the virgin birth in the Gospel cannot be decided, though the former
Elisabeth, and in the second place, it is merely subjective at the best. /pg. 58/
7. After the necessary changes have been made in 1:27 and 3:23, the Gospel of Luke knows nothing of the virgin birth, except in 1:34, 35. "After these few and easy deletions, which are required, as soon as we are convinced of the interpolation of vers. 34, 35, but which otherwise also obtrude themselves upon us, the narrative is smooth and nowhere presupposes the virgin birth."
As we have already shown, Harnack has no ground for saying that the removal of B"D2X<@H in 1:27 has its own reasons, apart from the theory about 1:34, 35. Other objections also have already been noticed.
8. The composition of vers. 34, 35, is easily discerned. Ver. 34 prepares for ver. 35 (very clumsily it is true); ver. 35 is to be explained from Luke 1:31, 32, and Matt. 1:18–25. It is rather suspicious that the redactor should be so clumsy in one point, and should yet exhibit positive genius in imitating (ver. 35) so admirably the style and spirit of the narratives.
9. So Matt. 1:18–25 becomes the starting-point for the representations of the virgin birth, which simplifies matters in the history of the legend. In our judgment, however it may be in Harnack's, this is merely begging the question.
10. Whether Luke himself subsequently or an interpolator inserted the virgin birth in the Gospel cannot be decided, though the former alternative is less probable. This does not seem to be intended as an argument at all, and so demands no answer. Against all these minuter arguments may be balanced the important consideration of the parallelism with the annunciation to Zacharias. In vers. 11–20 we have (1) the appearance of the angel, (2) fear of Zacharias, (3) promise by the angel, (4) surprised question of Zacharias, (5) reiteration of the promise, with a sign. To these details we have in the
full text of the annunciation to Mary striking parallels, and the details are there arranged in the same order. The general impression is very strong that this parallel was intended by the writer, so that it is very unlikely that vers. 34, 35, are to be removed; for without these verses the symmetry of the chapter is destroyed.51
Our conclusions may be formulated as follows:
(1) It is impossible that vers. 34, 35, should have been interpolated into the completed Gospel. That is excluded by the weight of external evidence. (Against Harnack.) /pg. 59/
(2) It is highly improbable that vers. 34, 35, are an addition made by the Evangelist to a Palestinian source that elsewhere he follows closely. On that view it is difficult to explain the peculiarly marked Semitic style and spirit which prevails in the two verses, so precisely in harmony with the rest of the narrative. (Against Zimmermann.)
(3) It is less improbable (but still far from likely) that in 1:34, 35, Luke is departing from a Palestinian source which he does not here follow closely but employs in so loose a way that we can seldom (as here) separate the source from the finished composition.52 Against this view of the matter, Wernle himself notices the objection that it fails to account for the apparent contradictions and roughness caused by the insertion, but he supposes that that contradiction was not apparent to Luke in the same way as to us. So Wernle holds that the birth narrative is the work throughout (even through 1:31–35) of one author, and is as closely knit as we can expect in a time of lively productiveness and variegated religious syncretism. But how, then, can we be confident of separating between author and source in 1:31ff.? Wernle would perhaps be more consistent if he were more skeptical about this point. Perhaps, too, the same line of reasoning as that of Wernle will allow us to attribute the whole to some writer other and earlier than the writer of the Gospel. At any rate, grave objections may be raised, for example, from style and diction, against
the large place which Wernle attributes to the Evangelist in the composition of chapters 1, 2.53
Before passing on, we must notice a remarkable modification of the interpolation theory we have just been considering—a modification which has recently (1900) been suggested by Kattenbusch54 and defended by Weinel.55 According to Kattenbusch, the birth from the Holy Ghost was originally thought of independently of the birth from a virgin, and it is to the former conception that Luke's narrative attaches the chief importance. Indeed, even 1:35, taken by itself, does not mean anything more than that the Spirit of God so overshadowed the mother that not merely was the child filled with the Spirit from the moment of birth, as in the case of John, but that which was begotten ((,<<f:,<@<) partook from the very first of the nature of the Spirit. That verse excludes the /pg. 60/ human father only when it is taken in connection with the last clause of ver. 34 (¦B, –<*D" @Û (4<fF6T). So that in order to remove the virgin birth from Luke's narrative and thus secure unity of representation, it is not necessary to delete the whole of vers. 34, 35, with Hillmann, but merely to remove the four words ¦B, –<*D" @Û (4<fF6T.
The special grounds that speak in favor of this new suggestion (as they are to be gleaned partly from Kattenbusch, but particularly from Weinel, who is more confident about the literary and critical question) seem to be derived largely from the comparison with the annunciation to Zacharias. As we there find no suggestion of the agency of Zacharias, because that was regarded as a matter of course, so the agency of Joseph is in this second annunciation similarly regarded as a matter of course. In the second place, the statement of ver. 35 about the B<,L$:" "(4@< cannot exclude the coöperation of the human father, because it is expressly correlated with the conception by Elisabeth (ver. 36). In the third place, the very giving of a sign (ver. 36) requires that a surprised or doubting question should have preceded. But this requirement is not satisfied by Hillmann's theory. And in the fourth place, the parallelism of
structure between the accounts of the two annunciations, which is destroyed by Hillmann, is preserved by this new suggestion.
As to this last argument, we observe that the parallelism is not preserved by Weinel's suggestion so well as by the maintenance of the integrity of the passage. For in ver. 18 Zacharias gives the reason for his doubt, to which reason there is nothing corresponding in the case of Mary unless the words ¦B, –<*D" @Û (4<fF6T are retained. Therefore this very argument of Weinel speaks very strongly against his own theory, as it does against the theory of Hillmann. The most attractive thing about the new theory is that it removes one difficulty about Mary's question, in that it makes her surprise centre about the greatness of her son, rather than about a hitherto unmentioned peculiarity in the manner of His birth.56 Furthermore, by retaining ver. 35, it procures the great advantage over the theory of Hillmann of not obliging us to attribute to a redactor such a marvelous genius in imitating the spirit and style of the original writing. Indeed, we are almost tempted to admit that the new theory is preferable to the old; at any rate, we gladly admit that the old has received a new wound from the fresh arguments of Weinel, especially the literary argument from the parallelism with 1:11ff. But these arguments oppose the older interpolation theory /pg. 61/ as much in the interests of the integrity of the whole passage as in the interests of the new theory. On the other hand, many of the arguments of Harnack, and arguments upon which the champions of the old theory were accustomed to stake their cause to no mean extent, fall to the ground if ver. 35 is retained. Furthermore, although Kattenbusch is correct in saying that ver. 35 does not require the virgin birth, yet it naturally suggests something of the kind, so that it is better in place if the clause ¦B, –<*D" @Û (4<fF6T has preceded. And then one great objection to the new theory (an objection which Weinel has not altogether ignored) is the extreme cleverness of the redactor. According to the new theory the redactor is too clever, as according to the old theory he displayed too much literary genius. On the whole the two theories are about equally improbable; for, after all, the
really fundamental objections apply to both alike, while the peculiar difficulties are about equally divided.
In Matthew, Hillmann supposes the first two chapters to have been no part of the original Gospel, while Hilgenfeld regards 1:18–2:23 as an interpolation. It is argued that the ¦< J"bJ"4H J"4$H º:XD"4H of 3:1 would not be natural if the third chapter was originally joined to what now goes before. According to Hillmann, probably some chronological note similar to that in Luke 3:1 was left off by the redactor who added chaps. 1, 2; for the redactor was so far from the time described that he would take no offense at applying the phrase ¦< J"bJ"4H J"4$H º:XD"4H to what really happened after an interval of thirty years. But this is a mere supposition. Perhaps the author of the Gospel would himself have been looking back over a long enough interval not to have objected to the phrase, especially in view of the loose way in which the incidents are coupled all through the Gospel. Nowhere is the chronological succession very clear.
Hilgenfeld supposes that the ¦< J"bJ"4H J"4$H º:XD"4H refers to the close of the genealogy, for it would be perfectly natural to mean by the phrase merely "in the time of Jesus," if it is taken in connection with the many generations indicated in 1:1–16. But this seems rather unlikely, for the genealogy is the expression of one idea, and has no chronological purpose. It would, therefore, be very unnatural to separate 1:16 from the rest by applying to it the phrase ¦< J"bJ"4H J. º:. That phrase requires that something in the nature of narrative should have gone before, and this requirement is not satisfied by the genealogy. Meyer argues further that 4:13 manifestly refers to 2:23. /pg. 62/
As to the content of the section, Hilgenfeld57 enumerates as marks of the redactor (1) the Old Testament pragmatism, (2) the friendly attitude toward the heathen, (3) the view of Christ as born Son of God. But the Old Testament pragmatism is rather a mark of the author of the whole Gospel, who is interested throughout in showing the fulfilment of Old Testament prophecy. The friendly attitude
toward the Gentiles proves nothing if the story of the Magi (Gentiles) is essentially true, for in the mere form of the story there is no evidence of a desire to magnify the Gentiles at the expense of the Jews. And it is not at all self-evident that the author of the rest of the Gospel should not himself have felt the contrast between the acceptance of the gospel by the Gentiles and its rejection by the Jews. Finally, why may not the idea that Christ was born Son of God have been the view of the author of the Gospel? Some one—i.e., the redactor at least—held to both the Davidic sonship and the virgin birth. Why may not the author have done so?
One piece of supposed external evidence must be mentioned, even though we consider it to be of little value. Conybeare58 and Hilgenfeld attribute considerable weight to a Syriac tract, extant in a sixth-century manuscript (British Museum, Add. 17,142), and published, with a translation, by Wright in the Journal of Sacred Literature, 1866, Vols. ix and x. The tract is attributed to Eusebius and purports to be an account of the Star and the Magi, the history having been written down in 119 A.D. According to Conybeare, "the Syriac author of this tract…had in his hands a pre-canonical Greek source of 119 or 120," to which belonged the colophon that gives the date. Conybeare's conclusion is that the date 119 or 120 is the terminus a quo of the introduction of Matt. 2:1–15 into the canonical text. The document is interesting, but the conclusions drawn from it seem to be best described as "problematical"—a word which J. Weiss aptly applies to Conybeare's Ephraem passage about Luke. And in view of the undisputed unity of style and diction between 1:18–2:23 and the rest of the Gospel—a unity far too perfect to be explained as due merely to a common redactor—we may safely agree finally with J. Weiss when he declares that there never were forms of the Gospels of Matthew and Luke without the infancy narratives.59 As to the sources of the infancy section in Matthew, nothing very definite can be said. It is mere speculation, for example, when /pg. 63/ Schmiedel makes 1:18–25 an addition later than chap. 2. Indeed, for all we can see, the two chapters might go back to the same source, for the failure to mention the place Bethlehem in 1:18 instead of in 2:1
proves very little;60 but, after all, the theory of merely oral sources can never be disproved. The ultimate home of the sources is far more likely to have been Palestinian than Gentile, for the section shows acquaintance with Jewish customs, and with the Hebrew text of the Old Testament; and perhaps is combatting Jewish slanders.61 The story of the Magi does not oppose this view of the matter, for the Gentile coloring, so far as it exists, might be due to the Gentile subject;62 and perhaps it is even a positive evidence for the Jewish character of the narrative, for it may represent the Jewish Messianic conception of a gathering of the heathen for worship to Mount Sion. If Matthew's Gospel is in general destined for Jews, then it is not necessary to suppose that 1:18–2:23 is a foreign element; or rather it is not necessary to do so until we have proved that the idea of the supernatural birth could not possibly have arisen on Jewish ground.63 As to the genealogy of Matthew, the attempt of Charles64 to prove that it is a later addition to the Gospel (about A.D. 170) is interesting only in showing how more usual critical theories can be reversed. Conybeare65 has shown how impossible it would have been for the genealogy to have been added at that late date, when interests other than the interest in the Davidic descent were predominant; and Badham has argued with some weight against separating 1:1–17 and 1:18–2 at all. At any rate, there can be no doubt whatever that the genealogy was part of the original Gospel, or, to sum up our results, that the whole of chaps. 1, 2, is genuine. The discovery of Syrsin in 1892 has made Matt. 1:16, from a textual point of view, one of the most extensively discussed verses in the New Testament, and has acted as a lively stimulus to the investigation of the genealogies in general. The bewildering mazes of the textual question66 must here, for obvious reasons, remain /pg. 64/ unexplored; nor do we need to explore them for our purpose. For after the first shock of discovery has passed away, the general consensus of scholarship seems to be leaning to the opinion that the readings of the new manuscript do not tell us as much as was at first supposed. As has been remarked, the reading at 1:16 merely intensifies difficulties already present; at any rate, it cannot prove that 1:18ff. was not a part of the original Gospel. Either one of two
lines of solution seems to me to be possible. In the first place, we may say with J. Weiss67 that the original form of the genealogy was "Joseph begat Jesus," though this was, of course, never the reading in the Gospel; the problem then being how to account for the variants after the change had once been made. This problem J. Weiss dismisses as insoluble. Wilkinson,68 in one of the most convincing papers which I have seen upon the subject, attempts something of a solution. He decides (and correctly) that our present Greek text is the original text of the Gospel. For the narrator of 1:18ff. had two motives: (1) to assert the miraculous conception, (2) to assert that the birth took place while Mary was Joseph's wife. The latter was the narrator's way of effecting a "compromise"[?] between the virgin birth and the Davidic Messiahship. Now 1:16 in our critical text is in exact accord with this purpose, as the reading of Syrsin is not, while Conybeare's reading from the Dialogue of Timothy and Aquila is manifestly conflate. The other readings, Wilkinson continues, were due to two causes: correction due to dogmatic sensitiveness, and corruption from the original sources (i.e., from the reading of the original genealogy, "Joseph begat Jesus"). There are many attractive features about such a construction of the history of the variations, but I am not quite convinced that "Joseph begat Jesus" was the reading of the original genealogy—if there was a genealogy of this peculiar type—before the author of the Gospel made use of it. For, in the first place, the compiler who inserted the names of women throughout the genealogy would have been likely to mention the mother of Jesus; indeed, it is not impossible that he inserted the women expressly in view of the fact that there was something remarkable about Mary—i.e., the virgin birth.69 We must simply refrain from trying to make a decision.
In the case of Luke, perhaps there was an original genealogy /pg. 65/ which made Joseph the father of Jesus without indication of anything peculiar in the relationship. At any rate, the ñH ¦<@:\.,J@ was added at least as early as the reception of the genealogy into the Gospel, and probably earlier. Indeed, I think we should not be too certain that the words of ver. 23 were ever without the ñH
¦<@:\.,J@, for it is not even so evident as is sometimes supposed that no one would have gone to work to compile a genealogy who was expecting to remove (apparently, at least) the very point of it by these words. For, to emphasize what we have mentioned many times, we know that there were some who were interested to prove both Davidic descent and virgin birth. Why may not the compiler of the genealogy have been one of these? And suppose the genealogy was not first compiled at all in order to show the Davidic descent of Jesus, but was a long-prized family record which was continued from the generation to generation. If it was to be continued at all after Joseph, it could be continued only in the form in which we now have it—that is, in case the virgin birth was a fact. So there would be no question of going to work to construct a genealogy of Jesus; the genealogy already existed as a genealogy of Joseph.
It must be remembered that our discussion of divisive theories about the infancy narratives, long and tedious as it has been, is merely a means to an end. The great problem for those who deny the historicity of the birth stories is to show how the idea of the virgin birth could have arisen in such a way and at such a time and in such a place as to find a lodgment in those stories. This problem would be much simplified if certain things about the character and date of the account of the virgin birth could be established by clear internal evidence. Now the result of our examination of the supposed internal evidence, we believe, has been to show that the propositions—which we enumerated as the four logical motives for divisive theories—have not been established. In the first place, the infancy narratives are not interpolations in the Gospels; so all the evidence for the early date of the Gospels is also evidence for the early date of the infancy narratives. In the second place, those portions of the infancy narratives which tell of the virgin birth cannot so be separated from the rest as to allow us to suppose that the Davidic descent could not in the early days be maintained by the same writer that also believed in the virgin birth. So if the other New Testament writers emphasize the Davidic descent, it is no proof that they did not also believe in the virgin birth. In the third place, one of the narratives of the virgin
birth—that of Luke—is pronouncedly Jewish-Christian and even Palestinian /pg. 66/ in origin; while the narrative of Matthew also bears marks of Jewish-Christian origin, and at any rate is contained in a Gospel probably destined for Jews. Finally, since the account of the virgin birth is part of the fundamental structure of both narratives, and since the narratives are manifestly independent of each other, it follows that our two testimonies to the virgin birth cannot be reduced to one. The narratives being of such a character, the problem now is to show how the virgin birth, unless it were a fact, ever could have found a place in them. We must not merely show how the idea of the virgin birth might have been developed during the first century; we must further show—and this is often neglected—how this idea was ever taken up by just those narratives in which we now find it.70
Since the narratives of the virgin birth are Jewish in character, it is most natural to suppose that the basis of the idea is to be found on Jewish-Christian ground.71 Within the limits of Judaism itself, two starting-points have been suggested for the development of the idea of the virgin birth. In the first place, certain great heroes of old—such as Isaac—being born by a peculiar exercise of the power of God, were regarded as begotten not 6"J( FVD6", but 6"J( B<,L$:" (cf. Gal. 4:29); and Luke even gives an account of such a birth in the case of John the Baptist. So since Jesus was considered greater than these spiritual children, it was only a short step to exclude the human factor altogether by making the Holy Spirit, in this case, not only an important factor, but the sole factor in His conception in His mother's womb (cf. the case of John, Luke 1:15). Not only was this "greater than the prophets" to be filled with the Spirit "from his mother's womb," but the Holy Spirit was to be the very constituting element of His personality. To this short step in advance the virgin prophecy of Isa. 7:14 would afford the necessary impetus. Of course, as Beyschlag says, all this is merely the formal factor of the representation of the virgin birth; the material factor was the belief in Jesus Christ as a new beginning in humanity, as the one who came down from above. The course of development has been fully
described by Lobstein:72 The disciples began with a profound impression of the uniqueness of Jesus' personality. This impression they interpreted at first along merely Jewish lines—they interpreted the title "Son of God" as applied to Jesus merely in a /pg. 67/ Messianic or theocratic sense. But as Christian thought began to seek for the underlying causes of what it had at first accepted without deep reflection, the simple explanation of the unique personality of Christ as rooted in His Messiahship was no longer able to suffice. Thus arose the Pauline doctrine of preëxistence, and finally, under the influence of Alexandrian philosophy, the more highly developed Logos Christology of the fourth Gospel. To the theocratic sonship was added the metaphysical sonship. But parallel with this theological development, or preceding it, a more popular development had been going on. To the popular mind—assisted by the stories of spiritual children such as Isaac, and by the prophecy of Isa. 7:14—the most natural explanation of the unique personality of Christ was that He was not born like other men, but begotten directly by God. So we have not only the theocratic and the metaphysical sonship, but also (inferior to the latter) the physical sonship.
Such a theory has an advantage over some that we shall presently consider, in that it does not call in elements which could not possibly have been included in Jewish-Christian narratives. Even here, however, we might with some reason object that the stage of mythical development required by Lobstein's theory is too advanced to be represented in a narrative reflecting so purely as that of Luke the spirit of the Old Testament and of Palestinian thought. But we waive this point, in order to emphasize even more serious objections. In the first place, Harnack is basing his theory upon a very unsteady foundation when he makes the passage Isa. 7:14 not only a necessary element in the development, but apparently the only determining cause for the peculiar form which the myth has assumed.73 For the word used in the Hebrew, %3-/%, would give no impulse whatever to the idea of a virgin birth; while there is no evidence that the Septuagint translation )B"D2X<@H) had ever as a matter of fact given rise to the inference that the Messiah was to be born of a virgin