The Origin Of The First Three Gospels

Part 1


It is a remark often and appropriately made, that Christianity is a historical religion. Fully understood, the epithet sets forth, not the mere accident, but the very essence of the life of our faith. What God has so wedded together cannot by any course of criticism be separated and either part remain vital. As on the one hand history is without deep meaning and peculiar charm unless the doctrine of a redeeming Christ be the thread on which every one of its bright and dark beads is strung, so on the other, the doctrines and morality of Christianity will avail us little when parted from the historical Jesus, who, in his real character, in the facts of his life, suffering, and death, is the doctrine and the embodiment of the moral law. History is unintelligible without the doctrine of redemption: the doctrine is unreal, is not, without the historical Redeemer.

But this historical religion rests for us, in the main, upon certain books which claim to be histories. No questions are then more intimate to our faith than those which concern these histories, and among such questions surely none is more fundamental than that of their origin. We shall feel this when we have weighed well what would become of our faith if it should be proved that our Gospels originated as Strauss and his school claim that they did. If such an origin is proved, the truth must indeed stand, but Christianity as we prize it could not stand also. By a figure most apt the hypothesis of Strauss has been called “a fable-spinning sybil,” who like a vampire sucks all the fresh life-blood out of each narrative of the evangelists one by one, and then tosses them over into “the death-kingdom of abstract thought.” If these books arose as his hypothesis maintains they have lost forever their high value for us. The same in substance may be said of the “tendency “criticism of Baur and his followers. What is so thoroughly true of these hypotheses is more or less true of all views which touch the origin of our Gospels, that their erection or overthrow is of the greatest concern to our faith. And indeed how can it be otherwise? For if one could answer every question about the sources of these writings and the use made of them, he could also tell why there is so much apparent discrepancy in matter and arrangement, and how far, if at all, there is real discrepancy. But these are the puzzling questions of gospel harmony.

Nor is the doctrine of inspiration far removed from the discussion, since to recognize and explore the human element in the compound product is the best preparation for a belief in the divine. And further, this inquiry does not fail to affect somewhat deeply the understanding of each narrative of the actions and discourse of Jesus. In the words of another, “misconceptions in the exegesis of these writings and in the treatment of their text will scarcely be shunned, so long as their genetic relation to one another remains not cleared up, or false decisions are adopted.” Not that one cannot be a good exegete, much less intelligent Christian, without a definite hypothesis on this subject. A definite and satisfactory hypothesis is quite likely impossible at present. But no one can hold with rigor to any hypothesis without having his views of harmony, inspiration, and interpretation influenced; nor can any one hold views on these points unexposed to disturbance without an intelligent opinion on the question of origin.

This inquiry thus intimate to faith is also one of the greatest scientific interest. The phenomena to be explained are altogether unique, and of such sort that the deeper and more detailed the examination, the more wonderful does the strange agreement blended with difference, the matchless unity amid variety, continually appear. What more inviting field for research? Nor are the first inducements to enter this field hid from all but very curious eyes. On the contrary, they appeal to the most thoughtless reader. It would seem that phenomena so rare in themselves and in their relations to faith should have excited early attempts at complete solution. It was however only within the last half of the eighteenth century that such attempts began to be made. This indifference to the problem was due, with little doubt, to that theory of inspiration which, though it manifested some uneasiness at alleged discrepancies, found in the “suggestio verborum” an adequate account for even the most remarkable verbal agreement.1

Since first fairly started the question has been discussed with painstaking and ingenuity which are really surprising. As would be expected, the supposition that the evangelists made use of each others’ writings was first tried, and in settling the order, that in which they stand in the canon first found favor. As new investigations brought new facts to light, new forms of hypothesis sought to satisfy the facts, and in time “all the domain of possibility was measured out.” But why not believe that “these three Gospels were in part sought out from similar or the same fountains, that is, from the memoirs of those who heard Christ’s various discourses,” was a question proposed by Clericus as early as 1716. Many were now found to show how this belief might be sustained. Here again, still governed by a belief in the priority of Matthew, Semler (1783) fixed upon Syro-Chaldaic documents, others, as Lessing (1784), Niemeyer (1790), Weber (1791), and Thiess (1804) upon the Gospel of the Hebrews; still others, as Corrodi (1792) and Schmidt, upon the Hebrew original of Matthew as the common source. The more complicated the facts needing explanation were shown to be, the more complex were the hypotheses, until the climax was reached in the later view (1804) of Eichhorn, who, having at first (1794) detected, besides the original source, four others, recensions of it, and being criticised and outdone by Marsh (1802), with his eight Gospels and parts of Gospels, in turn outdid his critic by increasing the number to twelve. To this hypothesis, variously modified, a number of names attached themselves, among which Gratz (1812) and Bertholdt are worthy of mention. The former reduced the number of processes to seven, and differed from Eichhorn also in that he ascribed priority of composition to Mark instead of Matthew. The latter drawing quite near to the view of Herder, supposed that a protevangel was planned by the apostles jointly while they were yet in Jerusalem, roughly sketched by one of them, and that copies of this writing were used by them all and by the early evangelists, to secure unity in the historic statement of the new doctrine. During the first half of the present century, while the protevangel hypothesis and one form of the so-called supplementary were winning great esteem, there was proposed, as a protest against the method of Eichhorn, a new way to solve the problem in the views of Dr. Gieseler, first promulgated at Leipsic, 1818. He attempted to justify historically the hypothesis of oral tradition. For this hypothesis the path had been broken by Eckermann (1796), Herder (1797), Paulus (1799), Schleiermacher (1817), and by the theory of Wolf touching the compositions of Homer. So many just considerations of general import as Gieseler’s book contained could not fail to have weight, and they brought forward into due prominence certain facts of history too much overlooked in the preceding hypotheses, although few have been found ready to accept his conclusion entire. Nearly all hypotheses of choicest and most recent research have founded themselves upon oral tradition. No one who refuses to give a considerable place to its influence need now attempt to solve the question. Different combinations of it in its essentials with other views have found favor with such scholars as Credner, De Wette, Olshausen, Meyer, and many others. But when does truth come without error following soon after? Dr. Gieseler had said: “Uniformly as the cycle of narrative was formed among the Palestinian disciples, yet it must be modified according to circumstances when the Gospel was preached abroad.” “Especially was that cycle changed in the case of Paul, who had gone over to Christianity with quite another education than that which the Palestinian disciples had received. Even if the narratives themselves were not altered, yet those must be made prominent which most corresponded to his views, while he left out others as less important. Matthew gives a genuine Palestinian Gospel, Mark one Palestinian though modified abroad, Luke a Pauline Gospel.”2 The statements that the first gospel was a spoken gospel, and that the evangelists present each a somewhat different form of it, are just, but have been wrested for unjust uses. With the work of D. F. Strauss, in 1835, a new epoch in the consideration of this question begins, and yet, as a matter of course, an epoch not wholly without preparation. Schleiermacher, in whose sight the synoptic Gospels were a conglomerate of short written sketches and bits gathered from oral tradition, put together without unity of purpose or sure temporal sequence, had pronounced the beginning and close of these books to be mythical. De Wette had given oral tradition not only transforming, but also creative, power over the evangelic narrative. It has also been said of a work of Lessing, written even so long ago as 1778, that it takes away from the book of Strauss every merit of originality. But this class of opinions is not complete when the pupil of Baur has argued that the Gospels had their origin in the myth-making tendencies of the early followers of Jesus; for the teacher himself and those who believe with him in the correctness of the “tendency “critique convert all the New Testament writings into mere records of the strife between different church parties. “The dot has thus been put upon the I.”

This short sketch of the history of this question has been given, not for its own sake, but that the reader may notice the nature of the progress which the hypotheses exhibit. For there has been a progress. It consists partly in this, that the many relations of the question have come into clearer light, and partly in the fact that, while these different attempts at solution have been making, some of them so unsatisfactory, and others both unsatisfactory and dangerous, the data for the true solution have been revealing themselves. It has grown to be an imperative necessity that at least certain solutions be shown to be false if a true and complete one cannot be found. The search after the true one has not gone wholly unrewarded. Observe how, as it has been proposed by each hypothesis to survey the whole field from a single point of view, and the field has been shown too large for this, the lenses have been complicated or their arrangement altered, until perhaps only a very incorrect image of the real domain was left. Then the instrument has been shifted, and another map, supposed more reliable, has been drawn. But each survey, partial and unfit to serve for a true plot of the whole, has contributed some true lines toward the perfect sketch, to which the approach is made, though itself be always impossible.

In the light of history the following statements concerning the subject are made clear. The question is an important one, for it is fundamental to Christian faith, both because of the dangers which certain answers carry with them, and because of the influence which any definite answer, or want of definite answers, must have upon important doctrines.

The question is a complicated and abstruse one. Some hypotheses for its solution are surely complicated enough. “Few subjects of theological science,” said Gieseler, even fifty years ago, “have been so often and so thoroughly cultivated as this.” “This problem,” says Davidson, “has engaged the attention of many, especially in Germany, and given rise to more treatises in connection with the Gospels than any other.” And further, the true hypothesis must acknowledge those true elements which have already entered into each view, perhaps not refusing everything that the mythical theory and “tendency “criticism have advanced. The proof of this lies in the entire history of the discussion, and in the consideration that such is the present method of the best scholarship.

Again, the true hypothesis must by all means acknowledge the claim of each evangelist to be considered as something quite different from a mere copyist. To this view the history of the discussion shows progressive approach.

Bearing in mind these truths, derived from a survey of the method in which the question has unfolded, we now turn to the question itself. The most satisfactory way to undertake its answer is that of immediate appeal to the phenomena themselves. It may be, it can scarcely fail to be, that dogmatic considerations which were so long deemed of chief importance should have some weight; but they cannot be allowed decisive or even considerable authority. We are not to be so certain how the evangelists ought to have written as to refuse to learn how they have written.

There are also certain general historic considerations, and in particular certain historic testimonies, which have more or less bearing upon the question. On such grounds, indeed, and almost entirely, Gieseler erected his hypothesis. For, although he admits that any conjecture which will make pretentions to probability, must fully explain the inner relation of the Gospels; yet, holding that different external connections of the writers may be thought of which will equally well explain this inner relation, he proceeds to make history decide between them.3 Well-known conditions under which these books were composed, ought doubtless to be kept ever in mind. But on the other hand, we ought to hold no evidence derived from surmised or obscurely-known conditions against the evidence from the phenomena which these books plainly show. “What influences the past furnished for the writers, we may be able, in certain cases, only to guess at; what the writings are, they themselves are present and can answer. As for the direct testimonies, though they are very valuable, their meaning is too much disputed; and could we be sure of this, they are far too meagre to afford any satisfactory solution. At any rate, it cannot be amiss to investigate the phenomena apart from the witness of history, and then if their testimony, cautiously taken, is found to accord with its far feebler voice, the conclusions arrived at will be strengthened. The final appeal must be to the books.

The present investigation will consist, in the first place, of a statement of the phenomena as full and fair as possible; after which we shall be in position to discuss the various hypotheses offered to account for them, and to derive such and only so many conclusions as shall seem plainly called for.

What, then, are the phenomena? In brief, those of agreement and those of difference — agreement due to sameness of thing known, sameness in mode of receiving and apprehending the thing known, and of design in giving it expression; and difference due to inequality in extent of the thing known, and to variety in the mode of its reception, apprehension, and expression. The general nature of this agreement and difference will be presented in two ways; first by an analytical statement of both, and afterward in a more concrete form, by means of certain selected examples. Certain phenomena, as for instance, those of citation and verbal characteristic, deserve a special treatment which will follow the more general. Let not the reader, unless he be already familiar with the subject, grudge the patient use of the Greek Testament.4

The following are some points in which our Gospels, Matthew, Mark, and Luke, agree:

I. They all, though differing somewhat among themselves, differ so much more from the fourth Gospel, and coincide so closely with each other in the general aspect which they present of Jesus’s character and work, that their presentations are peculiarly one. This unity is the point most easily seen among those of external and internal relationship, which are summed up in the single word “synoptic.” This word, though it has been deemed the “πρῶτον ψεῦδος of the rationalistic treatment of the Gospels,” embodies aptly a truth which no one will deny. A scene so rich in events as the life of our Saviour in his varied contact with men, from his first entrance upon his Messianic office to his ascension on high, must have presented widely various aspects, as caught from different points of view. How differently it might have been set forth the Gospel of the apostle John remains to tell. His view, while it is of the same grand original, is still so unlike that of the synoptic Gospels, that, to adopt a beautiful figure, they will not be arranged into the same stereoscopic picture. Which difference, when we consider how many-sided in attributes and activities Christ was, and how difficult it is to take two copies alike of the same face, or for two intimates to write memoirs of the same original without considerable discrepance in the presentation of certain phases of character, seems scarcely more surprising than this remarkable agreement. “Whoever,” says Marsh, “has compared Christ’s descent from the cross by Rubens, with his descent from the cross by a painter of the Italian school, knows how greatly the representations differ from each other.”5 But the views of the synoptic Gospels will form a stereoscopic picture together, though they here and there slide by each other. It is, then, scarcely a complete account of this striking similarity to say that the different impressions are of the same original.

II. When we examine the Gospels we do not find that they flow on like continuous history, but are largely made up of single narratives, fragments, so to speak, oftentimes in temporal sequence, plainly expressed or implied, while at other times such sequence cannot be traced. Of these narratives there is a tolerably well-defined cycle. It is not meant simply that all agree in giving a certain number of these fragments of Jesus’s discourse and work, but also, that where one or two of the writers introduce new bits of narrative, these have in general the same generic marks. The new miracle or the new teaching bears the synoptic character. The discourse of Christ is especially unique as given by John. The material in Matthew and Luke which precedes the baptism, and the two miracles peculiar to Mark (7:32–37; 8:22–26) seem to depart most widely from this rule. This point of agreement is closely allied with that first mentioned. But, besides the fact that the cycle of narrative bears tolerably well-defined characteristics, the writers of the synoptic Gospels agree in making, to a large extent, the same selections. The number of sections in which all three or two agree is variously given, owing to variety in division. Gieseler distinguishes forty-two common to all, twelve to Matthew and Mark only, five to Mark and Luke, and fourteen to Matthew and Luke.6 Marsh, adopting the division given by Eichhorn, discovers forty-two common to all, four to Matthew and Mark only, one to Mark and Luke, and twelve to Matthew and Luke, in all of which there is verbal agreement.7 According to still another division, of one hundred and fifty sections, sixty-five are common to all three, fifteen to Matthew and Mark, five to Mark and Luke, and twelve to Matthew and Luke.8 Why have these three evangelists confined themselves almost without exception to the same type of narrative, and in so large a degree to the same selections, though their writings are mere sketches of Christ’s life? When we think how many conversations like that with Nicodemus, how many deep spiritual discourses, how many an incident and miracle of quite another sort from any recorded by the writers of the synoptic Gospels must have taken place, and remember that the Apocrypha resulted in part from a feeling of this lack, though they meet it unsatisfactorily, and that the fourth Gospel shows how by the selection of other and generically different incidents this lack could have been satisfactorily met, we are forced to believe that this point of similarity does not find its full explanation in the history itself (vid. John 21:25).

III. These Gospels all have the same general progress and main divisions: (1) Entrance upon the Messianic office, baptism, and temptation, Matt. 3:1–4:11 = Mark 1:1–13 = Luke 3:1–4:13; (2) Ministry in Galilee, with a series peculiar to Matthew and Mark, Matt. 4:L12-18:35 = Mark 1:14–9:50 = Luke 4:14–9:50; (3) Journey to Jerusalem, Matt. 19:1–20:34 = Mark 10:1–52 = Luke 9:51–19:28, with 9:51–18:14, peculiar; (4) Entrance into Jerusalem, and activity there, Matt, 21–25. = Mark 11–13 = Luke 19:29–21:38; (5) Arrest, passion, death, and burial, Matt, 25, 27 = Mark 14, 15. = Luke 22, 13; (6) Resurrection, Matt, 28. = Mark 16 = Luke 28.9 Beyond this general agreement Mark and Luke agree very closely in their arrangement of the sections common to all three, if the interpolations of Luke are left out.10 But Matthew agrees only partially in his arrangement of the sections common to him and one or both of the other writers. For example, 3:1–4:17. From this point the sequence differs in the main, though with special points of agreement, until 14:1, where falling into the same order with Mark he keeps it during a series of narratives common to these two, of which Luke has only 9:7–17, until 16:13, where Luke joins them and all go on in company for a time. After giving 18:10–35, for the most part peculiar to him, he joins (19:1) with Mark and afterward (19:13) with Luke, from which point the common thread of sequence is quite unbroken to the end. Says De Wette: “Comparison with the Gospel of John shows that the pattern of this progress is not sketched throughout by the history itself.” This agreement in main divisions has been made by Lachmann the basis of a hypothesis to account for the origin of the Gospels.

IV. Certain narratives are found always closely tied together into the same groups, and that even where the preceding and following sections have a varied order. Such are the healing of the paralytic and the calling of Matthew (Matt. 9:1–13 = Mark 2:1–14 = Luke 5:17–28); the plucking of the ears of corn and the healing of the withered hand (Matt. 12:1–14 = Mark 2:23–3:6 = Luke 6:1–11); the hushing of the tempest and the healing of the Gadarene demoniacs (Matt. 8:23–34 = Mark 4:35–5:20 = Luke 8:22–39); Herod’s judgment and the feeding of the five thousand (Matt. 14:1–21 = Mark 6:14–44 = Luke 9:7–17). An especially interesting example is that day so full of activity as recorded by Mark 1:21–38, and Luke 4:31–43.11

V. There is verbal agreement, surprising for its exactness and extent. Only a few among the many examples can be indicated: (1) Narrative of the paralytic (Matt. 9:2–8 = Mark 2:3–12 = Luke 5:18–26), and notice particularly the passage commencing ῞Λνα δὲ εἰδῆτε ὅτι etc.; (2) Matt, 26:13–28 with vs. 17, 19 peculiar = Mark 8:27–9:1 = Luke 19:18–27; here notice Matt. vs. 24–26 and comp. parallel passages; (3) Matt. 21:23–27 — Mark 11:27–33 = Luke 20:1–8; in particular vs. 25-26, of Matt, with parallel passages; (4) Matt. 8:2–4 = Mark 1:40–45 = Luke 5:12–16 particularly vs. 3-4, of Matt; (5) Matt. 21:33–46 = Mark 12:1–12 = Luke 20:9–19, in particular the quotation from Ps. 68:22; (6) The eschatological predictions (Matt, 24 = Mark 13 = Luke 21.) are a very remarkable example. Says Marsh: “in Mark 13:13–32 there is such a close verbal agreement for twenty verses together with the parallel portion in St. Matthew’s Gospel that the texts of St. Matthew and St. Mark might pass for one and the same text in which a multiplication of copies had produced a few trifling deviations.” 12 Compare here particularly verses 6–8, 19, 30, 35 of Matt, with the corresponding passages; (7) Matt. 14:19–20 = Mark 6:41–43 = Luke 9:16–17, where is an example of very exact verbal agreement in the narrative itself. (8) Matt. 15:1–20 = Mark 7:1–23; especially vs. 7–9 in Matt., 6–7 in Mark; (9) Matt. 20:20–28 = Mark 10:35–45, to which passage, with its wonderful coincidences, the remark of Marsh quoted above will well apply; (10) Mark 1:21–28 = Luke 4:31–37, especially the passage ἔα τί ἡμῖν καὶ σόι etc.; (11) Matt, 8:5–13 = Luke 7:1–10.13

Bishop Marsh believed that throughout the common sections; “St. Mark never fails to agree verbally with St. Luke where St. Luke agrees verbally with St. Matthew.”14 This statement is disproved by the following among other examples: Matt. 3:11 = Lu 3; 16, where they coincide, and, unlike Mark, do not omit καὶ τυρί. Matt. 9:7 = Luke 5:25; ἀπῆλθεν εἰς τὸν οἶκον αὐτοῦ instead of ἐξῆλθεν ἐναντίον πάντων, as Mark 2:12. Matt. 9:20 = Luke 8:44; προςωλθοῦσα ὄπισθεν instead of ἐλθοῦσα ἐν τῷ ὄχλθͅ ὄπισθεν as Mark 5:27. Matt. 10:9 = Luke 9:3 ἀργύριον instead of χαλκόν, as Mark 6:8. Matt, 27:54 = Luke 23:47 ὁ ἑκατόνταρχος instead of ὁ κεντυρίων, as Mark 15:39.15 The somewhat similar statement of Meyer, that “in the parts where Mark does not stand with them they two depart furthest from each other, while they essentially agree where Mark forms the middle term,” is the correct one.16

It needs also to be noticed that the greater part of the verbal agreement is in the recital of words spoken by the characters introduced, and especially by Jesus. In Matthew the passages with verbal coincidences between him and the others, either one or both, amount to less than one sixth of Matthew’s contents, and seven eighths of this one sixth occur in giving the words of others and one eighth in the narrative. In Mark the coincident passages are about one sixth of his contents, and less than one fifth of these in the narrative; while in Luke the total of coincidence is about one tenth of the whole, and less than one twentieth in the narrative. But the narrative in Matthew is about one fourth of the whole Gospel, in Mark one half, in Luke one third, and therefore the verbal coincidences in the recital are more frequent than in the narrative, as two to somewhat less than one in Matthew, four to one in Mark, and nine or ten to one in Luke.17 These last-mentioned facts do not conflict with several different hypotheses. The words of Jesus would be less subject to change, whether in oral tradition or in the use of written sources.

VI. The first three Gospels agree in the use of rare words and infrequent turns of expression. The following are selected examples: Matt. 9:15 = Mark 2:20 = Luke 5:35; the word ἀπαρθῇ. Though the active is used often in the LXX. the verb is not met with elsewhere in the New Test., and here it is in the passive, which is not elsewhere found. Matt. 16:28 = Mark 9:1 = Luke 9:27; γευσώνται θανάτου (comp. John 8:52; Heb. 2:9). Matt. 19:23 = Mark 10:23 = Luke 18:24; δυσκόλως, not found elsewhere in the New Test, or LXX. Matt. 26:51 = Mark 14:47 = Luke 22:51; the diminutive ὠτίον, a word however of common life, and found John 18:26. Matt. 9:2, 5, = Mark 2:5, 9 = Luke 5:20, 23; Doric passive ἀφέωνται (vid. Winer’s Gram. 6th Aufl. p. 74). Matt. 12:13 = Mark 3:5 = 10; ἀπεκατεστάθη, with the double augment (Winer, p. 67). Matt, 27:12 = Mark 14:61 = Luke 23:9; ἀπεκρίνατο middle, instead of the passive, which is elsewhere in the New Test, universally used. There are also instances where two only agree in the use of a rare word. Such are Matt. 4:5 = Luke 9:9, πτερύγιον τοῦ ἱεροῦ, found nowhere else. Matt, 7:5 = Luke 6:42, καὶ τότε διαβλέψεις, a verb not used elsewhere in the New Test, or LXX. Matt. 24:51 = Luke 12;46, διχοτομήσει and καὶ τὸ μέρος αὐτοῦ.....θήσει. Matt 24:22 = Mark 13:20, κολοβοῦν twice, used only here in the New Test.; in LXX. (2 Sam. 4:12) employed to translate קָצַחּ, which Hebrew word, occurring sixteen times in the Old Test, is translated by nine different Greek words, thus giving one among many decisive proofs that the Gospels are not independent translations.18 Further examples of agreement in words seldom or never more than once employed by the synoptists are the following: 1. Matthew, Mark, and Luke, καταγελᾶν, καταμαρτυρεῖν, κλάσμα, κόκκος. 2. Matthew and Mark, ἄγειν as intransitive, found in the subj. Matt 26:46 = Mark 14:42, ἁλίζειν, ἀμφίβληστρον, γενέσια, δειλός, ἐκφύειν,θορυβεῖν, καρματίζειν, λαλιά, μνημώσυνον, ὄμμα, ἐντυλίσσειν, ἰχθύδιον, σκοτία. 3. Matthew and Luke, ἄρωμα, ἐκπνέειν, κεραμιον, προσαιτεῖν.19

VII. The synoptic Gospels agree in their wording of certain quotations from the Old Test., while they depart from both the Hebrew and LXX. In quoting Isa. 40:3 (Matt. 3:3 = Mark 1:3 = Luke 3:4) they all depart from the Hebrew and join ἐν τῇ ἐρήμῳ with βοῶντος, agreeing (so De Wette) or not (so Meyer) with the LXX. Isa. 29:13, quoted (Matt. 15:8-9, = Mark 7:6-7, is made to read διδάσκοντες διδασκαλίας ἐντάλματα ἀνθρώπων; but in the LXX it stands διδάσκοντες ἐντάλματα ἀνθρώπων καὶ διδασκαλίας. The subject of Old Testament quotation will receive a separate treatment.

To nearly all these points of agreement in the synoptic Gospels correspondent points of difference stand in contrast.

I. The synoptic Gospels, though agreeing among themselves in the general aspect they present of Christ so far as to be classed together, differ among themselves in presenting each one a characteristic aspect. This truth, hinted at in their very titles, εὐαγγέλιον κατά, long ago symbolized by comparing the different look of our Gospels to the four faces of the creatures seen in prophetic vision, has recently been made the basis for the distorted opinions of the “tendency “criticism. To dispute which face belongs to each Gospel shows perhaps only difference of taste; the fact of the dispute shows the truth we are concerned with. The view of Jesus which each evangelist has caught and holds up is so unlike any other as to establish beyond doubt the claim of each to individuality of character.

II. They all, though having each one many sections common to one or both the others, have sections peculiar to themselves, some of which depart more or less from what may be called the synoptic type.

The material peculiar to Matthew consists of the first two chapters entire, and some thirty-five or more fragments scattered through the Gospel, of which the following are the more lengthy: portions of the sermon on the mount; miracle recorded 9:27–31; Peter’s walking upon the sea, 14:28–32; 18:15–20; 20:1–16; 21:28–32; 22:1–14; portions of the eschatological discourse; fate of Judas, 27:3–10; 28:11–20. Especially to be noticed are the numerous citations from the Old Testament found in this Gospel alone, 4:14 sq.; 8:17; 12:17 sq.; 13:35; 21:4 sq., etc.

Four sections, two of them parables (4:26–29; 13:33–37), and two miracles (7:32–37; 8:22–26) are all that Mark alone has. But notice besides, 9:49, 50; 11:18, 19; 14:51, 52; 15:44-45, etc. To Luke the material contained in the first two chapters and some sixty shorter passages (3:23–38; 7:11–17; a large portion of the great interpolation 9:51–18:14; 19:1–10; 19:11–27; 19:39–44; 23:6–12; 24:13–49, etc.) are peculiar.

Gieseler gives, of special sections, five to Matthew, two to Mark, nine to Luke. On the basis of a division into one hundred and fifty sections, fourteen are peculiar to Matthew, two to Mark, and thirty-seven to Luke, or the percent of the whole contents is forty-two, seven, and fifty-nine, respectively. Some of this material used by only one writer will be examined more in detail below.

III. The synoptic Gospels, though they exhibit the same main divisions and general progress, and though the same thread of sequence may be traced in particulars through considerable portions of their narrative, yet differ in the arrangement of many sections. Especially do Mark and Luke differ from Matthew. For this various reasons are given by those who hold the various hypotheses of origin. In connection with that of oral tradition it may be held, that the variety of arrangement is due, either wholly or in part, to variations in the tradition itself; or it may be ascribed to subjective reasons, either on the ground that it was no part of the design of the writers to observe chronological sequence, or that their material being largely subjective, their disposal of it is necessarily so, both of which opinions give large play to fancy in determining the true principle of arrangement; or again, it has been held to result from mistake, or ignorance of the true order. In connection with the so-called supplementary hypothesis, the attempt will be to refer all differences to the supposed prior Gospel, and account for them by reasons generally or specially applicable, while in case all the writers drew from common written sources, those which are distinguished by each investigator will be thought to shed light upon the inquiry. It is a favorite view with some that Matthew’s peculiar arrangement hinges upon the early position he gives to the sermon on the mount. His plan is explained by one writer as follows: Matthew, having given in this sermon a “programme of the public activity of the Lord,” groups together such of the most characteristic miracles as suit his purpose (chap, 18; 9. Then follow “the founding of the kingdom,” in the discourse (chap, 10.); the doubts of the Baptist, and the complaints of the Lord (chap, 10.); the blasphemous charges of the Pharisees, and the defence of Jesus (chap, 12.); and finally (chap, 13.), certain parables which disclose the inner ground of this hostility, together with an account of his rejection at Nazareth. But why does Matthew now join and henceforth coincide with the other two? The answer is, because his peculiar purpose being accomplished, he has nothing left but to subjoin his remaining material in the order of his source.20 But it may well be doubted whether the thread of temporal sequence in each one of the Gospels is not far too strong to be thus easily broken.

IV. There are all those differences in the individual narratives with which gospel harmony has to deal. These are either such unimportant variations as may fairly be used to supplement and explain one account by another, or such discrepancies in details as do not admit of satisfactory solution, while they leave the general agreement undoubted, or, finally, such as tend to destroy themselves by making it disputed whether narratives so unlike can have the same original. Numerous examples of the first kind will occur to every reader. Such are Matt. 3:13–17 = Mark 1:9–11 = Luke 3:21-22 observe the προσευχομένου and σωματικῷ εἴδει of Luke. Matt. 12:1–8 = Mark 11:23–28 = Luke 6:1–5; observe the ὁδὸν ποιεῖν of Mark, and the δευτεροπρωτῳ of Luke, if it be not a gloss. Matt. 7:28–9:1 = Mark 5:1–21 = Luke 8:26–40; observe the δύο δαιμονιζόμενοι in Matthew, and vs. 3–5 in Mark. Examples of the second kind are: the call of the four disciples, Matt. 4:18–22 = Mark 1:16–20 = Luke 5:1–11; the sermon on the mount, Matt. 5:8 = Luke 6:20–49; the healing at Jericho, Matt. 20:29–34 = Mark 10:46–52 = Luke 18:35–43. An example of the last class is the anointing of Jesus, Matt, 26:6–13 = Mark 14:3–9 = Luke 7:36–50.

V. Besides such differences in narration as are due to difference in the thing told, there is such constant verbal difference as is indispensable if the writings are to be in any sense distinct. This consists, in substitution of one synonymous expression for another, as in the parable of the sower (Matt ὁ σπαρείς, Mark οἱ σπειρόμενοι, Luke τὸ πεσόν; Matt. Πονηρός, Mark σατανᾶς, Luke διάβολος); or of one title for another (while ὁ ᾿ Λησοῦς is the more common, Luke introduces ὁ κύριος into the narrative where it is not found with the others, and Matthew seems to prefer ὁ χριστός or ὁ υἱὸς Δαυίδ; and for the disciples οἱ δώδεκα, οἱ δώδεκα μαθηταί,οἱ μαθηταί, are favorite terms with Matthew and Mark, while Luke uses ἀπόστολοι more freely); or in change of verbal position, or new turn given to the sentence; or in explanatory addition, as Mark 15:21 τόν πατέρα ᾿Αλεξάνδρου καὶ Ρούφου; and Matt. 6:21 εἰς ῾Λεροσόλυμα ἀπελθεῖν, or in the use of the same word in a different construction. More light will be thrown upon this point while treating the subject of verbal characteristics.

Having seen these phenomena of agreement and difference as they appear when analyzed and classified, we are now ready to approach them in another way, by examination of a few passages which will exhibit them in a more concrete form. This method, while it shows the phenomena as they really occur, each kind of variation blended with each kind of agreement, will also help the transition to the hypotheses which they have called forth, for it will make known how different investigations have justified different views. Certain opinions may perhaps seem valuable psychologically rather than otherwise. It will however be impossible to do more than simply point out a few things of interest in each example. Whoever would arrive at an intelligent opinion must consult the text for himself in every instance.

1. Matt. 3:1–12 = Mark 1:1–8 = Luke 3:1–9, 16-17. In the υἱὸς θεοῦ of Mark, Holtzmann thinks he sees traces of later preparation by the same hand which substituted ὁ υἱὸς τῆς Μαρίας (6:3). The citation v. 2 of Mark (attributed to Isaiah, but really from Mal. 3:1), by whomsoever made, occasions some difficulty, and Meyer finds here an error of memory on the part of the original composer. Holtzmann considers it the addition of a later hand, since it is quoted from the Hebrew, while the following verse is from the LXX., and adduces 15:28, taken from Luke 22:37 as another example. Hengstenberg justifies the quotation by making Malachi only “auctor secundarius,” while Isaiah was really “auctor primarius,” but of this there is no proof. According to De Wette the “inadvertence of Mark is made natural by his dependence on Matthew and Luke.” The marked agreement, as seen above, in the quotation of Isa. xl. 3 proves a common, written (?) source, and the addition of Luke is from the LXX., either as the result of his own reflection (Holtz.), or because it was so given in his peculiar source, or because it was customarily so given in the evangelic tradition (Meyer). In κύψας is seen one of Mark’s characteristic touches. According to Holtzmann, Mark has shortened the account given in the common source, designing only to bring forward John’s relation to Jesus; and the somewhat remarkable omission of καὶ πυρί is due to the omission of the following verses.

2. Matt. 3:13–17 = Mark 1:9–11 = 21–22. It is claimed that the priority of Mark is shown by the gradual change from his form of presenting the vision and the voice, through Matthew and Luke to John.

3. Matt. 4:1–11 = Mark 1:12-13, = Luke 4:1–13. According to many, Mark’s narrative is here, without doubt, the older form, and was enriched by the others from oral or written sources. Proofs for this view are found in his brevity, which corresponds to the “yet undeveloped summary beginning of the tradition,” here given in its oldest, nearly germinant form “(Meyer, so also Eichhorn, Ewald, and others). The εἰ υἱὸς εἶ τοῦ θεοῦ spoken by Satan shows reference to the preceding σὺ εἶ ὁ υἱός μου. That, on the contrary, Mark has the later form, is thought to be shown by the “coloring of concrete situation,” which Matthew and Luke furnish, more like a “fresh product” than the “abstract fact” of the second Gospel. Besides, in the words ἦν μετα τῶν θηρίων is a trace of later origin, though they must have been added by Mark himself, since they would not have been omitted by Matthew and Luke had they been in their source (Holtz.). Mark has, says one, according to his custom, left out whatever is disconnected, surprising, abnormal. De Wette, however, who considers the account of Mark as an abbreviation of Matthew and Luke, sees in these words “an artist’s addition,” and finds the οἱ ἄγγελοι διηκόνουν αὐτῷ, introduced from Matthew, out of place, since Mark had said nothing about fasting.

4. If Jesus’s deliberate change of abode as recorded Matt, 4:13 had been in Luke’s source, would he have omitted to mention it when so much to his purpose, 4:31?

5. Matt. 9:2–8 = Mark 2:1–12 = Luke 5:17–26. Here Mark’s graphic description is thought to favor his originality (Holtz., Meyer). According to De Wette, Mark follows Luke here in sequence, supplementing him however with Matt, 9:1, and laying the scene in Capernaum.

6. Matt. 9:9–17 = Mark 2:13–22 = Luke 5:27–39. De Wette sees in the καὶ ἦσαν.....νηστεύοντες of Mark an “archaeological notice” (Winer, p. 312), and proof of his dependence, since he has combined the οἱ μαθηταὶ ᾿ Λωάννου and οἱ φαρισαῖοι in such way that the reply does not answer to the question. But ἦσαν νηστεύοντες etc. is better understood as referring to a definite time (Meyer), at which the Pharisees were also fasting, or Mark may have supplemented the subject of ἔρχονται out of the question, in which allusion was made to this sect.

7. Matt. 9:18–26 = Mark 5:21–43 = Luke 8:40–56. An advocate for the originality of Mark’s form of this incident detects here a common source which Mark has somewhat abbreviated (τοῦ κρασπέδου, τοὺς αὐλητάς, and perhaps v. 56 of Luke have been left out), Luke still more, and Matthew most of all. “If then Matthew has so manfestly and forcibly compressed the narrative of Jairus’s daughter, we ought so much the less to hold his account of the woman with bloody flux for original” (Meyer, and vid. Holtz.). The above touches however are just such as Mark would be least likely to omit from the picture (Ewald goes so far as to suppose that τοῦ κρασπέδου first fell out in the present Mark). Another, who can never suppose Mark to be anything but an epitomizer, sees in his account only a selection, now from Matthew and now from Luke (De Wette).

8. Matt. 10:1–14 = Mark 6:7–13 = Luke 9:1–6. The noticeable difference here is between the εἰ μὴ ῥάβδον μόνον of Mark (8), the μηδὲ ῥάβδους of Matthew (10), and μήτε ῥάβδον Luke (3). This is explained in several ways. The prohibition is genuine, and against staves for defence, but Mark, who probably read ῥάβδους in the common source, has softened it, conceiving it to refer to staves for support (Holtz.). Or the difference is an over-nicety in Matthew and Luke which has pressed its way in, but no misunderstanding (Meyer). Or, Mark, combining the texts of Matthew and Luke, has made alterations (De Wette).

9. Matt. 13:53–58 = Mark 6:1–6 and, assuming its identity, Luke 4:16–30. It is asserted by Holtzmann that the different forms of the question are best accounted for by supposing it to have stood οὐχ οὗτός ἐστιν ὁ τίκτων, ὁ υἱὸς ᾿Λωσήφ,, in the common source, which Mark changed because of his dogmatic point of view (so also Hilgenf. and Baur) into ὁ υἱὸς τῆς Μαρίας, while both the others took offence at the word τέκτων. and made alterations accordingly. But the identity of Luke’s narrative is very justly denied by perhaps the majority of modern commentators, and the other grounds of the argument are more than doubtful. “As if,” says Meyer, “Mark would npt have had opportunity and skill enough to bring his views definitely and significantly forward elsewhere.”

10. Matt. 16:13–28=Mark 8:27–9:1 = Luke 9:18–27. One writer (De Wette) sees here a proof of the priority of both Matthew and Luke, in the ὄχλος of Mark 8:34 held to be an expansion of 9:23 in Luke, who has himself failed to give the motive of Matt. 16:22, and has widened the circle of hearers. That Matthew, Mark, Luke, is the correct, order, another finds proof in 5:38 of Mark, supposed to be combined from Matt. 16:27 and 10:33, in which combination Luke has followed him (Hilgenf.). Still another thinks that there is proof of a common source apart from all three, and that Matthew (5:27) has generalized the expression found in it, because he had already given this dictum out of another source (Holtz.). According to still another and preferable view (Meyer), the ὄχλος of Mark is to be explained from his foregoing text, and is no proof of want of originality, nor does the ἐν τῇ γενεᾷ of v. 38 furnish such proof. That Matthew alone (17:19) has the promise to Peter is noteworthy, and according to Meyer a proof that he, having the λόγια for a source, is in this portion richer and more original, though the promise could not have been unknown to Mark. Why, then, has not Mark given, it? Out of respect to his class of readers, or, as was long ago said, ἵνα μὴ δόξῃ χαριζόμενος τῷ Πέτρῳ. Baur explains the difference as an addition in Matthew, due to the growth of the hierarchical spirit. Mark 3:16 has been considered by some to be a compensation for this omission.

11. Matt. 19:13–15 == Mark 10:13–16 = Luke 8:15–17. Holtzmann supposes that the common source read αὐτοῖς, for which, since it was liable to be misunderstood, Mark, or a later transcriber, put the gloss, τοῖς προσφέρουσιν,

12. Matt. 20:29–34 = Mark 10:46–52 = Luke 18:35–43. The ὁ υἱὸς Τιμαίου of Mark seems to have its ground in the notoriety of Timaeus. It is quite improbable that Matthew and Luke would have omitted the name if they had had Mark before them.

13. Matt. 26:26–29 = Mark 14:22–25 = Luke 22:19-20. The words καὶ ἔπιον ἐξ αὐτοῦ πάντες, which De Wette attributes to the fondness of Mark for changing his sources, Matthew and Luke, and finds purposeless, since the act is assumed in the following verse, furnish proof for others of the originality of Mark, and his greater independence of the later liturgical custom (Meyer, Holtz.).

14. Matt. 14:22–16:12 = Mark 6:45–8:21. The question of interest in this example is, why Luke should have omitted so long a series, given, in the same order and in such close sequence, by both Matthew and Mark. Various answers have been conjectured. The solution of Hug is perhaps as bold and unsatisfactory as any. In his opinion, this series did stand in the original Luke, but, concluding with the miracle of the seven loaves, was lost out very early by the transcriber, who mistook the foregoing miracle of the five loaves for it. Thus its omission is explained by homoioteleuton.21 Reuss supposes that Luke used a copy of the common source in which this series was wanting. Weisse, very absurdly, ascribes it to his carelessness, while Holtzmann has a special reason for the omission of each incident in the series. (Jesus walking on the sea, because it seemed superfluous in connection with the other narrative of a storm; the miracle following, because Luke in general does not like summary accounts, and thinks his earlier passages will suffice in this respect, etc.)

15. Matt, 5–7. = Luke 6:17–49. The question, which gives the original account of the sermon on the mount, has received different, answers. The proofs that the two are essentially one, and from the same source, are as follows:

(1) The sameness of accompanying circumstances (though Luke 6:17 presents the situation somewhat differently), both as to the departure into the mountain, and especially the subsequent miracle of healing, a marked agreement, as between Matthew and Luke, where Mark is wanting. (Ewald conjectures that there is a break Mark 3:19 before the words καὶ ἔρχονται εἰς οἶκον, which is to be repaired by introducing the sermon and miracle.)

(2) The sameness in the addresses themselves, both as to their “similar characteristic beginning and conclusion,” and as to their “manifold and essential agreement in contents.” To this opinion agree Bengel, Olsh., De Wette, Meyer, Holtz. and most modern commentators.

In favor of the priority of Luke is urged that—while Matthew’s deviations from Luke, consisting of remarks upon Jesus’s relations to the law, can be explained by supposing Matthew’s account to be a compilation after the habit of the first evangelist—we cannot understand why Luke should have “cut out” from Matthew’s account its peculiar kernel. It is, however, admitted by Holtzmann, who supports this view, that the use of words peculiar to Luke (πιμπλάναι, σκιρτᾶν, κατὰ ταῦτα ποιεῖν, etc.) shows a partial and verbal working over of his source; to which he added vs. 39 and 40, as is indicated by the introductory words εἶπεν δὲ καὶ παραβολὴν αὐτοῖς, because they seemed, owing to the nature of their contents, most fit to stand in this connection.

On the contrary, it is urged that in Matthew’s account, “rich circumstantiality, gnomological brevity, and want of connection occur together in such manner as is adapted very naturally to long discourse, actually held, spiritedly improvised, but not at all to the compiling art of one who simply transmits “(Meyer). It is admitted, however, that this essential originality must be looked upon only as a relative one, “in which is embodied, not only the influence that its repetition in writing, partly in the λόγια, partly in the later formation of the gospel, had already exercised upon much in the form and order; but also, much spoken by Jesus on other occasions was woven in here, in part involuntarily, in part by design.” Again it is urged that most of the passages given by Luke elsewhere, which are parallel to those in Matthew v.-vii., are either less aptly introduced in Luke (Luke 2:34 = Matt. 6:22 sq.; Luke 16:17 = Matt. 5:18; Luke 16:18 = Matt. 5:32); or are such as Jesus might have repeated (Luke 12:33 sq. = Matt. 6:19 sq.; Luke 13:24 = Matt. 7:13; Luke 14:34 = Matt. 5:13, and perhaps also Luke 11:1 sq. = Matt. 6:9 sq.). It may be remarked, however, that in this case the admission of Matthew’s claim to priority involves the supposition of much artificiality on the part of Luke, who must then have separated these utterances, given together in their original form, and interwoven them with so many different narratives.

16. Matt. 8:5–13 = Luke 7:1–10. A very specious argument is made, by a comparison of these passages, for a common source, differing somewhat from each, which the verbal agreement in the speech of the centurion shows to have been a written one. Thus, Matthew understands the ambiguous word παῖς to mean son (so Neander, Strauss, Holtz.) as is proved by the article, and by the anxiety of the centurion, which Luke attempts to account for by throwing in a word (ἔντιμος) peculiar to him among the writers of the synoptic Gospels. But the other evangelist interprets it δοῦλος, which variation a written form of the narrative, different from both, will alone explain. To this it is replied that Matthew does not so understand παῖς, and that the article only proves that he had but one servant, a view which is confirmed by the τῷ δούλῳ μου of v. 9 (Mey.).

17. There is considerable material common only to Matthew and Luke, but very differently arranged in each, about which, as a whole, something ought to be said. It consists for the most part of fragments of discourse. Holtzmann states that of such fragments the principal part is gathered by Matthew around five or six points, while the same material in Luke forms no close sequence with the portions between which it is interpolated, and falls, to a great extent, into chaps, 9:51–18:14. Matthew has unclothed these portions of the historical surroundings which Luke has given them, and so interwoven them with material from his other source, that the two are often most closely blended into one account. It would seem, then, that the form which this material takes in Luke is more likely to be original, that of short sentences, gnomes, properly introduced by such forms as ὁ δὲ εἶπεν αὐτῷ, εἶπε δὲ παραβολήν, etc.22 According to others, addresses of considerable length, which are fitly followed by the words ὁτε ἐτέλεσεν τοὺς λόγους τούτους. etc., found without historical connection, have been placed by Matthew in such connection, when he saw them apt to his purpose (chaps, 5–7, 10, 11, 13, 18, 23, 24, sq.). It has been maintained by Hilgenfeld that a collection of sayings of any sort, without a background of history, is not to be thought of. It is maintained on the other hand, that, while historical additions or introductory historical notices are to be conceded so far as they are demanded for an understanding of what is said, there was also a great number of sayings of Jesus, separated from the circumstances which induced them, “floating utterances,” which had no fixed place in the evangelical history, and thus appear in Matthew at another point of their wandering than in Luke.23 The source for this material is supposed to be the κόγια of which Papias speaks. That it was a written source is proved by such examples of agreement as follow: Matt. 11:4–6 = Luke 7:22-23. Matt. 11:11 = Luke 7:28, the rare expression ἐν γεννητοῖς γυναικῶν. Matt. 11:25 sq. = Luke 10:21, observe particularly the exactness in the repetition, πατέρ,.....ὁ πατήρ. Matt. 24:50 sq. = Luke 12:46.

18. Mark 1:21–28 = Luke 4:31–37. This Baur thinks a striking example of the dependence of Mark on Luke, since the words διδαχὴ καινὴ.....ὅτι κατ᾿ ἐξουσίαν, v. 27, show his inability to understand Luke’s higher view of the connection between the miracle and the teaching. Hilgenfeld, however, regards Mark’s view of attestation by miracle, the oldest. Unfortunately for the critics the correction of the text by leaving out on takes away what little force the argument may otherwise have had. The expression τίς ὁ λόγος αὗτος. v. 36 of Luke, with its twofold sense, is made an argument for Mark’s originality (so Meyer and Holtz.).

19. Matt. 1:1–17. The title βίβλος γενέσεως is so closely bound with the immediately following, while the section 18-25 is separated, that it cannot be held*to apply further than to the genealogy itself (Calvin, Bengel, De Wette, Meyer, Holtz.). But since so limited a use of the phrase is without parallel (vid. Gen. 5:1 sq.; 11:27 sq.; 37:2 sq., in which cases, history as well as, or instead of, mere genealogy follows), the question is started, did Matthew compile the genealogy himself and use the phrase thus singularly, or, finding it already compiled, appropriate it, with its title, to his purpose. To support the latter view, besides the title itself, is further urged in proof, that since Matthew, 5:20, announces the divine origin of Christ, sufficient motive for undertaking such a work was wanting, while the motive for introducing the genealogy, supposing it to have been already prepared, may be found in the expectations of the Jews (Meyer). Holtzmann, holding this view, still detects traces of Matthew’s preparing hand in the numerical arrangement, in the names Thamar, Rahab, Bathsheba, and in the ᾿ Λησοῦς ὁ λεγόμενος Χριστός of 5:16 (᾿Λησοῦς Χριστός and ὁ Δαυὶδ ὁ βαδιλεύς, however, are unlike Matthew). On the other hand, that Matthew was the original compiler, De Wette urges from the connection of v. 17, which he thinks was evidently written by the evangelist. Believing that Matthew received the genealogy essentially in its present form, some have ascribed to him an alteration (v. 16) of the supposed original form ᾿Λωσὴφ δὲ ἐγέννησεν ᾿Λησοῦν (comp. the ὡς ἐνομίξετο of Luke 3:23) (so Strauss, Hilgenf., Holtz.). According to Meyer the present form of this verse was prior to our Matthew. This section is at any rate of Hebrew origin, and the view that the evangelist took it from some written source seems best supported.

20. Matt. 1:18–2:23. In favor of the view that Matthew gave to the two traditions 1:18–25 and 2:1–23 their first existence in writing, are urged their verbal character and the Old Test, citations (1:23–2:15, 18, 23), made after Matthew’s fashion, and introduced with his accustomed formula (Holtz.). On the other hand, the traditionary character of these chapters, and the strange connection of the third chapter, which, although it joins on to 2:23 verbally, passes over the whole history of the youth of Jesus, show that “the elements are certain separate evangelical records.’’ The similarity of expression is, then, due to the translator, and how much in the form of Old Test, citation is to be ascribed to the first composer, or to the author of the Hebrew Gospel, or to the translator into Greek, cannot now be told (Meyer). These two chapters are certainly not a later addition.24

21. Luke 1:1–2:52. A proof for the order, Mark, Matthew, Luke is found in the consideration that, while in the first the appearance of John the Baptist is the ἀρχὴ τοῦ εὐαγγέλιον, the later formation of gospel history was ever going backward into a remoter domain, so that Matthew gives the narrative of the conception and birth of Jesus, and Luke, proceeding ἄνωθεν, adds that of his forerunner. The proofs of written sources for at least the first of these two chapters are quite conclusive, since in no other way can the artful form of the lyric, and the marked Hebraistic speech be accounted for. How many different written sources in all are to be detected is matter of dispute. The particulars recorded 2:43, 48, 50 have been thought to betray ignorance of the foregoing, and thus point to a separate source for 2:41–51 (Holtz.).

22. Luke 3:23–38. The genealogy of Luke, in the absence of a satisfactory reconciliation, would appear to prove, either that he was unacquainted with Matthew’s Gospel, or that he intentionally rejected the genealogy there given, and selected among several records at his disposal one better suited to his purpose.

It may seem that the examples just given, through the conflict of opinion which they exhibit, serve to obscure our inquiry, rather than throw light upon it. It is much gained, however, to have seen how all-pervading the data for the problem are, and how complex the balancings of judgment over minute points which it calls for; as well as how arbitrarily and ingeniously these points have oftentimes been handled. And let it be always remembered that the problem still remains. It cannot fail, then, to provoke effort as long as it is unsolved. Much of this complexity in treatment is necessarily caused by the complex nature of the problem, so that, though we may justly find fault with some special ways of proceedure, if we wish to reject the whole of this criticism of details, we must begin by refusing to examine the facts which have called it forth.

In order to a tolerably complete presentation of the phenomena, three classes of them deserve to be subjects each of a brief separate discussion. Among the three let us first consider that of,

Citation from the Old Testament.—To this weighty element of the data for a solution of our question Credner contributed much. The canon laid down by him was, that Matthew cites freely from the LXX., but according to a text which in the Messianic passages has been compared with the original Hebrew, and altered after it. De Wette, following Bleek, makes this distinction: The Old Test, citations are of two kinds; such as give proofs of the fulfilment of prophecy, and seem derived from the author’s own reflection, and such as occur in the narrative: the latter class are quoted from the LXX., sometimes literally, sometimes more freely, while the former are the author’s translation of the Hebrew. The only conclusion then, since the difference in method of quotation shows different sources, is that the author was a learned Jew, more familiar with the Hebrew than the LXX., but, composing in Greek, used certain evangelical writings also in Greek, in which quotations were made from the LXX.25 According to the results of another minute investigation, in the citations peculiar to Matthew which are to be considered as pragmatic contributions of his own, the Hebrew is the basis (two passages indeed (2:15, 23) agree with the Hebrew where the LXX. significantly differ),26 though in almost all cases the influence of the LXX. is felt. In citations found in the narrative, however, the influence of the LXX. is much more manifest (Matt. 2:6 seems an exception).27 That there is a marked difference in Matthew’s quotations, which admits of classification, all are agreed, though what is the most precise statement of the difference is not determined.

According to the authority last quoted, the phenomena of agreement between Matthew and Mark are, in the main, as follows: of seventeen citations, in ten they agree verbally, and in four others there are only insignificant variations. There are more important differences in two passages (Matt, 13:14-15 = Mark 4:12 and Matt. 19:18-19, = Mark 10:19), where Matthew has followed the LXX. more closely. Matthew (19:19) has added the words καὶ ἀγαπήσεις τὸν πλησίον etc. from Lev. 19:18, as well as τῆς ποίμνης (26:31 = Mark 14:27). In Matt. 22:24 = Mark 12:19 the form of a quotation is given to an allusion to the Mosaic law. Mark 1:2 is from the Hebrew, but the following verse from the LXX. The whole question of this citation is a difficult one. In Luke all citations, with one exception, are from the LXX.28 as is the author’s practice also in the Acts.29 This exception (7:27 = Matt. 11:10 = Mark 1:2) is freely from the Hebrew (Mal. 3:1), the LXX. having ἐπιβλέψεται instead of κατασκευάσει; and is explained, either by the dependence of Luke on Matthew (so Ritschl), or by difference in sources from which he copied this quotation (Holtz.), or perhaps better by the fact that the citation had in tradition taken this as its customary form (Meyer). Luke 10:27 is formed by joining together Deut. 6:5 cited from the LXX. (the text here not differing more from either text of the LXX. than the various texts of the LXX. differ from each other), and τόν πλησίον σου ὡς σεαυτόν from Lev. 19:18.

Doublettes, — The second class of phenomena which seems -worthy of special mention, is less important, and there is danger of giving it undue weight for purposes of destructive criticism. It comprises all those instances where the gospel history seems to repeat itself. Weisse was the first to use them in the interests of our question, and to bestow upon them the designation “doublettes,” which we shall retain. Most of them are utterances of Christ, but the attempt has been made to detect the same causes in repeated fragments of narrative. Matthew, it is said, has (12:22–24, 38) joined the motive for mention of the sign of Jonah, derived from the incident given Mark 8:11-12, with the motive for Christ’s defence against the blasphemous charge of the Pharisees, found Luke 11:14-15; though he has before given the same (Matt. 9:32–34). This may point to different sources, but does not imply error; for, questions, charges and answers like these can hardly fail to have been repeated in the life of Jesus. Under this head fall such uncalled-for suppositions as that Matthew, having left out the incident recorded Mark 1:21–28, doubles, in compensation, the number of demoniacs (8:28–34), and that the two blind men of Matthew (20:29–34) are due to the omission of Mark 8:22–26.

The following are instances of doublettes, where the same utterances of Jesus are given once in the same connection by all of the three evangelists, and occur a second time in Matthew and Luke in different connection. Matt. 23:12 = Mark 4:25 = Luke 8:18; comp. Matt. 25:29 = Luke 19:26. Matt. 16:24-25, = Mark 8:34-35, = Luke 9:23-24; comp. Matt. 10:38-39, = Luke 14:27, 33. Matt, 24:8–14 = Mark 13:9–13 = Luke 21:12–19; comp. Matt. 10:17–22 = Luke 12:11-12. There is little difficulty in supposing that Jesus repeated several times expressions so like proverbs as the first two cases. There are instances where the doublette occurs in Matthew alone (5:29-30, and 18:8, 9; 5:32 and 19:9), or Luke alone (19:3 and 10:4). While little can be made of these doublettes to throw doubt upon the credibility of the gospel narratives, they are valuable for our purpose chiefly from this remarkable circumstance, that they all occur in the first and third Gospels. They thus point to the conclusion that these writers had material common to them, but unused by Mark, and throw additional light upon the phenomena presented above, No. 17 of the concrete examples.30

We shall conclude the whole statement of the phenomena of agreement and difference with which our inquiry has to do by a brief separate examination of a third class, than which none gives more interesting or satisfactory results. This class comprises the phenomena of Evangelic Verbal Characteristics. No one who composes can fail to impress himself upon the language which he chooses to express his thoughts. The flow and weaving of the sentences, the turn of the expression, the arrangement and selection of the words, so that a certain order may be expected and certain terminology seems favorite, all these betray him whose personality lies back of them all, and works out through them. The way the stilus turns shows the hand that turns it. This whole effect is the resultant of so many and such delicate, impalpable forces, there is such a blending of the power of habit, working beneath consciousness, with the power of the will, only, if we may thus speak, in half-conscious play, that to control this effect so that it shall not reveal the forces is, even for the workman himself, well nigh impossible. Who will suppose that Mark is the epitomizer of Matthew and Luke if he must join with this supposition the other, that Mark has been able to trace and eliminate the authorial peculiarities of his sources, while at the same time retaining his own?

To this class of phenomena such questions as these may be put: First, Has each Gospel, in its verbal phenomena the mark of individuality so stamped upon it that it must be held to be one work, and that its author can have been no mere copyist of the other Gospels or of any number of supposed sources? But in the second place, Are there at the same time traces of the influence of sources, which, though not admitted in crude state, still betray themselves by peculiarities of their own, blended with those of him who made use of them? In search for the answer to these questions, the whole domain of gospel narratives has been most thoroughly surveyed. Wetstein, Gersdorf (1816), who first did creditable service here, De Wette, Credner, Zeller, Wilke, who went carefully through the whole text, section by section, Hitzig, who compared the second Gospel with the Apocalypse, and above all, Holtzmann, who occupies nearly a hundred pages of his work with results of the most minute and candid research, have all contributed to answer these inquiries.

What reply do the facts discovered give to the first question? A few among the great mass of them may answer for themselves:

1. Verbal Characteristics of Matthew. — The phrase βασιλεία τῶν οὐρανῶν occurs twenty-seven times in Matthew, for which the others have βασιλεία τοῦ θεοῦ. ῾ Ο πατὴρ, ὁ ἐν τοῖς οὐρανοῖς, or ὁ ἐπουράνιος is favorite; ἵνα πληρωθῇ is a frequent form of citation, with which or some other form, the expressions ῥηθείς, ῥηθέν, ἐῤῥέθη, are found at least twenty times in this Gospel, and not once in the others (Mark 13:14 is rejected by Tisch.). Frequent in the mention of names ὁ λεγόμενος. The particle τότε, often ἀπό τότε, which occurs in Mark six times and Luke fourteen times, is found ninety-one times in Matthew. He often adds to the words γραμματεῖς and πρεσβύτερον the words τοῦ λαοῦ, and is fond of verbs in - εύειν. Εἰς τὸ ὄνομα occurs four times, only in Matthew, where the others have ἐν or ἐπί. ᾿Από is often found with verbs where the others use ἐκ. ̓́Αγγελος κυρίου, αἷμα, δίκαιον, ἀναχωρεῖν in Matthew ten times, but elsewhere in the synoptic Gospels only in Mark 3:7; ἡγεμών, eleven times in Matthew, elsewhere only thrice; the plural θησαυροί, ὀμνύειν ἐν or εἰς, only in Matthew; προςέρχεσθαι in Mark only six times, and ten in Luke, but found fifty-one times in Matthew; συνάγειν twice as often in this Gospel as in Mark and Luke together; τάλαντον only here; ὕστερον seven times in Matthew but only three in both the others. These and many other similar examples, especially when taken together with the whole Hebrew coloring of this Gospel, will serve to answer the question so far as it is concerned.

2. Verbal Characteristics of Mark. — Most noticeable are those touches which show the lively circumstantial character of this Gospel. Such are καθίσας 9:35; 12:41; κύψας 1:7; προσδραμών 10:17 and δραμών 15:36, τολμήσας 15:43, and περιβλεψάμενος frequently, while the frequent use of the historical present marks the same peculiarity. With Mark no word is more a favorite than εὐθέως or, according to the customary reading of Teschendorf, εὐθύς, for which Luke has often παραχρῆμα. The sentence is often introduced with καὶ or καὶ πάλιν, referring back to the foregoing narrative. The frequent use of diminutives is also peculiar, such as παιδίον, κοράδιον, θυγάτριον, ἐχθύδιον, ὠτάριον. Mark is particularly fond of ἤρξαντο with a following infinitive. More Latin words are found here than in the other Gospels, κεντυρίων, σπεκυλάτωρ, etc. These and other similar verbal peculiarities, together with the singularly graphic style of description, and the absence of long speeches, sufficiently establish Mark’s claim to authorship.

3. Verbal Characteristics of Luke. — This evangelist uses the article with the infinitive (often διὰ τό, τοῦ with the infinitive of purpose twenty-five times, but in Matthew only six times, and Mark once; ἐν τῷ with the infinitive thirty-seven times, only three in Matthew) far more frequently than the others. He admits often the attraction of the relative. With him is found ἒλεγέν (εἶπε) δὲ παραβολήν. The very frequent use of iyivero is peculiar to him. Descriptive participles are often found, sometimes in pairs. Καὶ αὐτός and καὶ αὐτοί occur in Luke three times as often as in both the others. Δὲ καὶ is found twenty-nine times. The title λίμνη for the Galilean sea is peculiar to Luke. ᾿Αδικιά, four times in Luke, once in Matthew; ἅπας twenty times in the Gospel, sixteen in the Acts, and elsewhere in the New Testament only ten times; βαλάντιον, βρέφος, δεῖ, which occur oftener in Luke than in all the other New Testament writers together, διέρχεσθαι, δοξάζειν τὸν θεόν, τὸ εἰρημένον for Matthew’s τὸ ῥηθέν; ἐνώπιον twenty times in Luke, but in neither Matthew nor Mark; ἐπιστάτης only in Luke, but here six times in address to Jesus, and four times in place of διδάσκαλος or Ῥαββί; εὑρίσκειν, ζητεῖν, twenty-seven times in Luke; θημέλιον, ἱκανός, meaning great, many; κατανοεῖν, μακρός, μένειν in the sense of dwell; νομικός, six times, where Matthew has γραμματεύς;ὄπτεσθαι, σκιρτᾶν, ὑπάρχειν, seven times in Luke only; φωβεῖσθαι twenty-four times; χάρις only in this Gospel; ὡσεί, oftener in Luke than all the other New Testament writers together; these, selected out of the great number of examples, show what this examination has to say with regard to the third Gospel.31 To our first inquiry, then, the phenomena of verbal characteristic furnish an answer that admits of no doubt.

The reply to our second inquiry is not so satisfying. The following statements are made by Holtzmann (p. 280 sq.). The phraseology of those sections common to all three seems particularly to have the character “of popular and expressive circumstantiality.” This is seen in such artless repetitions as ἐὰν θέλῃς καθαρίσαι (Matt. 8:2 = Mark 1:40 = Luke 5:12), compared with the following, θέλω, καθαρίσθητε. See also Matt. 9:2 = Mark 2:5 = Luke 5:20, comp. Mark 2:7 = Luke 5:21; and Matt. 9:6 = Mark 2:10 = Luke 5:24. Of the same character is the fondness for strengthening the utterance by restating it in another form (Matt, 13:21 = Mark 4:17 = Luke 8:13, and Matt. 22:16 = Mark 12:14 = Luke 20:21). Mark, as we should expect of the Gospel which lies nearest in form to the original, common source, betrays this proximity by retaining many of these verbal characteristics, where the others have lost them in the process of remodelling. He joins most frequently synonymous expressions, a phenomenon such as we should expect to find in the popular language of this common source (ἀπῆλθεν ἀ̓π᾿ αὐτοῦ ἡ λέπρα, καὶ ἐκαθαρίσθη 1:42; ὅτε χρείαν ἔσχεν, καὶ ἐπείνασεν 2:25; ὀψίας γενομένης, ὅτε ἔδυ ὁ ἥλιος 1:32; τότε, ἐν εῇ ἐκείνῃ ἡμέρᾳ 2:20; εὐθὺς, μετὰ σπουδῆς 6:25). In these common sections the union of two negatives is especially frequent (Matt. 24:35 = Mark 13:31 = Luke 21:33; and Mark 1:44; 3:27; 6:5; 14:25; 16:8). These and similar phenomena are thought to indicate some document lying at the basis of these sections, the language of which was circumstantial and popular; and at the same time they confirm the opinion that Mark is proximate to this source.

It is asserted by the same authority (p. 335 sq.) that sections common only to Matthew and Luke also exhibit such peculiarities of verbal characteristic as prove a source for these writers, which both used, but which was unknown to Mark. The examples brought forward, however, scarcely suffice to establish the assertion.

Those phenomena of agreement and difference in the synoptic Gospels which form the data for solving our problem are now before us. By them every hypothesis and each explanatory remark must be tested. But while the hypothesis and the remark are valuable only so far as they serve to explain the phenomena, a knowledge of them is valuable of itself. It will be a constant stimulus, and it must be the only starting-point of well-directed effort. Whoever masters the phenomena is already well paid for his pains-taking, whether he finds any satisfactory supposition to account for them or not. He can at least intelligently pronounce the solutions offered false or inadequate, and, selecting such general conclusions as seem true, he can rest in them until yet more thorough investigation shall show results more complete and at the same time satisfactory. To examine some of the most important hypotheses, testing them by means of our previous work, and to gather from them whatever seems most like truth, will occupy us at another time.


1) Vid. Hoatzmann, Die synoptischen Evangl. p. 15; Meyer on Mattthew, fifth Aufl. p. 24; and Gieseler, Entstehung der Evangl. p. 31.

2) Die Entstehung, etc. p. 110.

3) Die Entstehung, etc. p. 2.

4) The text used is that of Tischendorf, ed. sept., but the general result will not be altered if any critical text be referred to.

5) Michaelis’s Introduction (2d ed.), Vol. v. p. 168.

6) Die Entstehung, etc. p. 3.

7) Michaelis’s Introd. Vol. v. p. 208 sq.

8) Westcott’s Introduction, p. 201 sq.

9) Vid. De Wette, Einl. 79; and Holtzmann, Die Synoptischen Evangl. p.10 sq.

10) Marsh in Michaelis, Vol. v. p. 169.

11) Vid. Holtzmann, p. 12.

12) Michaelis’s Introd. p. 170, note.

13) For a full list vid. Davidson’s Introd. Vol. i. p. 373 sq.

14) Michaelis’s Introd. Vol. v. pp. 317, 336. The same erroneous statement is to be found Westcott’s Introd. p. 203, note.

15) For further examples rid. De Wette, Einl. 80, and Holtzmann p. 61 sq.

16) Meyer on Mark, fifth Aufl p. 6.

17) Vid. Norton’s Genuineness of the Gospels, Vol. i. Appendix D., and Westcott’s Introd. p. 202 sq.

18) Michaelis’s Introd. Vol. v. p. 257.

19) Vid. Holtzmann, pp. 12, 289 sq., and De Wette, Einl. 79.

20) Holtzmann, pp. 99 sq., 169 sq. Vid. also Westcott’s In trod. p. 344 sq. with the notes.

21) Hug, Einl. Vol. xi. p. 156 sq. and p. 410 sq. of Fosdick’s trans.

22) Vid. Holtzmann, p. 126 sq.

23) Vid. Holtzmann, p. 133 sq.

24) Vid. De Wette, Einl. 92; Meyer on Matthew; and Holtzmann, p. 172.

25) Vid. De Wette, Einl. 97 b. and Westcott’s Introd. p. 232, note.

26) Matt. 2:15 from Hos. 11:1 τὸν υἱόν μου; LXX. τὰ τέκνα αὐτοῦ; Heb.לְבְנְי. Matt. 2:23: If the word Ναζωραῖος is to find its explanation in Isa. 11:1 it must be in the Heb. נֵצֶר.

27) Holtzmann, pp. 258 sq.

28) Vid. Holtzmann, p. 263

29) Vid. De Wette, Einl. 15 a.

30) Vid. Holtzmann. p. 254 sq.

31) Vid. De Wette, Einl. 91; Holtzmann, p. 271 sq.; and notes on cap. vii. in Westcott.