. A = 830 1 Tim. 230 — OAT = 226 2 Tim. 172 114 ¦ tin = 112 Tit. 97 116 CUJ3 = 116 Phil. 38 53 \X = 37 Imperfect as the table is, it shews that the Sinai and Malabar lists come from a common original ; in no other way can we explain the extraordinary agreements under 1 Thess., 2 Thess. and Titus. Moreover this common original appears to be the Euthalian reckoning ; for if we replace, in the count for 2 Tim. - by J*. we have the number 172 of the Euthalian list ; and if we suppose that in the reckoning for Titus, the sign ? has been translated by st we have easily the reckoning 96 for the arche type. The last four books would then stand Euthalius. Syriac renderings 1 Tim. 230 226 2 Tim. 172 172 Tit. 97 96 Phil. 38 37 which is sufficiently exact, especially in view of the slight variations shewn in Greek MSS. of Euthalius. The other APPENDIX. 25 confusions are more obscure ; in the original Syriac text in which the number for Philemon va became .^ as in the Sinai count, we see how easily the transcriptional errors could arise. The conclusion in favour of a Euthalian original is con firmed by noticing that the whole of the subscriptions in the Malabar text and not merely the numbers, are Euthalian. A single instance will suffice for a specimen : the colophon to Romans is . cniLSk r«'_M_^sAv&o .jjCSO rt'Jb.To r£ia&v& We will now give some reconstructions of an old Testament Stichometry, based chiefly on the three Syriac MSS. B.M. Add. 12,154. 12,178, and 7183. The Stichometries contained in these three MSS. are evidently derived from a common original, though there are some individual variations in the order of the books. The first three columns are in remarkable agreement, and the errors are corrected without much difficulty; but the score in the Sinai MS. diverges widely from the B.M. MSS.: yet it ought to be correct, if we may judge from the fact that the total for the Law comes out right upon addition of the separate members. As far as I can see at present the two systems are independent : neither of them shews any agreement with the Catalogue of Nieephorus. B. M. Add. 12,154. B. M. 12,178. B. M. 7183 Genesis reilco.i = 4631 4631 4631 Exodus Qft_l^= 3560 3560 3560 Leviticus GQ_2ftO\3 = 2445 2445 2445 Numbers CU^ = 3016 3016 3560 Deuteronomy .\^s-> = 2783 2783 2783 Total for the Law CQi-CU = 16925 [16435] [16979] actually [16435] Joshua IQnnra = 2167 2167 2167 Judges V»- 2079 2079 2079 Job is^_i= = 2553 2554 2553 Samuel oXioo = 5236 .anion = 5202 5206 David .^£¦"1 = 4083 4830 4803 Kings i^ajE.on = 5323 5323 5323 Proverbs cAirs> = 1236 •30O_*.r<'= 1762 1863 Coheleth V^.00 = 627 VA- = 927 Deest 627 Shirat Shirin a^.i = 296 296 Isaiah rda.i = 4081 4801 4801 Sinai. Syr. 10. toOS 4516 3378 2684 3481 2982 17041 1953 2088 >-fl ns H 3 APPENDIX. 27 o o tci in i-l (M -I" •* CO co CO o 0C CO CM CM >o r~ CO o CO ¦* CO OO T— t cm Ol IO CM CM CO ¦* ^f CM r—C CM i-H Q Q O IO —IT* CO Od CO O O oq cm || co cm co o o eo oo " CM CO Od io o .- ; „; ; M* rMrt-rilSCDH fifififi m 50 or; CO -t in i— i W M 3 I— » -* -* CO CM + o CM o CD CM CM T— 1 CD CM o CM CO •cf CO CO CM CM CO lO O CM CM j | | 1 CO •* "# CM o" o1— < CM CM CD r— I Illi li II II II II II II II II II II II II II 1 K I in 3 r c 3 • r9 j -l-l-l-l Y 1 c 9 J3 A c o mCD ao P5 'cil » -a P fi! g ¦a a CD cu i -a p £ o 73 GO es J3 O a o flCD 'ct02 j .a1 uth sther asan obit ebre Yamin H >-? W a H H e < !"? M H 02 H fi QLambriDgt -. PRINTED BY U. J. CLAY, M.A. AND SONS, AT THE ONIVRERITY PRESS. YALE UNIVbHSTY LIBRARY 3 9002 08844 5128