personal Logos, and his appearance upon the earth (the "beginning"
that John was in the desert is introduced incidentally, in a way which seems to imply acquaintance with Luke 1:80.19
4. According to Corssen,20 the Logos in Luke's prologue is the personal Logos, and his appearance upon the earth (the "beginning" of the Word) was the baptism, when God said to His Son, "This day have I begotten thee." With this agrees the absolute •DP`:,<@H in 3:23 and Acts 1:21, 22. F,\ (1:23) is to be taken in a strictly comparative sense: the Logos appeared in the form of a man of thirty years.
The difficulties connected with this view are of course apparent. In the first place, it rests upon the more than doubtful reading of the Western text in 3:22, "This day have I begotten thee." In the second place, to interpret ñF,\ as comparative is here impossible, because it comes in close conjunction with a numeral, where no one would think of any other meaning than the common meaning, /pg. 46/ "about." The •DP`:,<@H (3:23) indicates "the beginning of the Word" only if we allow Corssen's reading in 3:22 and his interpretation of the baptism. If we interpret the baptism as the beginning of Jesus' Messianic work, rather than as the beginning of His divine Sonship, then the •DP`:,<@H evidently refers to the same thing. So •DP`:,<@H proves nothing in itself. Nor does Acts 1:21, 22, give it any added force, for there it is a question merely of the conditions necessary for apostleship. To be an apostle a man had to have been a disciple of Jesus only from the baptism, because before that Jesus had had no disciples.
Nor does the elaborate attempt (Luke 3:1) to fix the date of the baptism necessarily prove (even in comparison with the method of Thucydides) that that was what Luke desired to fix as the "beginning" mentioned in the prologue. Perhaps the reason he did not so elaborately fix the date of what is recorded in 1:5ff. is that he did not there happen to possess such complete information. In any case, the baptism, even if not the beginning of the whole history, was surely an event important enough to lead a historian like Luke,
writing for Gentiles and Romans, to give as complete chronological details as his sources would permit. 5. In Acts 1:1, the Gospel of Luke is described as a treatise concerning all that Jesus began to do and to teach until He was taken up. In this B@4,4$< J, 6"Â *4*VF6,4<, says Hilgenfeld, the narrative of Luke 1:5–2:52 cannot be included; therefore those first two chapters were no part of the "former treatise."
But we must remember that Luke is at the beginning of Acts characterizing his former treatise as a whole and as contrasted (:X<) with the history to follow. From such a point of view, it might well be described in general terms as an account of Jesus' earthly activity, even though it contained some introductory matter necessary to explain that earthly life. In a modern biography, we do not think it strange to find at the beginning a description of the state of affairs at the birth of its subject, or an account of family-relations for some generations back. Furthermore, as Zimmermann points out, we cannot, even on Hilgenfeld's theory, interpret the B@4,4$< J, 6"Â *4*VF6,4< too strictly, for even the main part of the Gospel contains an account of events where Jesus was not the agent—e.g., the preaching of the Baptist. Finally, if Hilgenfeld's view is correct, it is rather remarkable that in Acts 1:1 we do not find the baptism mentioned as the terminus a quo as in Acts 1:22.21 /pg. 47/
Thus far we have not mentioned what at first sight seems to be the most striking indication that 1:5–2:52 was no part of the original Gospel—namely, the striking contrast in the style and diction of this section, both with the prologue on the one hand, and with what follows it on the other. It is one of the commonplaces of New Testament investigation that at Luke 1:4, 5, the most flowing Greek period and perhaps the most strongly Hebraistic section of the New Testament come together. Yet from this undoubted fact no conclusion can at once be drawn against the genuineness of the infancy section, for it is possible that in 1:5–2:52, Luke was so closely following a source that he refrained from changing its style and diction. This explanation is the more probable because the contrast
between 1:5–2:52 and what follows is by no means so great as between that section and the prologue. It is an undoubted fact that in the admittedly original part of the Gospel, the author has allowed the style of the source to color the narrative. Therefore, he may well be carrying out the same method a little more fully in the infancy section. The difference would be one of degree, not of kind. But this is not all. Harnack22 has argued that the Magnificat and the sections 2:15–20, 41–52 (the latter two being chosen because of the difference of the subject-matter from the rest of the Gospel and Acts) exhibit specifically Lukan characteristics of style; and Harnack's investigation has been completed for the rest of the infancy section by Zimmermann,23 with a similar result. Now with reference to the Magnificat, Spitta24 has undoubtedly pointed out a serious defect in Harnack's method. Harnack has picked out the Old Testament passages upon which he supposes the Magnificat to rest, and has then extracted from the song the fourteen words which were not given by these passages. These words, he argues, are Lukan. Spitta's general criticism is that we cannot be certain enough that just Harnack's Old Testament passages and no others were consciously or unconsciously in the mind of the author of the song. So that if we find that Harnack's fourteen words are common in the Septuagint, we can scarcely draw any sure conclusion as to the Lukan authorship. But even if we allow to this objection its full weight, it does not vitiate the whole argument of Harnack and Zimmermann; for the method objected to is not carried through the other passages examined, or at any rate is not /pg. 48/ fundamental there. Indeed, the cumulative evidence adduced for the linguistic affinity of the birth narrative with the other Lukan writings must, I think, be pronounced very convincing—far too convincing to allow us to stop short with the hypothesis of a common redactor merely. It seems highly probable that the writer of the Gospel and of Acts impressed his style upon the infancy narrative, although not so as to destroy the strongly Semitic character of the language of that section.
Of course there are several possible ways of explaining these facts. In the first place, we might say with Harnack that the Semitic coloring
and Old Testament spirit of 1:5–2:52 are due largely to the conscious art of the writer, rather than to a close adherence to Semitic sources.25 But I do not think we should by any means go so far as to suppose that Luke, in possession, on the one hand, of a certain unadorned tradition, and acquainted, on the other hand, in a general way with Jewish modes of expression, went deliberately to work artificially to mould that tradition into the language best suited to the time and place described. For example, it is highly improbable that Luke actually composed the Magnificat, as Harnack maintains. Rather should we say that in the first two chapters of the Gospel the author must be closely reproducing Palestinian tradition. It is not certain that that tradition was given to him in anything more than oral form; for it does not seem too much to expect that Luke should have had literary discernment enough to catch the charm of the beautiful Jewish stories and literary ability enough not to spoil that charm in the writing. But in view of the strongly Semitic character of the language, and the still more Semitic and strictly Jewish character of the thought, it is an impossibility to suppose that Luke was the actual composer of the stories, as Pfleiderer has contended. That would attribute to him too much historical sense and dramatic art for any historian of any time; much more for a historian possessing the characteristics of Luke and living at the time when he lived. Indeed, after all, the general effect of the section will probably always be such as to suggest to most minds that the author is using a written source, and a source which could have arisen only on Palestinian ground, and in circles where the ancient Jewish traditions and aspirations were preserved in their purest form. The linguistic data collected by Zimmermann point very strongly to the use of an Aramaic document, for how else but upon the theory of translation can we explain the distinctly Lukan character of the superficial coloring as against the yet more distinctly /pg. 49/ Jewish character of the warp and woof? This, however, we must leave undecided. The special arguments for the theory of translation as given by Zimmermann26 do not prove the matter, though they may show that that theory explains very satisfactorily at least some of the facts.27 However, we may regard it as proved that Luke 1:5–2:52 follows closely a Jewish Christian
source, which, if not written in Aramaic, was yet thoroughly Palestinian in character. But the linguistic characteristics of the section rather favor than oppose the view that the source was used by the author of the rest of the Gospel.
One other argument against the genuineness of our section remains to be considered—namely, the argument of Hilgenfeld that in the first two sections certain unPauline ideas are emphasized, such as the obligation of the law (2:22, 23, 39), righteousness of works (1:6, 15, 75, 2:25), the throne of David and the eternal kingdom over the house of Jacob (1:32, 33); things which could never have been added to the Gospel by the Paulinist Luke. But, in the first place, Hilgenfeld's objection rules out of court on purely a priori grounds the view that the author in writing his narrative may have consulted the facts or the sources as well as his own dogmatic prepossessions. It is not impossible that a Paulinist should have written 1:5–2:52, unless it is impossible that a Paulinist should have desired to tell the truth—and the latter proposition is not so self-evident as Hilgenfeld and others of his school seem to suppose. In the second place, Hilgenfeld supposes that the redactor who added the two songs (with certain Pauline alterations, 1:55b, 73a, 76–79), and joined the whole narrative to the Gospel, was himself a Paulinist. It is not clear why, if the second Paulinist could do that, the first one, or the writer of the Gospel, could not have done it just as well. So Hilgenfeld's theory, aside from its other defects, is hardly consistent.
The first question, then, we may regard as settled. There are no good solid reasons for regarding 1:5–2:52 as an interpolation. Furthermore, in settling this question, we have incidentally established the fact that the narrative in 1:5–2:52 is of distinctly Jewish-Christian origin28—a fact which we shall find to be of great importance.
The attempts to separate Luke 1, from Luke 2, or to separate their sources,29 may be at once dismissed as devoid of evidence. /pg. 50/ Holtzmann30 argues that Nazareth, Joseph and Mary are mentioned
in 2:4ff. as though these names were not already known from 1:26, 27; but really the manner of repetition is perfectly natural as taking up the narrative where it had been dropped. So Luke 2:4, 5, seems, if anything, rather to presuppose a previous mention of Joseph and Mary. Joseph's Davidic descent is introduced again in order to explain the journey. Moreover, the view in question is directly contradicted by 2:21b ("which was so called by the angel before he was conceived in the womb"), where 1:31 is referred to. So Schmiedel is obliged to regard this clause (2:21b) as added when the two chapters were put together—a purely artificial expedient to bolster up a baseless theory. The two chapters are closely connected so far as style and diction are concerned, and have other things in common. For example, the same character is attributed in both chapters to Mary, and in both she is given a peculiarly prominent position in the narrative.
Far more serious is the attempt to exhibit 1:34, 35, as an interpolation; indeed it is against these two verses that the chief attack of all has been directed. Among those who have argued against the original presence of the two verses in the context where they now stand may be mentioned Hillmann,31 Usener, J. Weiss (with a little hesitation), Harnack, Zimmermann, Schmiedel, Pfleiderer and Conybeare, to say nothing of others who less deserve mention because they make little attempt to ground their objections to the verses upon anything more definite than their general theories of mythical or legendary development. The integrity of the passage has been defended by Hilgenfeld and Clemen,32 as well as by "conservative" scholars.
First, we must remind ourselves that there is no external evidence whatever for regarding vers. 34, 35, as an interpolation. Conybeare, it is true, emphasizes the reading in MS. b which substitutes ver. 38 for ver. 34 and omits ver. 38 (,Æ$B,<…JÎ Õ0$:") from its proper place; but that may have been a mere blunder in transcription, especially as the two verses begin alike, "dixit autem Maria" (Headlam). Or perhaps the change might have been made by the
scribe to save Mary from the appearance of unbelief. The testimony of John of Damascus to the omission of the phrase "seeing I know /pg. 51/ not a man" in some Greek codices is too late to be of great importance. Conybeare33 claims the authority of Tischendorf (8th ed.) for the omission of vers. 34, 35, in the Protevangelium of James. But the facts are that the Protevangelium, though it omits ver. 34 in this context, substitutes what is rather an elaboration of that verse (+Æ ¦(ã FL88Z:R@:"4 ñH B"$F" (L<¬ (,<<($), and actually contains the greater part of ver. 35.34 That Conybeare can claim Tischendorf for his view about the Protevangelium seems to be due, as Headlam has pointed out, to a surprising misunderstanding of Tischendorf's notes, which arose from not looking under ver. 31 as well as under ver. 34.35
The evidence for the interpolation theory must therefore be purely internal evidence. In the first place, we must at once dismiss the argument36 that since the fatherhood of the Spirit of God [?] would suit very badly a purely Jewish Christian source (9&M(H being feminine, and the Jewish conception of God being transcendental(, and since the basis of Luke 1, 2, was such a source, therefore vers. 34, 35, could not have stood originally in their present place. This argument proves that a conception from the Holy Spirit, or a birth described in such terms as even to suggest the personal Holy Spirit as Father, would never have been invented on Jewish ground; but it does not prove that it may not have been recorded in a JewishChristian narrative if it were a fact. What we are just now trying to do is simply to lay the basis for future investigation by estimating the narrower and more solid grounds for supposing the whole or portions of the birth narratives to be interpolations—grounds which will hold firm upon any general theory of early Christian history. There are many who suppose the doctrine of the virgin birth, assuming it to be untrue, to have arisen on Jewish-Christian ground, and they may appeal, among other things, to the strongly Jewish character of the records. Against such scholars it is begging the question to say that since the doctrine of the virgin birth must be of
Gentile origin, therefore it must be an interpolation where it finds a place in Jewish narratives.
It is further urged that 1:34, 35, is not merely without corroboration from the rest of the infancy narrative, but is even contradicted by it; for the whole of the first two chapters except our two verses /pg. 52/ proceeds upon the supposition that Jesus was the son of Joseph and traces his Davidic descent through him. We freely admit (though in contradiction to B. Weiss) that in 1:27 ¦> @Ç6@L )"L,\* must almost certainly be taken with TFZN, rather than with B"D2X<@<, for on any other interpretation the manner of addition of J0$H B"D2X<@L is very hard to explain. So that when the angel (ver. 32) calls David father of the coming child, it seems most natural that his words should be understood of a descent through Joseph. The emphasis on Joseph's Davidic descent rather than on that of Mary in 2:4, however, proves nothing, for it was the man only who would be considered as determining the place of enrolment. But if the Davidic descent of Mary is presupposed, surprisingly little emphasis is placed upon it, for, as has been observed, in the only place where anything is clearly said about her family relations (1:36) she is called kinswoman of the Levite Elisabeth. The repeated occurrence of such words as (@<,4$H, applied to Joseph and Mary; and B"JZD, applied to Joseph, has already been noticed; but these two terms do not necessarily imply anything more than that there was really an adoptive relation between Joseph and Jesus, and that Jesus before the world was regarded as an actual son. The failure to refer to 1:35 in 2:21 proves absolutely nothing,37 for any such reference would have made the sentence extremely clumsy. Nor is the phrase "their cleansing" in 2:22 very convincing. It is quite in line with a good many things connected with the life of Christ, e.g., the baptism of a sinless man. As to the failure of Mary to understand, or her astonishment at what was said about the child by Symeon and Anna and by the boy Jesus Himself, even Zimmermann admits that this has little bearing upon the question of the original presence of 1:34, 35, in the narrative. The astonishment of the parents was due to the fact that Symeon and Anna and the boy Jesus were found to be
possessed of the secret of the Messiahship. Only thus, according to Zimmermann, can the passages be explained, whether the parents knew about the supernatural conception or only about the Messiahship of their son.
In general, we can say that it is unreasonable to expect that the account of the supernatural conception should be repeated again and again. In a narrative it is enough that it should be given once, whatever might be true of a dogmatic treatise. Yet, after all, we do not desire to depreciate the force of the argument against the two verses, derived from the silence or seeming contradiction of the rest of the story; for although that argument may not prove the /pg. 53/ verses to be an interpolation, it will do much to render us hospitable to other proofs. If we really find that in the rest of the first two chapters there is not the slightest hint that might point to the virgin birth, or that there is a good deal that seems almost directly to deny it, we shall be very much disposed to look with suspicion upon the only two verses that tell of such a remarkable event. As a matter of fact, however, this is not the case.
In the first place, 1:27 deserves the most careful attention. We there read in the clearest terms that Mary was a virgin when the announcement was made to her by the angel. Now, since there is no subsequent mention of a marriage to Joseph, the natural conclusion is that in 1:27 we have a preparation for 1:34, 35.38 To avoid this conclusion two expedients have been adopted. In the first place, Usener suggests that the redactor has left out a statement (which originally came after ver. 38) that Joseph took Mary to wife and that she conceived by him. But that is a mere supposition. In the second place, Harnack supposes that the word B"D2X<@H in 1:27 is an interpolation made by the same redactor who added vers. 34, 35. For, he says, the word ¦:<0FJ,L:X<® in 2:5 can only mean "wife," so that the same author could never have written a few verses back B"D2X<@< ¦:<0FJ,L:X<0<. One of the words must be removed, and the most natural one to remove is, of course, B"D2X<@<. But this really begs the question. For ¦:<0FJ,L:X<® in 2:5 means simply
"wife" only on the supposition that 1:34, 35, are to be deleted—which is exactly the thing to be proved. Nor is the removal of the mention of the virginity of Mary from 1:27 at all an easy task, for the word B"D2X<@H occurs twice (B"D2X<@<, B"D2X<@L), and is indissolubly connected with the very structure of the sentence.39 Whatever maybe said about the ease with which the two verses, 1:34, 35, taken by themselves, may be removed; if the removal of those verses necessarily requires another deletion, which, far from being equally easy, is so harsh as to be practically impossible, then the former deletion must be seriously reconsidered.
Harnack's argument has led us to the second chief reference to the two verses in question. In 2:5 we find the phrase J®$ ¦:<0FJ,L:X<® "ÛJ($ @ÜF® ¦<6bå—a phrase absolutely inexplicable unless 1:34, 35, is referred to. For, after all, if the author had meant "wife," he would certainly have said "wife"—at any rate, he certainly would /pg. 54/ not have used ¦:<0FJ,L:X<® in conjunction with ¦<6bå.40 So evident is this that most of those scholars who regard 1:34, 35, as an interpolation can overcome the difficulty only by choosing the reading (L<"464 instead of ,:<0FJ,L:,<0 in 2:5. The external testimony is briefly as follows: (L<"464 is omitted altogether by ! B C* D L =, 1, 131, e, f, q**. sax. sah. copt. syr.sch arm. (L<"464 is added after "LJT by A C² ' ) 7, 1, q*, vg. goth. syr.P aeth. (L<"464 is read without a preceding ,:<0FJ,L:,<0 or corresponding word by the Latin manuscripts a, b, c, ff², and by syrsin. The reading with both ,: <0FJ,L:,<0 and (L<"464 is evidently to be dismissed at once as a mixed reading. Now of course this leaves the overwhelming manuscript authority in favor of ,:<0FJ without (L<"464, and this authority has been followed by Tischendorf (8th ed.), WH, Baljon, etc. Some scholars, however, have argued that (L<"464 represents the original reading, on the ground that (L<"464 might easily have been changed into ,:<0FJ for dogmatic reasons, whereas there would have been no ground for an Ebionitic alteration of ,:<0FJ.41 But it is not necessary to think of an Ebionitic alteration, since ,:<0FJ might easily have given offense on account of the difficulty of conceiving of Mary as only betrothed when she made the journey with Joseph, as