It is possible, indeed probable, that the book originally had no title. The manuscripts give the title in several forms. Aleph (in the inscription) has merely "Acts" (Praxeis). So Tischendorf, while Origen, Didymus, Eusebius quote from "The Acts." But BD Aleph (in subscription) have "Acts of Apostles" or "The Acts of the Apostles" (Praxeis Apostolon). So Westcott and Hort, Nestle (compare Athanasius and Euthalius). Only slightly different is the title in 31,61, and many other cursives (Praxeis ton Apostolon, "Acts of the Apostles"). So Griesbach, Scholz. Several fathers (Clement of Alex, Origen, Dionysius of Alex, Cyril of Jerusalem, Chrysostom) quote it as "The Acts of the Apostles" (Hai Praxeis ton Apostolon). Finally A2 EGH give it in the form "Acts of the Holy Apostles" (Praxeis ton Hagion Apostolon). The Memphitic version has "The Acts of the Holy Apostles." Clearly, then, there was no single title that commanded general acceptance.

(1) The chief documents. These are the Primary Uncials (Codex Sinaiticus, Codex Alexandrinus, Codex Vaticanus, Codex Ephraemi Rescriptus, Codex Bezae), Codex Laudianus (E) which is a bilingual Uncial confined to Acts, later Uncials like Codex Modena, Codex Regius, Codex the Priestly Code (P), the Cursives, the Vulgate, the Peshitta and the Harclean Syriac and quotations from the Fathers. We miss the Curetonian and Syriac Sinaiticus, and have only fragmentary testimony from the Old Latin.

(2) The modern editions of Acts present the types of text (Textus Receptus; the Revised Version (British and American); the critical text like that of Westcott and Hort, The New Testament in Greek or Nestle or Weiss or von Soden). These three types do not correspond with the four classes of text (Syrian, Western, Alexandrian, Neutral) outlined by Hort in his Introduction to the New Testament in Greek (1882). These four classes are broadly represented in the documents which give us Acts. But no modern editor of the Greek New Testament has given us the Western or the Alexandrian type of text, though Bornemann, as will presently be shown, argues for the originality of the Western type in Acts. But the Textus Receptus of the New Testament (Stephanus' 3rd edition in 1550) was the basis of the King James Version of 1611. This edition of the Greek New Testament made use of a very few manuscripts, and all of them late, except Codex Bezae, which was considered too eccentric to follow. Practically, then, the King James Version represents the Syriac type of text which may have been edited in Antioch in the 4th century. Various minor errors may have crept in since that date, but substantially the Syriac recension is the text of the King James Version today. Where this text stands alone, it is held by nearly all modern scholars to be in error, though Dean Burgon fought hard for the originality of the Syriac text (The Revision Revised, 1882). The text of Westcott and Hort, The New Testament in Greek is practically that of Codex Vaticanus, which is held to be the Neutral type of text. Nestle, von Soden, Weiss do not differ greatly from the text of Westcott and Hort, The New Testament in Greek, though von Soden and Weiss attack the problem on independent lines. The text of the Revised Version (British and American) is in a sense a compromise between that of the King James Version and the critical text, though coming pretty close to the critical text. Compare Whitney, The Reviser's Greek Text, 1892. For a present-day appreciation of this battle of the texts see J. Rendel Harris, Side Lights on the New Testament, 1908. For a detailed comparison between the King James Version and the Revised Version (British and American) Acts see Rackham, The Acts of the Apostles, xxii.

(3) In Acts the Western type of text has its chief significance. It is the meet of the late Friedrich Blass, the famous classicist of Germany, to have shown that in Luke's writings (Gospel and Acts) the Western class (especially D) has its most marked characteristics. This fact is entirely independent of theory advanced by Blass which will be cussed directly. The chief modern revolt against theories of Westcott and Hort, The New Testament in Greek is the new interest felt in the value of the Western type of text. In particular Codex Bezae has come to the front in the Book of Acts. The feeble support that Codex Bezae has in its peculiar readings in Acts (due to absence of Curetonian Syriac and of the Old Latin) makes it difficult always to estimate the value of this document. But certainly these readings deserve careful consideration, and some of them may be correct, whatever view one holds of the Codex Bezae text. The chief variations are, as is usual with the Western text, additions and paraphrases. Some of the prejudice against Codex Bezae has disappeared as a result of modern discussion.

(4) Bornemann in 1848 argued that Codex Bezae in Acts represented the original text. But he has had very few followers.

(5) J. Rendel Harris (1891) sought to show that Codex Bezae (itself a bilingual MS) had been Latinized. He argued that already in 150 AD a bilingual manuscript existed. But this theory has not won a strong following.

(6) Chase (1893) sought to show that the peculiarities were due to translation from the Syriac.

(7) Blass in 1895 created a sensation by arguing in his Commentary on Acts (Acta Apostolorum, 24 ff) that Luke had issued two editions of the Acts, as he later urged about the Gospel of Luke (Philology of the Gospels, 1898). In 1896 Blass published this Roman form of the text of Acts (Acta Apostolorum, secundum Formam quae videtur Romanam). Blass calls this first, rough, unabridged copy of Acts (beta) and considers that it was issued at Rome. The later edition, abridged and revised, he calls alpha. Curiously enough, in Ac 11:28, Codex Bezae has "when we had gathered together," making Luke present at Antioch. The idea of two editions is not wholly original with Blass. Leclerc, a Dutch philologist, had suggested the notion as early as the beginning of the 18th century. Bishop Lightfoot had also mentioned it (On a Fresh Revision of the New Testament, 29). But Blass worked the matter out and challenged the world of scholarship with his array of arguments. He has not carried his point with all, though he has won a respectable following. Zahn (Einl, II, 338 ff, 1899) had already been working toward the same view (348). He accepts in the main Blass' theory, as do Belser, Nestle, Salmon, Zockler. Blass acknowledges his debt to Corssen (Der cyprianische Text der Acta Apostolorum, 1892), but Corssen considers the alpha text as the earlier and the beta text as a later revision.

(8) Hilgenfeld (Acta Apostolorum, etc., 1899) accepts the notion of two edd, but denies identity of authorship.

(9) Schmiedel (Encyclopedia Biblica) vigorously and at much length attacks Blass' position, else "the conclusions reached in the foregoing sections would have to be withdrawn." He draws his conclusions and then demolishes Blass! He does find weak spots in Blass' armor as others have done (B. Weiss, Der Codex D in der Apostelgeschichte, 1897; Page, Class. Rev., 1897; Harnack, The Acts of the Apostles, 1909, 45). See also Knowling, The Acts of the Apostles, 1900, 47, for a sharp indictment of Blass' theory as being too simple and lacking verification.

(10) Harnack (The Acts of the Apostles, 48) doubts if Luke himself formally published the book. He thinks that he probably did not give the book a final revision, and that friends issued two or more editions He considers that the so-called beta recension has a "series of interpolations" and so is later than the alpha text.

(11) Ramsay (The Church in the Roman Empire, 150; Paul the Traveler, 27; The Expositor, 1895) considers the beta text to be a 2nd-century revision by a copyist who has preserved some very valuable 2nd-century testimony to the text.

(12) Headlam (HDB) does not believe that the problem has as yet been scientifically attacked, but that the solution lies in the textual license of scribes of the Western type (compare Hort, Introduction, 122 ff). But Headlam is still shy of "Western" readings. The fact is that the Western readings are sometimes correct as against the Neutral (compare Mt 27:49). It is not necessary in Ac 11:20 to say that Hellenas is in Western authorities (AD, etc.) but is not a Western reading. It is at any rate too soon to say the final word about the text of Acts, though on the whole the alpha text still holds the field as against the beta text. The Syriac text is, of course, later, and out of court.

III. Unity of the Book.

It is not easy to discuss this question, apart from that of authorship. But they are not exactly the same. One may be convinced of the unity of the book and yet not credit it to Luke, or, indeed, to anyone in the 1st century. Of course, if Luke is admitted to be the author of the book, the whole matter is simplified. His hand is in it all whatever sources he used. If Luke is not the author, there may still have been a competent historian at work, or the book may be a mere compilation. The first step, therefore, is to attack the problem of unity. Holtzmann (Einl, 383) holds Luke to be the author of the "we" sections only. Schmiedel denies that the Acts is written by a companion of Paul, though it is by the same author as the Gospel bearing Luke's name. In 1845 Schleiermacher credited the "we" sections to Timothy, not to Luke. For a good sketch of theories of "sources," see Knowling on Acts, 25 ff. Van Manen (1890) resolved the book into two parts, Acta Petri and Acta Pauli, combined by a redactor. Sorof (1890) ascribes one source to Luke, one to Timothy. Spitta also has two sources (a Pauline-Lukan and a Jewish-Christian) worked over by a redactor.

Clemen (1905) has four sources (History of the Hellenists, History of Peter, History of Paul, and a Journey of Paul), all worked over by a series of editors. Hilgenfeld (1895) has three sources (Acts of Peter, Acts of the Seven, Acts of Paul). Jungst (1895) has a Pauline source and a Petrine source J. Weiss (1893) admits sources, but claims that the book has unity and a definite aim. B. Weiss (1902) conceives an early source for the first part of the book. Harnack (The Acts of the Apostles, 1909, 41 f) has small patience with all this blind criticism: "With them the book passes as a comparatively late patchwork compilation, in which the part taken by the editor is insignificant yet in all cases detrimental; the 'we' sections are not the property of the author, but an extract from a source, or even a literary fiction." He charges the critics with "airy conceit and lofty contempt." Harnack has done a very great service in carefully sifting the matter in his Luke the Physician (1907). He gives detailed proof that the "we" sections are in the same style and by the same author as the rest of the book (26-120). Harnack does not claim originality in this line of argument:

"It has been often stated and often proved that the 'we' sections in vocabulary, in syntax, and in style are most intimately bound up with the whole work, and that this work itself including the Gospel, in spite of all diversity in its parts, is distinguished by a grand unity of literary form" (Luke the Physician, 26). He refers to the "splendid demonstration of this unity" by Klostermann (Vindiciae Lucanae, 1866), to B. Weiss, who, in his commentary (1893, 2 Aufl, 1902) "has done the best work in demonstrating the literary unity of the whole work," to "the admirable contributions" of Vogel (Zur Charakteristik des Lukas, etc., 2 Aufl, 1899) to the "yet more careful and minute investigations" of Hawkins (Horae Synopticae, 1899, 2nd edition, 1909), to the work of Hobart (The Medical Language of Luke, 1882), who "has proved only too much" (Luke the Physician, 175), but "the evidence is of overwhelming force" (198). Harnack only claims for himself that he has done the work in more detail and with more minute accuracy without claiming too much (27). But the conversion of Harnack to this view of Acts is extremely significant. It ought not to be necessary any more to refute the partition theories of the book, or to set forth in detail the proofs for the unity of the book.

Perhaps the compilation theory of Acts is nowhere set forth more cogently than in McGiffert's The Apostolic Age (1897). See a powerful refutation of his argument by Ramsay in Pauline and Other Studies (1906, 302-21). "I think his clever argumentation is sophistical" (305). Harnack is fully aware that he has gone over to the rode of "Ramsay, Weiss and Zahn": "The results at which I have arrived not only approach very nearly to, but are often coincident with, the results of their research" (The Acts of the Apostles, 302). He is afraid that if these scholars failed to get the ear of critics "there is little prospect of claiming the attention of critics and compelling them to reconsider their position." But he has the advantage of coming to this conclusion from the other side. Moreover, if Harnack was won by the force of the facts, others may be. This brief sketch of Harnack's experience may take the place of detailed presentation of the arguments for the unity of the book. Harnack sets forth in great wealth of detail the characteristic idioms of the "we" sections side by side with parallels in other parts of Acts and the Gospel of Luke. The same man wrote the rest of Acts who wrote the "we" sections. This fact should now be acknowledged as proven. This does not mean that the writer, a personal witness in the "we" sections, had no sources for the other parts of Acts. This aspect of the matter will be considered a little later.

Assuming the unity of the book, the argument runs as follows: The author was a companion of Paul. The "we" sections prove that (Ac 16:10-17; 20:6-16; 21; 27; 28). These sections have the fullness of detail and vivid description natural to an eye-witness. This companion was with Paul in the second missionary journey at Troas and at Philippi, joined Paul's party again at Philippi on the return to Jerusalem during the third tour, and probably remained with Paul till he went to Rome. Some of Paul's companions came to him at Rome: others are so described in the book as to preclude authorship. Aristarchus, Aquila and Priscilla, Erastus, Gaius, Mark, Silas, Timothy, Trophimus, Tychicus and others more or less insignificant from the point of view of connection with Paul (like Crescens, Demas, Justus, Linus, Pudens, Sopater, etc.) are easily eliminated. Curiously enough Luke and Titus are not mentioned in Acts by name at all. They are distinct persons as is stated in 2Ti 4:10 f. Titus was with Paul in Jerusalem at the conference (Ga 2:1) and was his special envoy to Corinth during the time of trouble there. (2Co 2:1, 12:18.)

He was later with Paul in Crete (Titus 1:5). But the absence of mention of Titus in Acts may be due to the fact that he was a brother of Luke (compare 2Co 8:18; 12:18). So A. Souter in DCG, article "Luke." If Luke is the author, it is easy to understand why his name does not appear. If Titus is his brother, the same explanation occurs. As between Luke and Titus the medical language of Acts argues for Luke. The writer was a physician. This fact Hobart (The Medical Language of Luke, 1882) has demonstrated. Compare Zahn, Einl, 2, 435 ff; Harnack's Luke the Physician, 177 ff. The arguments from the use of medical terms are not all of equal weight. But the style is colored at points by the language of a physician. The writer uses medical terms in a technical sense. This argument involves a minute comparison with the writings of physicians of the time. Thus in Ac 28:3 f kathapto, according to Hobart (288), is used in the sense of poisonous matter invading the body, as in Dioscorides, Animal. Ven. Proem. So Galen, De Typis 4 (VII, 467), uses it "of fever fixing on parts of the body." Compare Harnack, Luke the Physician, 177 f. Harnack agrees also that the terms of the diagnosis in Ac 28:8 "are medically exact and can be vouched for from medical literature" (ibid., 176 f). Hobart has overdone his argument and adduced many examples that are not pertinent, but a real residuum remains, according to Harnack. Then pimprasthai is a technical term for swelling. Let these serve as examples. The interest of the writer in matters of disease is also another indication, compare Lu 8:43. Now Luke was a companion of Paul during his later ministry and was a physician. (Col 4:14). Hence, he fulfils all the requirements of the case. The argument thus far is only probable, it is true; but there is to be added the undoubted fact that the same writer wrote both Gospel and Acts (Ac 1:1). The direct allusion to the Gospel is reinforced by identity of style and method in the two books.

The external evidence is clear on the matter. Both Gospel and Acts are credited to Luke the physician. The Muratorian canon ascribes Acts to Luke. By the end of the 2nd century the authority of the Acts is as well established as that of the Gospel (Salmon, Introduction to the New Testament, 1885, 366). Irenaeus, Tertullian, Clement of Alexandria, all call Luke the author of the book. The argument is complete. It is still further strengthened by the fact that the point of view of the book is Pauline and by the absence of references to Paul's epistles. If one not Paul's companion had written Acts, he would certainly have made some use of them. Incidentally, also, this is an argument for the early date of the Acts. The proof that has won Harnack, the leader of the left in Germany, to the acknowledgment of the Lukan authorship of Acts ought to win all to this position.

The use of the Acts does not appear so early or so frequently as is true of the gospels and the Pauline epistles. The reason obvious. The epistles had a special field and the gospels appealed to all. Only gradually would Acts circulate. At first we find literary allusions without the name of book or author. But Holtzmann (Einl, 1892, 406) admits the use of Acts by Ignatius, Justin Martyr, Polycarp. The use of the Gospel according to Luke by Tatian and Marcion really revolves knowledge of the Acts. But in Irenaeus frequently (Adv. Haer., i. 23, 1, etc.) the Acts is credited to Luke and regarded as Scripture. The Canon of Muratori list it as Scripture. Tertullian and Clement of Alexandria attribute the book to Luke and treat it as Scripture. By the times of Eusebius the book is generally acknowledged as part of the canon. Certain of the heretical parties reject it (like the Ebionites, Marcionites, Manicheans). But by this time the Christians had come to lay stress on history (Gregory, Canon and Text of the New Testament, 1907, 184), and the place of Acts is now secure in the canon.

1. Luke's relations to Josephus.

The acceptance of the Lukan authorship settles the question of some of the dates presented by critics. Schmiedel places the date of Acts between 105 and 130 AD (Encyclopedia Biblica). He assumes as proven that Luke made use of the writings of Josephus. It has never been possible to take with much seriousness the claim that the Acts shows acquaintance with Josephus. See Keim, Geschichte Jesu, III, 1872, 134, and Krenkel, Josephus und Lucas, 1894, for the arguments in favor of that position. The words quoted to prove it are in the main untechnical words of common use. The only serious matter is the mention of Theudas and Judas the Galilean in Ac 5:36 f and Josephus (Ant., XX, v, 1 f). In Josephus the names occur some twenty lines apart and the resemblance is only slight indeed. The use of peitho in connection with Theudas and apostesai concerning Judas is all that requires notice. Surely, then, two common words for "persuade" and "revolt" are not enough to carry conviction of the writer's use of Josephus. The matter is more than offset by the differences in the two reports of the death of Herod Agrippa (Ac 12:19-23; Josephus, Ant, XVIII, vi, 7, XIX, viii, 2). The argument about Josephus may be definitely dismissed from the field. With that goes all the ground for a 2nd-century date. Other arguments have been adduced (see Holtzmann, Einl, 1892, 405) such as the use of Paul's epistles, acquaintance with Plutarch, Arrian and Pausanias, because of imitation in method of work (i.e. parallel lives of Peter and Paul, periods of history, etc.), correction of Ga in Acts (for instance, Ga 1:17-24 and Ac 9:26-30; Ga 2:1-10 and Ac 15:1-33). The parallel with Plutarch is fanciful, while the use of Panl's epistles is by no means clear, the absence of such use, indeed, being one of the characteristics of the book. The variation from Galatians is far better explained on the assumption that Luke had not seen the epistles.

2. 80 AD Is the Limit if the Book Is to Be Credited to Luke.

The majority of modern critics who accept the Lukan authorship place it between 70 and 80 AD. So Harnack, Lechler, Meyer, Ramsay, Sanday, Zahn. This opinion rests mainly on the idea that the Gospel according to Luke was written after the destruction of Jerusalem in 70 AD. It is claimed that Lu 21:20 shows that this tragedy had already occurred, as compared with Mr 13:14 and Mt 24:15. But the mention of armies is very general, to be sure. Attention is called also to the absence of the warning in Luke. Harnack (The Acts of the Apostles, 291 f) admits that the arguments in favor of the date 70 to 80 are by no means conclusive. He writes "to warn critics against a too hasty closing of the chronological question." In his new book (Neue Untersuchungen zur Apostelgeschichte, etc., 1911, S. 81) Harnack definitely accepts the date before the destruction of Jerusalem. Lightfoot would give no date to Acts because of the uncertainty about the date of the Gospel.

This date is supported by Blass, Headlam, Maclean, Rackham, Salmon. Harhack, indeed, considers that "very weighty considerations" argue for the early date. He, as already stated, now takes his stand for the early date. It obviously the simplest way to understand Luke's close of the Acts to be due to the fact that Paul was still in prison. Harnack contends that the efforts to explain away this situation are not "quite satisfactory or very illuminating." He does not mention Paul's death because he was still alive. The dramatic purpose to bring Paul to Rome is artificial. The supposition of a third book from the use of proton in Ac 1:1 is quite gratuitous, since in the Koine, not to say the earlier Greek, "first was often used when only two were mentioned (compare "our first story" and "second story," "first wife" and "second wife"). The whole tone of the book is that which one would naturally have before 64 AD. After the burning of Rome and the destruction of Jerusalem the attitude maintained in the book toward Romans and Jews would have been very difficult unless the date was a long times afterward Harnack wishes "to help a doubt to its lust dues." That "doubt" of Harnack is destined to become the certainty of the future. (Since this sentence was written Harnack has settled his own doubt.) The book will, I think, be finally credited to the time 63 AD in Rome. The Gospel of Luke will then naturally belong to the period of Paul's imprisonment in Caesarea. The judgment of Moffatt (Historical New Testament, 1901, 416) that "it cannot be earlier than 80 AD is completely upset by the powerful attack of Harnack on his own previous position. See also Moffatt's Introduction to the Literature of the New Testament (1911) and Koch's Die Abfassungszeit des lukanischen Geschichtswerkes (1911).

VII. Sources Used by Luke.

If we now assume that Luke is the author of the Acts, the question remains as to the character of the sources used by him. One is at liberty to appeal to Lu 1:1-4 for the general method of the author. He used both oral and written sources. In the Acts the matter is somewhat simplified by the fact that Luke was the companion of Paul for a considerable part of the narrative (the "we" sections, Ac 16:11-17; 20:5; 21:18; 27 and 28). It is more than probable that Luke was with Paul also during his last stay in Jerusalem and during the imprisonment at Caesarea.

There is no reason to think that Luke suddenly left Paul in Jerusalem and returned to Caesarea only when he started to Rome (Ac 27:1). The absence of "we" is natural here, since it is not a narrative of travel, but a sketch of Paul's arrest and series of defenses. The very abundance of material here, as in Ac 20 and 21, argues for the presence of Luke. But at any rate Luke has access to Paul himself for information concerning this period, as was true of the second, from Ac 13 to the end of the book. Luke was either present or he could have learned from Paul the facts used. He may have kept a travel diary, which was drawn upon when necessary. Luke could have taken notes of Paul's addresses in Jerusalem (Ac 22) and Caesarea (Ac 24-26). From these, with Paul's help, he probably composed the account of Paul's conversion (Ac 9:1-30). If, as I think is true, the book was written during Paul's first Roman imprisonment, Luke had the benefit of appeal to Paul at all points. But, if so, he was thoroughly independent in style and assimilated his materials like a true historian. Paul (and also Philip for part of it) was a host witness to the events about Stephen in Ac 6:8-8:1 and a participant of the work in Antioch (11:19-30). Philip, the of Paul's company (21:8) on the last journey to Jerusalem, was probably in Caesarea still during Paul's confinement there. He could have told Luke the events in Ac 6:1-7 and 8:4-40. In Caesarea also the story of Peter's work may have been derived, possibly even from Cornelius himself (9:32-11:18). Whether Luke ever went to Antioch or not we do not know (Codex Bezae has "we" in Ac 11:28), though he may have had access to the Antiochian traditions. But he did go to Jerusalem. However, the narrative in Ac 12 probably rests on the authority of John Mark (Ac 12:12,25), in whose mother's house the disciples were assembled. Luke was apparently thrown with Mark in Rome (Col 4:10), if not before. For Ac 1-5 the matter does not at first seem so clear, but these chapters are not necessarily discredited on that account. It is remarkable, as ancient historians made so little mention of their sources, that we can connect Luke in the Acts with so many probable fountains of evidence. Barnabas (4:36) was able to tell much about the origin of the work in Jerusalem. So could Mnason. Philip also was one of the seven (6:5; 21:8). We do not know that Luke met Peter in Rome, though that is possible. But during the stay in Jerusalem and Caesarea (two years) Luke had abundant opportunity to learn the narrative of the great events told in Ac 1-5. He perhaps used both oral and written sources for this section. One cannot, of course, prove by linguistic or historical arguments the precise nature of Luke's sources in Acts. Only in broad outlines the probable materials may be sketched.

VIII. The Speeches in Acts.

This matter is important enough to receive separate treatment. Are the numerous speeches reported in Acts free compositions of Luke made to order a la Thucydides? Are they verbatim reports from notes taken at the times and literally copied into the narrative? Are they substantial reports incorporated with more or less freedom with marks of Luke's own style? In the abstract either of these methods was possible. The example of Thucydides, Xenophon, Livy and Josephus shows that ancient historians did not scruple to invent speeches of which no report was available. There are not wanting those who accuse Luke of this very thing in Acts. The matter can only be settled by an appeal to the facts so far as they can be determined. It cannot be denied that to a certain extent the hand of Luke is apparent in the addresses reported by him in Acts. But this fact must not be pressed too far. It is not true that the addresses are all alike in style.

It is possible to distinguish very clearly the speeches of Peter from those of Paul. Not merely is this true, but we are able to compare the addresses of both Paul and Peter with their epistles. It is not probable that Luke had seen these epistles, as will presently be shown. It is crediting remarkable literary skill to Luke to suppose that he made up "Petrine" speeches and "Pauline" speeches with such success that they harmonize beautifully with the teachings and general style of each of these apostles. The address of Stephen differs also sharply from those of Peter and Paul, though we are not able to compare this report with any original work by Stephen himself. Another thing is true also, particularly of Paul's sermons.

They are wonderfully stated to time, place and audience. They all have a distract Pauline flavor, and yet a difference in local color that corresponds, to some extent, with the variations in the style of Paul's epistles. Professor Percy Gardner (The Speeches of Paul in Acts, in Cambridge Biblical Essays, 1909) recognizes these differences, but seeks to explain them on the ground of varying accuracy in the sources used by Luke, counting the speech at Miletus as the most historic of all. But he admits the use of sources by Luke for these addresses.

The theory of pure invention by Luke is quite discredited by appeal to the facts. On the other hand, in view of the apparent presence of Luke's style to some extent in the speeches, it can hardly be claimed that he has made verbatim reports. Besides, the report of the addresses of Jesus in Luke's Gospel (as in the other gospels) shows the same freedom in giving the substance exact reproduction of the words that is found in Acts. Again, it seems clear that some, if not all, the reports in Acts are condensed, mere outlines in the case of some of Peter's addresses.

The ancients knew how to make shorthand reports of such addresses. The oral tradition was probably active in preserving the early speeches of Peter and even of Stephen, though Paul himself heard Stephen. The speeches of Paul all show the marks of an eyewitness (Bethge, Die paulinischen Reden, etc., 174). For the speeches of Peter, Luke may have had documents, or he may have taken down the current oral tradition while he was in Jerusalem and Caesarea. Peter probably spoke in Greek on the day of Pentecost. His other addresses may have been in Aramaic or in Greek. But the oral tradition would certainly carry them in Greek, if also in Aramaic.

Luke heard Paul speak at Miletus (Ac 20) and may have taken notes at the time. So also he almost certainly heard Paul's address on the steps of the Tower of Antonia (Ac 22) and that before Agrippa (Ac 26). There is no reason to think that he was absent when Paul made his defenses before Felix and Festus (Ac 24-25) He was present on the ship when Paul spoke (Ac 27), and in Rome when he addressed the Jews (Ac 28) Luke was not on hand when Paul delivered his sermon at Antioch in Pisidia (Ac 13), or at Lystra (Ac 14), or at Athens (Ac 17) But these discourses differ so greatly in theme and treatment, and are so essentially Pauline that it is natural to think that Paul himself gave Luke the notes which he used.

The sermon at Antioch in Pisidia is probably given as a sample of Paul's missionary discourses. It contains the heart of Paul's gospel as it appears in his epistles. He accentuates the death and resurrection of Jesus, remission of sins through Christ, justification by faith. It is sometimes objected that at Athens the address shows a breadth of view and sympathy unknown to Paul, and that there is a curious Attic tone to the Greek style. The sermon does go as far as Paul can (compare 1Co 9:22) toward the standpoint of the Greeks (but compare Col and Eph). However, Paul does not sacrifice his principle of grace in Christ. He called the Athenians to repentance, preached the judgment for sin and announced the resurrection of Jesus from the dead. The fatherhood of God and the brotherhood of man here taught did not mean that God yanked at sin and could save all men without repentance and forgiveness of sin. Chase (The Credibility of Acts) gives a collection of Paul's missionary addresses. The historical reality and value of the speeches in Acts may be said to be vindicated by modern scholarship. For a sympathetic and scholarly discussion of all of Paul's addresses see Jones, Paul the Orator (1910). The short speech of Tertullus (Ac 24) was made in public, as was the public statement of Festus in Ac 26. The letter of Claudias Lysias to Felix in Ac 23 was a public document. How Luke got hold of the conversation about Paul between Festus and Agrippa in Ac 26 is more difficult to conjecture.