THE ORIGIN AND TRANSMISSION OF THE LUCAN NARRATIVE
ig HAS been shown in Chapter III that the whole narrative in Lk. i-ii is
Semitic and Palestinian in character, and displays a striking lack of acquaintance with events in the later history of the apostolic Church. These characteristics, it has been shown further (in Chapter IV), appear with special clearness in the two hymns, the Magnificat and the Benedictus.
But how is this Palestinian character of the Lucan infancy narrative to be explained? Various hypotheses suggest themselves.
The most obvious hypothesis, of course, is that in this section the author of Luke-Acts made use of a Palestinian source, written originally in Hebrew or Aramaic, which either had already been translated into Greek or else was translated by Luke himself.
This hypothesis has been opposed in recent years especially by Harnack.! By a detailed examination of the language of the infancy narrative he has sought to show that after subtraction has been made of the words and phrases taken from the Septuagint what remains is so characteristically Lucan that Luke must have been the author, and not merely the editor, of the narrative. He does not, indeed, exclude the possibility that Luke may have used an Aramaic source in these chapters, but he insists that if Luke did so he must have translated the source himself; he could not have found it already existing in Greek; the Greek form of the infancy narrative must be due to Luke alone. And although Harnack admits this possibility that Luke translated an Aramaic source, he is evidently inclined to favor the other view, that in the composition of this narrative oral tradition only was used, so that Luke was in the fullest sense the author, and not merely the editor or the translator, of the written narrative.
How then does Harnack explain the Semitic coloring of the language in this section? He does so very largely by insisting upon conscious imitation, on the part of Luke, of the style of the Septuagint. In the prologue, he supposes, Luke is writing according to his own natural style; in the following narrative he is imitating the style of the Septuagint.
At first sight the hypothesis may seem to be very unlikely, since it seems to attribute to the author of Luke-Acts a refinement of art which is hardly natural in an ancient writer. But possibly first impressions in this matter may have to be modified. As a matter of fact, imitations of the Septuagint, or, to say the least, detailed influence of the Septuagint, in the Lucan writings cannot altogether be denied. For example, despite his Greek literary affinities, Luke uses the Hebraistic phrase, “it came to pass,”* far more than it is used by any other New Testament writer. Evidently he had a keen appreciation for what may be called the “Bible style” of the Septuagint, and felt that that style was peculiarly fitted to be the vehicle of his own sacred narrative. Harnack’s contention would simply lead us to say, in effect, that in treating in a poetical manner the events connected with the Saviour’s birth, Luke carried the imitation of the Septuagint style somewhat farther than he did when he was narrating in a more matter-of-fact way the events of the public ministry, and very much farther than he did when he came to events that took place in a Gentile environment, like the events of the missionary journeys of Paul. In the case of the public ministry, we might say that Luke was hindered by his sources from carrying out his stylistic plans with perfect freedom; and in the case of the missionary journeys of Paul the subject-matter did not lend itself at all to imitation of the Septuagint. But in the infancy narrative there would be no such hindrances, and the style of the Septuagint could freely be employed if the author pleased.
Thus the hypothesis of Harnack cannot be regarded as inherently impossible. In support of it he has in a very careful way * gone through representative sections of Lk. i5-ii.52, pointing out Lucan peculiarities—that is, words or usages which occur only in Luke-Acts among the New Testament writings or else occur more frequently there than in the rest of the New Testament and especially in Matthew and Mark. The work of Harnack has received a valuable supplement from Zimmermann,* who examined in detail those portions of the narrative which were left unexamined in Harnack’s earlier discussion. In
8Harnack, “Das Magnificat der Elisabet (Luc. 1, 46-55) nebst einigen Bemerkungen zu Luc, 1 und 2,” in Sitzungsberichte der koniglich preussischen Akademie der Wissenschaften 2 Berlin, 1900, pp. 537-556; Lukas der Arzt, 1906, pp. 69-75, 138-152 (English translation, Luke the Physician, 1907, pp. 96-105, 199-218); compare also Neue Untersuchungen zur Apostelgeschichte, I91I, pp. 108-110 (English translation, The Date of the Acts and of the Synoptic Gospels, 1911,
‘ cy a des Lukas Kap. 1 und 2,” in Theologische Studien und Kritiken, \xxvi, 1903, pp. 247-290. Compare also Plummer, in his commentary (in the International Critical Commentary) and Stanton, The Gospels as Historical Documents, ii, 1909, pp. 291-295; Hawkins, Hora Synopticae, second edition, 1909, pp. 15-24, 27-29, 35~SI. Compare also various observations in Vogel, Zur Charakteristik des Lukas, 2te Aufl., 1899; and in Friedrich, Das Lukasevangelium und die Apostelgeschichte, Werke desselben Verfassers, 1890.
Harnack’s more recent work he has carried the examination through part of the sections that Zimmermann had already covered.*
These recent investigators were anticipated one hundred years ago by Gersdorf,° who defended the first two chapters of Matthew and Luke, as original parts of the First and Third Gospels, by an elaborate linguistic argument. The chapters in question were explored very much after the method that has been adopted by Harnack and Zimmermann. Lk. i-ii, for example, was traversed from beginning to end in order to exhibit those linguistic features which connect it with the rest of Luke-Acts as the work of the same writer. In thoroughness, Gersdorf was not one whit inferior to the more recent investigators. It is remarkable how very seldom Harnack or Zimmermann has detected a Lucan characteristic which Gersdorf had not already observed; and in a number of cases Gersdorf observed what his successors have overlooked. Gersdorf labored with insufficient textual materials and was too much inclined to emend the text in order to secure absolute uniformity of style; but such - faults do not affect the permanent usefulness of his work. The neglect with which he has been treated, in recent years, even by scholars who have been over exactly the same ground,’ is undeserved.
Harnack and Zimmermann agree in excluding a Greek written source for Lk. is-iis2. The style of the passage, they maintain, is found, after due allowance is made for peculiarity of the subject-matter and for imitation of the Septuagint, to be so totally Lucan that Luke must have been something more than the mere editor; he must have been the first to treat the material in a Greek narrative. If he had used a Greek source, they maintain, the style of the source would necessarily appear in the use of a non-Lucan vocabulary and phraseology.
So far, Harnack and Zimmermann agree. But they differ in what they substitute for the rejected hypothesis of a Greek source. Zimmermann supposes that Luke used an Aramaic source which he translated himself; Harnack,
5 The result is as follows:—i.5—15 has been examined both by Harnack and by Zimmermann; i.16-38, by Zimmermann; 1.39-56, by Harnack, supplemented by Zimmermann; 1.57-67, by Zimmermann; i.68-79 by Harnack, supplemented by Zimmermann; i.80-11.14, by Zimmermann; 1i.15-20, by Harnack, supplemented by Zimmermann; ii.21-40, by Zimmermann; ii. 41-52, by Harnack, supplemented by Zimmermann. It will be observed that the essay of Zimmermann was written after Harnack’s earlier work, but before the later work. In his later work, Harnack displayed no acquaintance with Zimmermann’s investigations. In i.5-15, therefore, Harnack and Zimmermann have investigated the same material independently. In view of their independence, the agreement of the investigators in many of the proofs urged with regard to this passage becomes significant.
8 Beitrage zur Sprach-charakteristik der Schriftsteller des Neuen Testaments, 1816.
Feine (Eine oorkanonische Uberlieferung des Lukas, 1891, p. 19) constitutes an exception in that he does recognize the value of Gersdorf’s work.
while admitting the possibility of an Aramaic source, apparently thinks it more probable that Luke depended merely upon oral tradition.
Harnack began his investigation with the Magnificat and the Benedictus, in the case of which, as we have already observed, he excluded the possibility, which he admits for the rest of the narrative, that Luke translated an Aramaic source. The language of the hymns, he maintained, is, after subtraction of the Septuagint element, so totally Lucan as to show unmistakably that the hymns are free compositions of the author of the whole double work. This contention has been refuted in the last chapter. Far from showing that the hymns were free compositions of Luke, Harnack has not even succeeded in showing that the hand of Luke has been at work on them at all; so far as the linguistic phenomena go, it is possible, though not obligatory, to hold that the hymns came to Luke in a Greek form and that he incorporated them in his work without change.
But if the arguments of Harnack are unconvincing with regard to the hymns, they are far from being without force for the rest of the narrative. In 1911 the present writer engaged in a rather detailed examination of the whole infancy narrative, in order to determine how many of the Lucan characteristics detected by Gersdorf, Harnack, Zimmermann and others, might more properly be regarded simply as characteristics of the Septuagint.® The result showed, indeed, that Harnack had not proved his point; many of the words and phrases which Harnack had listed as Lucan may equally well be placed in a different category. The phenomena are not such as definitely to exclude the use of a Greek written source in Lk. i-ii. Nevertheless, after all such deductions have been made, enough remains to show plainly that the hand of the author of the whole book has been at work in this section. That much, at least, has been demonstrated by Harnack. We do not, indeed, think that the Lucan characteristics, after the necessary deductions have been made, are sufficient to show that Luke was the first to put the narrative in a Greek form and sufficient to exclude the hypothesis that he used a Greek written source; but they are amply sufficient to show at least that if he did use a Greek written source he used it with freedom and made it conform in important particulars to the style of the rest of his great double work.
For one thing, the objection which Harnack raised to our method of examining his argument,’ applies with very much more force to the whole narrative than it does to the hymns. Harnack had said, in effect, that when
8 “The Hymns of the First Chapter of Luke” and “The Origin of the First Two Chapters of Luke,” in Princeton Theological Review, X, 1912, PP. 1-38, 212-277. j; ® See Harnack’s review of the two articles by the present writer (cited above), in Theologische the Septuagint element is removed, what remains is so characteristically Lucan that Luke must have been in the full sense the author, and not merely the editor, of the narrative. Our investigation showed, by a detailed use of the Hatch-Redpath concordance to the Septuagint, that the Septuagint element in the language of the narrative is vety much greater than Harnack supposed, so that many of Harnack’s Lucan characteristics are really just characteristics of the Septuagint, likely to be used by anyone who was influenced by the Septuagint style. In reply Harnack insisted, in effect, that even if the characteristics to which he points are also found in the Septuagint, that does not necessarily destroy their evidential value, since a choice of just the same usages from the rich store of Septuagint phraseology might indicate unity of authorship as well as it would be indicated by coincidence in other respects.
This reply, which had been anticipated by Ladeuze and to some extent by the present writer,’° is by no means without weight. And yet we think it does not quite go the whole way. When it is observed that similarity in language between the Lucan writings in general, on the one hand, and the Septuagint, on the other, is far greater than that between the Septuagint and the other New Testament writers, then some, at least, of the coincidences in language between Lk. i-ii and the rest of Luke become less significant. Where else in the New Testament can a section be found which approximates so closely as does Lk. i-ii to the narrative of the Old Testament? Part of the exceptionally close affinity of Lk. iii-Acts xxviii to Lk. i-ii may be due simply to an exceptionally close affinity to the Septuagint in general. Other New Testament writers may be found to diverge more than Luke does from Lk. i-ii simply because they diverge more from the Septuagint. Thus Harnack’s reply, although unquestionably it deserves to be borne carefully in mind, does not altogether destroy the damage to his extreme conclusion which comes from the discovery in the Septuagint of a large number of those words and phrases which he designated as specifically Lucan. The Lucan residuum being smaller, after subtraction of the Septuagint element, than he supposed, perhaps he ought to be content with the hypothesis of Lucan employment of a previously existing source, instead of insisting that Luke must have been the first to put the narrative in a Greek written form.
A second caution, moreover, needs to be borne in mind before Harnack’s conclusion can be accepted in its entirety. May not some of the parallels between Lk, i-ti and the rest of Luke-Acts be explained as due probably to dependence of Luke upon various sources of a common type? With what part of Luke- Acts is Lk. i-ii to be compared in order to exhibit its completely Lucan character? Obviously the comparison would best be made, if possible, with parts where Luke’s own style appears in its purity. But this requirement is not certainly satisfied, for example, in the former part of Acts, especially in the speeches which are there attributed to Jewish Christians. Very probably Luke is there using sources, and sources of a Jewish Christian kind very much like the source that has been posited for Lk. i-ii. Therefore the affinity of Lk. i-ii with these early chapters in Acts does not necessarily prove full Lucan authorship. It may conceivably prove nothing more than authorship by a Jewish Christian, whose linguistic and religious environment was the same as that of the author or authors who produced the sources of Luke’s account, in Acts, of the primitive Jerusalem Church.
Finally, even if Lk. i-ii is found to coincide in a certain usage with elements in Luke-Acts which may be admitted to be actually due to Luke himself, that still does not necessarily prove dependence of Lk. i-ii upon the Lucan usage in question. For the dependence may be the other way around. This possibility has indeed usually been neglected in the discussion of the subject. But such neglect is quite unjustifiable.
That Luke’s style has been colored by at least one phrase in the Benedictus (regarded as a previously existing hymn) has already been shown to be possible, and there is no good reason why the dependence may not have extended also to other parts of the infancy narrative.
One case of such possible dependence deserves special consideration. It is possible dependence upon the angelic doxology in Lk. ii14. With that verse, which reads according to the correct text, “Glory to God in the highest, and on earth peace among men of [His] good pleasure,” * Gersdorf compared the latter part of Lk. xix.38, where the words of the multitude at the triumphal entry read, “In heaven peace, and glory in the highest.” ** These words are not in the other Gospels, but take the place in Luke of the “Hosanna in the highest” ** of Matthew and Mark. But what is the relation between this Lucan addition and the angelic song in Lk. ii.14? The two have two things in common; the “glory in the highest” and the conjunction of this with “peace in....” With regard to the former point, Ryle and James can cite an interesting parallel in the Psalms of Solomon: “glorious, dwelling in the highest.” ** This sug-
2 gdta ev WWloros OG Kal emt yijs eloqyn ev 4vOpwmos ebdoxlas.
13 zy obpav@ elpqrn kal dota ev dploros.
18 Psalms of Solomon, xviii.11 (xix.1). See Ryle and James, YAAMOI ZOAOMONTOZ Psalms of the Pharisees, commonly called Psalms of Solomon, 1891, p. Xcil. Compare Bruce in Expositor’s Greck Testament, 1, 1897, on Lk. ii.14; Resch, “Das Kindheitsevangelium,” 1897, in Texte und gests the possibility that the “glory in the highest” of Lk. ii14 and of Lk.
xix.38 may be quite independent. The idea was a natural one, and also the expression of it. The writer who first put the angelic song into Greek (supposing it to have existed first in Aramaic or Hebrew) may have written independently of the one who performed a similar service for whatever Semitic original there may have been for the words attributed to the multitude in Lk. xix.38. In the latter passage the “in the highest,” at any rate, was already given, in the tradition lying back of the “Hosanna in the highest” of Matthew and Mark. The conjunction of “peace” with “glory in the highest” might also be explained as due to mere coincidence, especially since the order is reversed and since “peace” goes with “in heaven” in the first place and with “on earth” and “among men of [His] good pleasure” in the other.
However, in view of the rather striking parallel between the two passages, other solutions of the problem call for careful consideration. Wellhausen *® is quite confident that Lk. ii14 is dependent upon Lk. xix.38. This hypothesis, however, is apparently connected with the untenable *” view that Lk. i-ii is an addition to the completed Gospel. The reverse hypothesis, that Lk. xix.38 is dependent upon Lk. ii.14, is favored by Holtzmann,"® B. Weiss,’® J. Weiss,”° and Bruce.”* In Lk. xix.38, Holtzmann supposes, the wording of the triumphant cry was exchanged for a reminiscence of the Gloria (Lk. ii.14), which had already become a hymn of the Christian congregation. There is nothing inherently impossible in this hypothesis, though if a correct view be held with regard to the date of the Gospel, the Gloria must have become a hymn of the Church long before Holtzmann would suppose to have been the case. If there is dependence, may it not be—rather than the reverse—dependence of the author of the Gospel, either upon the Gloria existing in separate form or else upon the Gloria already inserted in a source underlying Lk. i-ii?
Such dependence would not be irreconcilable with the trustworthiness of Luke as a historian. The cry of the multitude at the triumphal entry of Jesus into Jerusalem was not recorded, perhaps, with verbal exactness. Indeed, it was perhaps not stereotyped when it was first uttered. There were many persons who accompanied Jesus as He descended from the Mount of Olives; some, no doubt, said one thing, others another.” It is not surprising, therefore, that
18 Das Evangelium Lucae, 1904, p. 109: “That Lk. xix.39 is the model for ii.14 cannot be doubted, although Weiss reverses the relationship.”
20 Die Schriften des Neuen Testaments, 3te Aufl., revised by Bousset, 1917, in loc.
the tradition of what was said is not exactly uniform. The characteristic cry of the multitude—what was heard again and again during the descent into the city—might have impressed itself upon a hearer with greater or less fullness. The translation into Greek, moreover, necessarily brought divergence from the exact words that were spoken by any one person among the multitude. One Aramaic word, “Hosanna,” has been preserved by three of the Evangelists. Luke, for the benefit of his Greek readers, may have substituted Greek words for it. His words, “in heaven peace, and glory in the highest,” reproduce perhaps the general meaning of the “Hosanna in the highest” which has been preserved by Matthew and Mark. That latter phrase is not altogether clear to modern readers, as a glance at the commentaries will show; and possibly it may not have been clear to the first readers whom Luke had in view. In order to bring out the meaning of the original cry—that is, its real significance in the feeling of the original speakers—possibly literal translation was insufficient. There are cases where a certain amplification is the truest translation.
We are far, indeed, from affirming that this is the true way of explaining the divergence of Luke at this point from the other two Synoptic Gospels. To those who have not accepted without modification the current view of Luke’s dependence solely upon Mark, it will seem more probable that Lk. xix. 38b is no mere amplification of “Hosanna in the highest,” but rather a reproduction of independent tradition.
But the decision with regard to that question does not affect essentially the problem involved in the parallel with Lk. ii.14. For the tradition was not in the first place fixed in a Greek form; if it came to Luke in Greek, it came in a translation, which might be modified with more freedom than would have been permissible in the case of the wording in the original language. In any case, therefore, Luke may have made use of the words of the angelic doxology in determining the form, though not the content, of the triumphal cry as it was to appear in his Gospel. He may have done so consciously or unconsciously, more probably the latter. The song of the angels, as recorded in the infancy narrative, was, it may be held, familiar to him; the wording of it would then come naturally from his pen when he was putting the similar praise of the multitude into a form which would be both intelligible and impressive to himself and to his Greek readers.
Such a solution of the problem would be a compromise between two hypotheses; the hypothesis that Lk. ii14 and Lk. xix.38 are completely independent of each other, and the hypothesis that the latter is dependent upon the former. The present form of Lk. xix.38, so the compromise hypothesis would hold, is partly due to a real coincidence in the praises originally rendered to God in the two cases, and partly to Luke’s familiarity with the wording of the angelic hymn.
At any rate, the reverse hypothesis, that Luke received Lk. xix.38 in its present form, and then conformed the wording of the angelic hymn to it, is less probable. In view of what has just been said, Luke must be pronounced less likely to have conformed to something else the wording of the angelic hymn than the wording of the cry of the multitude; for in the former case the various special considerations that we have just adduced would not apply. At any rate, the view that Luke conformed Lk. ii14 to the wording of Lk. xix.38 is very unlikely unless he was himself the first translator of the angelic song or of the whole narrative in which it is contained. In that case he may conceivably have allowed himself such freedom.
Against the view that he actually composed the angelic hymn himself various objections present themselves. For instance, the very bold Hebraism, “men of good pleasure,” ** would be difficult to explain.
Enough, then, has been said to show that the similarity between Lk. ii.14 and Lk. xix.38 does not prove that Luke was the actual composer of the former. The most that could be held is that he was the first to put the angels’ song into Greek. That would perhaps help to explain his familiarity with its wording and his natural employment of it in Lk. xix.38 without artificial imitation. But such an hypothesis is not necessary.
This instance may serve to show, not, indeed, that dependence of the rest of Luke-Acts upon a source underlying Lk. i-ti is certain, but that it is at least possible; so that for this reason, as for other reasons, even striking similarities of style between the infancy narrative and the rest of the work do not by any means indubitably prove that the author of the whole work composed that narrative without the use of sources. Some of those similarities may be due to dependence of the author in the rest of his work upon the language of the source that he used at the beginning.
It is true, the example which we have just discussed at some length is taken from a hymn—the angelic doxology—which may conceivably have existed in separate form; so that dependence upon that hymn would not necessarily prove dependence upon the infancy narrative as a whole. But who can say that this latter possibility must necessarily be ruled out of court? There is no real reason why the dependence of Luke should not be extended even to the narrative portions of Lk. i-ii, if he was really using a written source in that section of his Gospel. He was attracted, not unnaturally, by the simple grandeur and poetic dignity of the “Bible style,” and followed, therefore, the Old Testament model, as we have seen, in his own sacred narrative. But why should such linguistic influence have been exerted by the Old Testament alone? If Lk. i-ii is what at first sight it would seem to be, a poetic narrative produced on the very native soil of the Old Testament, at a time when the Old Testament type of language was still a living thing, and if this narrative fell into the hands of Luke, what would be more natural than that it, as well as the Old Testament, should impress itself permanently upon his mind and heart? Surely it is hardly inferior in beauty to the best of what the Old Testament can offer; and it is concerned with the events most stimulating to the Christian imagination. It may well have taken a place side by side with the Old Testament in moulding the literary gifts of the Greek historian for a sacred use.”
However, after all deductions have been made, the Lucan residuum in the style of Lk. i-ii remains amply sufficient to prove that the author of Luke-Acts certainly had a part in the production of the present form of the infancy narrative. Nothing that we have just been saying should be allowed to obscure that very important fact. Some of the “Lucan characteristics” discovered by Harnack were shown by our investigation of the Septuagint usage not to be also characteristic of the Septuagint. Many of them are to be found, moreover, not only in the former part of Acts, but also in the latter, and presumably more purely Lucan, part. Finally, some of them can be shown not to have been taken by Luke from Lk. i-ii, because they belong not to Semitizing Greek or even to popular Greek, but rather to the literary form of the Koiné. Hence Luke was certainly at least the editor of Lk. i-ii. That fact was established by Gersdorf one hundred years ago. It has been confirmed and not disproved by recent investigations—confirmed not only by Harnack, but also, we think, by our criticism of Harnack.
Thus the first (and perhaps the most important) result of the detailed examination to which in 1911 we tried to subject the argument which Harnack and Zimmermann had developed for detecting Lucan style in Lk. iti was a clear confirmation of that argument. In the birth narrative, we found, the hand of Luke has certainly been at work. Some of the supposed indications of Lucan style were eliminated; but the severity of the test exhibited only all the more clearly the cogency of the proof that remained.
2% Compare Stanton, The Gospels as Historical Documents, ii, 1909, p. 291: “Luke may have been led to use the expressions in question [éyévero, etc.] partly from his own familiarity with the LXX, partly from his having become accustomed to them in copying this document at the beginning.” Compare also Feine, Eine vorkanonische Uberlieferung des Lukas, 1891, p. 20: “The source which we are tracing is used only in this one Gospel; we have therefore no standard for judging it outside of the Third Gospel. Moreover, if this source bore a Hebraistic stamp, Luke himself also introduced Hebraizing turns of expression in his two writings, and also knew how to assimilate the linguistic material of his sources and use it further in an independent way.
On the other hand, however, the investigation resulted in a deepened impression of the affinity of Lk. i-ii for the Septuagint. Many of the supposed Lucan characteristics were shown to be merely characteristics of the Septuagint. Of course Harnack himself admitted the presence of a large Septuagint element in the language and style of Lk. i-ii. But our investigation showed that element to be much larger than Harnack supposed.
This remarkable affinity of Lk. i-ii with the language of the Old Testament points clearly to a Jewish Christian origin of the narrative. We are not forgetting that from the very beginning the Old Testament Scriptures were accepted as authoritative by the Gentile Christian Church as well as by the Jewish Christian Church. But an examination of patristic literature will show that despite the fact that the Old Testament was universally accepted as authoritative, and despite the fact that it is quoted at great length by the patristic writers, yet the style of the Gentile Christian writers remained their own and was by no means so suffused with Septuagint usages as is the case in Lk. iti. The close similarity of the language of Lk. i-ii to the language of the Old Testament does, therefore, still require some special explanation.
Moreover, our detailed examination only served to exhibit anew the Palestinian character of the narrative. That Palestinian character does not appear merely in those words and phrases that are also found in the Septuagint, but runs through the whole section. It appears, moreover, as we have seen,” not merely in the language but also in the thought. Recent researches have not in the slightest served to modify the first impression made upon an unprejudiced reader that we have in this section a genuinely Jewish and Palestinian narrative.
Even Harnack does not altogether deny the correctness of that impression; for although he holds that the narrative was composed by the Gentile Christian, Luke, he admits at the same time that in composing it Luke made use of Palestinian tradition. It must be remembered that according to Harnack the author of Luke-Acts was actually Luke the physician, a man who was demonstrably in Palestine at the beginning and end of the two years that Paul spent in prison at Caesarea, and who, because of personal associations, would have had abundant opportunity to acquaint himself both then and at other times with what Palestinian disciples were saying. Harnack holds that the stories about John the Baptist which appear in the Lucan infancy narrative must have originated in a circle of John’s disciples; and that the stories about Jesus must have come to Luke with the claim of having been derived from Mary, and therefore must certainly have come from Palestine.?®
What, then, is our conclusion regarding the hypothesis that Lk. i-ii was composed by the author of Luke-Acts himself without the use of written sources? That hypothesis, as we have already seen, must emphatically be rejected so far as the hymns are concerned. But with regard to the bulk of the narrative, the answer is not so simple. One form in which the hypothesis might be held must indeed certainly be excluded. Luke certainly did not compose the narrative simply by artificial imitation of the Septuagint and artificial adaptation of the details to Palestinian conditions of which he had no firsthand knowledge. Such a refinement of art is almost inconceivable in an ancient writer. If, therefore, the author of Luke-Acts actually composed the birth narrative himself, he must have been himself in close touch with Palestinian conditions. This supposition is by no means unnatural. Harnack ™ can even suggest that before he united himself with the Christian community the author of Luke-Acts had belonged to the disciples of John the Baptist and had already pursued investigations which he afterwards used for his Gospel. However that may be, he certainly—supposing him to be the person who speaks of himself in the first person in the we-sections of the Book of Acts—came to Palestine at the time of Paul’s last visit to Jerusalem. At that time, or on previous occasions of which nothing definite is known, he may have acquainted himself intimately with Palestinian conditions. And certainly he may have come into possession of definite Palestinian tradition with regard to the birth of John and Jesus. If so—if he had himself been in Palestine, if he had come into possession of Palestinian oral tradition—and if, moreover, he put that tradition into written form without any long delay, then conceivably he might have composed the birth narrative without the aid of written sources. In this form, the hypothesis is perfectly possible. It was by no means rejected categorically when we examined the narrative in detail. All that was maintained is that the hypothesis has not been proved; all that we maintained (at this point) against Harnack is that the non-Semitic Lucan element in Lk. i-ii is insufficient to exclude the possibility, at least, of written sources.
The important thing to observe is that if Luke was the first to put the Lucan birth narrative into written form, even then the genuinely primitive Palestinian character of Lk. i-ii cannot be denied. If Harnack is right, if the linguistic phenomena show that Luke was the original author of Lk. i-ii, that proves not that Lk. i-ii is late, but that the whole of Luke-Acts is early. If it can be proved that this section was composed by the author of Luke-Acts, then we have simply one more weighty argument in favor of the Lucan authorship and early date of the entire work. The hypothesis that Luke composed
Lk. i-ii on the basis of primitive Palestinian oral tradition, aided by first-Land acquaintance with Palestinian conditions, is very attractive. It explains admirably many of the facts. It may well be correct. But it has not been proved.
A second hypothesis to be considered is that for Lk. i-ii Luke himself translated an Aramaic written source2® This hypothesis would explain on the one hand the thoroughly Palestinian character of the content of Lk. i-ii and on the other hand the undeniable Lucan elements in the style. That Luke should have known sufficient Aramaic to translate an Aramaic document can hardly be pronounced impossible. Thus Harnack, who himself is inclined to reject the hypothesis of an Aramaic source for Lk. i-ii, supposes that Luke may have translated an Aramaic document in the early part of Acts, and adds that knowledge of Aramaic sufficient to translate a simple Aramaic text may well be attributed to a native of Antioch (as Harnack thinks Luke to have been) and a companion of Paul.” This hypothesis is possible. But like the first hypothesis it cannot be proved.*°
The third hypothesis is that Luke employed a Greek written source. Our examination of the linguistic phenomend showed that this hypothesis, like the two others, is possible. After deducting from the language of Lk. i-ii what is characteristic of the Septuagint and what is natural in a Jewish Christian document, the Lucan element that remains is insufficient to prove anything more than Lucan revision of a previously existing document, though on the other hand it is sufficient to render actual Lucan composition—with the qualifications that we have noted above—also possible. The hypothesis of a Greek written source, like the other two hypotheses, is possible, but not certain.
If Luke was using in Lk. i-ii a Greek written source, then the source may have been composed originally in Greek, or the Greek form in which Luke used it may have been produced by previous translation from Hebrew or Aramaic.
80 The hypothesis that Luke himself translated a Semitic document in Lk. i-ti is upheld by C. C. Torrey (“The Translations Made from the Original Aramaic Gospels,” in Studies in the History of Religions Presented to Crawford Howell Toy, 1912, pp. 285-295). But Torrey thinks that the Semitic document was Hebrew, not Aramaic. This suggestion is worthy of consideration. That the author of Luke-Acts should have known Hebrew is perhaps less likely than that he should have known Aramaic; but in view of our ignorance concerning his birth and education, it cannot, perhaps, be pronounced impossible.
According to Resch,** Gunkel,*? and C. C. Torrey,*® the source was originally composed in Hebrew. That view would no doubt explain admirably the linguistic phenomena; for in the opinion of an expert like Dalman* a number of the Semitisms in Lk. i-ii—for example, the familiar narrative use of the phrase, “it came to pass’—are Hebraisms, not Aramaisms. But is it likely that such a narrative should have been composed in Hebrew in the first century after Christ? At that time Hebrew seems to have long ceased to be the ordinary language of Palestine.®® Yet it also seems to have remained in use as the language of certain kinds of literature®® Perhaps we may say that if Lk. i-ii was originally written in Hebrew, then it was probably intended from the first to be in some sort a sacred narrative, for which the sacred language would be the fittest vehicle. In view of the lofty, poetical tone of the narrative, comparable to the best parts of the Old Testament, such a supposition is by no means impossible. The necessary knowledge of Hebrew would not have been lacking; for, despite linguistic changes among the people, the Old Testament continued to be read and studied in its original language. It is true, however, that the earliest Christian community, despite the presence of many priests,*” was probably composed chiefly of persons who, in the sense that was true of the apostles,®® were “unlearned and ignorant men.” And in such a community the employment of the ordinary language of the country for such a work as the source of Lk. i-ii may perhaps be regarded as more natural than the employment of Hebrew.
The Hebraisms (as distinguished from Aramaisms) of Lk. i-ii have not always been regarded as overbalancing these a priori considerations. Indeed, they have sometimes been regarded as furnishing an argument not for, but positively against, any Semitic original for the narrative. For, it is said, Hebraisms are not Aramaisms; they might establish a Hebrew original, but certainly not an Aramaic original. But if a Hebrew original, it is argued, is impossible in view of the linguistic conditions prevailing in Palestine in the first century after
% Zum religionsgeschichtlichen Verstandnis des Neuen Testaments, 1903, p. 67. According to Gunkel it is “very probable” that the narrative goes back to a Hebrew original.
Loc. cit. It will be remembered (see above, note 30) that Torrey thinks that the Hebrew source was translated not by some previous translator, but by Luke himself.
88 Acts iv.13. That sense is by no means exactly the one that we attach to the corresponding
Christ, then the Hebraisms still require explanation. According to Dalman, they can be explained only as due to the Septuagint; in other words, they are not really Hebraisms, but “Septuagint Graecisms” or “Greek Biblicisms.” °° Of course the Hebraisms might conceivably be explained as due to an influence exerted upon the Aramaic of the source by the Hebrew Old Testament. The apparent Hebraisms would then be Aramaic Biblicisms. Thus, although the familiar Hebraistic phrase, “it came to pass,” has no equivalent in the living Aramaic, yet in the (Aramaic) Targums the Hebrew usage is imitated.*° It might have been imitated also in the Aramaic source of Lk. i-ii. But this Dalman regards as improbable, because (he evidently thinks) the imitation of a Targum is hardly to be attributed to Christian writers.“ In general, says Dalman,*? “the Jewish Aramaic, as it lived among the people, displayed even less tendency to adopt Hebrew expressions than did the Greek of the Synoptic Gospels.” Hence Dalman can even enunciate the principle for the literary criticism of the New Testament: “the more Hebraisms, so much the more activity of Hellenistic redactors.” “? The Hebraisms in Lk. i-ii are thus made to afford merely another support for the view of Dalman that the narrative was originally composed in Greek.
Such questions certainly cannot be settled without first-hand knowledge of the dialects of Palestine. But even if one accepts most of what Dalman says with regard to these dialects, the possibility of an Aramaic source “* does not seem to be altogether excluded. In the first place, it might be maintained (against Dalman) that the Aramaic source imitated the Hebrew of the Old Testament. That supposition, which certainly does not seem altogether unlikely, would explain some or all of the non-Aramaic Hebraisms of Lk. 1-i1. Or else, in the second place, these non-Aramaic Hebraisms might be explained by imitation of the Septuagint on the part of the Greek translator.“
At any rate, if Lk. i-ii was composed originally in Greek, it was composed by some one thoroughly familiar with Palestinian conditions and in all. prob-
“That the source was originally composed in Aramaic was designated by Klostermann, in the first edition of his commentary (Handbuch sum Neuen Testament, II. i, 1919, p. 364), as being the majority opinion. He cited in favor of it Plummer, Bousset, Gressmann.
46 What has just been said about the possibility that Luke’s source was originally written in
Aramaic has obvious importance also for the hypothesis (discussed above) that Luke himself translated an Aramaic document.
ability in the very early days of the Jerusalem Church.“® So much has been established by what has been said above about the Palestinian character both of the form and of the content of the narrative. That such a primitive Jewish Christian narrative should have been written in Greek is by no means impossible. The earliest Christian community at Jerusalem, which is described in the first chapters of Acts, was composed, to a very considerable extent, of Greekspeaking Jews.*?
Some scholars have attempted to establish a Semitic original for Lk. i-ii by pointing out mistakes in translation in the extant Greek narrative. Gunkel * regards “Christ the Lord” *° in Lk. irr as the clearest indication of a Hebrew original. It looks, he thinks, like a mistranslation of “the Lord’s anointed,” se which is certainly mistranslated thus in the Septuagint of Lam. iv.20. “Christ the Lord”—so no doubt Gunkel reasons, is an unprecedented phrase on Semitic ground; whereas a Christian translator, to whom the title “Lord” seemed natural as applied to Christ, would very naturally take the Hebrew phrase in this way. But the phrase, “Christ the Lord,” in Greek occurs also in the Psalms of Solomon, xvii.36. It is true, some scholars have regarded it there also as a mistranslation. Nevertheless, every additional occurrence of the phrase makes the theory of mistranslation less likely.°' Dalman, who has been followed by Loisy,®* supposes that “the Lord” was added by the Evangelist to explain “Christ,” which is used here for the first time in the Gospel. The matter is problematical at best. “Christ the Lord” may well be a correct translation of the source. That the tense of the verb in the clause, “seeing I know not a man,” ®4 in Lk. i.34 is due to a mistranslation has also not been proved. That passage will be discussed in the next chapter.
Our general conclusion is that many questions with regard to Lk. i-ii must be left undecided. It is uncertain whether the narrative was composed by Luke himself merely on the basis of Aramaic oral tradition (though we think that less likely), or whether he himself translated a Semitic document, or whether
46 If, indeed, Luke himself was the composer of the narrative—a possibility which we have not entirely excluded—the date would have o be put a little later, unless Luke should be supposed to have composed the narrative during some early stay in Palestine and then incorporated it at a later time in his Gospel. But if Luke was not the composer, there is not the slightest reason to put the date of composition later than the very early days of the Jerusalem Church.
«8 Zum religionsgeschichtlichen Verstandnis des Neuen Testaments, 1903, P. 67.
50 As the phrase appears in the Hebrew Old Testament.
51 See the defence of the correctness of the translation in Ps. Sol., xvii.36 by Ryle and James, in loc.
he used a Greek written source. It is also uncertain whether the source, if it came into Luke’s hands in a Greek form, was composed originally in Greek, or in Hebrew, or in Aramaic. But in the midst of so much uncertainty, two facts stand out clear. In the first place, the birth narrative formed an original part of the Third Gospel; and in the second place, it is genuinely primitive and Palestinian. These two facts are quite independent of the disputed questions. And they are the really important facts.