CONTENTS

Page

I On   IN SCRIPTURE AND TRADITION 1

NOTE A The details of early Greek Patristic Evidence 30

NOTE B The details of Latin Evidence 43

NOTE C Some details of Æthiopic Evidence 46

NOTE D Uncius and unigenitus among the Latins 48

NOTE E On   in the Nicene Creed 54

            

            

            

             

           

         

BASILIUS

ON THE WORDS

 

IN SCRIPTURE AND TRADITION

The purpose of this Dissertation is to investigate the true reading of the last verse in the Prologue to St John’s Gospel (i 18). The result, I think it will be found, is to shew that   should be accepted in place of the received reading   , alike on grounds of documentary evidence, of probabilities of transcription, and of intrinsic fitness. The reading of three primary Greek MSS. has been known only within the last half-century; so that naturally this verse has not shared with other disputed texts of high doctrinal interest either the advantages or the disadvantages of repeated controversial discussion; and thus it offers a rare opportunity for dispassionate study. The history of the phrase   in ear1y Greek theology, of which I have attempted to give a rude outline, has also an interest of its own.

The verse stands as follows in the better MSS.:

              

.

1

2

The Documentary Evidence for   consists of Manuscipts אBC*L 33 (א* omits the following

 ; א* and 33 prefix ).

Versions: the Vulgate (‘Peshito’) or Revised Syriac; the margin of the Harclean Syriac; the Memphitic; and one of the two Æthiopic editions (the Roman, reprinted in Walton’s Polyglott), in accordance with one of the two earlier British Museum MSS., a third of the MSS. yet examined- having both readings1. The article is prefixed in the Memphitic rendering. The Thebaic and the Gothic versions are not extant here.

   is found in

Manuscripts; ACcEFGHKMSUVXI  and all known cursives except 33.

Versions: the Old Latin (q has u.filius Dei); the Vulgate Latin; the Old Syriac; the text of the Harclean Syriac; the Jerusalem Syriac Lectionary; the Armenian; and Mr Pell Platt's Æthiopic edition, in accordance with many MSS.

The Patristic evidence, though remarkable on any possible view, admits of various interpretation on some points. The grounds for the chief conclusions here stated will be found in a note at the end: it must suffice here to mark the limits of doubtfulness as clearly as the circumstances permit.

The reading  , with or without , in direct quotations from St John or clear allusions to his text, is attested as follows. Two independent reports of VALENTINIAN doctrine furnished by Clement of Alexandria (Exc. ex Theodoto, p. 968 Pott.: a paraphrastic allusion a little later has  by a natural combination, see p. 32), and Irenæus (p. 40 Mass.: corrupted in the inferior MSS. of both Epiphanius, who supplies the Greek, and the old translation, which in this allusion is faithfully literal). IRENÆUS himself at least once (256), and I strongly suspect two other times (255, 189): in all three places the original Greek is lost. CLEMENT himself twice (695, 956: in the second place, where the language is paraphrastic,

I It is impossible to convey a true impression of the Æthiopic evidence in a few words. Some particulars will be found in Note C.

3.

Clement has  .  , as in a still looser paraphrase at p. 102 he has  …   ). ORIGEN at least three times (on John i 7 [the commentary On i 18 itself is lost], iv. P.89 Ru.; [on John i 19, p.102, the reading of two MSS. only is recorded and they vary suspiciously between  .   and

    ; in an indirect reference shortly afterwards  . stands without a substantive;] on John xiii 23, p.439; c. Cels. ii 71, p.440, certainly in two MSS., apparently in all except two closely allied MSS., from which De la Rue introduced ). Eusebius twice, once as an alternative not preferred by himself (De Eccl.Theol. p67,      ), and in one other exceptional but seemingly unsuspicious place, p. 174. EPIPHNIUS three or four times (Ancor. p. 8 [the clear statement here confessedly leaves no doubt as to the quotation at p. 7, hopelessly mangled in the printed text]; Panar. 612, 817). BASIL at least twice (De Sp. Sanct. 15, 17, pp. 12, 14 Garn., quotation and statement confirming each other; as the Benedictine editor notes, adding that earlier editions, unsupported by any of his six MSS., read ; the quotation with  at p.23, which has no note, may therefore be only an unwary reprint).. GREGORY OF NYSSA ten. times, always somewhat allusively, as is his usual manner in citing Scripture, (c. Eunom. ii p. 432 [469 Migne]; 447 [493]; 478 [540]; iii 506 [581]; vi 605 [729]; viii 633 [772]; ix 653 [801]; x 681 [841]; De

vit. Mos. 192 [I 336]; Hom.xiii in Cant. 663 [i 1045] on the other hand  is printed twice, c. Eun. ii 466 [521]; Ep. ad Flav. 648 [iii 1004]). The (Homœousian) Synod of Ancyra in 358 (in Epiph. Pan. 851 c: the allusion here is reasonably certain1). DIDYMUS three times (Die Trin. i 26 p.76; ii 5, p. 140 [cf.i 15, p. 27]; on Ps. lxxvi 14, p. 597 Cord. [with absolute certainty by the context, though  is printed]: an allusion on Ps. cix 3, p.249 Cord. or 284 Mai, drops the substantive). CYRIL OF ALEXANDRIA (ad 1.

1. The laxity of a reference to Prov.viii 25 ( for  ) in the same sentence was unavoidable, and it was guarded by ample previous exposition (852 BC, 853 B—D): here it would have been gratuitous and misleading.

p. 103 [without ] by Mr Pusey’s best MS. and repeated references in the following comment), and in at least three other places (Thes.137, [without ] 237; Dial. quod Unus, 768: twice (Thes. 365; Adv. Nest. 901) Aubert’s text has  which will probably have to give way, as it has had to do in the commentary 2. To these might perhaps be added the emperor. JULIAN (p. 333 Spanh), for though the full quotation and one subsequent reference have , another has , which the argument seems on the whole to require.

The patristic evidence for []   has next to be given. Irenæus twice, but only in the Latin translation (see above), and exactly in the Old Latin form, with nisi inserted before unigenitus, and once with Dei added to Filius, so that we seem to have the reading of the translator, as often, not of Irenæus. HIPPOLYTUS (c. Noetum 5) without : all depends on Fabricius’s editing of a modern copy of a single Vatican MS., and the context is neutral. An EPISTLE, from certain bishops at ANTIOCH (260—270 A.D) to Paul of Samosata (Routh, B. S. iii 297), again dependent on a single MS., unexamined for some generations, and with the detached phrase       occurring not long before. The Latin version of the “ACTS” of the disputation between ARCHELAUS and Mani, c. 32, where again the inserted nisi shews the impossibility of deciding whether author or translator is responsible. EUSEBIUS OF CÆSAREA six times, De Eccl. Theol. p. 67 (with  as an alternative, see above), 86, 92, 142; in Ps. lxxiv. p. 440 Mont.; in Es. vi. p. 374. EUSTATHIUS,

1. In this case the text is also Pusey’s (p.170); but it rests on a single MS. of the fifteenth century: it is followed in a few lines by

              .

   
   

2. In the ‘Dialogues’ of an unknown CÆSARIUS (Inter. 4, post Greg. Naz. Iv 864 Migne), probably of the fifth if not a later century, the context implies  though  is printed. The apparent conflict of text and context has been 1ate1y pointed out by Prof. Abbot, who still regards the reading as only doubtful. The possibility of reconciling with the actual language an inferential argument from John i 18 containing  seems to me infinitesimal: but I am content to leave Cæsarius in a note.

De Engastr. p..387 All. ALEXANDER of Alexandria, Ep. ad Alex. in Theodoret; H. E. i 3; but with the detached phrase    on the next page. ATHANASIUS seven times (Ep. de Decr. Nic. 13, 21, Or.c. Ar. ii 62; iv 16, 19, 20, 26). GREGORY OF NAZIANZUS, Orat. xxix 17. Basil of Caæsarea, Ep. 234, p 358, besides one of the three places in the De Spiritu Sancto already mentioned, where at least one Moscow MS. has : but the evidence adduced above casts doubt on both places. Gregory of Nyssa twice (see p. 3); but the reading is most suspicious. TITUS OF BOSTRA (adv. Man. p. 85 Lag.: but p. 93  .  ). THEODORE of MOPSUESTIA (ad l. bis in Mai, N. P. B. vii 397f.). CHRYSOSTOM ad l., and later writers generally. On Julian see p. 4.

It is unsatisfactory that so much of the patristic testimony remains uncertain in the present state of knowledge; but such is the fact. Much of the uncertainty; though not all, will doubtless disappear when the Fathers have been carefully edited. In familiar passages scribes, editors, and translators vie with each other in assimilating biblical quotations to the texts current among themselves; and from the nature of the case the process is always unfavourable to ancient readings, whether true or false, which went out of use comparatively early. It would therefore be absurd to treat the uncertainty as equally favourable to both readings. Where we have a Greek original, without various reading noted, and without contradictory context  has a right to claim the authority provisionally, in spite of private suspicions: but it would be unreasonable to concede to  any appreciable part in Origen, Gregory of Nyssa, Didymus, or Cyril — I ought to add, in Irenæus or Basil— notwithstanding the variations already mentioned. Serious doubt must also rest on an isolated  in a neutral context, when, as in the case of the Epistles of the Antioch bishops and of Alexander;   is found at no great distance, though without any obvious reference to John i 18: the doubt is not removed by the

fact that one or two Latin Fathers1 have unigenitus Filius in their quotation. and unigenitus Deus often elsewhere.

To gather up the documentary evidence with the usual abbreviations, we have

 אBC*L 33

Memph. Syr.vulg. Syr.hel.mg [?Aeth.].

*VALENTINIANI. Iren.. *CLEM. *ORIG. [Euseb.]

†Syn.Anc. *Epiph.*DID; *Bas *GREG.NYSS.. *CYR.AI. Cf. Caes.

 AX &c. &c. [?D]

Latt.omn. Syr.vet. Syr.hel. Syr.hier. Arm. [Aeth.codd] [??Iren.(lat.)] ?†Ep.Ant. ?†Act.Arch.(lat.) *EUSEB.

*ATH. †EUST. ?†Alex.Al. [??Bas.] Greg.Naz. [??Greg.Nyss.] †Tit.Bost. *THEOD.MOPS. *CHRYS., &c.

Testimonies marked with * prefixed are clear and sufficient; those marked with † depend on a single quotation, with a neutral context. The Latin Fathers, as almost always, attest only what was read in the Latin versions; all Latin authorities have unicus Filius or unigenitus Filius: q adding Dei.

Against the four best uncials  has no tolerable uncial authority to set except A and X, of which even A is in the Gospels very inferior to any one of the four, much more to their combination, and it is here deserted even by Syr.vulg., its usual companion, while 33 is approached by no other cursive. Manifestly wrong readings of AX and their associates abound hereabouts as everywhere: see i 16, 21, 26 bis, 27 quater, 30, 31, 39, 42, &c.: when

D is added, wrong readings still recur, as iii 34; iv 2, 21, 25, 36, 37, 39, 42, 52, &c. The solitary position of 33 among cursives here arises from the peculiarity of its position generally, and not merely from its comparative excellence, great as that is. The good readings supported by the

1 Hilary and Fu1gentius. The latter twice quotes the text with unigenitus Deus, but doubtless not from a Latin copy of the Gospels.

other good cursives of the Gospels are, with rare exceptions, found likewise in the authorities called ‘Western’, such as D and the early Latins; that is, their ancient element is almost wholly ‘Western’, for good and for evil: the ancient element in 33 on the other hand can be only in part ‘Western’, for it abounds in true ancient readings which, as here, have little or no ‘Western’ authority. That the Old Syriac has  is quite natural, when it has so many early ‘Western’ readings: what is really singular is the introduction of  at the revision, when few changes came in at variance with the late Antiochian text (Theodore, Chrysostom, &c); and as  is not an Antiochian reading, its support by the Syriac Vulgate acquires especial weight. Among early versions this and the invaluable Memphitic more than balance the Old Latin and Old Syriac, which so often concur against BCL Memph. in wrong readings of high antiquity, as i 4, 24, 26, 38, 42; iii 8, 25; iv 9. In the later versions  has no doubt the advantage.

The Ante-nicene Fathers follow the analogy of the versions. With the exception of the Antioch epistle,  occurs in writers with a predominantly Western type of text, Hippolytus and Eusebius (compare the gloss in iii 6 at p. 72 of the De Ecc Th.); while Irenæus leaves their company to join Clement and Origen in behalf of

. After Eusebius the two readings are ranged in singular conformity with the general character of the respective texts generally. Cyril of Alexandria, Didymus, Epiphanius, are almost the only Post-nicene writers in whom we find any considerable proportion of the true ancient readings of passages corrupted in the common late text, while Basil and Gregory of Nyssa have also a sprinkling of similar readings, a larger sprinkling probably than Athanasius or Gregory of Nazianzus, certainly than Theodore, Chrysostom, or their successors. Thus it comes out with perfect clearness that  is one of the numerous Ante-nicene readings of a ‘Western’ type (in the technical not the strictly geographical sense of the word) which were adopted into the eclectic fourth century

text that forms the basis of later texts generally. As far as external testimony goes,  and  are of equal antiquity: both can be traced far back into the second century. But if we examine together any considerable number of readings having the same pedigree as , certain peculiar omissions always excepted, we find none that on careful consideration approve themselves as original in comparison with the alternative readings, many that are evident corrections. No like suspiciousness attaches to the combination of authorities which read

. Analysis of their texts completely dissipates the conjecture, for it is nothing more, that they proceed from an imagined Egyptian recension. The wrong readings which they singly or in groups attest can be traced to various distant origins, and their concordance marks a primitive transmission uncorrupted by local alterations. Such being the case,  is commended to us as the true reading, alike by the higher character of the authorities which support it, taken separately, and by the analogy of readings having a similar history in ancient times.

External evidence is equally decisive against the insertion of , omitted by the four uncials, one passage of Origen probably (c. Cels. ii 71), and two of Cyril (ad l. and Thes. 257). On such a point the evidence of versions and quotations is evidently precarious.

Probabilities of Transcription will doubtless be easily recognised as favourable to .   is an unique phrase, unlikely to be suggested to a scribe by anything lying on the surface of the context, or by any other passage of Scripture.   (the reading of Hippolytus and of Eusebius once, in Ps.), and still more    , is a familiar and obvious phrase, suggested by the familiar sense of

 in all literature, by the contrast to  in the same verse (and   in 14), by two other early passages of this Gospel (iii 16,      , and iii 18,

          ),

and by a passage of St John’s first Epistle (iv 9,

           ). The always questionable suggestion of dogmatic alteration is peculiarly out of place here. To the Monogenes in the Ogdoad of the Valentinians, among whom by a mere accident we first meet with this and other important verses of St John,

 could be only an awkward appendage: the Valentinians of Clement take it up for a moment, make a kind of use of it as a transitional step explaining how St John came to give the predicate  (in i 1) to Logos, whom they anxiously distinguish from Monogenes (= Arche,), and then pass on to their own proper view, in which Sonship alone appears as the characteristic mark of Monogenes; while the Valentinians of Irenæeus content themselves with reciting the bare phrase ( …      [sic]

  ,        ,       

 ) and leaving it, justifying i 1 by the general remark

      , but not otherwise referring again to any  except Him whom St John, they say, distinguishes in i 1 from Arche (= Son) and Logos. Neither in the Valentinian nor in any other known Gnostical system could there have been any temptation to invent such a combination as

 . Nor is it easy to divine what controversial impulse within the Church could have generated it in the second century; for the various doctrinal currents of that period are sufficiently represented in later controversies of which we possess records, and yet there is, I believe, no extant writer of any age, except that very peculiar person Epiphanius1, who makes emphatic controversial appeal either to  per se, or to  as coupled with , or (with a different purpose) to  as coupled with , whether in this verse or in the derivative detached phrase mentioned hereafter. The whole verse, with either

1Also Cæsarius, if the printed  is wrong. The emperor Julian may be added, as finding matter of accusation against St John in this verse, if I am right in surmising that   was the reading before him.

reading, soars above the whole extant theology of the second century antecedent to the great Catholic writers at its close: but I could almost as easily believe that that age invented St John’s Gospel, as some learned persons say it did, as that it invented  . Once more, assuming   to have obtained a footing in MSS., we cannot suppose that it would gain ground from    in transcription, unless we trust modern analogies more than actual evidence. The single fact that   was put to polemical use by hardly any of those writers of the fourth century who possessed it, either as a reading or as a phrase, shews how unlikely it is that the writers of our earliest extant MSS. were mastered by any such dogmatic impulse in its favour as would overpower the standing habits of their craft.

The only other possible explanation is pure accident. The similarity of YC to C, though doubtless greater than that of the words at full length, is hardly strong enough to support a word forming a new and. startling combination, though it might be able to cooperate in a transition to so trite a term as  . But a still more serious objection to this suggestion is the absence of the article in what we must consider the primitive form of the reading,  . Supposing for the sake of argument that YC might pass into

C, the change would still have left  standing ten letters back, and there would have been as little temptation to drop  before  as before  , as is shown by the profuseness with which the Fathers (and their scribes) supplied it subsequently. On the other hand the known boldness of ‘Western’ paraphrase would have had little scruple in yielding to the temptation of inserting  after changing  to , whether immediately or after an interval in which the article remained absent.

Thus, on grounds of documentary evidence and probabilities of transcription alike, we are irresistibly led to conclude that   was the original from which    and   proceeded. More than this no evidence from without can establish: but in a text so amply attested as that.

of the New Testament we rightly conclude that the most original of extant readings was likewise that of the author himself, unless on full consideration it appears to involve a kind and degree of difficulty such as analogy forbids us to recognise as morally compatible with the author’s intention, or some other peculiar ground of suspicion presents itself.

This is perhaps the best place to mention a third reading to which Griesbach was somewhat inclined (it must be remembered that BC were as yet assumed to agree with most MSS, in reading , and א was unknown), and which at one time seemed to me probable; namely   without either substantive. It is supported however by neither MS. nor version except the Latin St Gatien’s MS., but by a few quotations in Greek and. Latin Fathers, almost wholly writers who use one or other of the fuller readings elsewhere; the only considerable exception being Cyril of Jerusalem (Cat. vii 11). It is doubtless common to find different authorities completing an originally elliptic or condensed expression in different ways, But the stray instances of   and Ungenitus are sufficiently explained by the extreme frequency of this simple form of phrase in the theological writings of the fourth and fifth centuries. Nor, on an attentive scrutiny, does it commend itself even as a conjecture, these unsubstantial shreds of authority being discarded. To those indeed who justly recognise the conclusiveness of the evidence which shews that   cannot be a corruption of   , yet are unable to believe that St John wrote it,   affords the best refuge. In sense it suits the immediate context, having in this respect an advantage over

  ; though it seems to me to fail in relation to the larger context formed by the Prologue, and to lack the pregnant and uniting force which I hope to shew to be possessed by  . But serious difficulties as to transcription have to be added to the want of external evidence. It is as inconceivable that

 should have been supplied to complete   in the second century, with the further omission of the article, as that   

should have been. altered to  . Nor is the case improved by supposing accidental errors arising out of similarity of letters, CO becoming CθCO, and 0 being lost after E. It would be an extraordinary coincidence either that both slips of the pen should take place at the same transcription, though separated by MONOΓENHC; or that two corruptions of the same clause should take place at different times, yet both before the earliest attested text of the New Testament. And again to suppose  without  to be the true reading would only change one difficulty for another:  without either article or substantive, followed by  , and caught up by , would be harsh beyond measure. Thus the conjectural omission of the substantive produces no such satisfying results as could for a moment bring it into competition with the best attested reading. Except, on the assumption that the best attested reading is impossible.

Accordingly the field of criticism is now in strictness narrowed to the alleged impossibility of

 . It will however be well for several reasons to examine the readings on their own positive merits, without reference to the strong assertions of private and overpowering instinct by which criticism is sometimes superseded. We have therefore, thirdly, to consider Intrinsic Fitness.

St John’s Prologue falls clearly and easily into three divisions:

(a) 1. The Word in His Divine relations in eternity antecedently to creation.

() 2—13.. The Word in His relations to creation, and especially to man, chiefly if not altogether antecedently to the Incarnation.

() 14—I8. The Word as becoming flesh, and especially as thereby making revelation.

(The two digressions 6—8, 15, in which the Baptist’s office of witness is put forth in contrast, do not concern us here.)

The first division ends with the simple affirmation that the Word, who was    was Himself

. In the

13

second division, after the initial  which reintroduces the second clause of verse 1, His original name is not repeated: He is presented as the universal Life, and as the Light of mankind; coming into the world, and ignored by it; visiting His own special home, and receiving no welcome there, though in a manner accepted elsewhere: so ends the history of the old world. The third division pronounces at once the name unheard since verse 1, but now as part of the single stupendous phrase     , and adds the visible sojourning of the Word ‘among us’, whereby disciples were enabled to behold His glory. This glory of His is further designated, by a single phrase which is a parenthesis within a parenthesis, as being “a glory as of an only-begotten from a father”. Neither the Son nor the Father, as such, has as yet been named, and they are not named here: there is but a suggestion by means of a comparison (the particle  and the absence of articles being mutually necessary), because no image but the relation of a  to a father can express the twofold character of the glory as at once derivative and on a level with its source. Then the interrupted sentence closes in its original form with the description    , followed, after the interposition of the Baptist’s testimony by a notice of this fulness of grace as imparted to Christians, and its contrast with the preceding Law. Finally verse 18 expounds the full height of this new revelation. Now, as truly as under the Law (Ex. xxxiii 20; Deut. ix 12), Deity as such remains invisible, although the voice which commanded has been succeeded by "the Truth” which was “beheld”. Yet a self-manifestation has come from the inmost shrine: One of whom Deity is predicable under that highest form of derivative being which belongs to a , not one of imperfect Deity or separate and external place but He who in very truth is

    ,—He, the Word, interpreted Deity to the world of finite beings.

Part of this meaning is undeniably carried by the common reading   ; but incongruously, and at best only

a part. Here as in v. 14 special force lies in  in contrast to the share possessed by one among many brethren; and for this purpose  adds nothing, if indeed it does not weaken by making that secondary which was meant to be primary, for other ‘children of God’ had just been mentioned (vv. 12, 13). There would also be something strangely abrupt in the introduction of the complete phrase   , as a term already known; which ill suits the careful progress of St John: the leap from     would be too sudden; the absence of any indication identifying   with the Word would be dangerously obscure, while the article would mar the integrity of the Prologue by giving its crowning sentence a new subject in place of  ; and in any case a designative name would serve the argument less than a recital of attributes. This last point comes out more clearly as we follow the exquisitely exact language of the whole verse. The ruling note is struck at once in , set before  in emphatic violation of the simple order which St John. habitually uses; and further  has no article, and so comes virtually to mean ‘One who is God', ‘God as being God’, and perhaps includes the Word, as well as the Father1. In exact correspondence with  in the first sentence is   in the second. The parallelism brings out the emphasis which the necessary nominative case might otherwise disguise, and a predicative force is again won by the absence of the article. St John is not appealing to a recognised name, as an inserted article would have seemed to imply, but setting forth those characteristics of the Revealer, already described (v.14) as ‘the Word’, which enabled Him to bring men into converse with ‘the Truth’ of God, though the beholding of God was for them impossible. It needed but a single step to give the attribute  to Him whose glory had been already called a glory as of a

 from a father. It needed no fresh step at all to give Him the attribute , for He was the Word, and the Word had at the outset been

1 Cf. Greg. Naz. Ep. 101 p.87A      .

declared to be  . The two elements of the phrase having thus been prepared, it remained only to bring them together, associating Deity with Him as Son (for that much is directly involved in the single term ) as expressly as it had been already associated with Him as Word; and then the combination is fixed and elucidated by the further description        1 It begins with the article, for now that One has been called   — and in One alone can both attributions meet,— there is no longer need for generality of language; we exchange “One that is—” for “He that is—”. In like manner now that He has been set forth as actually  as well as , it has become right to speak definitely of

 . The connecting phrase     is a repetition of      

translated into an image appropriate to the relation of Son to Father.

Thus St John is true to his office of bringing to light hidden foundations. The name ‘The Word’, in which he condenses so much of the scattered teaching of our Lord and the earlier apostles, leads gradually, as he expounds it, to the more widely current idea of Sonship, which after the Prologue he employs freely; and yet is not lost, for  suggests at once the still present middle term of v. 1. through which  has become linked to . The three salient verses of the Prologue are 1, 14, 18.These by themselves would suffice to express the absolute primary contents of St John's ‘message’: the intervening verses are properly a statement of the antecedents of the Gospel, and of its meaning as illustrated by its relation to its antecedents. Verse 1. declares the Word. to have been ‘in the beginning’ ; verse 14 states that the Word, when He became flesh, was beheld to have a glory as of a ; verse 18 shews how His union of both attributes enabled Him to bridge the chasm which kept the Godhead beyond the knowledge of men. Without

  the end

1 Cf .Cyr. 1l. ad 1 p. 107B                   

       ...

of the Prologue brings no clear recollection of the beginning:  is the luminous word which recites afresh the first verse within the last, and in its combination with  crowns and illustrates the intervening steps.

It is therefore vain to urge against the phrase that it is unique in the New Testament. The whole Prologue is unique, and   seems to belong essentially to a single definite step in the Prologue. No writer except St John applies  to our Lord at all, and he only in the three other closely connected places already cited. In each of them there is a distinctly perceptible reason why  should be introduced; and moreover there were obvious objections to the employment by St John of the definite title

  , that is, with the article. If we examine the combination dispassionately, it is hard to see in it anything inconsistent with the theology of St John, unless the idea of an antecedent Fatherhood and Sonship within the Godhead, as distinguished from the manifested Sonship of the Incarnation, is foreign to him. This idea is nowhere enunciated by him in express words; but it is difficult to attach a meaning to

       on any other view, and it is surely a natural deduction from the Prologue as a whole (with either reading) except on the quaint Valentinian theory that the subjects of vv. 14 and 18 are different, while it seems impossible to divine how he can have otherwise interpreted numerous sayings of our Lord which he records. The paradox is not greater than in the other startling combination

   , the genuineness of which no one affects to question, though its force has been evaded in different directions in all ages.

The sense of  is fixed by its association with  in the other passages, especially v. 14, by the original and always dominant usage in Greek literature, and by the prevailing consent of the Greek Fathers. It is applied properly to an only child or offspring; and a reference to this special kind of unicity is latent in most of the few cases in which it does not lie on the surface, as of the Phœnix in various

authors, the   of Plato (Tim.31B) as made by the 'Father’ of all (28c), and the

  of writers who follow him. Instances are not entirely wanting in which  is used of things that are merely alone in their kind (as if from , and in its widest sense); but this rare laxity of popular speech, confined, if I mistake not, to inanimate objects, cannot be rightly accepted here. It finds indeed some support from Gregory of Nazianzus (Orat. xxx 20 p. 554 A) and Ammonius (on iii 16 in the catenæ): but Basil’s simple rendering (adv. Eum. ii 20 p. 256 A)   , put forward in opposition to Eunomius’s arbitrary invention    , (compare Athanasius’s negative definition, Or.c. Ar. ii 62 p. 530A,

         ,) expresses the sense of the greater writers of different ages1, though they sometimes add   to . While however the idea conveyed by the verb itself in the paraphrase   belongs essentially to the sense, the passive form goes beyond it, as perhaps even in unigenitus, and the narrower sense of the English verb in ‘only- begotten’ departs still further from the Greek. If  .  were the true reading, it would on the whole be a gain to adopt ‘the only Son’ from Tyndale in iii 16, 18, and from the English Apostles’ Creed, where ‘only’ represents the  of this or the other like passages, as ‘only-begotten’ represents it in the ‘Nicene’ Creed of the English Communion Service. But no such. expedient is possible with  ; and so the choice lies between some unfamiliar word, such as ‘sole-born’, and the old rendering which certainly exaggerates the peculiarity of the Greek phrase, though it may be defended by imperfect analogies from other passages of the New Testament.

1 A few out of the many somewhat later patristic illustrations of the true sense are collected, not without confusion in the appended remarks; by Petau de Trin. ii 10 10 ff.; vii 11 3 ff. Cyr.Al.Thes. 239 f. is specially clear:

 …      : again    : again

   :

again                  

  ;

18

A change of a different kind however seems absolutely required, either the insertion of ‘One who is’, or the resolved rendering ‘An Only-begotten who is God, even He who &c’: without some such arrangement the predicative force of   is lost, and. the indispensable omission of the English article becomes perilous.

But these matters of translation do not affect, though they illustrate, the primary question as to St John’s own Greek text. I have, I trust, now given sufficient reasons for concluding not only that   presents no such overwhelming difficulty as to forbid its acceptance notwithstanding the weight of evidence in its favour, but that the whole Prologue leads up to it, and, to say the least, suffers in unity if it is taken away.

All these considerations are entirely independent of the truth of any theological doctrines which have been deduced, or may be deduced, from St John’s text. When it is urged that certain words are incongruous with the context and with St John’s teaching generally, it becomes legitimate and perhaps necessary to discuss their genuineness on grounds of sense; and not the less legitimate where, as in this case, the sense is manifestly theological, the criterion for the present purpose being not doctrinal truth but doctrinal congruity. Since however it is matter of fact that a fear of theological consequences is acting in restraint of dispassionate judgement, and that in opposite quarters, I feel justified in appending to the critical discussion a few remarks on the treatment of   in ancient times, which may at least suggest some diffidence in relying on the infallibility of modern instincts.

The list already given of Fathers who read []   in their text of John i 18 takes no account of the much more widely diffused use of the phrase []   without a biblical context. Professor Ezra Abbot justly points out that

19

the phrase in itself affords no sufficient evidence as to the reading of St John followed by those who employ it, since it is a favourite with one or two who undeniably read    when they quote the Gospel1. Yet it is equally true that this widely spread usage bears an indirect testimony which may be fitly noticed here, partly by its mere existence, partly by its probable connexion with public formularies.

Origen’s voluminous remains contain the detached phrase   eight or ten times, usually softened, by the addition of  or in some other way. It lurks in one place in the Antioch Epistle against Paul of Samosata (           , p. 292), and ought, I suspect, to be restored to another (   ,   ϯ  ϯ,     ,

…        ,   , p. 290), where the second  cannot be sustained by any punctuation, but must either be omitted or, with better reason, exchanged for . With these exceptions it is, I believe, absent from the extant Ante-nicene literature, notwithstanding the diffusion of the corresponding biblical text. The absence of this reading from good secondary MSS. and from almost all the later versions shews how rapidly it was superseded in the fourth and fifth centuries; yet we encounter the phrase itself on all sides in this period, and certainly not least abundantly in the latter part of the fourth century. Without attempting an exhaustive list, it may be useful to set down the following names and references, partly taken from Wetstein and other critics, partly from my own notes. Atha- nasius (c. Gent. 41 p. 40 c,

                 

     ; c. Apoll. ii 5 p. 944A,

            

      .

1 The few Greek writers coming under this description, all of whose quotations with  are either solitary or otherwise doubtful, cannot properly be taken into account.

20

[i.e. One who is God, even   ]

           

           

); Arius (ap.Ath. de Syn. 15 p. 728 E,   …   ; Epiph. Haer. 732 A,

 …            

1); Alexander the bishop of Alexandria with whom Arius came into conflict (1. c. p. 734 Noess.

     ); Marcellus (ap. Eus. c. Marc. i 4 p. 19 c2); Asterius (ap. Ath. Or. c. Ar. ii 37 p 505 c [v.1.]; de Syn. 18 p. 732 B); Theodorus of Heraclea (on Isaiah in Mai, N.P B. vi 226); Eusebius [of Emesa, by Thilo’s identification] (de fide &c. [Latine] in Sirmondi Opp. i 3B, 16 D, 22 A); Rufinus of Palestine (Latine in Sirmondi Opp. i 274 ff. cc. 39, 52, 53, and with Verbum often); the Synod of Ancyra (ap. Epiph. Haer. 854 C); Epiphanius (Haer. 755 C, 817C , 857 A, 912 A, 981 A); Cyril of Jerusalem (xi 3,   ); Eunomius (Apolog. 15, 21, 26; Expos. Fidei 2 bis); Basil (Ep. xxxviii 4 p. 117 C; de Sp.19 p. 16 C; 45 p. 38 B; c. Eun.ii 1 p. 238 C; also  .   , i,15 p. 228; 26 p. 237 B); the Apostolic Constitutions (iii 17; v 20 § 5; vii 38.§ 3; 43 § 1; viii 7 § 1, 35); the interpolator of the Ignatian Epistles (ad Philad. 6); Gregory of Nazianzus (Ep. 202 p. 168 C); Gregory of Nyssa repeatedly and in various writings (Professor Abbot counts 125 examples in the treatise against

1 It has been urged that  invalidates the reference. On the contrary the sense is that before  and  the Son attained that full height, subject to no change, which is expressed by   .

2 Marcellus seems to be quoting a Creed, but in such a manner as to make its language his own.  , says Eusebius (c.

Marc; 19 C)

    ,        ,     

           . Quite different in form is the Creed presented by him to Julius of Rome (Epiph. Haer. 836), the suspiciously Western character of which is well known. In the epistle to Julius (835 D) he uses the phrase        , where the added  probably implies , itself excluded by .

Eunomius alone); Didymus (de Trin. i 25 p. 68 Ming.; i 26 p. 72; with  , i 18 p. 53; 26 p. 76; with

  interposed, i 16 p. 40; with , i 26 p. 75); the ‘Macedonian’ interlocutor in an anonymous Dialogue on the Trinity (Ath. Opp. ii 509 B1); Isaac ‘ex Judaeo’ (Sirmondi Opp. i 406 ABC); Cyril of Alexandria repeatedly; Andrew of Samosata (ap. Cyr. Al. Ap. adv. Or. p. 290 Pusey [ix 333 Migne]); Theodoret (Repr. xii Capp. Cyr. 12 with 2; c.Nest. iv 1047 Schulze); Theodotus of Ancyra, once with , once without3 (post Cyr. Al. x 1336 f: Migne); Basil of Seleucia (Hom. i p. 5 A; cf. xxv p.139 D); Isidore of Pelusium (Ep. iii 95); even John of Damascus in compound phrases4, perhaps following the Henoticon of Zeno (see p. 24 n.1); Hilary in peculiar abundance in different writings (a single typical instance will illustrate his use: “Deus a Deo, ab uno ingenito Deo unus unigenitus Deus, non dii duo sed unus ab uno,” de Trin. ii 11); the fragments of a Latin Arian commentary on St Luke (in Mai S. V. N. C. iii 2 191, 199) and of Latin Arian sermons (ib. 217: cf. per filium unigenitum Deum in the Arian Primus capitulus fidei catholicae, ib. 233); the Latin Opus Imperfectum on St Matthew a few times (e.g. i 20 bis, 25) &c. The chief apparent exceptions are the later Antiochian school of Greek writers, and Ambrose and his disciple Augustine among Latin writers. Yet the subsequent theologians of North Africa by no means eschew the phrase, and it is of frequent occurrence in the

1The ‘Orthodox’ interlocutor neither objects to the term nor uses it himself.

2 So in Pusey’s text of Cyril (Apol adv. Theodoret. p. 492) with (apparently all) the Greek MSS. and the Syriac and Latin versions. Prior editions (as Schulze of Theodoret v 66 and Migne of Cyril ix 449 c) substitute   for , apparently without authority.

3 In his Exposition of the Nicene Creed. But the context leaves it doubtful whether he assumed the combination to be already in the Creed, or only took its elements from the Creed.

4          (De fid. orth. i 2 p. 792 c Migne; iii 1 p. 984 A);

 .      (iii 12 p. 1029 B);  .    (i 2 p. 793 B). In the third passage  might be independent of ; not so, I think the context shews, in the others.

writings of Fulgentius in particular. Even in the days of Alcuin and Theodulphus it is not extinct.

In the later times the tradition doubtless passed directly from writer to writer: but this explanation will hardly account for the wide and various acceptance found by   in the fourth century, combined with the almost complete absence of attempts to argue from it by any of the contending parties. This remarkable currency arose, I cannot but suspect, from its adoption into Creeds. We look for it of course in vain in Latin Creeds1, for Latin Christendom from the earliest times known to us did not possess the fundamental reading in the Gospel: Hilary must have learned it, as he learned much else, from his Greek masters. Among the very few Greek Creeds belonging clearly to the second or third century of which we have any knowledge, we can identify   only in that of Antioch, incorporated with the remarkable exposition of Lucianus (Sozom. H. E. iii 5 9; vi 12 4), who suffered martyrdom about 311. Here we read

     ,      ,    , 

         ,    ... (Graece ap. Ath.de Syn. 23 p. 736 A; Socr. H.E. ii 10; Latine ap. Hil. de Syn. 28 p. 478C: cf. Bull Def. Fid. Nic. ii 13 4—7). The word  after  was perhaps not in the earliest forms of this Creed (see pp. 24, 26): but there is no reason to doubt that it stood there in the time of Lucianus, of whose amplifications there is no sign till further on. In the passage of Marcellus of Ancyra referred to by Eusebius (about 336), in which he apparently follows some Creed (see p. 20), we have already found the identical Antiochian phrase

     . The exposition of Lucianus was one of the four formularies brought forward at Antioch in 341: another, perhaps a modification of the local Creed of Tyana, the see of Theophronius who recited

1 One elaborate private formulary, long attributed to Jerome or Augustine the Confession of Pelagius (Hieron. Opp. xi 202 Vall.), has verum Deum unigenitum et verum Dei filium.

it, has in like manner,

        ,   ,    

 ,    ,          

  ,        ... (ap. Ath. de Syn. 24 p 737 B). Once more the formulary of the Synod of Seleucia in Isauria held in 359 declares,

           ,    

     ,  ,    , , , 

, , ,      ... (ap. Ath. de Syn. .29 p 746 C; Epiph. Haer. 873 B, C; Socr. H E. ii 40). The influence of the two latter documents would probably be limited and temporary: but the details of their language, so far as it was not shaped by current controversy, must have been inherited directly or indirectly from formularies now lost, matured before the outbreak of the Arian disputes. Nay the original Nicene Creed itself appears to embody the phrase, though in a form which admits of being interpreted either as a deliberate retention or as a hesitating and imperfect obliteration of an earlier statement of doctrine (see Note D). Indeed it occurs once without any ambiguity, as a friend points out, in what purports to be a copy of the Nicene Creed included in a memorial from Eustathius of Sebastia and other representatives of the Asiatic Homœousians proffering their communion to Liberius of Rome, and expressly accepted by him as the Nicene Creed, shortly before his death in 366. This copy differs in nothing but two or three trivial particles from the usual ancient form except in the words      

 ,    , and the omission of  from its accustomed place in the next clause (ap. Socr. H.E. iv 12). In the familiar Creed usually regarded as the Constantinopolitan recension of the Nicene Creed   was undoubtedly wanting, for reasons explained in Dissertation II. But finally in 451 it stands included, though with the old Alexandrine addition , in the carefully chosen last words of the Definition of Chalcedon       ,

    ,     ,    (“sed unum eundemque Filium et unigenitum Deum Verbum Dominum Jesum Christum,” in Mansi’s primary old version),

             

,       . It is true that Evagrius (H. E. ii 4), Agatho (in Mansi Conc. xi 256), and the third Council of Constantinople in 680 omit  so as to bring  and  into combination, as also most Latin versions omit et, some further making transpositions: but the reading of the best authorities is sustained not only by its less obvious character but by the unquestionable separation of  from  a few lines above, in the sentence

         ,     

              

   ,     , , , .1

At this point a possible suspicion requires notice, whether   may not owe its origin to Creeds, and have passed from them into the text of St John. The authority of a Creed might doubtless succeed in importing a difficult and peculiar reading, the introduction of which in any other way would be inconceivable. But the facts already stated are as fatal to this as to all other suggested explanations of a change from

   to  ; and the evidence of Creeds does but corroborate the other evidence. I do not press the late date, the close of the third century at Antioch, at which we first find

  actually standing in a Creed. The Creed of Antioch in that form might be of earlier date: and the same may be said of any Creeds which may have supplied materials at Nicæa in 325, at Antioch to Theophronius in 341, and at Seleucia in 359, though these might also belong in their corresponding form to Lucianus’s or even to the next generation. But

1 The Henoticon of the emperor Zeno, promu1gated in 482, begins its final confession with the words

        ,  ,.. (Evagr. H. E. iii 14).

25

conjectures of this kind will not avail unless we are prepared to go so far as to say that   stood in several distant Creeds towards the close of the second century, or that it stood in some one leading Creed near the beginning of the second century, for nothing less would account for its presence in such various biblical texts. Ptolemæus (see p. 30) speaks either from Italy for himself in the third quarter or at most a few years later, or from Alexandria or Rome for his master Valentinus in the second quarter of the century; Irenæus from Asia Minor or (less probably) Gaul; Clement and the Memphitic version from Alexandria; Origen a little later from Alexandria and probably also Palestine. It would not be easy to trace these scattered texts to Alexandria, the only imaginable single centre, at that early period: but if it were, we should find ourselves still confronted by two weighty facts. First, there is not a trace of theological activity at Alexandria, except that of the ‘Gnostic’ chiefs, till the Catechetical School of the Church (Athenagoras, Pantænus, Clement) arose in the last third of the century, which is too late for our purpose: if such existed, some record of it must have been preserved by Eusebius, who had a special interest in Alexandria, and has given us a tolerable roll of contemporary writers from other parts of the East. Secondly, little as we know of the Creed of Alexandria, it happens that that little suffices to shew that it did not contain  . There is no trace of the words in the rule of faith expounded in Origen’s early work De Principiis (Preface to Book i § 3 f.), though in various places where he speaks in his own name (as in i 2; ii 6) there are suspicious signs that the translator Rufinus had them before him. But even in the days of Arius   is clearly absent from the Alexandrian Creed as recited by Alexandria, notwithstanding his own use of the term; for the evidently ancient words run

     ,      ,  ... Thus all external evidence fails to sustain a derivation from Creeds in the second century: if we are to consider intrinsic probabilities, it must be repeated that the invention of the phrase in the first half (and more) of the

century is at variance with all that we know of any of its theologies: and as for the Creeds of the Church, that in those early days of elementary simplicity they should admit such a combination without direct Scriptural warrant would contradict all that we know of their manner of growth. Whether it could have been so admitted in the third century, with the theology of which it easily associates itself, is highly questionable; but that is not the period with which we have to deal. Yet even in the third century, as has been shown, the usage is cautious and tentative, by no means such as we should expect with words freely pronounced in Creeds. Origen quotes the verse almost half as often as he employs the phrase, and in a majority of cases he adds to the phrase some tempering word. At Αntiοch, where alone else it appears, it is conceivable that the Creed had an influence, though hardly if unsupported by Greek MSS., in changing the reading of the Syriac version; but the converse is equally possible. It is only in the fourth century that the phrase pervades the greater part of the extant literature: and the cause surely is that, though   as a reading was being swept out of biblical MSS. by the same accidental agencies of transcription which removed hosts of Antenicene readings of no doctrinal moment, as a formula it had at last established itself in widely known Creeds. We cannot look to Creeds as the sources of the reading without inverting history.

The one historical demerit then, if demerit it be, which attaches to the combination   is that each of the great parties in the fundamental and necessary controversies which began in the days of Constantine was willing to pronounce it, and that it has never itself become a watchword of strife. It was not avoided by Arius or his successor in the next generation, Eunomius, though neither of them inserted it in his own shorter Creed (see the letter of Arms and Euzoius to Constantine, in Socr. Η. E. i 26; Sozom. H.Ε. ii 27, without even

; and the Confession in Eunomius’s Apologeticus,c. 5,

      ,  ), by the

Latin Arian commentator on St Luke, on by the author of the Opus Imperfectum, usually classed as an Arian. It appears sporadically in various quarters in the intermediate movement, commonly called Semi-Arianism, which, however inconsequent in thought, retained much of the letter of Antenicene language; while on the other hand it was not used spontaneously by Eusebius, who habitually followed his MS. or MSS. in reading  in St John. It is uttered but sparingly and guardedly by Athanasius, once in youth and once in old age, probably for a similar reason1 ; for he seems hardly likely to have shrunk from it on grounds of doctrine or feeling, when we remember that he speaks of     (Or. c. Ar. i 28 p. 432 C) and that the phrase in which he most loves to clothe his characteristic teaching is      . Once more we find   in Marcellus, the blind violence of whose antagonism to Arius conducted him to a position of his own. Hilary, the wisest as well as the most successful champion of the cause of Athanasius in the West, employs it with startling freedom, evidently as the natural expression of his own inmost thought. Among the greatest of the theologians who continued and developed the same line of tradition in the East are confessedly Basil, Gregory of Nyssa, Didymus, and Cyril of Alexandria; and to none of these, widely as they differ from each other, is   strange, while with two of them its use is habitual. Finally, with an accompaniment which guards but does not neutralise it, it obtains a place in the definition of the last of the ‘four’ primary Councils.

This great variety of belief among those who have received   into their theological vocabulary suggests at once that its utility is not that of a weapon of offence or defence. Experience has shown that it is possible to affix a

1 Sometimes (as de Decr. 16ρ. 221 E; Or. c Ar ii 47 p 515 E; Ep. ad Αfr.5 p 895 A, C) he has the derivative form []  , which occurs in a passage of Origen quoted by him de Decr. 27 p. 233 c, and is not rare elsewhere.

28

considerable range of meaning to words which simply express either Deity or Sonship, and even, as here, to a combination of the two predicates in the same subject. But it is rarely by the literal and apparent cogency of single texts that deliberate convictions have ever been formed: power in producing belief is not to be measured by convenience in argument. Understanding as I do both terms in the highest sense, and holding that the doctrine of perfect and eternal Sonship within the Godhead, for which Origen and Athanasius contended, and which the Nicene and 'Constantinopolitan’ Creeds explicitly set forth is fundamental truth, I cannot affect to regret that a reading of St John’s words which suggests it, though it does not prove it, is established as genuine by a concurrence of evidence which Ι could not disregard without renouncing critical honesty. Perhaps the words may prove in due time instructive, thus much may be said without presumption, both to us who receive the doctrine and to those who as yet stumble at it.

It does not however follow that good results would now arise from a resuscitation of the ancient formula detached from the context of the Gospel. To employ it with the article prefixed would open the way to serious evil; while without the article it requires arrangements of diction which could seldom be contrived in common usage, and which incautious writers would be perpetually tempted to discard. The danger of the article is somewhat less in Greek than in English: nevertheless it must have been a dread of possible misuse that induced the Greek theologians so often to temper the article, as it were, by adding afterwards , , or some other term which fixed the denotation of  without lowering its sense or suggesting ‘division’.

Yet these considerations can have no place in determining the text of St John. Taught by himself to “believe on the name of the Only-begotten Son of God”, we do well to adhere to the name thus entrusted to us: but we need not shrink

from accepting and trying to interpret his other language in the single instance when he is led — not to put forward another name but — to join two attributes in unwonted union, that he may for a moment open a glimpse into the Divine depths out of which his historical Gospel proceeds.

NOTE A

The details of early Greek Patristic Evidence

The earliest known Greek reference to John i 18 occurs in two independent accounts of Valentinian doctrine, furnished by Irenæsus and Clement respectively1. The Valentinianism sketched by Irenæsus in his first book is commonly recognised to be that of Ptolemæus, who apparently belongs to the generation succeeding the middle of the second century. He cannot at all events be later than the episcopate of Eleutherus, about 175—190, under which Irenæus wrote (p. 176 Mass.). “They further teach” Irenæus says (p. 40), “that the First Ogdoad was indicated () by John the Lord’s disciple, these being their words: ‘John, the Lord’s disciple’, intending to give an account of the genesis of the universe whereby the Father put forth () all things2, supposes a certain , the first thing gendered by God (     ), which he has also3 called () Son and  , in

1 The recent criticisms of Heinrici (Die Valentinianische Gnosis und die heilige Schrift) and Lipsius (Protestantische Kirchenzeitung of Feb. 22 1873, pp. 182 ff.: cf. Quellen d. ältesten Ketzergeschichte 90) have not thrown so much light on the mutual relations of these two accounts as might have been hoped for from such otherwise instructive investigations. It seems clear that neither Clement drew from Irenæus nor Irenæus from Clement, nor both from a common immediate source. More than this it would he rash to assert at present.

2 The text followed up to this point is that of the Greek extract preserved in Εpiphanius (ρ. 198 Pet.), which shews no sign of amplification here. The old Latin version has omitted some words, including those which mark the quotation as verbal; while at the end of the quotation it adds “Et Ptolemaeus quidem ita,” omitted by Epiphanius. But both texts imply a Valentinian appeal to “John the Lords disciple” for what fellows.

3 There is no reason to change quod etiam nunc (al. q. e. me) of the MSS. to quod etiam Nun with Εrasmιιs,

whom (or which) the Father seminally put forth all things1.” The Valentinian writer proceeds to treat St John’s Prologue, clause by clause, as a commentary on his theory that  was derived from , and  from , all three being nevertheless intimately united; and endeavours to extract the personages of his Ogdoad from St John’s terms. From i 14 he obtains the first Tetrad, Pater and Charis, Monogenes and Aletheia; and there he stops, the second Tetrad having been already found in i 1—4, so that i 18 is not quoted in so much of the passage as Ιrenæus transcribes. But the simple term Monogenes, required as a masculine synonym of Arche to make a syzygy with Aletheia, is distinctly taken from i 14; so that when the writer parenthetically attributes to St John two other designations of Arche, Son and  , neither of which is convenient for his present purpose, he cannot mean only that they are fair deductions from language used in i 1—14, but must have in view some literal use by St John elsewhere; that is doubtless i 18; iii 16, 18.

The same result presents itself at once in the Valentinian statements of doctrine, partly copied, partly reported by Clement of Alexandria in the Excerpta found at the end of the Florence MS. of the Stromates, and now reasonably supposed to belong to his lost Hypotyposes (Βunsen, Anal. Antenic. i 159 ff.). “The Valentinians”, he says, (p968 Pott.; p. 210 Buns.) “thus interpret” Jo. i 1: “they say that Arche is the Monogenes, who is likewise called () , as also in what follows he [John] expressly signifies Him to be

whose conjecture is adopted by later editors. Quod etiamnunc (or etiaimnum) is a natural rendering of   : and though  occurs in Clement’s parallel exposition, and has been noticed already by Irenæus (p. 5), it could have no place among the terms enumerated as taken from St John, and it is absent from the context which follows.

1 So in the Venice MS (the best) of Epiphanius        ; the common text inverting

 and . The true order is retained in the Latin, "et Filium et Unigenitum Deum", though in some of the inferior MSS. and in the editions Domini (Dni) has been substituted for Deum (Dm), as read by others, including the Clermont and Arundel MSS., the two best, and representatives of different families.

 (        ), saying “           .” The word 'expressly' was doubtless used because the writer considered the Deity of Arche, though not explicitly stated by St John, to be obviously included in the attribution of Deity to Logos (   ), since Logos was derived from

 not directly but through Αrche1: but this preliminary inference only throws into clearer relief the coupling of the Monogenes with  by the Evangelist himself in i 182. When then in what follows reference is made to the Father’s ‘putting forth’ of the Monogenes, who is further identified with the Son (   ,      ), we have at once in the combined designations a sufficient explanation of the appearance of  in a succeeding allusion to i 18 (              

   ,       ), without supposing  to have stood here in the writer’s text of St John. The Ηyροtyposes were probably written in the early years of the third century, certainly not later3. If all the Valentinian Excerpts belong to the 'Eastern School' mentioned in the obscure title (cf. Hippol. Haer. vi 35), the coincidence with the Valentinianism in Irenæus would bring the evidence as to St John’s reading far back, perhaps to the second quarter of the second century; for Ptolemæus is named by Hippolytus (l. c.) as belonging to the

1 So the writer in Ireæus (p. 41).

         ,          .    

    .        ,           

,                 .    

   ,     .

2 The next sentence appears to contain a retrospective argument justifying the ascription of Deity to the Logos, as in i. 1, by the subsequent ascription of Deity to the Monogenes (=Arche = ), as in i 18, which would imply the presence of  in each verse. But in other respects the language is obscure, and probably corrupt.

3 Without referring to the Hypotyposes, which must be a late work, Heinrici (1.c. 12 f. ) places the Excerpts and the cognate Eclogae Propheticae in Clement's youth, about 170-180. His argument is not convincing.

other or ‘Italian’ School, and thus the coincidence would have to be traced to Valentinus as the common source of both schools. But this assumption cannot be trusted, and we must be content to take Clement’s author as probably belonging to the same period as Ptolemæus.

Ιrenæus himself thrice quotes i 18, "Deus qui fecit terram ... hic et benedictionem escae . . . per Filium suum donat humano generi, incomprehensibilis per comprehensibilem et invisibilis per visibilem, cum extra eum non sit sed in sinu Patris exsistat. Deum enim, inquit, nemo vidit unquam nisi unigenitus Filius Dei qui est in sinu Patris, ipse enarravit. Patrem enim invisibilem existentem ille quia in sinu ejus est Filius omnibus enarrat” (p. 189). "Deus.. .qualis et quantus est, invisibilis et inenarrabilis est omnibus quae ab eo facta sunt, incognitus autem nequaquam, omnia enim per Verbum ejus discunt … quemadmodum in evangelio scriptum est, Deum nemo vidit unquam nisi unigenitus Filius qui est in sinu Patris, ipse enarravit. Enarrat ergo ab initio Filius Patris, quippe qui ab initio est cum Patre, &c.” (p. 255). “Manifestum est quoniam Pater quidem invisibilis, de quo et Dominus dixit, Deum nemo vidit unquam. Verbum autem ejus. … claritatem monstrabat Patris... quemadmodum et Dominus dixit, Unigenitus Deus qui est in sinu Patris, ipse enarravit” (p. 256). The Greek original being lost, the text may be due either to Irenæus or to his translator, who frequently transcribes an Old Latin version of the New Testament when he comes to a quotation, even in cases where the extant Greek shews that Ιrenæus had other readings. Now the two former quotations coincide exactly (waiving Dei1) with most Old Latin anthorities2, even to the insertion of the characteristic nisi: the Deus of the third quotation is unknown to Latin texts of St John, and therefore doubtless represents the Greek. The only question that can reasonably arise is

1 Itself found in q.

2 Not it is true the oldest. But this is of no consequence except on Massuet’s groundless theory that Irenæus was known to Tertullian through the translation. There is no real evidence, as Dodwell has shown, for an earlier date than the fourth century.

whether Irenæus followed different texts in different places, or Filius was introduced by the translator. But the close proximity of the two latter quotations is unfavourable to the supposition of a variation in the original Greek, and the addition of Dei after Filius in the first passage savours of a corrective combination of a Latin Filius with a Greek 1. In neither case is the context available as evidence; for though it contains references to sonship, they are such as might easily be founded on the single word . Irenæus therefore read

  at least once, and there is no solid evidence that he ever read otherwise.

Hippolytus the disciple of Irenæus, in the fragment against Noetus now generally recognised to be the close of a larger work, which is almost certainly the lost early Syntagma against Heresies2, has the following sentence:

               

            , 

           (c. 5 p. 47 Lag.). It is to be regretted that the text depends on Fabricius’s editing of a modern copy of a single Vatican MS.; and the context is neutral. There is however no sufficient reason for doubting that Hippolytus read , but without the preliminary article. The Syntagma must have been written in the last decade of the second century3: the later Hippolytean remains are barren of evidence.

Clement himself quotes the whole verse once only (Strom. v p. 695), and then reads   . He adds that St John gives the name   to    , and this remark explains the combination of     with

1 Compare the similar case of Origen pp.35 f., 38.

2 See especially Lipsius Zur Quellen kritik d. Ερiphαnios, 37 ff.; Die Quellen d ält. Ketzergesch. 128 ff.;

3 So Lipsius, Q.Ep. 33-43, and much better Q. Ketz. 137 ff. Harnack (Zeitschrift f.d. hist. Theol. 1874 101 ff.) places it in the following decade: but after Volkmar, he refers the fragment against Noetus to a supposed treatise against all Monarchians, for which, if I understand him rightly (p.183), he accepts the date assigned by Lipsius to the Syntagma.

1 in a sentence in his tract, De divite salvando (p. 956),     ,

      ,       

                

  ... Here  and  stand side by side, and it may be that the two readings are combined: but it is more likely that  was inserted simply to soften the peculiar combination   ; just as elsewhere Clement (Exc. Theod. p. 969), in controverting the Valentinian interpretation already cited, inserts

, perhaps from the familiar Alexandrine form   founded on John i 1:

         ,

        , , ,

             …

       . And the process is carried a step further in an allusion which drops  but retains  (Paed. i p. 102):

              

; It will be observed that there is no trace of  except in the passage from the tract De divite, where the subject, , would have rendered the introduction of ,, inappropriate

Origen’s extant quotations of the verse are confined to his commentary on St John’s Gospel and his treatise against Celsus. Commenting on John i 7, he transcribes the whole passage 15—18 (iv 89 Ru.), reading ό

  . Unfortunately we do not possess his exposition of the passage itself, his third, fourth, and fifth tomes being lost. The sixth tome begins, after the preface, with i 19, treating the ‘witness of John’ as a second witness of his, that is, of the Baptist, and arguing against Heracleon who had attributed v. 18 (though strangely not 16, 17) to the Evangelist. He thus sets up a former witness of John, as

          ,

       

1 The same combination occurs, as we shall see (pp. 43 f.), in early Latin authorities.

 (or  )          (iv 102). The variation of reading is here significant. The Benedictine text adopts    from the Bodleian MS.1, while Huet reads  2 with the Paris MS. It is hard to believe that in a verbal citation of this kind Οrigen would have inserted the superfluous   and    is quite like a scribe’s correction of

 ; while thisphrase is too peculiar to have been substituted for   , yet might easily be written by Origen, either as a combination of the two alternative readings which certainly existed in his time, or to provide against possible misinterpretation. No inference can be drawn from the loose form of expression a few lines further down, when he pleads for the consistency of supposing

            (the Baptist)

       . In his 32nd tome the description of St John as reclining      occasions the remark that he

     ,

         ,

              (iv 438), where the selection of the term  confirms what appears to be the reading of all the MSS. Again in the second of the books against Celsus (c. 71 i 440 Ru.), which are transmitted in a different set of MSS. from those of the commentary on St John, we find:

                 

                 

       

      . Such is the reading of one of

1 Prima facie the lost Venice MS. used by Ferrari for his Latin version might appear to have read the same, as Ferrari has Filius Dei. But it is morally certain that he would have rendered   likewise by Filius Dei; since in the two other quotations, where there is no  to help him, he gets rid of  by simple omission, adding nothing after Unigenitus.

2 The silence of the collator of the Barberini MS. Favours this reading, as he can have had no other standard than Huet's edition. But the collation is evidently too imperfect to be trusted negatively.

Höschel's two MSS., confirmed by Gelenius’s Latin version, Unigenitus quippe Dei Deus; Höschel's other MS. merely substituting

  for ό  . The Benedictine text has the received reading ό   , but only on the authority of the Basel and Paris MSS., two closely related representatives of a single archetype, abounding in excellent readings but also in manifest corruptions. The silence of De la Rue as to his other MSS. (about six) implies the absence of at least any recorded difference from Höschel's readings. The combination of  with    in the closing paraphrase moreover suggests the presence of  following on the initial 1. To these four quotations may be added the following places,— the list is doubtless not exhaustive, — where the detached phrase is used.

             

  (Cels. iii 37 p. 471 Ru.).         ,

    (Cels. vii 43 p. 725).    ,

     ,    (Cels. viii 17 p. 755).

               

 2

                

    ,   ,

           

1  …  singles out . Or ..

2 Origen can hardly be introducing here the language of an actual hymn, as the context shews. Celsus has been rebuking the Christians for their scruples against consenting to join in a pæan to a heavenly body or a goddess,

         ,

               

. The reply is       ,

        ,

          ,     ,

     ,    (Ps.exlviii 3),       (so read for

  and   of the MSS.)

           ..….    

      ,        ,

  . Then follows the passage in the text, as an answer to Celsus's second sentence.

 (Cels. viii 67 p. 792) : for    Hoeschel has  , probably rightly. “Qui enim &c., et qui in medio etiam nescientium se consistit, Unigenitus Dei est Deus Verbum et sapientia et justitia et veritas &c.: secundum hanc divinitatis suae naturam non peregrinatur &c.: and after a few sentences,“Speciem autem dicimus Verbi et sapientiae et veritatis et justitiae et pacis et omnium quidquid est Unigenitus Deus” (In Μαtt. Com. Ser. 65 iii 883). “Unigenitus ergo Deus1 Salvator noster, solus a Patre generatus, natura et non adoptione filius est. . . Sed [Deus] . . . factus est Verbi pater, quod Verbum in sinum Patris requiescens annuntiat Deum quem nemo vidit unquam, et revelat Patrem quem nemo cognovit nisi ipse solus, his quod ad eum Pater caelestis attraxerit (quoted from the second book on St John in Pamph. Apol. pro Onig. c. 5) Lastly the most plausible instance of a seeming testimony to the reading  in any form of Origen’s writings is in Rufinus’s version of the commentary on Canticles: “Possumus. ... etiam hoc addere quod promurale (Cant ii 14) sinus sit Patris, in quo positus unigenitus Filius enarrat omnia et enuntiat ecclesiae suae quaecunque in secretis et in absconditis Patris sinibus continentur: unde et quidam ab eo edoctus dicebat Deum nemo vidit unquam: Ugenitus Dei Filius qui est in sinu Patris ipse enαrravit (iii 81) [iv.91]. Yet here too the evidence doubly breaks down. Had Filius stood alone, the Greek quotations would have suggested, that as in many undoubted cases of doctrinal phraseology, the translator’s very free hand introduced the Latin reading. But we have Dei Filius, that is, one more instance of a disguised .

I Two pages earlier Ρamρhilus quotes from the fifth book on St John the single sentence, “Unigenitus Filius Salvator noster, qui solus ex Patre natus est, solus natura et non adoptione filius est." If, as seems probable (for the manifestly incomplete state of our second book renders superfluous the natural suggestion that ii may be a corruption of v), the two passages are distinct, no allusion to John I 18 is perceptible here. If they are identical, the words that follow in the longer quotation suggest that Unigenitus Deus rather than Unigenitus Filius is the true reading, though     is also possible; in any case their own reference to i 18 contains not Filius but Verbum, which implies .

39

The first five books of Origen on St John were written about the second decade of the third century, the sixth not long afterwards, the later books, including the 22nd and therefore doubtless the 32nd, after 235, the treatise against Celsus between 244 and 249. Thus our quotations cover a long period, and proceed alike from Alexandria and from Palestine.

The epistle addressed to Paul of Samosata by certain bishops assembled at Antioch between 260 and 2701 quotes the verse with  and the article (ap. Routh R. S. iii 297). The doubts which have been raised as to the genuineness and age of the epistle appear to be unfounded. Its theology fits well into the third century; while the text of its quotations from the New Testament is mostly good, and entirely free, John i 18 excepted, from early ‘Western' readings. As in the case of Hippolytus, the text of the epistle appears to rest on a single Roman MS. Two other passages probably contain the phrase   , as has been already noticed (p. 19): but it has become detached from John i. 18; and there is at present no sufficient reason to doubt that

   as read there.

The Acts of the disputation alleged to have been held in Mesopotamia between Archelaus and Mani should perhaps be noticed here, though it is doubtful whether they belong to the last quarter of the third century or the first quarter of the fourth. The ancient Latin translation has (c.32) “Dominum nemo vidit unquam nisi unigenitus Felius qui est in sinu Patris”; where once more the presence of the Latin insertion nisi throws some doubt on the whole reading: elsewhere the quotations shew clear traces of modification, though not of transcription, from Latin texts of the New Testament. This part of the Acts has been printed only from a Vatican copy of a Monte Cassino MS.

In Eusebius of Cæsarea we have the last virtually Ante-nicene writer, that is, whose training belongs to the days before

1 It is unnecessary here to attempt greater definiteness, the chronology of the proceedings against Paul being singularly difficult.

Constantine. The clearest evidence for our purpose is furnished by two of his latest treatises, those against Marcellus, written in 336. Both treatises abound in the detached phrase‚   ; but there is no reference to John i 18 till a few pages after the beginning of the second and longer work, De ecclesiastica theologia, where Eusebius says

        ,    

     ,   ,   (ρ. 67D). No one can doubt that Eusebius here adopts the reading : but it is wholly arbitrary to reject the clause

   as a gloss of scribes1. It would be difficult to find any similar interpolation of theirs in a scriptural quotation, especially if it introduced for once a reading which elsewhere they persecute. It is more likely that Eusebius, familiar as he must have been with the reading  through his Origenian lore, took advantage of this first quotation to indicate in passing that, while he adhered to his own reading, he did not care to rest his case upon it2. Accordingly, having thus appealed to “the evangelist”, he goes on at once to claim the yet greater authority of “the Saviour Himself” whom he supposes to have spoken John iii 16, which contains

    . At p. 86A he again quotes the verse, with a context which confirms , and again at p 142 c, with a neutral context; and  recurs for the fourth time in a clear allusion at p. 92 D. On the other hand in a solitary passage the sentence

               

   

1 It has been urged in favour of this conjecture that in a quotation of 1 Tim. i 15 by Origen (c. Cels. i 63 p.378 Ru), Hoeschel's text has

             . such a wild collocation as the supposed "gloss" is evidence of nothing. It an be only a blunder of a scribe or the editor, probably  С  for С.

2 Marcellus (see ρρ.20, 22) used the phase    (Eus. c. Marc p. 19 c); and his theological tendency was to evade the idea of Divine Sonship. On both grounds there would be force in a refusal of Eusebius to haggle about the various reading.

  κ.τ.λ.       (Eph. iv 6) is continued by

              ,

           ,

        (ρ. 174 f.). It is vain to urge that   is not the same as    , where the title maintained for the Son is found verbally in a single verse of Scripture and where the preceding title is likewise transcribed from Scripture (2 Cor. i 3 &c.) with the exception of the word  used just above.1 Corruption of text is also unlikely, as  could hardly stand here in both subject and predicate, to say nothing of intrinsic improbability2. Doubtless therefore Eusebius did on this occasion for a special purpose avail himself of the reading3 to which he habitually preferred another. It probably never occurred to him that one of the two must be right, and the other wrong: an inability to part absolutely with either of two respectable traditions is not unusual in his writings. Lastly  stands, with neutral contexts but probably rightly, in two of Eusebius’s Commentaries, on Psalm

1 Indeed  as so little force here, as an adjunct, that it becomes suspicious. It may represent  (CC for C); or Euecbius may have written     [1Cor.viii 6, quoted p. 93]        κ.τ.λ., the intervening words ό     being lost by homœoteleuton.

2 The concluding words     are probably all in antithesis to the second clause    . . . and, if so, they imply ; whether they refer to the alternative readings (as at p. 67D), or simply take up  from the beginning of the clause. But it is not impossible to take   as in antithesis to the first clause   ... .

3 Passages like the following shew that it could not have been a stumbling-block to his own mind on the score of doctrine, though

   had a sharper edge against Μarcellus: indeed the first (on which more hereafter) substantially contains it.

   ñH    ,         ,

        (Dem. Eu. Iv 3 p. 149A).

        ,

           ,    ,

    (Eccl. Th. P. 62A).

                  

 (Com. In Es. 375D).

42

lxxiv (lxxiii)) 111 without the article, and on Isaiah vi 12 with the article.

Ι In Montfaucon, Coll. No. Patr.I 440. A freely condensed extract in Corder’s Catina, ii 535, has the article.

2 In Montfaucon, ib. II 374. comment of Procopius, p. 91, founded here chiefly on Eusebius but perhaps also on Origen, has

       ...

43

Note B

The details of Latin evidence

The Latin patristic evidence is properly speaking only a branch of the evidence of Latin versions, So far as it refers clearly to St John’s own text, it supports  exclusively. Tertullian’s citations, all occurring, as is not unnatural, in the single treatise against Praxeas, are in no case quite verbal; but they leave no reasonable doubt. He says (not to quote references to the first clause only), Apud nos autem solus Filiiis Patrem novit, et sinum Patris iρse exposuit, et omnia apud Patrem audivit et vidit”, &c. (c.8); “Deum nemo vidit unquam: quem Deum? Sermonem? Atquin, Vidimus et audivimus [et contrectavimus] de sermone vitae, praedictum est: sed quem Deum? scilicet Patrem apud quem Deus erat Sermo, unigenitus Filius qui sinum Patris ipse disseruit” (c.15, some early editors for sinum reading est in sinu, and Rigaut [1634, ? on MS. authority] simply in sinum); “Hujus gΙοτiα visa est tanquam unici a patre, non tanquam Patris: hic unius (? Unicus1) sinum Patris disseruit, non sinum suum Pater, præcedit enim, Deum nemo vidit unquam” (c. 21). Cyprian does not quote the verse; but had he read Deus, he would probably have used it in his Testimonies (ii 6) under the head Quod Deus Christus, the texts of which from the New Testament are Matt i 23, Jo. i 1, (x 34—38,) xx 27ff; Apoc xxi 6.f . The same may be said of Novatian (de Regula Fidei 11, 13, 14, 18, &c.), and is probably to be inferred from the only passage

1 Ραmèle’s reading unus, which is probably likewise conjectural, deserves mention, as it might represent  (see next note): but

Unicus makes as good sense and was more likely to be altered.

.

in which he alludes to this clause, being part of an argument to shew that Christ is idem Angelus et Deus: “Manifeste apparet non Patrem ibi tunc loquutum fuisse ad Agar, sed Christum potius, cum Deus sit; cui etiam angeli competit nomen, quippe cum magni consilii Angelus factus sit, angelus autem sit dum exponit sinum Ρatris, sicut Joannes edicit: si enim ipse Joannes hunc eundem, qui sinum exponit Patris, Verbum dicit carnem factum esse, ut sinnm Patris possit exponere, merito Christus non solum homo est sed et angelus; nec angelus tantum sed et Deus per scripturas ostenditur, et a nobis hoc esse creditur” (c. 18). It will be observed that to both Tertullian and Novatian the last words of the verse must have stood as sinum Patris [ipse] exposuit (Τert.1 Νov.3). or sinum Patris ipse disseruit (Τert.2, perhaps his own rendering, as it occurs nowhere else), and we have the same construction with a different Latin verb in α, the oldest of existing Old Latin MSS., which reads "Deum nemo vidit umquam nisi unicus Filius solus sinum Patris ipse enarravit1” These primitive forms of the Old Latin rendering were smoothed away by degrees. The inserted nisi2, probably derived from vi 46, vanishes only in the Vulgate and one or two other late revisions (fq). Unicus3 is exchanged for unigenitus, and sinum for qui est in sinu, with hardly an exception. Solus lingers only in

1 Tischendorf calls attention to the coincidence of this part of the rendering of a (he might have added Tertullian and Novatian) with the omission of   in א*, suggesting that  was read as : and apparently with good reason, for א* has readings here-abouts in common with what must have been the original of the Old Latin in an early form, and solus stands for  in many authorities in Mark ii 7, and several in x 18, both passages having a similar turn. The correction was probably suggested by , for transitive verbs used absolutely are always a distress to scribes and translators. As we have seen, Clement likewise supplies

    in interpretation.

2 There is no Greek authority of any kind, as far as I am aware, for nisi: it might of course be introduced from vi 46 in Latin as easily as in Greek.

3 Retained only, it would seem, by the Manichean Adimantus as cited by Augustine (c.Adim. viii 2 t. viii p.120 bis), Sinum Patris gives place altogether to in sinu Patris (in Patre c).But negative statements as to the Latin quotations could not be made quite confidently without disproportionate labour.

mm, and probably other revised MSS. of the same group. The final verb is represented pretty constantly1 by enarravit, varying occasionally (after ipse, it will be remembered) into narravit. The final form, as it stands in the present MSS. of the Vulgate, answers exactly to the prevalent Greek text: "Deum nemo vidit umquam; unigenitus Filius, qui est in sinu Patris, ipse2 enarravit." This statement includes the Latin Fathers of the fourth and following centuries, and it is needless to give references: various types of Old Latin are represented, as the names of Victorinus, Vigilius, Hilary, Ambrose, and Augustine will sufficiently shew.

1 Adimantus (1. c,) has adnuntiavit: Victorinus once (adv. Ar. i 2) exposuit with Tertullian and Novatian, elsewhere enarravit.

2 Ipse similiarly represents  in ix 37., and in scattered authorities elsewhere. Like , which is to be found in Greek quotations but not MSS., it was evidently suggested by the apparent sense.

NOTE C

Some details of-Æthiopic evidence

Dr Wright has most kindly ascertained the texts of the two MSS. at Cambridge, and of the nineteen in the British Museum. They singularly illustrate the truth of Dr Treggelles's account of the Æthiopic version (Horne's Introduction iv 139f.), which has been questioned of late, being all paraphrastic, and exhibiting no less than 12 combinations of readings, owing in part to the addition of pronouns, and the insertion of conjunctions in

various places. Nineteen MSS. are of the 17th century or later: of the remaining two, ascribed to the fifteenth, one (B.M.Or. 525) agrees prima manu with the Polyglott. The accusative particle is here prefixed to

  doubtless owing to a misinterpretation natural in a language incapable of expressing

 otherwise than by a word like unicus (wahed), since it was not to be supposed that "the only God” denoted the Son. To   (or--),six other MSS. add  followed by wahed, which in this second place probably

stands for  or ; two of them (including the other 15th century copy, B.M. Or. 507) having

 , the other four the accusative form. This interpolation supplied another possible construction for the accusative unicum Deum: it could be taken either simply in apposition to the previous  (Deum nemo vidit unquam, unicum Deum: [Filius unicus] qui &c.), or as the object of  (unicum Deum [Filius unicus] qui est in sinu …enerravit), or as the object of an intermediate clause (unicum Deum [sc. vidit] Filius unicus (or unis):qui est &c.): all three constructions seem to be indicated by punctuation and conjunctions in different MSS. An eighth MS.

omits , retaining   wahed. The remaining thirteen likewise omit . The probable sequence was as follows, the position of the second wahed in all known MSS. being fatal to other interpretations of the facts which might be suggested. The original text (preserved now, as far as the MSS, yet examined shew, only with the accusative modification) had   , the Memphitic reading. With this was next combined the alternative reading , accompanied by wahed, either a relic of the early reading mentioned in Note B or a like but independent interpolation: similar couplets of readings originally alternative are not uncommon in this version1. The first wahed would then be dropped as a needless superfluity in MSS. which escaped the accusative prefix: and lastly the further omission of   would reduce the phrase to a familiar shape. The evidence is not very important; but its history is instructive.

The verse is closed by a gloss from Heb. i 2 in one of the seventeenth century MSS. which omits 

 (B.M.Or. 521).

1 It is possible, but much less likely, that the Æthiopic had originally the double reading, and that  wahed was then omitted in some MSS.

NOTE D

Unicus and unigenitus among the Latins

The varieties in the Latin rendering of  in the New Testament are sufficiently interesting. to be given in full. Sabatier's references have of course been freely used.

I Passages referring to our Lord

John i 14     .

1. unici (a patre) Tert.½ (Prax. 21) - Fr.Arian. (Mai, S.V.N. C. iii 2 228) Hil.½(Trin. i 10 in comment.).

   unici (patris sic) e.

   unici filii (a patre) a.

   unici nati (a patre) Oros.1 (Ap. de arb. lib. 613 Hav.).

   
   

2. unigeniti (a patre) b c f vulg. Tert .½ (Prax.16) Novat. (Reg. Fid. 13) Hil.½ (Trin. i 10 text) Amb1(i 1204 F)

Iren. lat.2(42, 315) Aug(ad l. &c.) Hieron.1(Eph. v 33) &c.

John i 18 03          .

A unicus (filius) a Adimant.1(ap. Aug. viii 120).

unigenitus (filius) b c e f Tert.1(Prax.15: cf.7) Hil.(Ps.138 § 35 &c.) Victorin. Iren.lat. Amb. Aug. &c.

John iii 16      .

A (filium suum) unicum a b d e m g1 gat mm mt Tert.1(Prax.21) Rebapt.1(13) Fr.Aria'n.(226) Lucif,1(151 col.) Hil.cod. al.3

B (filium suum) unigenitum c f ff vulg. Hil.1 (Trin. vi 40 ed.) Amb.(ii 406, 626) Aug. &c.

John iii 18       .

1. unici (filii Dei) a d Tert.(1,c.) Cyp.(Test. i 7; iii 31) (Fr. Arian. 226) Lucif (I.e.)

   
   

2. unigeniti (filii Dei) b c e ff m vulg. Iren.lat.(325) Amb. (i 762) Aug.(ad I.) Vig.(Trin.213 Chif.) &c.

1 John iv 9        .

1. (filium suum) unicum m Lucif.(140).

   
   

2. (filium suum) unigenitum vulg. Aug.(ad I.)

II Other passages

Luke vii 12   (or . .)   .

A (filius) unicus all, including Amb. (waiving order).

Luke viii 42    .

A (filia) unica all, including Amb. (waiving order).

Luke ix 38   ,     (or  ). A unicas (mihi est) all (waiving order).

Heb. xi 17       .

A unicum (without filium, or summ) d Ruf.[Orig.] (In Gen. Hom. i 1, ii 81 Ru.) Aug (C.D. xvi 32). B unigenitum vulg.

In the canonical books of the Old Testament + b z I b , the only Hebrew original of , is uniformly rendered by unigenitus in the Vulgate where an only son or daughter is meant (Gen. xxii 2, 12, 16; Jud. xi 34; Prov. iv 3; Jer. vi 26; Am. viii 10; Zech. xii 10). Singularly enough the LXX has

 ( Prov.) in all cases but that of Jephthah's daughter, though  was used by one or more of the other translators in at least five of the other places (no record being known for Gen. xxii 16; Zech.). But at least some form of the LXX must once have had  for Isaac1 (the

1 Gregory of Nyssa (De Deit. F. et sp. S. iii 568 Migne) has Gen. xxii 2  , ,     ,

  where , if only a gloss on , must at least have been found by Gregory in his MS., for he remarks in his comment          ,

      ... This case

Vatican MS is wanting here), for we have clear Old Latin authority accidentally preserved for unicus in Gen.

xxii. 2, 12 and Judges, though most Old Latin quotations follow . Unicus is also the Old Latin word in three of the four remaining passages, all peculiar, Ps. xxii (xxi) 21; xxxv (xxxiv) 17 (solitarius Hier.); xxv

(xxiv) 16 (solus Hier.). In the Apocrypha the uniform unicus of the Old Latin was not disturbed by Jerome; Tob. iii 15; vi 10 cod.; viii 17 or 19 (duorum unicorum, Tobias and Sarah) ; and even Sap. vii 22.

Thus throughout the Bible unicus is the earliest Old Latin representative of ; and unigenitus the Vulgate rendering of ۰ת۰ר however translated in Greek, except in St Luke and the Apocrypha, where Jerome left unicus untouched, and the four peculiar verses from the Psalter (lxviii [lxvii] 7, and the three already mentioned), in which he substituted other words. But unicus had been previously supplanted by unigenitus in one or more forms of the Old Latin in all the five passages where it has reference to our Lord, all occurring in St John’s writings; and in the Prologue of the Gospel the change took place very early.

These facts would prove, if any proof were needed, that  was the reading of the MS. or MSS. from which the Old Latin version was originally made; for unicus Deus1 could never

renders it not unlikely that Irenæus is following a similar double reading when he speaks of Abraham. (233) as

        ,      ...

        ... In Jud. xi 34 the Alex. and other MSS add to 

without a conjunction  , and others  ,  .

1 In Dr Swainson’sa History of the Creeds attention is called to a “not infrequent punctuation" of MSS. by which unicus is strangely separated from the preceding Filium ejus and joined to the following Dominum nostrum (pp. 163, 166, 365). He points out that this construction occurs in two sermons wrongly attributed to St Augustine: in one (240 in t. v p. 394 Ap.) it is at variance with the interpretation, and must be due to a scribe; in the other (t. vi p. 279 Ap.), a very late cento, it belongs to an extract from Ivo of Chartres, a pupil Of Lanfranc. It is indeed, I find, as old as Rufinus, for he labours (Com. In Symb. 8 p.71) to justify it, though evidently preferring (6 ff.) to take unicum with Filium. But unicum Dominum nostrum can hardly be more than a Latin

have been a designation of our Lord, and moreover it was actually applied to the Father in the Creed of Carthage in Tertullian's time (De Virg. vel. 1; Adv. Prax. 2 f.). But they also give additional interest to the almost uniform rule that unicus belongs to native Latin Creeds, unigenitus to comparatively late Greek Creeds translated into Latin, both alike having but one original, the  of St John's third chapter, if not also his first. It is needless to enumerate the various forms of what we call the Apostles’ Creed, which have been several times collected. They all have unicus1, (mostly in the order Filium ejus unicum as John iii 16, but the Aquileian form given by Rufinus2 unicum Filium ejus as iii 18, and the Poictiers form used by Venantius Fortunatus [Hahn, Bibl. d. Symb. 33; Heurtley, Harm. Symb.55] unicum Filium only) with the exception of two peculiar Gallican documents, closely related to each other, which have unigenitum sempiternum (Hahn, 35f.; Heurtley, 68f.)3. In Tertullian we have seen unigenitus (cf. De An. 12; Scorp. 7), possibly a word of his own coinage, side by side with unicus. But the influence of the Creed remained strong: a century and a half later Lucifer seems to have only unicus, which he repeats incessantly. Augustine vacillates between the Creed and his Latin MSS of the 'Italian' revision. Writing de Fide et Symbolo in 393 he puts unigenitus into the Creed but promptly explains it by the equivalent to which his hearers were more accustomed

blunder, arising from the separation of unicum from Filium by the genitive ejus and the immediate proximity of Dominum, together with the latitude of sense in unicus. In some Spanish Creeds the insertion of Deum et before Dominum (Swainson 164, 323) brings unicum and Deum into contact: but the resemblance to   can be only fortuitous.

1 So also the Latin original of the Sirmium formulary of 357 (Hil. De Syn. 11 p. 466A), notwithstanding the Greek cast of its language.

2 This order cannot be safely assumed for the Roman and 'Eastern' forms to which he sometimes refers.

3 In the Te Deum we have veram et unicum Filium in the common text, probably rightly. but in the present state of knowledge unigenitum must be admitted as an alternative reading. The Gloria in excelsis has Domine Fili Unigenite Jesu Christe, without appa rent variation.

(“credimus etiam. in Jesum Christum Filium Dei, Patris unigenitum., id est unicum, Dominum nostrum c.3 t. vi

p. 153A), and twice afterwards repeats unigenitus. Nearly thirty years later in the Enchiridion he employs unicus (34, 35, 36 bis) till he has to quote John i 14, when he takes up for a moment the unigenitus of his version (36 s.f.), but in the next sentence slips back to the Creed by again combining both words, unigenitus id est unicus: and in the rest of the treatise he uses only unicus when commenting on the Creed (38, 56), unigenitus only with Verbum, (41) or else absolutely (49, 56, 103, 108). But the influence of the Greek controversies of the fourth century upon Latin theology, the convenience of the antithesis to ingenitus, and the revision of Latin biblical texts secured the ultimate victory for the more explicit term unigenitus, except in the Creed itself. It is the word adopted in several private formularies, all imbued with the results of Greek thought; those of Pelagius (but with Deum, Hieron. Opp. xi 202 Vall.), Auxentius of Milan1 (Hil. Lib. c. Aux. 14: cf. Caspari, Quellen u. s. w. ii 301), and Ulfilas (in Caspari 303)2. And from the fourth century onwards it is the constant rendering of  in all the Latin translations of Greek Creeds or other formularies, with hardly any exceptions and those in secondary authorities. Thus ten out of the eleven versions, or recensions of versions, of the original Nicene Creed collected by Walch (Bibl. Symb. 80 ff.) have natum ex Patre unigenitum, the eleventh3 omitting the word: and five4 out of the seven ver-

1 The closely related formulary of Germinius of Sirmium has however unicus (Hil. 0p. Hist. Xiii-xv: of. Caspari 302).

2 Another attributed to Damasus and several other Fathers (Hahn 185) has unigenitus, but it appears to be a translation.

3 As given by Lucifer (De non pare. p. 204 Col). Singularly enough unicus occurs in what can be only a quotation from the Nicene Creed following on the already cited use of unigenitus by Augustine in the De fide et symbolo (6 p. 154E): “naturalis ergo Filius de ipsa Patris substantia unicus natus est, id exsistens quod Pater est, Deus de Deo, lumen de lumine.” So also Gregory of Eliberis, if he is the author of the treatise De fide orthodoxa in the Appendix to Ambrose's works (ii 345).

4 Dionysius Exignus omits; the Code of Canons &c. of the Roman Church printed with Leo's works substitutes unicum.

sions or recensions of the 'Constantinopolitan' Creed, as quoted by Hahn (113), have Filium Dei unigenitum. The two renderings of  were unconsciously retained by Latin Christianity in the two Creeds throughout the Middle Ages, and the double tradition is still preserved by corresponding renderings in our own tongue.

NOTE E

On   in theNicerie Creed

The second part of the original Nicene Creed begins thus:      ,

   ,     ,/       ,/

  ,   ,     , ,  ,   .

.

Then follows the recital of the Incarnation.

If now we withdraw the parenthetic clause       , the words  and

 become contiguous. Is this contiguity accidental, so that  alone goes with , and a new clause in apposition is formed by   , or should the eight words

        be all read continuously, so that  belongs to ? Neither alternative presents any grammatical difficulty; and thus the question must he decided by analogy and sense. The first step evidently is to investigate the probable origin of the passage. The enquiry must occupy a space disproportionately great if   alone be considered: but it has to do with matters of sufficient historical interest to reward minute examination on other grounds.

It is certain (1) that the bulk of the Nicene Creed was taken from earlier formularies, one or more; and (2) that the three1 clauses       ,   , and    were novelties introduced by the Council with the special purpose of excluding ambiguity,

1 Three for some purposes, howsoever the second and third may be grammatically related.

Athanasius in his old age, nearly half a century later, explained how the introduction of the new phrases had arisen (De, Decr. Nic. Syn. 19ff; Ad.Afr. 5 f.), and justified them, as he or others had evidently done at Nicæa, by reference to similar language of Theognostus, Dionysius of Rome, and Dionysius of Alexandria respectively (De Decr. 25 f): and this anxious appeal to theological writers sets in strong relief the absence of authority de- rived from public Creeds. In a different quarter the unwonted language of the three clauses elicited from Eusebius a somewhat reluctant apology in the epistle which he addressed to his own diocese shortly after the Council (Ep. ad Caes., preserved by Athanasius De Decr. pp. 238 ff. and Socrates H.E. i 8). The testimony thus doubly borne renders it highly unlikely that the Nicene Creed contained other novelties not mentioned; and however modified in arrangement, the whole of its remaining contents may be assumed to have been taken from Creeds already in use.

The scattered and confused memorials of the Council afford little information as to the Creeds brought forward in the course of the discussions. Theodoret (H.E. i 6) mentions an exposition (   ) which was presented to the assembly by the small group of bishops comparatively friendly to Arius, led by Eusebius of Nicomedia; and which was at once torn up. Eustathius of Antioch, an eyewitness, cited in Theodoret's next chapter, tells the same story of " the writing () of Eusebius's blasphemy," meaning evidently the same document1, which was probably an elaborate private statement of doctrine. From the above-mentioned pastoral letter of Eusebius of Cæsarea, the leader of the middle party, we learn more. Its purpose is to explain the circum-

1 Identical also, it would seem, with the “epistle” of Eusebius of Nicomedia from which Ambrose (De Fide iii 125) cites a sentence as having furnished the term  to his opponents. What is said by Philostorgius (H.E. i 7), or rather by Photinius abridging his words about the winning over of Hosius and other bishops by Alexander at Nicomedia before the Council has no necessary reference to the term itself.

stances which had led him after some hesitation to subscribe the Conciliar Creed, as he was afraid that incorrect rumours might cause misunderstanding1. "We first," he says, "transmit to you the writing concerning the faith which was put forward by us, and then the second, which they have published after putting on additions to our expressions2. Now the writing presented by us, which when read in the presence of our most religious emperor was declared to have a right and approved character (     ), was as follows. As we received from the bishops before us both in our first catechetical instruction and when we were baptized, and as we have learned from the Divine Scriptures, and as we both believed and taught in the presbyterate and in the office of bishop itself, so now likewise believing, we offer to you our faith; and it is this.’” Eusebius then transcribed a Creed, to which he added a few lines of explanation and protestation3. When “this faith”, he tells his diocese, had been set

1 This is not the place to examine the characters and beliefs of the actors in the great Council. But it is worth while here to observe that though Eusebius differed on a grave point of doctrine from Athanasius, and probably yet more from Athanasius's non-Alexandrine allies, the difference which determined the attitude of the two men respectively in regard to the proceedings of the Council was not of doctrine but of policy. When the policy of Eusebius had at length been clearly overruled, he had to decide how he could most nearly conform to its spirit; by giving in his adhesion to the conclusion of the majority, or by recording his protest against it. He decided that the former course was the best now open, provided that he could receive sufficient assurance that the new terms were not meant to carry a sense inconsistent with his own belief, misgivings having perhaps been raised in his mind by wild language on the part of such men as Marcellus. The assurance was given, his conscience was relieved, and the accession of his name furnished a guarantee that the new Creed was not to be understood as a rejection of the elder theology. It was quite consistent with this decision that he should desire, on public and on private grounds, to be known as still regretting the eclipse of the policy which he represented.

2            ,   ,

      .

3 The defensive tone of this document implies accusations flung about in the previous debates. The later controversy with Marcellus may well have had a prelude at Nicæa; nor is it likely that the animosity of Eustathius (Socr. i 23) began after the Council.

forth by him (     ), there was no room for gainsaying. The emperor, followed apparently by others1, declared his entire agreement with it, and "urged all the bishops to give their assent to it and to subscribe to its articles and to express concurrence with them in this very form, with the insertion of the one single word "; which word he proceeded to interpret by rejecting various erroneous senses2. Such, Eusebius says. was the wise discourse of the emperor; "but they, under pretext of the addition of , have made the following writing3," i.e. the Nicene Creed. He then relates how, as soon as the Creed had been propounded, he or his party (the pronouns 'we' and 'they' are throughout ambiguous) enquired minutely about the intended meaning of the new phrases, and on receiving satisfactory answers thought it right to give consent, having peace always in view.

From this narrative it plainly appears that Eusebius presented a declaration of his own faith as his namesake of Nicomedia had done; that the kernel of this private declaration was a public Creed, the same with which he had been conversant in his own Church at all stages of his life; the Creed therefore of Cæsarea from at least the latter part of the third century; that

1 This seems to be involved in the words    …, although no second corresponding clause is extant. The shape of Constantine's proposal was probably suggested by the debates which had followed the reading of the exposition by Eusebius of Nicomedia. But much may have been due to the advice of Hosius, who enjoyed his special confidence, and who, whatever may have taken place at Nicomedia (see p. 55 n. 1), had doubtless not returned without instruction from his previous confidential mission to Alexandria (Eus. V. Const. ii 63-73; Soer. i 7 1; Soz. i 16 5).

2 Such must be the force of the evidently careful though ungainly language

             

,      . Following  and joined with

   must as usual denote some express act of agreement or compact.

3

               

    . Late usage would allow  to express the mere connexion of facts without implication of motive: but the equally common stricter sense is suggested by the context, as also by the form of the sentence.

Constantine advised the Council to be satisfied with adopting this Creed as it stood, inserting only the term

, this addition being evidently proposed in consequence of a previous discussion; that the Council, under colour of following the advice, did in effect go much further in the way of composition, so that the resulting document could be called a “writing” which they "made"; and yet that it might with equal correctness be described m the Creed of Caesarea with additions.

The truth of the principal statements is confirmed by historic probability and by internal evidence. An appeal to a venerable existing document, such as the traditional Creed of Cæsarea, was exactly in the spirit of the conservative policy espoused by Eusebius; nor could he easily find a better resource in endeavouring to draw to his side the greater part of the Council. In like manner the adoption of this Creed as a basis by the Council would naturally ensue, in approximate compliance with the emperor's recommendation. The Creed which Eusebius transcribes is simple in form, unlike the personal profession which encloses it1. Echoes of its phrases can moreover be distinctly identified in references made by Eusebius elsewhere to a testimony of “the Church [of God]” which must be a public Creed, and is not the Nicene2. Its verbal coincidences with

1 By a curious oversight Hahm (46 ff.) has included in the Creed part of this personal profession, and so been led to unfounded doubts as to the public character of the Creed as it stands.

2 These coincidences appear to have been overlooked. The variations are only of order, and that among complete clauses, and they have no perceptible significance. The passages as follows:

           ,

            ,   ,

   (De Eccl. Theol. P. 62c)

[five lines missing from :    …..]

  ,       , …

       ,

       ,      ,

     ,    ,   ,    ,

    (p.66 A, B).    [   ]

    ,       ,

     (p. 108 B). Another probable trace occurs in the Demonstratio Evangelica, p. 215 B.

the Nicene Creed, as is well known, are at least too large to be accidental1.

But it is equally certain that one or more other Creeds furnished their quota to the result. Prominent among the leaders of the majority were the representatives of important sees, as Eustathius of Antioch, Hellanicus of Tripolis, Macarius of Jerusalem2, and Marcellus of Ancyra, not to speak of Alexander of Alexandria; and there would be an obvious fitness on such an occasion in combining with the Cæsarean confession well chosen forms of language consecrated by the use of other great churches. Indeed two of these sees possessed rights which their bishops could not willingly compromise by allowing Cæsarea to furnish alone a standard for universal use. merely because Eusebius was in favour with the emperor: all Palestine was subject to the supremacy of Antioch; and the metropolitan jurisdiction of Cæsarea over the rest of Palestine was balanced by privileges peculiar to Jerusalem, which were ratified by the seventh canon of the Council. The silence of Eusebius as to the employment of any additional Creeds by the Council is of little moment, for his narrative is palpably incomplete, though sufficient for his purpose of shewing first how he had made the best stand he could for the old Creed of his church, and then how it was that he had nevertheless in good faith subscribed the Conciliar Creed. It is at least possible that the omission of certain phrases used at Cæsarea, as elsewhere,

   (Col. i 15) and     (1 Cor.

            : and doubtless others might be found.

1 At the end of these Dissertations will be found the Creed of Cæsarea in full and also the Nicene Creed printed so as to shew its coincidences with the Cæsarean base by diversity of type. The concordances and differences are exhibited in another way by Dr Swainson, pp. 65 f.

2 The prominent part taken by Macarius against the Arians in the Council is attested by Theodoret (H. E. i 18; cf.ii 2, 4) and Sozomen (H.E. i 13 2; ii 20). he was moreover apparently on terms of friendship with Constantine and Helena (Sozom. ii 1 7; 4 7; Theodoret i 15 f.; Euseb. V. Const. Iii 29 ff.).

ii 7:cf.Eph.iii11; Heb.i 2), arose from a dread of their lending themselves too easily to suspected interpretations. But the insertions and alterations in the latter half of the Creed all correspond with fair exactness to extant phraseology of Syrian and Palestinian Creeds1, though they cannot be traced to any one of the very few extant formularies. It is of course possible that other lost formularies of a similar type may likewise have supplied materials2.

These facts enable us to understand the manner in which the Council changed those articles of the Creed that touched on the immediate subject of controversy. The Caesarean Confession ran,

         ,   ,   ,   ,

 ,   ,        .

Not only were the phrases mentioned above omitted, and

1 Apostolic Constitutions and Jerusalem (compare Antioch in all forms)    for    ; Ap. Const. insertion of          ; Antioch (at least Cassianus and Eusebius of Doryæum have

 ) insertion of    ; Ap. Const. and Antioch (Lucianus and Eus. Dory1) insertion of

; Jerusalem  for   ; Ap. Const., Jerusalem, and Antioch (Lucianus and Cassianus)    for   ; Jerusalem  for   (  being likewise omitted by Cassianus); and Ap. Const. and. Antioch (Lucianus)    (at least these Creeds have

   ) for   . In the above enumeration ‘Eusebins of Dorylæum’ means the author of the

 against Nestorius, printed in the Acts of the Council of Ephesus (Mansi Conc. iv 1109): see Caspari, Quellen u.s.w. and Dissertation 11.

2 It would be rash to assume that there were no clauses on the Church, Baptism, &c. in the Cæsarean or other similar formularies. It is more likely that Eusebius presented only so much of his native Creed as related to the Persons of the Godhead, as sufficient for the special purpose of the Council; and. that the Council kept within the same lines. Compare the language of the ‘First’ Formulary of the Synod of Antioch in 341 (ap. Ath. De Syn. 22 p. 735 E),    ,

       . The Anathematism (doubtless suggested by a precedent in the closing exposition of Eusebius, as Mr Lumby points out, p. 50), being evidently intended as part of the Creed, rounds off what would otherwise be an abrupt termination.

with them     and   , but the surviving language reappeared in a different arrangement, including a new phrase1     , in addition to the three clauses which were the special creation of the Council. This arrangement bears no trace of having been devised with the sole purpose of carrying the new clauses. The rather loose and clumsy order of the Caesarean Formulary might seem to invite the substitution of a compact and methodical paragraph supplied out of other existing Creeds: and such a procedure would be in analogy with the course seen to have been pursued in the later articles. The first step would be to set the simple fact of our Lord's Divine Sonship2 in the forefront immediately after His name, in accordance with most precedents. Next would follow the declaration of the nature of His Sonship. Here even our imperfect evidence suffices to exhibit in outline what probably took place. The construction by which

    is followed by a predicate, in this case  [] is borrowed from the Jerusalem Creed, which has in like manner

          3. Probably the con-

1 New, that is, in relation to the Cæsarean Creed, but doubtless taken wholly or in part from another source,

for otherwise it would probably have been mentioned as new by Athanasius and Eusebius. The complete phrase

occurs in the Expositio Fidei of Athanasius himself (c. 1 p. 99 B: cf. Or. c. Ar. iii 9 p. 558 c,

      ); but so do similar forms not adopted at Nicaea, as

  ,    ,     . On the presence of   in the Jerusalem Creed at this time see note 3.

2. The extrusion of the clause setting Him forth as the Word, and the transfer of the following clauses to the Sonship, would find justification in almost universal precedent.

3 Touttée, the editor of Cyril of Jerusalem, in an excellent dissertation on the Creed of Jerusalem (p. 80), conjectures

  to have been introduced into the Creed from the Nicene Creed between 325 and the time, some quarter of a century later, when Cyril's lectures were delivered. The suppositopn is surely gratuitous. The presence of

    affords no grammatical argument, as our other evidence shews; the suggestion is sustained by no other Nicene echo in the Creed of Jerusalem; had any thing been interpolated from the work of the great Council, it would hardly have been a phrase so little conspicuous or characteristic; and any early Creed might easily take it at once from 1 Jo. v 20.

struction is the same in the Antiochian Creed of Lucianus1,

          . But at all events the Antiochian diction passes with great facility into the Nicene. It stands thus:-

  ,   ,    .

          ,   ,

  ,    ...

When once the evidently premature clause    had been deferred till the place which it held at Cæsarea and Jerusalem alike, and the inconvenient2 phrase    had been omitted, it was an obvious gain to shift   from its isolated position, now rendered doubly conspicuous by the removal of    , deprive it of its dangerous article, and employ it, in strict analogy with St John's own usage, as the chief predicate to    , combining it with the already present

   into the single phrase    3.

The other alternative now claims attention. The simple   of Jerusalem may have been preferred to the 

The exact date of Cyril's lectures cannot, I think, be determined, but it seems to lie shortly before 350: see Pearson De Succ. ii 21 2; Tillemont viii 779 f.; Tonttée Diss. exx ff. The most probable year is 348, which is preferred by Touttée, though partly on untenable grounds.

1 The doubt of course arises from the bare possibility of taking    as the sole predicate (   being excluded from direct pre dication by the sense), in which case    would become an addition in apposition. But this construction is virtually condemned, if 1 mistake not, by the order of the words. In both the local Creeds    seems to hold a weak place, as a secondary predicate only, though theplaces are not identical/ The ommission of these words at Nicæa, whether suggested by dogmatic prudence or not, was an undoubted gain as regards grammatical clearness. It may also be owing to a grammatical impulse that Hilary omits them in his version of Lucianus' Creed (De Syn. 29 P.478c).

2 See last note.

3 What follows hardly needs comment.    is succeeded by two clauses of similar form, as in both the Cæsarean and the Antiochian Creeds; but no actual phrases are borrowed from Antioch, and but one,   , retained from Cæsarea. The other,     , whether then first put together or not, had the advantage of taking up for better use what at Jerusalem had stood after    .

  of Antioch; and  may have been intended, when transposed, to stand alone after

   , with    as a fresh clause in apposition. It is impossible to disprove this rival supposition: but it is weighted with several improbabilities. First, it involves a somewhat wide departure from the real force of both the assumed precedents: in both of them the primary predicate to

    is a strong term containing , in the one case  , in the other,

  . It is not likely therefore that both these phrases would be deposed into a secondary position, and their room occupied solely by an adjective not in itself implying Deity. Secondly, the bare phrase

     is redundant and artificial1, if  retains its true usual sense of an only son or offspring. The rare secondary sense (see p. 17) in which it casts off the idea of parentage, and comes to mean only “unique”, receives no support from Athanasius or, as far as I can discover, any writer of the Nicene generation2. Thirdly, it is difficult to believe that a collocation so naturally suggesting the combination   to the many ears already familiar with it would have been chosen or retained except with the deliberate intention that it should be so understood3. On the other hand the one tangible ground for supposing the

1 The circumlocution would be all the more improbable because the obvious form    ( )  

was not only directly Scriptural (John iii 16; 1 Jo. iv 9) but stood already in the Creeds of Jerusalem and (by the easy omission of

) of Antioch. But in the case of   there would be no circumlocution, partly on account of the sense and the weight of the phrase, partly because of the need of introducing it only in a predicative position.

2 This seemingly stronger sense would in effect have served the purpose of the Council less; for no Arian would have hesitated to affirm the uniqueness of our Lord's Sonship. The point for which at least Athanasius repeatedly contends, as involving all else, is the strict and primary sense of the terms Father and Son; and this argument would have received no help from  as a scriptural designation of the Son, if it did not by recognised usage imply actual parentage.

3 The transfer of unicum from Flium to Dominum by transcribers of Latin Creeds (see p. 50 n. 1) can afford no real analogy for the skilful Greek theologians of Nicæa.

two words to have been intended to belong to different clauses, namely the position of the Nicene parenthesis, requires careful consideration. But first, a few more words must be said in illustration of the continuous construction        .

Apart from the unfamiliarity of  , the prevalent habit of treating    as a complete and independent formula may probably at first disincline a reader to accept its suspension, so to speak, on a preliminary participle. The absolutely independent use of    has undoubtedly sufficient authority in ancient theological writers; but on the other hand this use is virtually unknown in Creeds; for popular in the telligibility the help of     or some equivalent was apparently felt to be needed. Setting aside the Creed of Cæsarea, where    follows     with probably the same effect as to sense, and perhaps the Creed recited by Charisius of Philadelphia at Ephesus in 431, where

   follows     1, I can find no exceptions; for it is impossible to count as such the

highly technical Confession of Gregory Thaumaturgus (ed. Paris 1622 P. 1 A,  ,   ,

  ,     ,   ...), or the still more elaborate Exposition of Athanasius (p. 99 B), in which      is isolated among texts of Scripture2. On the other hand the rule is observed by the Antiochian baptismal Creed in all its extant forms3; the 'Third' Formulary of the

1 It is at least equally probable that here too      should be taken together; and then  would have the same effect as a participle.

2 A similar Exposition of uncertain authorship (ad calc. Greg. Naz. i 906 &c.: cf. Walch, Bibl. Symb. 172 ff.; Hahn, Bibl. der Symbole 185 ff.), has " Patrem verum qui genuit Filium verum, ut est Deus de Deo, lumen deo lumine, vita ex vita” &c .Yet here too the aid is given by the context, though not formally by the grammar.

3 As represented by Lucianus, Euse bius of Dorylæum, Cassianus. The last two writers doubtless represent the same form, which shews signs of Nicene influence: see Dissertation ii. I venture to cite Eusebius of Dorylæum, although the words in question precede his express quotation from the  of Antioch. He certainly began to interweave the diction of

Synod of Antioch, by Theophronius1; the 'Fourth' of the same (ap. Ath. De Syn. 25 p. 737 E, &c.; the 'Fifth' (A.

D. 345 known as  (ap. Ath. ib. 26 p. 738 c &c.); the Formulary of the Synod of Philippopolis, miscalled 'Sardica', in 347 (ap. Hil. De Syn. 34, p. 482 D: the only probable construction in the lost Greek is a little disguised in the Latin version); the 'First' Formulary of the Synod of Sirmium in 351 (ap. Ath. ib. 27, p. 742 A &c.); the 'Second' in 357 (ap. Hil. ib. p. 466 A &c.); the 'Third' in 358 (ap. Ath. ib. 8 p. 721 c &c.), with the peculiar form   ,     ,   ,

    , which was copied, with variations of perfect and aorist only, at the Synod of Nicé in Thrace in 359 (ap. Theodoret. H. E. ii 16 [al. 21]) and at that of Constantinople in .360.(ap. Ath. ib. 30

p. 747 A)2; and lastly by what is known as the 'Constantinopolitan' Creed3. Hence abundant analogy leads to the conclusion that   , whether forming part of the direct predicate to     or not in the Nicene Creed, is at least dependent on it, so that on either construction   presupposes : and when thus much is established, there can be no intrinsic difficulty, , and the parenthesis apart, in the closer construction which makes    part of the main predicate.

The chief external evidence for joining to  a

the Creed before he made formal appeal to it. The words are,

                ,

. . .

1 Cf. pp, 22 f. The words are,             ,

      ,       . . . The position of

   allows   . . . to be taken either predicatively or in apposition, though the former is the more probable construction, as two other participial clauses follow at once. For the present purpose the difference is immaterial.

2 We are not here concerned with the theological position of these various Synods, but solely with their incidental testimony to a traditional habit of language.

3 That is, in the clauses   ,     , as this Greed does not contain the simple

  . In all the other Creeds cited, that of Theophronius excepted (note 1),    stands unmodified.

predicate containing  has been already given, namely the probable analogy of the Creeds of Antioch and Jerusalem. To this must be added the Epistle to Paul of Samosata by the bishops assembled at Antioch in 260-270, if the correction already suggested is right1. The whole sentence must be quoted here.

   ,    (read ),     ,

  ,

        ,      ,

  ,          . As soon as  is substituted for the unmeaning second , the two preceding words acquire a clear force, the verbal

 being equivalent to a passive participle. Possibly however this ought not to be accounted independent evidence, but only as a reproduction of the Creed of Antioch2. The second required combination, that of

 with   , had undoubtedly an actual existence. In the Demonstratio Evangelica (p. 149 A) Eusebius speaks of our Lord as

               ,

       . The position of  proves a reference to two distinct forms, the familiar  , not seldom used by Eusebius (as by Athanasius), and    : the only other grammatical construction, that which makes  and  two distinct terms, would give

 an inappropriate position, imply an arbitrary distribution of the conjunctions, and enfeeble the

1 See pp. 4, 19, 39. Even if  is right, which seems incredible, we should still have as the predicate of  a combination of  with a substantive.

2 The construction of the Nicene Creed here advocated receives illustration, rather than direct confirmation, from the language of the Third Sirmian Formulary (quoted above, p. 65), adopted at Nicé in Thrace and at Constantinople in the two following years: it will be observed that   , an accepted gloss on  (see p. 17), occupies the place of the Nicene parenthesis. The parallel language of Cyril of Jerusalem (iv 7) is instructive,      ,     ,

    ,       (iv 7):

       ,    ,    ,   ,

   (xi 18: of. 4),

whole of the last clause as a climax. The same form, slightly resolved, occurs a little earlier (p.147 B),

       ; and, slightly extended, in the Panegyric on Constantine (xii 7: cf. Theophan. i 24),       1. It reappears in the Formulary of the Synod of Seleucia in Isauria (A.D. 359)  ,    , ,  . . . (ap. Ath. De Syn. 29 p. 746 c; Epiph. Haer. 873 c). And in the next century it is employed by Cyril in his commentary on St John,  ...              (Viii 35 P. 541 C), and again,   [ ]       (X 15, P. 653 c); as also in his Third (Second Œcumenical) Epistle to Nestorius (p. 24 Pusey)

        . It is immaterial whether these forms of speech were derived from the Nicene Creed or independent of it2. In either case they shew the naturalness of the combination in the eyes of theologians of the fourth and fifth centuries. Doubtless it was felt that each of the two elements associated with  in     would sustain and illustrate the other.

Thus far the discussion has left out of account the Nicene parenthesis       . Were it absent, the evidence would all, as far as I can see, be clearly in favour of taking     as an unbroken predicate of    . It remains to consider whether we are driven to a different conclusion by the position of the

1 The added  increases the resemblance to the Nicene language, though inverted in order

2 Yet it can hardly be doubted that at least Cyril had the Nicene Creed definitely in view; for in his Ep. 55, which is a commentary on the Creed, he says that the Fathers of Moses,   [the Paternity]   ...   ,

       (p. 178): and again,

    …       , …

           ...      

  ... (p. 180). Both passages lose their force if    was not part of the main predicate.

parenthesis. It matters little for our purpose whether the Nicene Fathers were here simply copying an earlier (lost) Creed, or, as the extant language of Jerusalem and Antioch has rather suggested to a certain extent modifying in combination and arrangement the traditional materials. In either case the sense and the place of their own entirely new parenthesis, must be taken into account in order to ascertain the meaning which they attached to their completed work.

A reader examining the passage merely as a piece of Greek, unaided by extraneous knowledge, could hardly fail to take  as the one weighty word interpreted by the parenthesis. Yet this supposition cannot be more than partially true at most, if we are to trust the concurrent testimony of the two men who had the best means of knowing the facts, who moreover regarded them from different points of view. Eusebius and Athanasius represent      as the interpretation of   1. Eusebius passes

 over altogether, and Athanasius alludes to it with a slightness and indirectness which throw it completely into subordination2.

1        ,

             ,

         ,

                 

       ,

                

  ... Eus. Ep. ad Caes.    [of Nocomedia] …

         [i.e. mankind]

…     …   ,

         ,

                . Ath. De Deer. 19 p 224 DR.

And so in the parallel narrative Ad Afr. 5 P. 895 B,

     ....      ,

         .

2 The possible allusions in the Ep.de Decretis to (represented by ) are in the two sentences   ,

   ,       ,

             ,

     , and

               ,

           ,       (225 A-C). The.Ep. ad Afros has likewise the word itself, but in an ambiguous context,       

But the more the stress is shifted back from  to   , the less reason is there to regard the clause as so terminating in  as to make    a fresh clause in apposition. It would seem in fact that  was put to double duty, combined alike with    and with

  ; just as we have already found reason provisionally to recognise  as doing double duty, combined alike with  and with   . Thus there would be no real pause between the seven words        . Yet the parenthesis had to be inserted somewhere. It could not be placed at the end, for   was too distant; nor before   , partly for the same reason, partly because    could not be severed. If placed before , it would have been close to    , but at the cost of depriving    of any additional force or clearness which it could derive from association with , including perhaps the reminiscence of John i 14 (    ). Placed as it actually was, the parenthesis, while chiefly limiting the sense of   , limited also the sense of , as against the Homœousians, and at the same time compelled  into a subsidiary limitation of   , as against the Anomœans. No doubt in the process   was disguised: but it was not possible to introduce the parenthesis without some sacrifice somewhere. Probably it was thought that   was too well known and accepted to lose instant recognition despite the parenthesis. But at all events its acceptance by Arius himself deprived it of controversial value for the special purpose of the Council; whereas in the eyes of at least Athanasius it must have been of primary importance to secure to the interpretation

    

 ,          (895 c). These incidental references are of no force as compared with the express statements of fact cited in the last note. Indeed elsewhere (De Syn. 51), assuming

   as the universal criterion of true parentage and filiation, Athanasius argues from Jephah's daughter and the son of the widow of Nain that a child is not less  with its parent because it is likewise .

the utmost possible force1. Thus  , though retained like other traditional forms too little stringent for the present need2, might have to suffer partial obscuration through the necessity of the case.

No other explanation than this appears to account for all the facts, and to do justice alike to the language of the Creeds of Antioch and Jerusalem, to the statements of Eusebius and Athanasius, and to the actual order of words in the Nicene Creed. There is the less difficulty in accepting a single long clause made up of closely combined terms, if we remember the evident purpose to give continuity of form to the entire declaration respecting the nature of the Divine Sonship, the other Creeds having been more or less disjointed hereabouts, the Creed of Cæsarea to an extreme degree3. Where all the clauses

1 Innumerable passages of his writings shew that the form of language adopted in this clause was the test on which he relied above all others for the exclusion of Arianism. On the other hand, loyally as he defends  when needful, he shews no great inclination to use it when left to himself: Dr Newman has noticed its almost total absence from the great treatise made up by what are called his first three Orations against the Arians (Sel. Treat. of Ath. 500, 210 d, 261 g), as also his use of the term   (210 a: cf. 136 g): cf. Tracts Theol and Eccl.291. The final result in the Creed may have been a combination of the expedients proposed by different sections of the majority in the Council.

2 Athanasius dwells on the desire of the Council to use only scriptural terms, till it was found that the party of Eusebius of Nicomedia was ready to accept them all (De Decr. 19 ff. p. 294 ff.; Ad Afr. 5 f. p. 894 ff). Among such terms he includes the following, evidently described somewhat vaguely,          , ,

    ... (895 A).

3 To this purpose must probably be ref erred the omission of  before the first , and the emphatic repetition of

, first to set forth the contrast  , and then to carry    without another participle. Then comes a fresh start on the relation of the Son to created things,    ; and the added clause

         , wanting at Cæsarea, Antioch, and Jerusalem (it is found in the Apostolic Constitutions), at once gives weight to this division of the second article of the Creed and constitutes a parallel to the first article, on the Father,      . The resumptive force of the second , as connecting

  with the earlier clause, is distinctly recognised in the later Antiochian Creed (Cassianus), which has been modified by Nicene influence, ex eo natum ante omnia saecula, et non factum, Deum verum ex Deo vero; as also, by exactly the same collocation,

bearing on a single subject are so carefully shaped into a whole, it is only natural that the series of terms relating to one portion of the subject should be knit together with unusual closeness. The arrangement may be exhibited as follows: -

     ,

   

     -

      –

  ,

  ,

    ,

,  ,

  ,

    ,

         

     ...

We have, it is to be feared, no means of knowing with any certainty how the sentence was understood in the following years. The remarkable form of the Creed noticed above (p. 23) as employed by Eustathius and others in 366 might be due either to an attempt to express more clearly the assumed sense of the Nicene language, or to a conscious reintroduction of a combination assumed to have been set aside. The concise Philadelphian Creed recited by Charisius, in borrowing the Nicene phraseology, omits the Nicene parenthesis, and thus removes the only hindrance in the way of reading         continuously: but the other construction remains possible; and again the authors of this Creed may have intended to improve rather than to interpret. Yet the growing favour of the phrase   with the friends and successors of Athanasius, in spite of its controversial uselessness, during the time that the distinctive terms of the Nicene Creed were the watchwords of every struggle, suggests the operation of some

1 in the (Syriac) Mesopotamian Creed examined in the following Dissertation, which rests on an Antiochian in the foundation.

more potent and universal cause than the influence of scattered local Creeds, or of Synods of doubtful orthodoxy which borrowed their language. The Nicene Creed itself would evidently be such an adequate cause, if it was understood as containing  : and if such was the retrospective view taken in the fourth century, such also, we may not unreasonably believe, was the intention of the Council.

Against this evidence there is, as far as I am aware, nothing to set. A Cappadocian Creed formed on the base of the Nicene Creed at a date not far from 370, of which some account will be given in the next Dissertation, merely repeats this part of the Nicene language unchanged. No other known Creed can be said with any propriety to be a revised form of the Nicene Creed. That the 'Constantinopolitan' Creed had no such origin, it is easy to shew: but a position so much at variance with commonly received views requires to be illustrated in some detail, and must therefore be treated separately. It is enough here to say that the history of   in ancient times virtually closes with the gradual supersession of the Nicene Creed. As its primary apostolic sanction had been lost long before through the increasing degeneracy of biblical texts, so its ecclesiastical sanction, such as it was, died out by an equally fortuitous process. Neither in 381 nor at any other date was the phrase   removed from the Nicene Creed. If it had a place there from 325, as we have found good grounds on the whole for concluding, it was never displaced while the authority of the Nicene Creed was in force. It passed away only when the Nicene Creed itself completely yielded place to another Creed which never possessed it.