I. A Suggested Theory
So far as we ourselves are concerned, there is only one hypothesis open to us, and it is not far to seek. It will be best if we first state it somewhat baldly, leaving obvious difficulties to be considered later. The theory is as follows:
In the first instance St. Luke wrote his Gospel, either in whole or in part, without any knowledge of the Virgin Birth. To him, as to the compiler of the Lukan Genealogy, Jesus was the son of Joseph and of Mary. St. Luke's estimate of Jesus was not less high than that of St. Paul and St. Mark, but, as was probably true in the case of each of these writers, no tradition of the Miraculous Birth had reached him. He looked upon Jesus as [pg 073] the Child of Wondrous Promise, and for his analogies he turned to the Old Testament to the stories of Isaac and of Samuel.
In contrast to earlier writers St. Luke had an excellent Birth-tradition at his disposal. According to his sources the coming and future Messianic greatness of Jesus had been divinely foretold. His birth was heralded by angelic choirs, and humble shepherds brought their meed of worship and of praise. By an insight divinely given, men like Simeon and women like Anna saw in Him the child of promise. He was to be a light for revelation to the Gentiles and the glory of His people Israel. We need not stay to look more closely into the story, which doubtless has been worked up as regards its form by the Evangelist's hand. Suffice it to say that St. Luke's picture is that of a Wondrous Birth, supernaturally foretold; not a virgin birth.[70]
Some time after he had penned his narrative, possibly after it had been dispatched to Theophilus, but at any rate before the Gospel gained a wider circulation, St. Luke received the tradition of the Virgin Birth. At what time and from what source the story reached him we are quite unable to say; possibly it was from some reader or readers to whom he had submitted his narrative; possibly the story travelled along some independent channel. In any case the probability is that the tradition was imparted to St. Luke by some one who claimed to possess a fuller and a better account, and whose claim the Evangelist respected and admitted. Having regard to St. Luke's standing and methods as an historian, we prefer to believe that the tradition reached him through a definite and personal channel, than to suppose that of his own initiative he freely altered a valuable source out of deference to a growing theory.
The historical value of the new information is a question we are not now considering. It is part of our theory, however, that it satisfied the mind of St. Luke; to him the Virgin Birth was historic fact. Probably the story appealed to him at once as a fitting explanation of the unique personality of Jesus. It was [pg 074] a tradition rich in doctrinal possibilities; it provoked reflection, and it answered questions.
The Evangelist saw at once that the story must find a place in his narrative. Fortunately it was not too late, and fortunately again there was a point where it could be included without entailing the necessity of rewriting cc. i, ii entirely. He had only to insert the words we have now in i. 34 f. into the address of the angel, and to add to the opening words of the Genealogy the phrase “as was supposed”, to obtain a narrative in which truths previously unknown to him found sufficient statement. If we can suppose that the adaptation of what he had previously written was not drastic enough, we obtain a hypothesis which at least does justice to every result we have yet secured.
The view that Lk. i. 34 f. is an interpolation made by St. Luke himself was put forward by Zimmermann in Studien und Kritiken (p. 273 f.) in 1903. His treatment (cf. Moffatt, INT., p. 269 n.) differs in several respects from that outlined above. Zimmermann posits an Aramaic Jewish-Christian source which described a natural birth, and suggests that it was in the course of translating this document that St. Luke added i. 34 f. The Evangelist is also credited with having altered i. 27 and ii. 5, so as to describe Mary as betrothed to Joseph. Zimmermann also explains ii. 22 (αὐτῶν) as a mistranslation, and ascribes to St. Luke the parenthesis of ii. 35 a, and the chronology of iii. 1-2, which he holds is inaccurate.
According to this hypothesis St. Luke must have been acquainted with the Virgin Birth before he began to translate the supposed Aramaic document. This view is encumbered with difficulty; for, if Zimmermann is right, we should certainly expect a much more drastic editing of the document than can be shown. The extent to which this difficulty appears in the case of our own theory is one for which we think that justification can be given.[71] In the case of Zimmermann's hypothesis the obstacle is too great. On this view we cannot understand how the Evangelist allowed himself to write down those expressions which are incompatible with the Miraculous Conception.[72]
The view we have preferred agrees with that of Zimmermann [pg 075] in positing a source or sources which described a natural birth. It differs from it in denying that the Evangelist knew of the Virgin Birth at the time when he made use of those sources. We prefer to think that it was after cc. i, ii had attained what is substantially its present form in Greek, that St. Luke came to hear of the Virgin Birth, and that it was then that he inserted i. 34 f. This supposition includes the positive advantages of Zimmermann's theory, and it agrees better with the existing literary phenomena of Lk. i, ii.[73]
II. Literary Conditions Under Which the Gospels Were Written
In holding the view we have outlined, we have no thought of running with the hare and hunting with the hounds. Our theory is not intended as an eirenicon. It is not an attempt to make the best of two worlds, the critical and the dogmatic. If we appear to have introduced the Virgin Birth into the Third Gospel by the back door, after we have bowed it out at the front, this is simply because the evidence leaves us no alternative. Our theory makes room for the twofold fact, as it seems to us, (1) that the Virgin Birth is not an original element in the Third Gospel, and (2) that St. Luke wrote the one passage in the Gospel which asserts the doctrine; thus for us it is inevitable.
If, from another point of view, our hypothesis seems a bold venture, we may justly claim that the facts are such as to demand a bold treatment. Nor is it a sufficient objection to say that the theory is complex. Life is a complex thing, and there are few times when we need to remember it so much as when we are thinking of the production of an historical work.
Apart from other claims which can be made, our theory has [pg 076] one important, if general, advantage; it takes account of the elementary facts of human composition. Have we given sufficient thought to the fact that a writer like St. Luke may well have turned back to review, and even to alter, in the light of further information, what he had already written? Is not this what nearly every one does who writes or relates anything at all? Is it not indeed rather a rare than a usual proceeding to write a story from start to finish without insertion, omission, and revision?
In his “Introduction” Dr. James Moffatt has drawn attention to these things, and especially as they concern St. Luke's two works. He shows that interpolation may take place “either (a) at the hands of the author himself, or (b) by subsequent editors of the volume, after the writer's death, or (c) by scribes (or editors) of the text” (p. 36). Under (a) he refers to instances in Aeschylus, Herodotus, Virgil, Juvenal, Martial, and Lucretius. “Several passages in the De Rerum Natura (e.g. ii. 165-83) are also to be explained most naturally as additions made by Lucretius himself to the original draft, and in the case of the Third Gospel or its sequel it is not unlikely that Luke may have re-edited ... his work” (p. 37). Dr. Moffatt gives a very interesting modern example in the case of Northanger Abbey, which was first composed by Jane Austen in 1798. “In the fifth chapter, however, we have an allusion to Miss Edgeworth's Belinda—a novel which did not appear until 1801. This proves that Miss Austen's work lies before us in a revised form; the first draft was gone over by the authoress before its final publication some years later” (p. 37).[74]
It will scarcely be denied that the possibility of interpolation by an original author has often been overlooked by many critics. They are not slow to find the insertions of later readers and scribes, but often it seems tacitly to be assumed that the original writers must have written with logical and almost unerring precision. Curiously enough, something like the Verbal Inspiration of Scripture is required to justify some of the critical results reached. This is a doctrine long since discredited, but being [pg 077] dead it yet speaks. It will have to be allowed, we think, that mechanical theories of Inspiration have not yet left us free to perceive those ordinary conditions of writing under which the New Testament writers wrote. The aftermath of Verbal Inspiration still blinds us to the commonplaces of composition.
Of all New Testament authors St. Luke is perhaps the last to have issued his works without modifications. The high art which is self-evident in a modern writer like Robert Louis Stevenson was not attained without corrections, substitutions, redrafting, and rewriting. Without drawing the parallel too closely, and without impugning his real inspiration, we may well credit some of these processes to St. Luke. This, however, is an argument we cannot press too far, for, as will be seen in the following section, there is good ground for the belief that St. Luke's revision of his work was never complete. It is sufficient for our hypothesis to find room for a measure of revision and for the presence of modifications required by new information.
The nature of St. Luke's task is an added reason for expecting these processes. In his Preface (i. 1-4) St. Luke shows a desire to produce a full and accurate record, and claims to have traced the course of all things from the first. Any new information bearing upon the Birth and the hidden years of the Infancy would be especially welcome to him. Any one, moreover, who has had anything to do with collecting memoirs knows that not infrequently new facts come to hand just when the task seems well-nigh completed, facts for which a place must be found, however great the difficulties may be.
We are not indeed left entirely to conjecture. We can examine St. Luke's treatment of the Markan record. The modifications which he introduces are manifest, and they arise in different ways. Many of them are stylistic, others are intended to clear up difficulties, while it is in every way probable that others again are corrections introduced as the result of new information. If, from such causes, St. Luke does not hesitate to modify the statements of St. Mark's Gospel, it is inconceivable that he would have refrained from altering his own narrative if occasion should arise.
We have at least one definite example, within St. Luke's works, of a story which has been modified in the light of further [pg 078] information. In Lk. xxiv there is good ground for thinking that the final parting of Jesus from His disciples is not described as an Ascension, and apparently it takes place at the close of Easter Day. In Acts i we have the story of a forty days' interval, during which the Risen Christ teaches His disciples the things concerning the Kingdom of God (i. 3). The Ascension is described as an act of visible levitation. Jesus is taken up into heaven and a cloud receives Him out of His disciples' sight (i. 9). As they stand gazing upwards two men appear by their side clothed in white garments, who declare that Jesus shall return in like manner as they beheld Him going into heaven (i. 10 f.). The disciples then return to Jerusalem. It can hardly be denied that this is a totally different story from that which is told in Lk. xxiv. Whatever its historical value may be the presumption is that it rests upon a tradition which had come to St. Luke's knowledge after he had completed his Gospel. Apparently he acquired his new information when it was too late to alter his earlier work. Otherwise we may believe that the story would have appeared in the Gospel and not in the Acts.
It may freely be granted that the foregoing considerations are of a purely general character. Admittedly they do not prove that Lk. i. 34 f. is a specific instance of modification. Our justification of this hypothesis is the results we have reached in Chapters II and III. What we have just urged, however, is sufficient to show that our theory is not by any means inherently impossible, but is consonant with St. Luke's procedure and methods as a writer.
III. The Objections to Which the Above Theory is Exposed
We have now to consider what is perhaps the strongest objection to which our theory is exposed. It may be stated as follows:
If the Virgin Birth is a later element in the Third Gospel introduced by St. Luke himself, the Evangelist's revision of cc. i, ii might reasonably have been expected to be much more thorough than it is. Why, for example, does he leave untouched the references to Joseph and Mary as “the parents” of Jesus? Why does he not qualify his ambiguous reference to “their” [pg 079] purification? Why is he still untroubled by their astonishment, and by their failure to understand the words of Jesus at Jerusalem? Why does he not insert some clearer reference to the Davidic descent of Mary, or at least give us reason to believe that he looked upon Jesus as the adopted, and therefore legal, son of Joseph? Why does he leave the Sonship mentioned in the first part of the angel's speech (i. 31-3) apparently of a purely Messianic character? Why does he not provide occasion in the Annunciation for the terms of Mary's question in i. 34? In short, are we not back again face to face with the same difficulties with which our investigation opened? These are some of the difficulties which our theory raises.
In reply to this objection there are two preliminary considerations to be borne in mind. They are not arguments in the sense of things which can be proved; they are rather possibilities which ought seriously to be taken into account.
(1) In the first place it should be recognized that we may not have all the details of St. Luke's actual reconstruction before us. Something may have been altered or excised; we have the result; we may not have all the stages. Usener (EB., col. 3350) has asserted that statements of fact have actually been omitted from the original narrative; he is even able to tell us what they are! He thinks that we can “infer with certainty” that in the original form of the narrative after i. 38 stood the further statement that Mary was then taken to wife by Joseph and that she conceived by him. Usener suggests that this statement was “judged inadmissible” by the redactor who interpolated i. 34 f., and that in consequence it was expunged. There can be little doubt that reasoning such as this requires omniscience as well as intuition! And the same criticism would be just in reply to any one who should elect to tell us exactly what St. Luke himself has altered or omitted. These are things which we do not know, and which we cannot know; we cannot even “infer with certainty” that St. Luke has omitted anything at all. But the broad possibility that he may have effected transformations and modifications in cc. i, ii, which we cannot now trace, is quite another matter, and, indeed, is by no means improbable. And if this is so, must it not affect the judgement we pass upon the skill or lack of skill which, on the theory proposed, St. Luke has shown? We may not know all. [pg 080] Obviously, we cannot prove this, but it is a consideration which we ought to have in mind.
(2) A second thing to be remembered is that, if our theory is true, we do not know anything of the actual circumstances under which the new tradition was introduced into the Gospel; it may have been in haste. Did the story reach the Evangelist at the last moment? Or, if not, was there a process of sifting and testing of the new information, which left little time when at length the fateful decision was taken, and the Evangelist took up his pen? Again we cannot prove these things, but again we cannot deny them. And if we cannot deny them, we must not ignore them. Only if we do ignore these possibilities, are we at liberty to insist that the reconstruction should have been more drastic. If, as we ourselves think, the supposition is reasonable, that i. 34 f. was added when the Evangelist had only just heard of the Virgin Birth tradition, we have clearly a good answer to the objection we are considering.
The foregoing arguments are speculative; there are, however, more positive considerations to urge. In addition to what has been said, we may point out (3) the fact that St. Luke's writings left his hand without a painstaking final revision, and (4) the different effect upon the mind of a new piece of information as compared with a belief, which has been held for some time, and has already become an intellectual presupposition.
(3) That St. Luke's writings left his hands without a final revision is strongly supported by the literary phenomena of the two works. The clearest evidence is found in the Acts, in which we probably have a closer literary parallel to the Birth Stories of Lk. i, ii than in the rest of the Gospel itself. Writing on the Acts (Acts of the Apostles, Eng. Tr., pp. 203 ff.) Harnack gives a list of more than two hundred “instances of inaccuracy and discrepancy”. Harnack does not accept them all, and shows that they are of different types, many of them being comparatively trifling and unimportant. Some are cases of anacoluthon and of transition from indirect to direct speech and vice versa. There are also “cases where St. Luke introduces persons with a certain unconcern, or in other places seems to forget that he has already introduced them” (p. 230). Harnack points out that “the details of a story are here and there inserted [pg 081] later or again earlier than their proper place” (p. 227), and he asserts that “instances of redundancy, of awkward repetition, of silence upon important points, and of extraordinary brevity, can be adduced from different parts of the book” (p. 230). He finds “instances of discrepancy” in the three accounts of the conversion of St. Paul, the letter of Claudius Lysias, the report of Festus, the last speech of St. Paul at Rome, and in other passages (p. 231).
Adequately to enter into this very interesting question would take us too far beyond the limits of our main subject. It is perhaps not unfair to suggest that Harnack's long list, as given in pp. 203-25, is capable of very considerable reduction. There is great force in Ramsay's remark: “He who reads Luke without applying practical sense and mother-wit and experience will always misunderstand him”, and in his caution: “When you think you find an ‘inconsistency’ in Luke, you should look carefully whether you have been sufficiently applying these qualities, before you condemn the supposed fault” (Luke the Physician, p. 55). Ramsay himself admits, however, that there are inconsistencies which cannot be denied, and holds that they show that “the work never received the final form which Luke intended to give it, but was still incomplete when he died” (ib., p. 24). In his earlier work, St. Paul the Traveller and the Roman Citizen, Ramsay has made the same suggestion, illustrations of which he finds in Acts xvi. 19, 20 and xx. 4, 5.[75]
We may describe the impression which St. Luke, as a writer, makes upon us by saying that, while his work is marked by great literary art, and while it is characterized by many striking instances of historical accuracy, yet, at the same time, the [pg 082] Evangelist shows a certain unconcern in matters of detail (Harnack would call it “a certain literary carelessness”), the results of which would probably have disappeared had he subjected his works to a close final revision. If this view is just there is little weight in the objection that, on the theory we have stated, St. Luke's reconstruction might have been expected to be more drastic than it is. The inconsistencies he has left are like those which we find elsewhere and are a feature of his works as they stand.
(4) Our final argument is of a psychological kind. It rests, as we have said, upon the difference between an intellectual prepossession and the first effect upon the mind of new information.
The previous argument might seem to point in another direction. Will not the character of St. Luke's writings sufficiently explain the literary phenomena of Lk. i, ii, on the view that he taught the Virgin Birth from the first? In the light of the discrepancies which occur in the Gospel and the Acts, can we not believe that after all the Virgin Birth is an original element in the Gospel? This contention would be an example of what Harnack has called attempting to gather apologetic figs from sceptical thistles.[76] We do not think that in this case the harvest would realize expectations.
It must be remembered that the two cases are not parallel. In the one case we begin with a writer whose mind is filled with an intellectual presupposition, with a knowledge, that is to say, of the Virgin Birth presupposed. Under these circumstances the miracle must be “a necessary stone in the structure”, and its effect determinative. If the Virgin Birth had been known to St. Luke for some considerable time, we cannot think that Lk. i, ii would have possessed the features to which we have called attention in Chapter II. In the other case—that of our hypothesis—the Virgin Birth is a piece of new information, and, if this is so, we submit that inconsistencies left in the adapted narrative wear a different hue. It is one thing to introduce into a narrative what is inconsistent to one's presuppositions. It is quite another thing not to perceive inconsistencies at once, when our knowledge is enlarged by a totally new fact. A presupposition is much more despotic than a subsequent discovery.
It is common knowledge that the implications of a new point of view are not always immediately recognized. For a time old and new live together. It is not by any means an easy task to introduce into a narrative, constructed under the guidance of alien presuppositions, a fact of an entirely new order. That St. Luke should have performed his task so well argues no little skill in literary craftsmanship. That his work was not completely done is after all no more than we might expect. From the standpoint of literary exactitude, no doubt the better plan would have been to rewrite the narrative, or at least to subject it to a rigorous pruning. But we ought not to complain if these things have not been done. St. Luke was probably too much of an artist to feel the merciless logic of his new information; and the result is a compromise.
In connexion with our theory we do not think that this is an unreasonable view to take. The difficulties are certainly much greater upon the theory that St. Luke knew of the Virgin Birth from the first. Granted certain presuppositions, and we can say with good reason what a writer like St. Luke would not be likely to do. Assume the entrance of a new fact, transforming by a whole world of difference the writer's point of view, and who can say just what he would do? We can say, of course, that he would introduce his new knowledge, if persuaded of its truth; but when we come to the details of reconstruction, we are face to face with the uncertainties of the personal equation. The logical procedure is drastic revision. If the writer stops short of this, as he may very well do, and attempts to fuse his material, seams must show and markings remain. This is precisely what we find in Lk. i, ii. In i. 34 f. and its context we can detect the seams; in c. ii we can see the markings.
It will be recognized that the situation is quite different on the view which credits i. 34 f. to a later Christian editor. Against this theory the objection we are considering has much greater force. For it is unlikely that the redactor would approach the Gospel with a knowledge of the Virgin Birth but lately gained. On the contrary, it would probably be a doctrine with which he had long been familiar. Accordingly, in addition to the other objections that we have raised against the theory of late interpolation, it would be legitimate to ask, Why has the redactor [pg 084] not done his work better? Our own hypothesis—that St. Luke had only just entered into a knowledge of the new tradition—is, indeed, the one theory where we have the least need to ask this question.
For the reasons given, we believe that the objection that St. Luke's revision should have been more drastic is not insuperable. That there is difficulty we allow. But there is probably no solution of the Lukan problem, not even the correct one, which will not leave difficulties of a kind. The problem is complex and the facts often elusive. It is on the ground that the theory we have sketched leaves least difficulties, and does justice to the facts as they appear, that we venture to find in it a reasonable solution of the problem of the Virgin Birth in relation to the Third Gospel and to St. Luke.