TRACES OF SYRIAC ORIGIN OF THE OLD-LATIN DIATESSARON *)
The origia of the Old-Latin Diatessaron is still a matter of dispute. Until quite recently the assumption that the original Diatessaron was a Greek composition and that the Old-Syriac Harmony on one side, and the (Old-)Latin version on the other, were more or less free translations, was looked upon as axiomatic. It is true that nearly half a century ago Zahn argued in favour of a Syriac original translated about a'. d. 500 from Syriac into Latin but he found no approval, and since it was shown first by himself and then by Vogels, that behind the Vulgate form of the Codex Fuldensis lies an Old-Latin Harmony of which only scanty traces were preserved, and since Vogels proved that the Harmony was the first attempt to clothe the Gospel in Latin dress, a Syriac original seemed to be excluded.
In 1923 however the present writer published, in a preliminary study 8), the results of an examination of the text of a mediaeval Dutch Harmony preserved in a Liège Ms. as a Flemish "Life of Jesus". The text of this Harmony seemed to show undoubtedly, that it was translated from a pre-vulgate text, prior not only to Victor's Codex Fuldensis but also to the OldLatin Gospels, all of which more or less betrayed influence of the Old-Latin Harmony. And the Old-Latin Harmony itself contained a number of Syro-Latin readings and of Syriasms, which seemed to prove that it was translated not from a Greek original,
') The present study is a somewhat enlarged form of a paper read in the Amsterdam Academy on April nth, 1927.
*) Forschungen, Tl. 1, Tatia*?s Diatessaron, Erlangen, 1881, S. 238 310 ff.
3) A primtive Text of the Diatessaron. by Dr. D. Plooij with an introduction by Dr. J. Rendel Harris, Leyden, 1923.
tot



2
but from the Syriac. This thesis was strengthened in another study published in 1925 l), in which there was adduced a good number of new instances all pointing in the sarae direction. The Liège Text affords frequent close parallels to unique readings of the Old-Syriac Diatessaron, in Aphrahat, Ephrem and the OldSyriac Gospels, parallels for which no Greek eyidence is extant, and which by themselves would suggest some direct relation between the Syriac and the Old-Latin tradition. Besides these parallels however there are cases found first in the Liège Text and afterwards discovered also in the Old-Latin Gospels in which the wording of the Old-Latin (resp. mediaïval Dutch) seemed to be explicable only as a literal translation of a Syriac idiom. I. cannot reproduce here the cases pointed out in the studies quoted above, but I may be allowed to mention one very interesting and convincing case discovered by Rev. Phillips in the Old-Latin Gospel tradition:
In John xii. 13 the Arabic Diatessaron translates rü /3xtx t&v (poivlauv by 'heart (or pith) of the palm'. This 'heart' is the word K&ccn\ which the Pesitta uses in Lev. xxiii. 40 in rendering the Hebrew kappoth, for which the Targurn uses the reduplicated form lulab. The same use of the word Ib is found in the Aethiopic Book of Jubilees, in the Apocryphal Story of Jeremiah in Arabic and in a few cases more, for which cf. the Bulletin of the Bezan Club for Nov. '26 and April '27. It is clearly a genuine Semitic, if not originally Syriac, use of the word. Now Rev. Phillips has drawn attention to the fact that whilst the Vulgate renders the word (iafa by ramos, the OldLatin codd. a b e ff2 aur render it as f/ores, and the cod. Usserianus (r) as medullas palmarum. This means that the OldLatin translators had before them the Semitic and not the Greek word, and seems convincing evidence of a Syriac origin of the Old-Latin Harmony *). This is only one instance clearly illustrating the unexpected confirmation which the thesis has found in remote
. ') A further study if the Liège Diatessaron, Leyden, 1925. \?>}
2) For further references cf. the notes of Rev. Phillips, Dr. Rendel Harris, Dr. van Wijngaarden and the present writer in the Bulletin of the Bezan Club, xa.YHiy. suboitjfri
102



3
quarters, and whatever the ultimate result of the investigations may be, the new facts fully account for the interest which they have aroused and for the vivid discussion which has ensued.
Dr. Dibelius1) has expressed some regret that the attention of scholars has been directed mainly to the problem of the origin, whether Greek or Syriac, of the Old-Latin Diatessaron and suggested that it would be preferable to try a reconstruction of the Old-Latin text on the base of the Liège Harmony. But it cannot be emphasised too strongly that for this purpose our evidence, even since the discovery of the Dutch text, is wholly insufficiënt. What we possess of evidence in Latin is limited to very scanty remnants and to a number of vocables which the Dutch translater incorporated in his work (turbeeren = turbare etc). A retranslation of the Dutch text into Latin, even with the help of survivals of the Old-Latin in the Old-Latin and Vulgate Gospels, in some Capitularia and in the Comments of Zachary, would pro vide us only with an artificial text, all the more misleading, because it would make us believe that we possessed really trustworthy evidence. In many single cases we are certainly able to recover the Old-Latin wording, but we are far from being able to reconstruct anything like a complete text. The present evidence however, until a real Old-Latin copy comes to light, can be used mainly as a means of reconstructing the history of the Diatessaron in the West.
A fortiori we must confess that a reconstruction of the Greek' Diatessaron is equally impossible. It was a plan of the late Prof. Preuschen, which his death has frustrated. But even when the very existence of a Greek original was'less open to serious doubt, the attempt could only have resulted in an entirely artificial product. It seems useful to remind ourselves that, whilst of the Syriac Diatessaron as well as of the Latin Harmony complete texts have been transmitted, though in more or less revised form, and also bothof the Old-Syriacandofthe Old-Latin Harmony fragments have been preserved, not a single line ofa Greek Diatessaron has been transmitted to us2). The ar-
>) Theol. LU. Ztg., 1927, Nr. 5 (for March 5).
2) Cf. Zahn, Forsch., I. S. 26 ff., and Rendel Marris, The Diatessaron of Talian, London, 1890, p. 16 ff.
103



4
guments for the existence of a Greek original are derived mainly from its Greek name and from the fact that late Syriac tradition speaks of Tatian the Greek as its author. But $i*Te<rirsipuv is a musical term in the sa me category as harmony, concordia and accord, and it is inconceivable that this term Uke other Greek words of the same kind, vufitpuvla, xid&px etc. could not be simply transliterated into Syriac. Moreover the term, according to Eusebius' letter to Carpianus, was not an invention of Tatian at all: it was Ammonius' Synopsis which bore this name, and a Synopsis - of Ammonius or of somebody else - must have preceded the Harmony. If Eusebius had really known a Greek Diatessaron, he, would have mentioned it in his letter to Carpianus nor would he have spoken of Tatian's work as vaguely as he actually does in the wellknown passage in H. E. iv. 29, 6. (avv&pei&v riva xxi fvi/etyoybv oök oli' Stu? rün EÖayysX'iuv trimdeie rb SictTeirvótpuv touto Tpotruvóficurev). If Eusebius had seen a Greek copy of the Tatianic Harmony he could scarcely have been ignorant about the method in which the Gospels had been interwoven.
More important, and certainly more interesting is the fact that the prologue of the Arabic Diatessaron in the Borgian Ms. speaks of the Harmony as a work of Tatian, 'the Greek'. Under this name and qualification the Harmony certainly was not known from the beginning. Neither the Doctrine of Addai nor Aphrahat nor Ephrem ever mention the name of the author 'of the 'Gospel' used by them nor do they call it by the name 'Diatessaron'. It is simply: the Gospel. It is only since the work was condemned, that its author was named and at the same time stigmatized as a heretic. Burkitt has drawn attention to the remarkable fact that "the Syfkc-speaking Church preserved no tradition about 'Tatian'; they only identified the author of their Diatessaron with the Tatianos about whom Eusebius had written. On the other hand, they had no-doubt as to who brought the Gospel to Edessa. It was Addai"1). Burkitt's suggestion as to the identity of the two names is probably right. But at all events, it is clear that the qualification of Tatian as a 'Greek' can be explained only as a scholarly remark dating from a
') Journal of Theologieal Studies, 1924, January, p. 130.
104



5
time when both his Apostleship and his excommtinication had been forgotten. He was known only as the author of the "Address to the Greeks" and as the skilful composer of a Harmony which had lost both its apostolic aureole and its heretic doom. Accordingly the note that it was Tatian 'the Greek' who wrote the Diatessaron, has no bearing at all on the language in which it was originally written. The only really important argument in favour of a Greek original, is the fact that in a distinct group of Greek Mss. Tatianic readings occur. On this question more preSentry *).
However, the thesis that, somehow, a Greek Diatessaron is at the back of both the Latin and the Syriac tradition has been the common opinion and it is so still with the exception of Lietzmann, (who not only accepted the thesis of a Syriac original but brought independent arguments in favour of it), Dr. Rendel Harris, Rev. Phillips and a few others, so that the theory of a Syriac original will need very sound credentials before being taken into consideration at all. Not only would it change fundamentally our views of the history of the text of the Gospels; it would have also unexpected consequences with regard to the history of the Church and of dogma both in East and West; and we should be compelled to concede to Tatian an influence far greater than we have been accustomed to admit.
It is, of course, impossible to give on the present occasion a complete review of the recent discussion which the question hag raised. But I may be allowed to quote Baumstark whose opinion as a Syriac scholar may stand for many others. In a review in Oriens Chrisiianus Which as far as I know has not yet appeared, but of which he has kindly sent me an advancè print, Baumstark says (S. 195): "Für eine sprachlich Syrische Grundlage jenes Diatessaron's beweisen aber derartige sachliche "Syriasmen" (Baumstark means syro-latin readings), so gross ihre Zahl auch sein möchte, natürlich nicht in das Mindeste,
') Lately the Greek origin has been defended by Pott in his posthumous Introduction to Preuschen's Translation of Tatian's Diatessaron, Heidelberg, 1926, S. 25 if. I do not think that his argument will be found convincing.
i°5



6
da sie sich auch unter der VoraussetfcBBggleichmasziger Abhangigkeit einer lateinischen und einer syrischen Version von einem griechischen Original vollauf erklaren. Hier waren nur unbedingt sichere sprachliche "Syriasmen" beweisend, Wendungen des mittèiniederlandischen Wortlautes, die nicht anders als durch eine latëinische Vermittelung bestimmter spezifisch syrischen Wendungen oder Ausdrücke sich verstandlich machen lassen. Ich verkehne nun keineswegs dasz P. manches beigebracht hat, was aüch in diesem Sinne auf der ersten Bliek völlig schlagend zu sein scheint. Es ist mir jedoch, nicht zuletzt von liturgie^ geschichdichem Gebiete her, allzusehr die Tatsache vertraut, wie leicht Endergebnisse vollkomnlen unabhangiger Entwick* lungslinien sich derartig zu berühren vermogen, dasz man, wenn man jene Entwicklungslinien nicht kannte, mit der gröszten Bestimmtheit ein in der einen oder andern Richtung verl&ufendes direktes Abhangigkeitsverhaltnis zwischen den Endpunkten annehmen würde. Das mahnt methodisch zu allerattsi serster Vorsicht".
I need not say that I fully agree with Baumstark's warning with regard to the necessity for the utmost caution. As a matter of fact I myself not only began my study of the Dutch text with the conviction that a Greek original1 was at the back ofthetradidon, but, when the facts seemed to point in another direction, I feit very sceptical about what those facts seemed to say. The theory of Zahn to which we have already referred, seemed not only in itself rather incredible, but as he had in mind a Latin translation made from the Syriac in the fourth or fifth century, his theory was unable to solve the problem that the Old-Latin Diatessaron has influenced the text of the Old-latin Gospels from the very beginning, and accordingly must be at least two centuries older than his theory would allow. Accordingly an early Greek text seemed to be unavoidable!
But the facts would not yield to the theory, so the theory had to yield to the facts. We must however be perfecdy clear as to what the theory of a Syriac origin of the Old-Latin Harmony involves and what it does not involve. It does not involve that a Greek Harmony never existed. As a matter of fact there are some traces of the existence of an early Harmony before Tatian.
106



7
I do not believe that in our Greek textual tradition much has been preserved of such an archaic Harmony. Whatever harmonistic traces occur in our Greek texts, they are — all of them it seems — Tatianic in origin. But both in the text of Justin and in that of Marcion, are harmonistic readings which neéd to be explained. They suggest the existence of a very early Harmony or a Synopsis probably for liturgical use, which may have been the prototype of Tatian's Harmony. But these are for the present mere speculations without any real basis of actual facts, or rather, the facts are insufficiënt for really useful debate.
On the other hand if the thesis of a Syriac original of the Old-Latia Diatessaron is accepted, it does mean that neither the Latin nor the Syriac tradition requires the assumption of a Greek Diatessaron for its explanation; that accordingly the origins of the Latin Church were much more closely connected with the Syriac-speaking Christians than with the Christians of Greek tongue; and that, if the existence of a Greek Harmony is to be proved, the proof mustbe found elsewhere than in the Syriac or the Latin Diatessaron tradition.
What is necessary first is a solid basis of textual facts, clearly stated and satisfactorily explained. Beyond doubt large unexplored fields lie open for the labourer and a good harjvest is waiting for him who enters into the work. For the present I shall confine myself to pointing out two or three remarkable facts illustrating the relation between the Old-Latin and the Syriac with respect to the Greek. It has been said that selected cases do not prove the general thesis. Well, if so, we shall have to wait for a good while before anything can be said at all, for a complete collation of all texts, or even a complete collation of the Liège Harmony alöne will take years. And even then we shall have nothing but fragments, unless in the meantime a complete Old-Latin text comes to light. At present our work is a kind of excavating; we are digging out fragments, outlines of foundations, traces of early strata. I wonder how Vogels, who is quite familiar with our kind of work, can speak of the Dutch text in this somewhat depreciating way: "Freilich, die Suche nach Tatianlesarten im Bergsmatext gleicht einer Jagd auf Eichhörnchen, 107



8
dié inan bald einzeln bald nestérweise in ihren Verstecken aufstöbern rhuss. Sechzehnender gibt es hier nicht zu erlegen"l). I do not know what in this hunting-field Vogels would call "Sechzehnender". We would expect a remark of this kind from the average theologian, but not from a textual critic as my Bonn colleague Vogels undoubtedly is. Vogels will have found from his ownstudies ön the Pafcrtine Gospels, that it is just in this «ort of forest that one may be out tp get snapshots of squirreis and unexpectedly start some big game. As a matter of fact, the apocryphal additiohs and alterationslnTatkn's work amounted probably to very little. When we find in the Pepysian Harmony the Light at the Baptism,' this is probably one of the most remarkable apocryphal additions in the Tatianic texts. Yet its presence in the Pepysian text is important not so much for its intrinsic value, as for the fact that it represents a reading quite familiar in the East, but extremely rare in the West. Sticking to the simile of excavation I should say: a fragment of a dated inscription, or a brick hearing the name of the builder;is worth quite a corner of the excavated building itself. When in a late Version or in some Vulgate Ms. a reading turns up bearing the clear mark of Eastern 'origin, its value is not determined by its more or less startling contents, but by its textual history.
It has seemed worth while to put these things clearly, becaüse it is nécessary to realize exacdy wbat we are after. Now with regard to the genera! objections of Baumstark. He is of course quite right in suggesting that Syro-Latin readings would be easily explained by the assumption of a common Greek origin, on one side for the Latin, on the other fór the Syriac. This has been the explanation which everybody, myself ktcluded, accepted as axiomatic. But it will not explain the long list of characteristic, often very slight variaots, which occur only in the Syriac and in the Latin tradition without any tracé of them being extant in the Greek also. It is too easy to declare this as merely accidental. We have to bear in mind that explanation by accident is no explanation at all, but merely a testimonium paupertatis. Besides these readings there is another series, equally Syro-
') Theol. Revue, 1926, Nr. 11, Kol. 494^ 108



9
Latin, which are found also in the Greek. But their Greek attestation occurs in a quite distinct group of Mss.: first of all Codex Bezae, then the Ferrar-group, in some passages the Washington Codex, the Koridethi Gospels and a few minuscles of Von Soden's Apparatus. For those that are acquainted with the characteristics of these Mss. it is clear that theiü peculiar readings can be used only with great caution as witnesses for the Greek text. Codex Bezae has already been suspected of syriacizing by Chase, and conyicted of latinizing by Rende! Hwris; the Washington Codex is so much under Latu*influence that Vogel» has called it in some parts a twinbrother of the OldLatin Palatiaus, the Ferrargroup has been shown by Rendel IJarris to be of Syriac ancestry; the Koridethi text has been located by Streeter in Caesarea. There is reason therefore for extreme caution before using the Syro-Latin readings of tfiese witnesses as evidence for a Greek text. It is safe, I think, to say that the Syro-Latin element in these texts, so far from giviag an explanation, itself requires one; we may be sure too that the Syro-Latin element in these Mss. is not in every case of the same origin. Even the Greek Sinaiticus contains a good number of readings of this-kind which hitherto have not been satisfactorily accounted for.
All these facts then can scarcely bear the assumption of a Greek Diatessaron. This is certainly the case with exclusively Syriac readings like that in John i. 8 where only L goes with the Old-Syriac. The Greek text is:
a/U.' ïvx [icipTiipijtry rept rov (purós
The Vulgate:
sed ut. Antimonium perhiberet de lumine
L. (p. 2519) reads exactly as the Syriac: mar hi was getuge van din lichte, sed erat testis de lümine, Sy^: rtficnco redrC K'ienoj.i oen (cf. A Further Study, p. 52 f.)
Another instance of the same kind in John iv. 39 where L. (p. us18) reads:
om dis wyfs ghetugnesse die seide, exactly the reading of Sycur:
109



because of the testimony of that woman who was saying, instead of the Greek
Sid rbv Aóyov rij? yvva.iv.hq pttprvpoOays (cf. A Further Study, p. 58).
I quote these instances because these readings are not simply accidental variants but stylistic alterations. They remind us of Eusebius' saying about Tatian's redaction of the Pauline epistles based, by the way, only on hearsay information, (peurt*, as is also in all probability his knowledge about the Diatessaron. In cases like these to speak of accident would be on the same line as speaking of the Homeric Epos as a merely accidental arrangement of letters and lines! If direct dependence of the Latin on the Syriac is rejected, a better explanation must be given instead.
Baumstark's second objection deserves close attention. Comparative study of religions has taught us to be very cautious with regard to parallels, and their derivatión from some supposed common origin. But the case is different here. Baumstark speaks of lines of evolution which we do not know. Here however we are in the position of being able, at least in a considerable number of instances, to foUow all the stages of the evolution. I would refer to just one example, that of the paraphrastic formula (which I have discussed in A Further Study P. 49) «took and led" instead of the simplex "/ed" »). Not once or twice, but in all the nine or ten independent cases where the formula occurs we find it not only in L. and Ephrem, but also in the Old-Syriac Gospels and the Arabic Diatessaron representing the Syriac tradition, and in ZaChary and the Pepysian Harmony as representatives of the Latin branch. In some cases the Old-Latin Gospels also are represented. Accordingly we cannot speak of the ends only of different lines as known to us: we can follow the lines all along, right up to the very point of bifurcation.
n HisU Eccl. IV, 29.6: t<* Si ImuniHm <paalv roerat rnat o&tJv
2) I may correct my statement 1.1. about Lk. iv, 29, where I thought the formula did not occur in the Syriac as in L. 9710- As a matter of fact it occurs there also in Ephrem, Comm., p. 129.
no



II
Baumstark speaks, in connection with the arguments brought forward in favour of a Syriac origin of thé Latin Diatessaron, of "manches das auf den ersten Bliek schlagend zu sein^scheint". We can only hope that Syriac scholars rhay be willing to give their attention to the problem, and determine the substantial value of these "striking first impressions" for it is mainly on the linguistic evidence that the decision in this case depends.
It is, however, not only on the part of Syriac scholars that a friendly warning has been extended to the present writer but the Germanists also have admonished him to caution. Prof. Frings of Bonn has written an exceedingly important review*) of my first study (really an independent study of the Germanist problems of the Dutch Harmony and its allies). It is satisfying to find that his conclusions with regard to the priority and independent position of the Liège Text, as compared with its allies S, H, etc, square entirely with the results obtained by my researches. His acute remarks lead him however to the conciüsion that a new edition' of the Liège Text is necessary and that: "alle Erörterungen der Theologen unnütz seien bevor diese grosse germanistische Vorarbeit geleistet ist". This was really somewhat disappointing for the 'Theologen" who had believed that their study of the Liège Text and its family had contributed a little to elucidate a fascinating problem. As a matter of fact the mediaeval Dutch is not so difficult as to be forbidden ground for any but a specialist, and in doubtful cases the help of my colleagues De Vooys and Van Ginneken has always been at my disposal. Moreover the difficulties of the Dutch coincide very seldom with those difficulties of textual criticism with which we are concerned.
The critics I have quoted are by no means the only ones whose opinion deserves attention, but this may suffice for denoting the widespread feeling of scepticism with regard to the thesis of a Syriac origin of the Latin Diatessaron, and may justify the attempt to put the question as clearly as possible in a few typical instances, hoping that both Syriac scholars and Germanists may
') In: Litcraturblatt für germamsche und romanische Philologie,_l^i6, Nr. 5—6, kol. 150—155.
Itt '



12
bè willing to put thèm to the test. It does not matter which theory will prevail, the only thing is to find the theory that is right, I only hope that the objection may not be repeated that the instances are merely selected cases. We have no other choice: and it Really does not matter. The main thing is not to colkct all instances, though the accumulative force of such evidence is certainly strong, but to explain the origin and to find the history of the texts.
Our first example will be the word "convent" in the Liège text on p. 4911 used in Lk. xii. 32. The Vulgate
noliie timere pusillus grex is the exact rendering of the Greek
aiij Qofioü, rb /iMpbv Tolftviov with the exception only of the plural nolite timere (which is read also by a b c e ff i and by Clem. Al. w (pofaiate). The ace. pusillum gregem, read by d e i, is probably due to the influence of the Greek ending -ov. I have discussed this very curious rendering of the Dutch in A Further Study p. 291, 42 f. As my explanation, however, according to Vogels1), only shows "was sich mit einiger Phantasie alles hinter einem harmlosen Worte vermüten lasst", it seems only fair to give his explanation first, in order to see whether it will explain the facts. Vogels is of opinion that the use of the word 'convent' is merely one of the "Freiheiten des Fkmen", "der in seiner Heimat genug kleine und grosse Konvente fand um mit solchem Ausdruck auf Verstandniss rechnen zu können". The explanation is easy, the question, however, is whether it is right. The burden of "Freiheiten" which has been laid to the Dutch man's charge, isalready so heavy that it seems unchristian to increase it unnecessarily, especially as in many cases it has been shown that the charge was not justified. Vogels evidently is of opinion that the mediaeval translater in this rendering of the Latin grex, was thinking of the monasteries of his days to which the consolatiqn would be especially applicable. Well, we may be allowed to put the question: Is it probable that a mediseval Dutch translater of circa
<) In his review Theol. Revue, 1926, Nr. 11, Kol. 404.
112



13
a. d. 1200, i. e. at the time of the beginnings of preaching to laymen and of the translation of the Psalter and the Gospels into the vernacular tongue — we think of course of Lambert le Bègue and the Flemish translation of the Psalms he mentions1) — that,at this time in a translation evidently intended for the benefit of lay people, a translater of Lk. xii. 32 would be thinking so exclusively of monasteries and of their members as inheritors of the Kingdom of God, that he could substitute so naively the word convent for grex ? I hardly think so. And if he did, could he expect to be readily understood? Evidently not: for the scribes of S and H have altered this convent either after the Latin, which may have been at their disposal, or merely from Flemish purism, into menigte and sameninghe'. This question deserves special attention. Though certainly the Dutch texts originate in one early Harmony of about a. d. 1200 (see the diagram in Frings' study, Lit. BI., Kol. 152), some cases of divergence seem to suggest an early revision after the Old-Latin copy used by the first translater. An Amsterdam Ms., however, evidently revising after
') I may quote from his Apology preserved in the Glasgow documents the following passage, which is also important for the history of the Dutch Harmony (printed for instance by A. Fayen, in: Compte Rendu des séances de la Commission Royale d'Histoire, Rome 68, 1899, p. 352 f.): "Unde et ego bonis eorum studiis cooperans virginibus vitam et passionem beate virginis et Christi martyris Agnetis, omnibus vero generaliter Actus Apostolorum ritmicis concrepantes modulis, ad linguam sibi notiorem a latina transfuderam, multis loco congruo insertis exhortationibus, ut videlicet haberent quo diebus festis, mundo in rebus pessimis exultante, a venenato ipsius melle sese revocare potuissent.
Et hoe est, quod preter scripti sui accusationes queritur iste, me scripturas sacras indignis aperuisse, non remorans quod predixit Dominus regnum Dei dandum genti facienti fructus eius, neque notans quod quem portabat nee manducabat puer panem discipulis suis tradidit ut illum turbe apponerent. Est preterea apud eos liber psalmorum cum omnibus glosulis suis et auctorita tibus eas roboran t ibus, in vulgarem linguam a quodam magistro flandrersi translatus. Quare de eo non queritur? Propterea forsitan quia nemo propheta accept us est in patria sua. Ille vero magister de patria eius non mit".
It seems to me that this important testimony deserves more attention than it has hitherto received.
113



14
the Vulgate, corrects convent into herte, grex. So it seems e*cluded that the Dutch translater sponte sua would render a Latin grex by a 'Dutch' convent, and we shall be obliged to assume that he found in his Old-Latin text conventus, and, as in a great number of similar cases, simply transliterated it int© a 'Dutch' convent.
We are then, however, confronted with the task of explaining how and when this Latin conventus came into the Latin text as a rendering of the Greek roifiviov. Conventus in mediseval Latin may mean as in mediaeval Dutch: a monastery, and also the brotherhood living in it. Du Cange (Glossarium, vol. ii, 184a, col. 582 s. v. conventus quotes a very remarkable passage from Ordericus Vitalis, lib. 5, p. 542: "Ut greges duorum coenobiorum permitterent adunari, Deique ad laudem sub uno Abbate et sub una regula unum conventum effici". This passage shows that in monasdc language the words grex and conventus were sometimes synonymous. But even so it is inconceivable that the scribe should change the grex of his original. There is however another consideration which makes it very unlikely that a twelfth or thirteenth century scribe (this is, as we have already remarked, the time of the revival of Harmony use and transcription) can be responsible for the alteration: all the Harmony texts in Latin that we possess show a deliberate tendency to correct the text after the Vulgate. Ever since the sixth century Victor's Vulgate Harmony had existed, and it was not the only copy of a Vulgate text in his days. I think we may safery say that after the sixth century and certainly in the thirteenth century the reverse process — deliberate alteration of a Vulgate text — is impossible. Accordingly the word must be pre-vulgate and if so, we may say it must be early christian. The easy explanation of Vogels is unsatisfactory and we shall have to look round in early Christianity for facts to account for this variant.
I suggested in A Further Study p. 29 f. that conventus was the term for tife Christian community in which the Latin Diatessaron was read, and I think we may now go a little further into the matter. If the suggestion is right, we shall find traces of the use of the word as an equivalent of ecclesia in, let us say, Tatianic circles. On the other hand, if we really find the word
114



there, this will account for the fact that in the West, where early Latin Christianity has been influenced in considerable degree by the Tatianic encratism, the word convent came to be used in the sense of monastery,— the celibates there, like the b'ne qHama in Syria, being the real ecclesia1).
We have remarked that conventus, the equivalent in the Vulgate of the (rvvtvyuyvj of the LXX and of James ii. 2, has in Latin. a flavour of heresy or schism, like our 'conventicle'. In early Christian writings however vwayuiyiij occurs a few times as a synonym of ExxX^ffi*, without any reference either to Judaism or to heresy: so in James ii. 2 (who uses iyxXvpia. in v. 14), in the Pastor of Hermas, Mand. xi. 9, 13, 14 and once ortwice in Ignatius, Justin, Irenaeus and Theophilus, Ad Autolycum. Perhaps Zahn is right in suggesting for these cases Semitic influence (Einleitung, 3 Aufl., 1906 S. 67). At all events the use is extremely rare in Greek. In Latin we find synagoga used once, by Commodian, Jnstructiones xxiv. 11 as a synonym of ecclesia:
Quid in synagoga decurris saepe bifarius? Ut tibi misericors fiat quem denegas ultro? Exis inde foris, iterum tu fana requiris?
It may be noticed that Commodian probably is North African and this might account for his use of the word. But at all events, these are only isolated cases.
It is however entirely different in the East. Epiphanius says (Hoer. xxx. 18) that the Ebionites call their communities not êtacKvjslx, but trvvsvywyv), i. e. NnB03. In Mt. xviü. 17 Sy5"1 uses
rC'crvx.cv.Vik for the Greek IxxA^tria. This can only mean that K9\x.cu& was the official term for the Christian Church. It is scarcely admissible to regard (with Burkitt in Ev. Da-Meph., II, 275) the use of the word here "as denoting not the Holy Church but some less august assembly". It is a pity that Mt.
') On this subject I have made some remarks in the Ztsch. f. d. Neutest Wiss., 1923, S. 7. See further in an article on the Celtic ascetes by Louis Gougaud: Muiierurn Consortia, étude sur les syneisaktisme chez les ascètes celtiques, in Eriu, the Journal of the School of Irish learning, Dublin, p. 147—156, to which Prof. van Hamel drew my attention.
"5



xvi. 18 is not preserved in Sysin. But both verses are extant in the Palestinian Lectionary ed. Lewis and Gibson, which reads >èuuA not only in Mt. xviü. 17 but also in Mt. xvi. 18 where there cannot be a shadow of doubt as to the soleran meaning. That it has really been in common use is shown by the SyroPalestinian fragments edited by Land in his Anecdota Syriaca, Vol. IV, where it is not only used in Mt. xviü. 17 but in four other, not Biblical passages; for instance rV&utiA&a KL&coo is simply: "the people in the Church".
That the use of the word, however, is not confined to the Syropalestinian Church is shown.not only by the instance of Sy««», but also by the inscription of the Marcionite Church in Lebaba and by the Baptismal Creed of the Aethiopic Church.
The earliest inscription of a Christian Church we possess is that found by Le Bas and Waddington at Deir-Ali (Cf. Inscr. Grecques et Latines recueillies en Grèce et en Asie Mineure, Vol. Hl, 1870, p. 582 f. nr. 2558):
CTNAmrH MAPKinNlCTON KflM(j$s)
AEBABflN TOT K(vpio)T KAI C(urj?)P(o?) IH(<r«w) XPHCTOT
ÜPÓNOIA nATAOT nPECB(w^eü) || TOT AX ETOTC
Harnack who discusses this inscription, dating from a. d. 318/19 in his book on Marcion, 1 Aufl. S. 263 f., rightly observes that the designation of a Church as o-vvctyuyv is "doppelt auffallend bei den Judenfeindlichen Marcionitischen Christen". But it is explainable by the Syriac usage of rVcWx-cUA and the Marcionites must have been very influential in Syria, or Ephrem would not have been so intent on opposing them.
In the Aethiopic Statutes of the Apostles there is preserved a very archaic form of a Baptismal Creed which Von der Goltz bas done 'into Greek thus: x/oretfw. .. sis rijv iylav a-vvayayijv, fiiav xafloAixijj/ ixxAJjfftej/. I have argued 1. 1. that the addition ptav xafloAfx^y £'xxA^<rl«j/ must be secondary and that the original form only contained a profession of faith in the Holy Gathering. At the same time I remarked that the original language can scarcely have been Greek. The foregoing facts suggest a Syriac origin, which in
116



17
connection with other facts (cf. Bulletin of the Bezan Club, nr. IV, p. 18) is quite probablex). To conclude:
i° The word conventus in the Liège Diatessaron is not of mediseval but of early Christian origin;
2° its original as a name for the Christian Church is not in the Greek avvovyuryvi which is very rarely used in this sense and where it is so used in early writings is suspect of being under Syriac influence;
3° its Syriac equivalent is found widely spread in early Eastern Christianity as denoting the Holy Church, That it was used by Tatian may be gathered from the testimony of Sywn. The Greek <rw«y«yjj in the Marcionite Church inscription can be explained only as a translation from the Syriac.
') Later the word seems to have acquired the connotation of forbidden assemblies, cf. the Aethiopic Didascalia, ed. and translated by Th. P. Platt, London, 1834, p. 103:
"XI. That it is unlawful for Christians to enter the Assemblies of the
Heathens neither associate with the assemblies of the Gentiles.... It be-
hoves the faithful to separate themselves from the assemblies of the wicked", thus developing a meaning similar to that in orthodox Latin circles. In the Syriac Didascalia the combination as in the Aethiopic Creed is found in cap. ix (ed. Achelis u. Flemming p. 44, Syriac of Lagarde p. 36): „Hort dies also auch Ihr Laien, du auserwahlte Kirche Gottes, zwar auch das frühere Volk ist Kirche (rVc^Ili») genannt worden, ihr aber sei die katholische heilige und vollkommene Kirche, das königliche Priestertum, die heilige Versammlung (r<iz>.Yo rOO&). Cap. xiii (p. 70 of Achelis, p. 58 Lagarde): „Belehrung an das Volk dass es an der Versammlung der Kirche eifrig teilnehmen soll («WSl^.n
das Volk dass es bestandig in der Versammlung der Kirche sei .V>-1 netlij) sei und sich- nicht zuruckhalte, sondern Dass es fort-
Wahrend versammelt sei (^xXIa), dass niemand die Kirche (rCW.T^-) vermindere, indem er keine Versammlung besucht (%&1^)". The word is used also of Gentile and Jewish "congregations" (Ach. p. 71, Lag. p. 59: "Sie (the Gentiles) kommen bestandig zusammen (*aSl&OGa)
und ebenso auch die ohne Grund so genannten Juden feiern
einen Tag nach je sechs und versammlen sich in ihren Synagogen (^ QcnèKx-CUAra ^atia&oo)".
117



i8
It may be that we did not expect this resuit, but I do not see how it can be avoided. For a more detailed history of the word conventus in the West the new Du Cange will possibly furnish us with the necessary material. Perhaps we may find one link in Sermo xii of Caesarius Arelatensis quoted by Du Cange, Vol. II, s. v. conventus: quia vos ad audiendas lectiones divinas video ad conventum fratrum vel ad ecclesiam fideliter accurrere.
We proceed to another instance:
The Liège Diatessaron omits Mt. i, 25, the Greek of which runs thus:
kxI oók tyu/tbaxev oivtviv ïug ov êrsKtv vlóv, x&l ix&teosv to SvopM aóroO 'lyiroüv 1). The Vulgate is:
et non cognoscebat eam donec peperit filiutn suum primogenitum et vocavit nomen eius Jesum. The omission in L is scarcely accidental. The Strassburg Ms. (S) gives as usual a literal rendering of the Vulgate:
ende en becander niet tote si gebaer haren eersten gebornen sone ende hiet sinen name Jhesus. There has been extant however a different rendering in media? val Dutch as may be seen from Maerlantfé' Rymbybel which reads here (ed. David, II p. 392 1. 21. 185): Hi trouwedse na der wet sede Ende bleef met hare in suverhede. We should be inclined to see in this rendering merely a poetical paraphrase if we did not notice that the text of Ephrem and of the Curetonian Syriac reads exactly as Maerlant: et sumpsit eam (in uxorem) et in sanctitate habitavit cum ea We notfce also that' not only the second verse but also the first of Maerlant correspond exactly with the Syriac and represents the Tatianic view of the casusposition as we shall see presendy;
Maerlant has known the Dutch Diatessaron as I pointed out in A primitive Text, p. 31. Accordingly not finding any other source for Maerlant's rendering (Comestor has a similar para-
') I doubt very much whether it is right to omit vfUTÓroKOy as spurious. 118



»9
phrase but too different in wording: accipiens sponsam in uxorem cum virgine virgo permansit) we are justified, I think, to regard the lost verse of L as Maerlant's source. It is important to notice that this Tatianic rendering: et in sanctitate habitabat cum ea, extant only in the Old Syriac and in Maerlant, is wanting not only in Greek but in the whole of the rest of the Latin tradition, in spite of its obvious doctrinal value. It would be interesting to know whether more traces of the same rendering are extant in mediasval Dutch.
The first part however of Maerlant and the corresponding renderings in the Syriac deserve special attention. The Greek says:
In connection with the angel's message in vs. 20 this can only mean that Joseph who had married Mary, did not send her away but took her with him. In Maerlant however Mary is Joseph's bride and he marnes her after the angel has so instructed him. That is also the meaning of vs. 24 in Comestor: the angel has bidden Joseph 'ut acciperet eam in conjugem' and Joseph obeys the command: 'accipiens sponsam in uxorem cum virgine virgo permansit'. We notice the expressions acciperet in conjugem and accipiens in uxorem, and shall come back to them presently. Zachary of Besancon is of the same opinion: he explains his Vulgate text: noli timere accipere Mariam conjugem tuam, in his note (Migne, P. L. vol. 186, col. 71 C), said to be taken from Rabanus, thus: tak est hoe ac si dicat: Accipe tibi in conjugem Mariam sponsam tuam. And he explains vs. 24 (in col. 72 B): Accepit igitur eam ad nomen conjugis.
This is the Tatianic view. Ephrem explaining the Diatessaron rendering of Mt. i. 24f says (Moes. p. 25): "In sanctitate habitabat cum ea. Praepostere dicta sunt verba. Nam prius sumpsit eam et postea habitavit cum ea in sanctitate; sed ita jegitur: Habitavit cum ea in sanctitate et sumpsit eam." Ephrem evidently means to say that what was done first is said last. Joseph took (i.e. married) Mary first and in his further life lived purely with her. Accordingly the irctpéAecfisv of the Greek is understood as married. This agrees with the Old-Syriac and the Pesitta. In vs. 20 the Cureton Syriac (and 119



20
the Harclensis) read óa.isa, the Sinaitic and the Pesitta ^.aoa. In vs. 24 all Syriac versions (Harcl. included) read coim . Both these words mean simply to take but are used constaritly for to marry. We see how obvious and simple the process was in Syriac, but when we find the same exegesis in the West we shall have to ascribe this to Tatiariic influence, especially as the Vulgate dimittere in vs. 19 is evidently contrary to the Tatianic view. These views of his are widely responsible for textual variation and there is a great confusion in the texts between the words sponsa or uxor said of Mary, but Tatian himself is quite consistent: he says sponsa before, uxor after the message of the angel. Accordingly Sycur and Liège read in vs. 20 instead of yvvctïKct: desponsatam • in vs. 24 SyCttr reads instead of ri)V ywaTm cuiroü 'Mary', a reading which, conflated with the original text, is found in sah boh r* and in In Lk. ii. s instead of the conflate reading : rjj peimfotevfiivy aörtS yuvcuxi, Sy*c> reads only Mary his wife.
We found in Zachary and in Comestor the expression accipere in conjugem (u xo rem) and Maerlant makés itprobable that also Liège in Mt. i. 24 read originally: nam Marien te wive, instead of: nam Marien met hem.
Mindfui of Frings' warning against rash and groundless reascning I consulted our fellow-member Prof. Van Ginneken as an expert of mediaeval Dutch and asked him whether in mediaeval Dutch the expression: wif nemen or te wive nemen, is the only possible rendering of the Vulgate: uxorem ducere, nubere, etc. He replied: "that uxorem ducere should be translated by wif nemen, te wive nemen is conceivable; but that this should be the only form, I see no reason at all. It would be equally possible to translate: te (huis)vrouwe nemen, huwen, trouwen". I may add that the Strassburg Ms. sometim es uses bruden, and also the literal rendering wif te huus voeren, the heimführen of the German parallel texts.
Well, the Liège Text uses always, without any exception wif nemen or te wive nemen. For this exclusive use the Dutch idiom does not account. Let us compare the Syriac, and consult the table on the opposite page.
120



Mt. 5.32
Mt. 19. 9
Mt. 19.9
Mt. 19. 10
Mt. 22. 24
Mt. 22. 25
Mt. 22. 30 Mt. 24. 38 Mc. 6. 17 Mc. 6. 18 Mc. 10. tl Mc. 10. 12 Mc. 12. 19 Mc. 12. 20 Mc. 12. 21 Lk. 14. 20
Lk. 16. 18
Lk. 16. 1U Lk. 20. 28 Lk. 20. 29 Lk. 20. 30 Lk. 20.31 Lk. 20. 34 Lk. 20.35
GREEK
ya/xijfy)
ya.jj.dv
ya.p.V)0~ix,i
êTriyxjj.(3psvo-gi
yvif/M?
ya.fj.ovo~tv
yafiovvrsg
syd/j-V/trsv
£%£iv TVjV yvv.
yay.yo~y
ya,fj.Vjtry
\ó,(3y TVjV yvv.
sKa/3sv yuv.
ïAot/Jfj' aóryv
yvv. gyvifiix,
yx/JLÜv
yajxSiv
X&fiyj-TVjV yvv. Aa/3w> yvv. 'dXufiiv aöriijv kXafisv aÖTyv ya.[j.ov<Tiv yxfiovirtv
VULG.
duxerit duxerit duxerit nubere
I
ducat j I
uxore ducta
nubent nubentes duxerat habere uxorem duxerit nupserit accipiat-ux. accepit ux. accepit eam uxorem duxi
ducit
ducit
accipiat-ux. accepit ux. accepit Mam accepit Mam nubunt nubunt
OLD-LAT.
k + capi p. uxorem
\ ff: accipiat
\ d: nubat
\ e: accepta uxore
X d: nubens
a: nupsisset k: nu bet
\ a: nubit ( d: nubens
SYCUR
iox. Aflnl
om.
K'cScAjk' .rxseiLSi
| K&ivsr*'
«lü finaal
\ -
s
ri&èVuK' .naai etmaal
CÓ3M1
SYSIN
■3 cal ■zicni
om.
nfèSèxlK' .3.03=*)
.cracujrc'A . .KöeocS
rdiü tradia
cruaoal rc'cSèuK' Aflaïxj r^Twri'A Kiocn AQai
K'cSèuK' .non Aflas
rc&èure' .3. OU cinsflal cnAflai rOtl ^a-flol rCLxi ^*3.flai
j SYPEs
I AaX.
.nfla»
.3.021 Anx.
rdïl ^lAflas Aflas
.3.00 è>
Aflas "A rctocnèVv .3.021 rrffoèurx' Aflai 023.021
aau ja.au
re'èxèxiK' .. .3021 .3021 ct13.q2j 023021 Klxi (iaau
L
456
1399 «<?w* 1399
I3910 ^ i73i8 nam en wyf
215U om. — (S: brudende) 99is ghenomen 1 97IS nemene .. wyf
17319
1711* «V «y/ genomen
17319 17321
1738' nemen wyf
1751 selen. . wyve nemen
HARCL.
(conjungitur) -i\o.ivi -\o.iVi -^p.-ivi
CYlxlcWjJ&V»
vA r^oeoèS
\ -i^o-ui
Aflas AOU 02=021
Aflas Aflas
Aflas Aflas 023021 can om






21
The above Hst is highh/ instructive. It shows with complete clearness the continual agreement of the Dutch and the Syriac. The most usual word used in Syriac is .nflai, to take; once in gycur an(j twice in the Pesitta Aax., equally meaning to take, is used instead. In Mt. xxii. 24 the Old-Sjrriac, evidently wishing to render ÉTryxfifJpefoei by a different word uses ,cnc\jjre'A . . . KbcoèS, but the Pesitta keeps to the natural ■aail. That .aflai is really the vernacular word is clearly shown by the Harclensis, who — wanting to render the Greek with a painstaking exactness — uses-\c\.iW, conjungitur;but sometimes (Mk. x, 12; Lk. xiv. 20; Lk. xvi. 18) falls back on the native usage and writes .aoal. We may be sure that in those passages where the Greek Gospels use Aaufi&vew (Mk. xii. 20 sq.; Luk. xx. 28 sq.) they reproduce the original Aramaic.
The Liège Text has always the rendering accipere uxorem or in uxorem *). This cannot be accidental, as is shown clearly by the twin translations of L. which in several cases give different renderings instead. We might think, as Prof. Symons remarked, of a personal idiosyncrasis of the translator, who having once chosen the expression, sticks to it consistently, but this is excluded by the fact, first that he uses the expression where the Latin in agreement with the Greek has habere uxorem viz. in Mk. vi. 18. Here the Dutch has like the Syriac nemene wyf. And secondly by the fact that traces of the early Latin rendering have been preserved in the Old-Latin: in Mt. xxii. 24^ reads accipiat 1. ducat, in Mt. xxii. 25 e reads accepta uxore 1. uxore ducta. The most interesting case however is Mt. v. 32 in k. The second part of this verse is omitted in Mt. v. 32 by Codex Bezae, k a b. In Mt. xix. 9 by a considerable number of witnesses. 1 can explain the omission only as due to the belief that any second marriage was essentially
') Prof. Wensinck has remarked in the discussion that in Syriac there is also found the construction with ca. It is quite possible that this construction has left its tracé in the te wive nemen of L and in the in uxorem accipere of Zachary and Comestor, but as the construction is 'also Latin, no stress can be laid on this colncidence.
121



22
condemned as carnal lust, and so a case like that presupposed in the omitted clause must eo ipso be excluded. The scribe of k, however, has read the omitted clause in his Latin text for he writes thus:
praeter causam fornicationis facit uxorem capi mechari
I have restored the stichometry which, according to the dots in k, the scribe found in his original. It is clear that he wanted to write:
praeter causam fornicationem facit eam mechari.
But he found evidently in his original (or in the Latin text he was using at te same time) the addition
et qui dimissam uxorem capit adulterat.
He overlooks however a line and has #ritten uxorem capi when he sees that he is wrong and that the second part of the verse is to be omitted altogether. He leaves the uxorem but cancels capi. The mistake is instructive for two reasons: first because it shows that the Old-Latin (we may say now the Old-Latin Diatessaron) had here the expression uxorem capere which we are discussing; secondly because we find in the omitted clause the addition uxorem, after dimissam, as in L:
ende so wie dat wyf nemt die en ander ghelaten heeft.
I think the proof is conclusive. To speak of accident here is to give up scholarly research. We must leave a margin for accident even in scholarly work, but here it seems precluded. If the thesis of a Syriac origin of the Old-Latin Diatessaron is rejected, we are justified in asking for one that better accounts for the facts, but we cannot accept that from accident.
One instance more may be given. It will show that the problems are more complicated still, though the fundamental solution has been found.
In John vii. 35 the Jews are discussing the quesdon wither
122



«3
Jesus is going that they will not be able to find him. They suggest that he will possibly go "to the Siaa-Top* r<2* 'EaajJmü» in order to teach the "EAAifves". We are not now concemed with the question as to what is the meaning of the Fourth Gospel in this passage. It is the Latin translation of the Vulgate which puzzles us. The Vulgate reads:
numquid in dispersionetn gen Hum iturus est et docturus gentesl
Accordingly "EaAjjwj is rendered by gen/es. In John xü. 20 "EAAijyes is rendered as gentiles in Vg. and the Syrophoenician woman (Mk. vii. 26) is also called gentilis. The Liège text renders John vii. 35 by heidine and heidene volk. In John xii. 20 it has a curious reading which states that they were "vremde liede ... die nin behoerden tin yodschen volke", foreign people ... not belonging to the Jewish nation, This looks more like a rendering of alienigenae, aXKóCpvAoi and should certainly be compared with the Pesitta where Acts £ 28 is rendered as rïöco reL.i^cu rdUK1
cn&xaix. ïa> "a foreign man who is not a son of his race." We leave this for the present, only remarking that we cannot say that the rendering gentilis is such an obvious one. In both passages of the Gospel of John it seems that Jews are meant, and not Gentiles; they are evidently Greek Jews from the Diaspora, the so-called Hellenists. Even if pagan Greeks were* meant, the rendering gentilis is not a matter of course. In the Pauline Epistles even when "Jews and Greeks" are contrasted clearly as "Jews and Gentiles", the Vulgate, with only two or three exceptions (Gal. ii. 3; I Cor. x. 32; xii. 13) uses graeci and not gentiles. In this as in other respects, the Latin tradition of the Acts is decidedly on the side of the Gospels as distinct from the textual tradition of the Paulinae. In Acts the rendering is generally gentilis, and graeci is in the Vg. an exception.
Accordingly we are justified in putting the question : where and when was the contrast Jew-Greek feit as the contrast Jew-Pagan in such a degree that the word Greek became simply a synonym for Pagan?The answer cannot be doubtful: it began during, and persisted after the Maccabaean wars, when a strong reaction
"3



24
against the peaceful and afterwards forcible invasion and infiltration of Greek civilizadon arose in the religious and nationalist circles of the Jewish people. Hellenism menaced not only old national customs and traditions but also the religion of the chosen people, and since that time "EAAj^ and 'EAA^wayifc ave equivalents of 'Pagan' and 'Paganism' in Palestinian Aramaic and in Syriac generally. A convincing instance is furnished by the comparison of the Syriac translation of II Macc. (ed. Lagarde) with the Greek:
II Macc. vi. 8: eig Txg .. 'EXXvjviSxg róteig, sy.: K&l&lJSaA II Macc. vi. 9: fierc^aiveiv iri to, 'EAA^wxi, sy.: ^_OAAcni II Macc. xi. 24: t% tov wovrpbg sig ra. 'EAAjfwxA pLSTt&évsi, sy.,: TI Macc. iv, 10: Trpbg rbv 'EXtyvuubv %*peatTfjpct, sy.: rAiVT-J
II Macc. iv. 13: &xfiij tov 'EXtyvio-iJ.ov kxI irp6<7f3ouriq oiKKoCpvXitr/JLoO, sy.: rVèiCUJCUrt rCèxcu^x-O K&<X3hlMM re'èSojyJVu
III Macc. iii. 8: "EAAjputc, sy.: fOarëuw.
An eariy instance of this meaning of "EAA^v is found in Mk. vii. 26 where certainly the Syrophoenicissa is meant to be denoted as a pagan woman. Accordingly the Sinaitic Syriac reads neèxlsoÏK' (a scribal error for rCèusoirc) and the Pelitta reads rVcH^ijj. I think we may be sure also that in Lk. iv. 26 Wellhausen is right in suggesting K&usaifV instead of rxèusaïr^, X^pn, but not to denote 'Syrian' but 'pagan' *). This must also be the meaning in the next verse where the point is not that
') A tracé of this exegesis in the Liège Text p. 977: buttnslands, and in Zachary, col. 238b: exlerae gentis vidua.
124



«5
Naaman is a K^jsaire', a Syrian, but that he is r<L*-sair<
a Pagan.
The constant contrast in the Pauline Epistles of 'lovSxïo? and "EAAi^as: 'Jew and Pagan' is a witness to Paul's Aramaic education and mother-tongue.
It «ray be useful to append the complete comparative list for the use of the word in Greek, Latin and Syriac. I have added also the Gothic to which Prof. Symons drew my attention. It is important because it has the Vulgate gentilis in the Gospels. So it is either a witness for the existence of this reading in the Old-Latin texts by which the Gothic has been influenced, or it shows that the Gothic is under Vulgate influence. If Jerome is to be made responsible for the Vulgate gentilis in Gospels and Acts, where did the Gothic, which shows the same difference between Gospels and Paulinae as thé Vulgate, get its gentilis from?
The table, printed at the end of this paper, is very instructive: in one case only is "E./Xkv)ves rendered in the Pesitta by ndiLicu (Acts xiv. i). In one other case where this rendering is given, Acts xvii. 4, it is qualified by the addition trsffópevoi, which evidently excluded the translation 'pagan'. In all other cases netiicu is the rendering for 'EAA^woraf evidently understood as Greek-speaking people, cf. Acts ix. 29 Pes.: ^A<rV K&ioco» AuKli cv» ooca (-Uk-.T»!, where 'EAA^j/wriff is explained by Chrysostom as: roii? 'EAA^wari (pBsyyopévous. In all cases where the word "Eaaj^sj is really understood as meaning 'pagan' it is rendered by ré&iu, rCÜoïrV, gentiles, or rdsosoa»., gentes.
It is certainly curious that in all these instances as far as the Gospels and Acts are concerned, the Vulgate follows the Syriac very closely. With the Pauline Epistles the case is different. The contrast is so clearly marked that a different textual tradition must be assumed.
One thing however is very puzzling: we should expect to find in the Old-Latin tradition the rendering gentilis and in the Vulgate the more literal graecus. The case is just the reverse; and occurs in the Acts as well as in the Gospels. Even in
"5



26
the Western Text of Acts xx. 24 we find in d gig Luc the rendering graecis, D: "¥./Xkqaiv. This observation deserves, I think, close attention. I do not think it is inexplicable, though for the present we simply state the fact, only mentioning that the case is not without parallel: compare for instance the beautiful rendering gratia plena in the Vulgate preserved from the Syriac whilst the Old-Latin e gives the literal gratificata. It shows that, though the rnain solution has been found — and the overwhelming evidence does not seem to admit the possibility of any other conclusion — the tradition of the Latin Gospels, especially its earliest history, is so complicated, that it still requires a research, that will be both extensive and exact.
126



GREEK VULG. OLD-LAT. SYS™ SYCUR SYpes HARCL. GOTH
Mc. 7.26 'EAA^e gentilis graeca: da b cff* q rVèvVwirC hiat rCAva.1*, rCcWw haipno
Joh. 7.35 (SictiTT. t.) 'EAAi^cw (disp.) gentium graecorum: a b deP l r q retoaWil ré^.i\ rAsaÏK'.i Kl^iï rt^asóat. r£&*'M piudo
"EAA^j/aj gentes graecos: abff^lq (eas: <?) w». «w. relA'ü, relSvï*, piudos
Joh. 12. 20 "EAA^s? gentiles graeci: a b c deff* r s KikW hiat KÜQrV rtLsaSOA. JS3 nï'AÜ, piudo
Act. 6. 1 'EAAjjv/orfih/ graecorum graecorum: A ex graecis </ reiici» rtxici»
Act. 9.29 'EAAjwsras } graecis ) ^ graecos: gentibus: c Aurficu oocn ^.1,1 rtiieL. r-u. „ >••>•> . /*a.
7 ( gentibus Vgcodd c \ b r 5 «u^jtu aoro ^..vi r^Mcu Chrys. rovg iW^yiar] (pSeyyopévovs
i 'EAAj^isra? )
Act. 11.20 < "EAA^a? J2C $4 } graecos craecos: d relicu r&icu
( ^5* Icl arm. '
Act. 14. 1 'EKkyvav graecorum graecorum: d | rci.10... rtiicu
Act. 16 1 Tarpo? $s"EKXijyog patre gentili graeco: d gig cdicu rC^Ai
Act. 16.3 "EAA^y gentilis graecum: d gig rctoaW reiau.», Ephr.: gentilis
Act. 17.4 (os/3.) 'EAAj4>co> gentilibus \ Sraecorurn • ^ . « .. ^ ,
( de grecis: <? ^ reiiCV» rcxicu Ephr.: gentihum
Act. 17. 12 'EAA^>töwv gentilium ne'cVüicu
'EKhVjVuv: $5 I graecorum: d, grece: £phr_. e graeds
Act.18.4 "EU*** j ^ ^aeC0S j graecos: rf, ^, grecis ', ^ü,
Act. 18.17 "EAA^ om. graecis ^ ^ * p^ cf. Ephr., «/.
Act. 19.10 "EAAj^j/aj gentiles graeci: d gig reisoïrf KiSa.
Act. 19. 17 "EAA/ir/ gentilibus graecis: <f . ■• . ..
y ' ' 8 00 rtóoirC rC.2Hjj Ephr.: gentiles
Act. 20.21 "EAA^<n gentilibus graecis: d gig Luc. reïsaïrc' re's»!
( "EAAll(7/> f • j ■ t
Act 2°- 2M (only cod. Bezae) f graCC1S! d *** Luc' - Ephr.: gentilibus
Act. 21. 28 "EAA«v«s gentiles graecos: d gig .. , ..
Rom. 1. 14 "EAA^a-/ graecis
Rom. 1. ,6 "EAA^, graec0 ^..^
Rom. 2.9 "EAAj^o? graeco * • * .
Rom. 2. 10 "EAA^x graeco
Rom. 3. 9 ^"ff graecos
Rom. 10. 12 "EAAi^cs graeci graeci atque gentilis: Fmt rtiSairC retSii,*» krekis
iCor. 1.22 "EAA^j graeci graeci reSwïrV r<AÏM krekos
lCor' ''23(v.l. 5 gentibus graecis: Ambr.? retoaW r^ï« biudom
iCor. 1.24 "EAA^tr/ graecis gentibus: Aug. retoaïrC piudo
iCor. 10.32 "EAA^i7< gentibus graecis: Hil. pCtoaïrC rCSkijj
iCor. 12.13 "EAAj^sj gentiles graeci: D d e f1 g péaWtK' r<i2*ï*»
Gal. 2.3 "EAA^ gentilis rttoaW Klaa« kreks
Gal. 3.28 "Y.Kkyv graecus reisaiK" r<ai« kreks
Col. 3. 11 "EAA^> gentilis KtoiirC r<la.l*»