10. Date
10
Because works were sometimes released in different drafts, as already noted, the question of Acts’ date may be more complicated than that of most modern writings; that is, when we speak of Acts’ date, which draft of Acts do we have in mind? For our purposes, we are interested in the author’s final completed form of Acts, which is similar to the form to which we have access today (as best we can reconstruct it despite its complicated subsequent textual history).
Views on the date of Acts range widely, and scholars should be charitable in their disagreements here. Scholars have argued for various positions, and it should be clear that I respect the arguments of those with whom I disagree. Indeed, various proposals enjoy varying degrees of certainty (e.g., the approximate date I suggest could be raised or lowered a decade without conflicting very significantly with the strongest evidence I offer). Some scholars have assigned others to partisan “camps” to concisely and effortlessly dismiss their position rather than considering their arguments seriously. Aside from the fact that many scholars do not fall predictably into “camps” on every issue, “camps” often hold particular views because of a conglomeration of beliefs that rest on prior arguments or assumptions. In this case, views about date cannot be separated completely from views about authorship, views about the degree to which Luke’s reports cohere with extant sources besides Luke, and the like.
No particular proposed date for Acts (among those between 64 and 90) is absolutely compelling, though the weight of evidence might incline us more toward one date than another. One cannot separate the question of date entirely from the question of authorship, since one who attributes Acts to a companion of Paul would need to date Acts within the lifetime of a companion of Paul. Still, a companion of Paul could have written as late as three decades after Paul’s death, as some scholars have, in fact, come close to arguing. Historical precedent certainly allows a substantial delay in writing; Xenophon wrote a vivid enough account of the Greek retreat from the heart of Persia long after the events he narrates, undoubtedly on the basis of notes he took during the events. Because we do not know the author’s age, the question of authorship is not absolutely decisive for the date (apart from excluding the rarer second-century proposals, which Acts’ reliability on numerous key points might also exclude). Unfortunately, neither are any of the other factors (in terms of certainty).
Nevertheless, having worked through Acts examining patterns that suggest what Luke desires to emphasize, I believe that a date after 70 is likelier than an earlier date. Because charges against Paul and his death in Roman custody appear to remain a live apologetic issue (see the discussion of apologetic for Paul in above), however, I am also inclined to date Acts much sooner after 70 than are some other scholars. I believe that the value of Paul’s legacy (and not merely its content) remained in dispute especially in the years immediately after his death.
Of the four positions surveyed below, the centrist position (70s–80s), which I hold, has by far the most adherents (perhaps four times as many adherents as supporters of a second-century date); probably the early date (60s) is second in number of adherents (Pervo cites more than thirty scholars); a date in the 90s ranks third; and the second century (clustering toward its beginning) boasts the fewest adherents. To my knowledge, no one dates Acts outside these ranges; the conclusion of Luke’s narrative precludes an earlier date, and textual and patristic evidence a later one.
Various factors support the first-century date, and in my opinion a date closer to 70 rather than later. First, as I have mentioned, authorship by a companion of Paul is, given ancient usage, by far the most natural way to construe the “we” citations in Acts, despite the reluctance of many NT scholars to grant this observation (see comment on Acts 16:10). Second, the massive correspondences between Acts and first-century historical events, shown by Talbert and others, reflect early memory (or at least heavy reliance on early sources). Third, the Pauline apologetic in Acts responds in detail to memories of fairly detailed local charges, memories that would have mattered most in the early period. As pointed out in of this introduction, to have produced such detailed charges merely for entertainment would have undercut Luke’s apologetic for Paul.
Many scholars have argued for an early date, before Paul’s death. This position has sometimes been advanced even by scholars skeptical of some of the contents of the book; for example, late nineteenth- and early twentieth-century liberal scholar Adolf von Harnack gradually recognized the appeal of this position, and Erwin Goodenough (much more skeptical of Acts’ content) also dated it early.
Usually scholars arguing for a pre-70 date contend that Acts ends where it does because the events had unfolded only this far at the time of Luke’s writing, that is, about 62 C.E. Why else would Luke devote a quarter of Acts to Paul’s trial and appeal yet not record the outcome? Some ancient interpreters came to the same conclusion.
Although this argument seems convincing on the surface, it is open to considerable challenge. An argument from the abrupt ending in Acts need not mean that Luke knew no more about Paul, any more than Mark’s abrupt ending (Mark 16:8) means that Mark knew no more about Jesus’s resurrection appearances (cf. 1 Cor 15:5–8). Some ancient writers criticized historians who ended their accounts prematurely, but these very criticisms confirm that the practice was known. No less renowned a historian than Thucydides, who lived to see the conclusion of the twenty-seven-year Peloponnesian War, ends his account of the war five years before its conclusion. Valerius Maximus ends his work abruptly with no genuine conclusion (Val. Max. 9.15.ext. 2). Herodian ends his history suddenly at the accession of Gordian III (Hist. 8.8.8), but he does not leave off at this point simply because he is writing then; what he writes he would not have dared write until after Gordian’s fall. Second Maccabees, though written after 1 Maccabees, both starts and ends at a period earlier than 1 Maccabees (and ends on a happy note); clearly, it did not bring events up to its author’s day.
Granted, if Luke recorded Paul’s Roman hearings in Jerusalem and Caesarea for use at a trial before Caesar, this material was clearly compiled before Paul’s death. One can, however, affirm such a purpose for this material and yet deny an early date of our book of Acts; Luke undoubtedly incorporated considerable earlier material into a later, final finished product. Also, this material, in its current literary (as opposed to purely forensic) form, could remain valuable in the years immediately following Paul’s execution, especially if it had been associated with a broader persecution of Christians on similar charges or if Paul’s reputation had been challenged in some circles (See comment on Acts 24:5).
Further, Luke could have suppressed explicit mention of Paul’s death (even if his audience shared knowledge of the event) because it did not suit his larger narrative purpose. Martyrdom would not match the pattern of vindication in Roman courts that he has provided (like Josephus’s focus on favorable precedents), the triumphant progress characterized in his narrative, or the parallel with Jesus’s resurrection. Or (on a different view) perhaps he omitted it because the imprisonment itself parallels Jesus’s passion in his Gospel but Luke and his audience knew that Paul was released at the end of two years. The period in which the story in Acts breaks off does not provide decisive evidence for when Luke finished writing it.
Thus although the arguments supporting a date before Paul’s death are stronger than most other specific arguments for dating Acts, they are hardly compelling by themselves. In fact, if we are correct that Acts is a historical monograph rather than a biography (see above), Luke is under no obligation to narrate Paul’s execution; his climax is the gospel reaching the heart of the empire.
Luke seems to specify a period of “two years” at the end of his record (Acts 28:30) because he and his audience share knowledge that something happened at the end of the two years—whether Paul’s death or his release, to one of which his narrative points. Luke also seems aware that Paul did in fact appear before the emperor (27:24). Some scholars also think that 20:25 predicts Paul’s death, leaving the reader to understand that it will follow in the narrative world, just as the Iliad predicted, without narrating, Achilles’s death (Hom. Il. 20.337; 24.538–40); if Acts 20:25 does not predict Paul’s death, it at least warns that he will not return to Ephesus.
Luke probably could not have extended Acts much beyond its current length if the symmetry between his two volumes is deliberate; the number of Greek words in Acts is nearly the figure for the Gospel, which may have been the fullest length for Luke’s normal scrolls. Although some writers may have completed more volumes and then revised them, usually each volume in a multivolume work was complete at the time of its “publication” (Pliny Ep. 5.5.5). Narrating Paul’s death would fit the parallel with Jesus minus the resurrection—an omission that would not end Acts on the triumphant note of the gospel’s progress, consistently portrayed earlier in Acts. Luke thus need not narrate it regardless of the date assigned to his work.
Scholars supporting a pre-70 date have also advanced other arguments favoring this position, drawn especially from the setting the book seems to address. One argument is based on Luke’s failure to describe the temple’s destruction as past; but Luke certainly implies Jerusalem’s destruction (past or not; Luke 19:43–44; 21:20–24), and this omission is a weak argument from silence for a narrative that ends well before the time of the temple’s destruction. (As Pervo rightly maintains, such an approach would require us to date James Joyce’s Ulysses before World War I, since it does not mention the war.)
At greater length than most, Richard Longenecker, known for his expertise in early Judaism as well as the NT, notes the following:
Some of these arguments are stronger than others, though I would argue that none of them is completely compelling. First, Luke’s portrayal of the influence of the Jewish elite within the narrative world need not reflect his own era; but even if it does, it does not necessarily support a pre-70 date. Judaism did apparently retain influence in some cities in Asia and presumably Greece after 70 (Rev 2:9; 3:9); further, Paul usually escapes Jewish accusations in Roman courts, which one could argue reflects the very post-70 setting that Longenecker doubts (though for other reasons I believe that this picture primarily reflects Luke’s historical information). Indeed, the very hostility of a number of Jewish opponents makes sense after 70 as an apologetic to Roman readers (though it was also historically true at an earlier period; see 1 Thess 2:14–16; 2 Cor 11:24; Gal 1:13; Phil 3:6).
Second, the association of Christianity with Judaism reflects historical reality, and it would profit Christians in some parts of the empire even when it might not help them in others; Revelation, in the 90s, probably offers the same association. By representing the Christian movement as a pacifist form of “true Judaism,” Luke could retain the most positive associations with Judaism while dissociating the movement from negative ones. (Luke also has salvation-historical interests relevant to teaching even a largely Gentile church, like the church in Philippi, about its heritage.)
Third, expectation of Roman justice might be difficult during Nero’s persecution but not necessarily afterward in stabler parts of the empire (e.g., Greece and Asia). Fourth, Acts may be accurate without being so early, especially if Luke did research and recorded it (Luke 1:1–4). Finally, Paul’s letters were not collected and circulated in his churches immediately after his death and probably not until some time after 70; postulating a date before the collection of Paul’s extant letters does not therefore require a pre-70 writing date. (For that matter, even if they were available, Luke, possessing other information, might not feel compelled to use them.) Luke would not have needed to depend on the letters at any date so long as he himself had known Paul (and hence lacked reason to depend on mere letters).
Other scholars point to the careful detail in some narratives as a sign of an early date, but such detail might mean only that Luke kept good notes (or, in some more skeptical scholars’ views, where extrinsic comparative data is lacking, had a vivid imagination). In favor of an early date, one might also argue that Luke speaks of the Beautiful Gate (Acts 3:2) and the synagogue of the Libertini (6:9) as if both remained standing; the present participles fit other continuing geographic realities (1:12). But they could be present simply in terms of the main verb (cf. 1:3, 6; cf. also the same passive participle in 8:6; 27:11; 28:24).
The pre-70 date is not the exclusive domain of “conservative” interpreters. Goodenough, who thinks Luke’s portrait of Paul largely fictitious, nevertheless dates the argument of Acts to the early 60s because Luke would need to argue this way for Paul’s greatness only in a period when there was a controversy about him. But the nature of Paul’s legacy continued to be disputed long after the 60s, and it would be no less possible, in any case, for Luke to use Paul as a ministry model because he was widely accepted by Luke’s time than it would be that he would idealize and defend him.
Some scholars have suggested a date after Paul’s death but before the fall of Jerusalem. T. W. Manson argued in 1944 that Luke distinguished Christianity from Judaism because most Gentiles in the empire associated Judaism with the revolt that began in 66; such a distinction, he thought, would prove less needful after 70 C.E.
One wonders, however, if Manson’s observation might not be stronger than the conclusion he draws from it here; given the special tax on Jews and lingering suspicions of them in the Flavian period (some of which Josephus writes to counter), Luke would still have reason to distinguish the two groups after 70. But in light of intensified demands for civic participation in the imperial cult in some Asian cities, probably presupposed in some passages of Revelation, the reverse temptation, to identify with Jews exempt from the cult, might prove more likely after 90. Manson’s suggestion could then point us toward a date anywhere between 66 and 90, less promising for narrowing the range of dating than Manson undoubtedly hoped. (Luke’s emphasis on the Christian movement’s continuity with its biblical heritage is even less helpful in pinpointing a date, since theological and apologetic reasons could account for this regardless of the date.)
Because Luke writes when Christians and the imperial authorities could have positive relations, some argue that the best range is between 70 and 90. Although this conclusion is probable, it should be frankly admitted that reconstructing the date from the setting of the implied audience is precarious, especially when conditions varied from one location to another even in the same period (see, e.g., Rev 2–3).
Some scholars insist that Acts contains too much legendary material to stem from apostolic times, but as I have argued in addressing historical questions, this premise is not only difficult to demonstrate but undoubtedly mistaken. Nevertheless, more defensible reasons than this exist for dating Acts after 70.
A stronger argument for a date after Paul’s death follows from the usual premise that Mark wrote before Luke. Papias reports that Mark wrote what he had heard from Peter; if this language suggests that their relation is past, it probably points to a date after Peter’s death. If Luke wrote after using Mark, he presumably would be writing after Peter’s death as well and presumably after Mark’s Gospel had begun circulating among churches in urban centers. One could respond by positing again a form of proto-Luke, but without Mark it would not even have the current Gospel’s structure. Lukan priority, though proposed, has so far attracted few advocates. One could date Mark before what Papias suggests or date Peter’s death before what early Christian tradition suggests, but then one is dispensing with the limited evidence we do have, in support of a hypothesis extrapolated from Acts’ ending, which can be explained in other ways. Luke’s dependence on Mark thus suggests a date after the deaths of Peter and Paul, that is, probably after the Neronian persecution.
At the same time, one should not use Luke’s dependence on Mark to press for a very specific range of dates in the post-70 era. Luke need only write at a time after Mark had become widely circulated (at a minimum, several years later), and one cannot easily assign Mark a date late in the first century. Mark had to write early enough that Matthew would use his work extensively well before Matthew became the favorite of the second-century church. Indeed, Luke’s dependence on Mark does not by itself require us to date Luke long after 70. Possible association of Mark with Roman persecution under Nero, and an apparently fresh memory of Claudius’s attempt to set himself up in the temple for worship, could support dating Mark before 70. The close connection between the temple’s demise and Jesus’s parousia also suggests a date within a few years of 70. Mark could easily have been available for Luke’s use in the early 70s.
Most scholars believe that Luke 21 (e.g., Luke 21:20; cf. 23:29–31) reflects the accomplished fall of Jerusalem and argue that Luke wrote Acts after completing his Gospel. Granted, most of the discrete elements in Luke 21 could date from before the war of 66–70; various Jewish figures predicted judgment on Jerusalem and its temple before its demise, and the language recycles Septuagintal descriptions of Jerusalem’s earlier sufferings. Certainly, Jerusalem’s judgment was in view as early as Q, almost certainly before 70 C.E. (Luke 11:50–51; Matt 23:35–36). The description does, however, more closely resemble the details of events in 66–70 than the description found in Mark (Mark 13:14–23). The wars and disturbances (Luke 21:9) might be those following Nero’s death, since 21:7 sounds less eschatological than the parallel texts in Matthew and Mark; certainly in 21:20, Luke clarifies Mark’s “abomination” (Mark 13:14) by inserting Jerusalem’s being surrounded. Granted, an eschatological section (Luke 21:25–28) seems to follow this, but the gap between Jerusalem’s fall and the end (and consequently the duration of the Gentile era in 21:24) remains indefinite.
One could argue that assumptions in such a description might seem inevitable, once Roman armies surrounded Jerusalem, for someone who accepted Jesus’s prophecies of judgment; hence one could date Luke to the period after 66 yet before 70. The mention of Jerusalemites being carried away captive among the nations (21:24), however, though echoing the OT, may suggest specific knowledge that the parousia did not occur when Jerusalem fell. In view of Luke’s use of earlier biblical language, this argument for dating Luke-Acts after 70 is not certain, but it is more secure than most of the other arguments.
C. S. C. Williams suggests that if Acts was written after 70, it (like Luke) would surely mention Jerusalem’s fall. But Jerusalem’s fall may be a critical presupposition of Luke’s plot development in Acts (culminating in the temple establishment’s rejection of Paul and his message of embracing Gentiles). Paul’s repudiation in the temple (Acts 21–22) parallels Jesus’s arrest for his act in the temple (although that arrest differed by its conspicuously private character, Luke 19:47–48; 22:2–6) and has the same function in the unified work. If Luke writes after 70, his concluding quotation from Isa 6 (Acts 28:26–27) seems particularly appropriate: Isaiah’s prophecy (Isa 6:9–10) stands in the context of Isaiah’s call (6:1–8) and the announcement of Israel’s judgment and devastation (6:11–13). This context might highlight two approaches, one leading through Jesus to the Gentile mission, the other leading to resistance against Rome without divine sanction, and the holy city’s devastation.
F. F. Bruce, who in earlier editions of his Acts commentary argued for a pre-70 date, shifted to a post-70 date in his final edition. While conceding that not all prophecies that come true are composed after the fact (e.g., those of the Dominican reformer Savonarola were not), he points to Luke (Luke 21:20) recasting Mark’s “abomination” (Mark 13:14 ) in terms of Jerusalem’s being surrounded by armies, arguing that the form in which Luke recasts prophecies of Jerusalem’s demise (Luke 19:41–44; 21:20–24; 23:28–31) “presupposes their fulfilment.”
Luke seems to understand Jesus’s prophecies of Jerusalem’s fall as akin to Jeremiah’s prophecies and views Jerusalem’s destruction against the historical template of the Babylonian captivity (a template widely shared among Jews and Christians alike). The terrible fate of Jerusalem stands behind not only Luke 21 but also 19:42–44, and probably also the warnings in 13:1–5 and perhaps also 11:50–51 with 13:34–35 (though Luke has not clarified this meaning as much as Matthew did). This commentary at some passages (e.g., Acts 22:22–23) will suggest that its shadow also stands behind Acts. It is therefore difficult to argue that Luke lacks interest in these events. He simply cannot narrate them as past because they remain future within his narrative world.
Further, would Luke relate a true story about a young man who betrayed a plot in Jerusalem when this story could endanger the young man (Acts 23:22)? Perhaps Luke knew that the young man left Jerusalem before Luke left for Rome with Paul; otherwise, however, the account makes better sense after 70 or at least after the flight of Jesus’s followers from Jerusalem, usually estimated to have been around 67.
A few scholars argue that Acts was written before the finished version of the Gospel and hence date the completed Gospel after 70 but Acts before Paul’s martyrdom. Such a view requires either an earlier proto-Luke Gospel (missing Luke 21 or at least the historically accurate divergences from Mark 13) or that Luke wrote the Gospel later, adding the connection between volumes (Acts 1:1–12, esp. 1:1) only to a later edition of Acts. But although one might add one’s introduction only to one’s finished work, it seems strange that Acts would sit for nearly a decade before being sent to the dedicatee. (Then again, Luke could have sent him the various editions, as ancient writers sometimes did.) The consistent length of the two volumes also tentatively argues against their finished composition at radically different times. Occam’s razor (the principle of logic that favors the simpler solution) also argues against it, though some have rightly warned that Occam’s razor is too dull for cases where evidence is too slender.
Most scholars today suggest dates between 70 and 85, with some as late as 90. This range of dates in the early Flavian period, before Josephus’s publications but when works similar to those of Josephus (and his competitors) were flourishing, is not unreasonable as a guess. Harnack originally suggested a composition date in the time of Titus or early in Domitian’s reign.
Many scholars who maintain that a companion of Paul wrote Acts nevertheless hold that he wrote after Jerusalem’s destruction, hence within this range. Fitzmyer is among the scholars who prefer a date in the 80s, although his explicit arguments specifically support only a post-70 date. Witherington thinks that Luke depends on fresh eyewitness sources but that Acts also reflects the benefit of sufficient hindsight, and hence he suggests a date in the 70s or early 80s. Dunn suggests a date in the 80s; although he believes a former companion of Paul wrote Acts, he dates it after the Gospel and the Gospel after Mark, which he dates in the late 60s or early 70s. He also believes that Acts reflects “concerns of the post-Pauline generation.”
If we date Mark before 70, however, and if Luke would have had early access to this Gospel in Greece or Rome, we need not date Luke’s Gospel much after 70; the church was not so large that a work such as Mark would necessarily take long to circulate, once it achieved some popularity. (Matthew, probably writing in Syria a few years after Luke, had access to many of the same sources.) Nor need we allow a great lapse of time between Luke’s Gospel and Acts, especially if Luke had been collecting notes for several years. Hengel and Schwemer argue that Luke mediates between Paul “and the supporters of the earliest [Jerusalem] community” (cf. Gal 2:9–12); this would hardly be necessary “long after 70.”
Theissen suggests that the historicizing language about Jerusalem’s fall in Luke 21 addresses the false eschatological expectations that surrounded Jerusalem’s fall (cf. Luke 21:8–9; Mark 13:14, 21). Thus he dates Luke’s Gospel to shortly after 70. Because he, curiously, reads Luke’s prologue as claiming to belong to the third generation after Jesus (in fact, it suggests instead that eyewitnesses remained available for Luke’s investigation), however, his “shortly” would be longer than ours: Theissen suggests that the Gospel should be dated in the 80s or 90s. Without this unusual reading of the prologue, however, Theissen’s own evidence could point us to a date in the early 70s, in the very wake of Jerusalem’s destruction.
Luke probably writes after Nero’s death and publishes shortly thereafter—after the war had begun and, in my view, probably after 70. This would make sense of the emphasis on Jewish (as opposed to Roman) rejection of the message; although Luke emphasizes continuity with Israel’s heritage, a distinction from current Judean politics would be helpful in a largely Gentile setting after 70. Granted, Luke’s notes were probably available much earlier, and the work probably went through various “editions” (originally of one or a few copies each) with the feedback of audiences (fitting the normal process of “publication”), some of them possibly before 70. Nevertheless, we lack evidence that compels us to accept a completed, published work before 70.
Luke favors more realized eschatology than some other early Christian authors, such as Paul, Matthew, or the author of Revelation (see , ). Nevertheless, pace Conzelmann’s view of Luke’s eschatology, Luke might believe that Jesus’s generation will see the end (Luke 21:31–36); one might speak of the generation remaining so long as some were still living (as some presumably were; cf. the beloved disciple in John 21:23; perhaps also Luke 9:27, but like Mark, Luke applies this saying to the transfiguration). Luke appears to expect the prophesied events fairly soon after the temple’s destruction (Luke 21:5–28, esp. 21:20–28), which would fit dates proposed for his work between 66 and 75 C.E. His approach appears different from that of Matthew (see the clear differentiation of questions in Matt 24:3), who may be writing later. Nevertheless, Luke’s “captive among the nations” (Luke 21:24) suggests a date after 70 or 73. A date in the mid-70s thus seems reasonable.
Given Luke’s emphasis on Jerusalem’s judgment, his portrayal of the large Jerusalem church vis-à-vis Paul’s mission, and various other elements treated in the commentary, I am inclined to think that the destruction of 70 is past but well remembered. More important for dating, Luke’s strong apologetic for Paul (see , ) makes most sense when his memory burned most brightly yet remained most susceptible to challenge—that is, not many years after his death or after the Neronian persecution, with which it was likely connected. I thus view Haenchen’s estimate that Luke prepared Acts about 75 C.E. as a reasonable one, give or take a few years.
Those who date Acts after 90 usually do so because they argue for dependence on Josephus. Thus, for example, Acts’ depiction of Paul employs some apologetic motifs also found in Josephus—though these are conventional enough apologetic motifs to undercut any need for dependence. It would be difficult, however, for Luke to have made direct use of Josephus; Josephus’s Antiquities probably was issued about 93/94, and the copies of its twenty books would have been quite expensive and not readily available soon after their publication to nonelite persons.
In the final analysis, it is highly unlikely that Luke depends on Josephus. The strongest argument for dependence is Luke’s citation of Theudas and Judas. Yet it seems unlikely that Luke would have read Jos. Ant. 18–20, used barely any of it, and then gotten wrong the one point most likely drawn from it, if he was dependent at all. The other matters that he could have taken from Josephus were widely known. Josephus did not compose his reports of Theudas or Judas from thin air, and oral or written reports about these leaders would have been available to Luke as well as to Josephus. Josephus completed his Antiquities no earlier than 93 C.E., but Luke’s sometimes positive portrayal of Roman administrators probably would make less sense in the later years of Domitian’s reign.
Bornkamm dates Acts in the 90s because it is later than Luke, which is later than Matthew and Mark, which, he avers, are both post-70. But even if Mark is post-70 and Luke is the latest of the Synoptics (I prefer a pre-70 date for Mark and think that Matthew, like John, addresses a Jewish Christian situation slightly later than Luke), this sequence provides us no exact chronology for pushing Luke-Acts so late. If, as many commentators (including myself) believe, Luke was a companion of Paul, a date before 90 is more likely (though a post-90 date would not be rendered impossible).
A smaller minority of scholars date Acts to the second century, sometimes as late as the mid-second century. Some do so because they believe that Luke depends on, or was writing a preface for, Paul’s collected letters, but this premise is unlikely and would not, in any case, necessarily demand a second-century date (see above). J. C. O’Neill thinks that Luke’s presuppositions place him in the generation of Justin Martyr—a view that I believe either underestimates the continuity with Judaism in Luke (see of this introduction) or overestimates it in Justin. John Knox thinks that Marcion shortened the earliest form of Luke whereas our present Luke-Acts (which he dates after Marcion) expanded it.
Such arguments have not, however, persuaded many, and such a late dating hardly seems likely. In addition to the relevance of arguments already lodged against a date in the 90s above, one would hardly expect a mid-second-century work to recollect so many historical details (such as first-century offices in various locations) so accurately, especially in a period that produced many novelistic acts rather than historical monographs. C. K. Barrett observes that Acts reflects the character of the provinces in the first rather than the second century; the early second-century church may not have known what to make of Acts, but it did know the work. Luke’s very probable independence from Matthew is also difficult to explain with a second-century date, when Matthew flourished as the most popular Gospel. Luke’s lack of emphasis on a number of Pauline theological themes (see the discussion in above) that must have become widely known once Christians began collecting Paul’s letters (cf. allusions in 1 Clement; 2 Pet 3:15–16), his pre-Ignatian ecclesiology, his lack of interest in Gnosticism, and other factors also militate against a later date.
In the first drafts of this commentary, I ended my cursory treatment of second-century proposals here, focusing on the primary views of pre-70, 70–90, and the 90s, since very few scholars today date Acts in the second century. Since that time, however, two noted Acts scholars (Richard Pervo and Joseph Tyson) have published monographs, roughly simultaneously, arguing for a second-century date. They are themselves aware that they are defending a relatively rare position; Tyson considers 70–100 the “consensus dating” of Acts scholars, complaining that few scholars today bother to question or defend it “except among those who would favor a date in the sixties.” Although this view attributing Acts to the second century remains relatively rare and just a few years ago could have been viewed as relatively idiosyncratic outside this circle of scholars, a view is not rendered incorrect simply by virtue of being rare, and hence it merits at least some concise engagement here.
Some scholars draw attention to patristic evidence and date Acts no earlier than its first attestation in those sources. Admittedly, extant sources before Irenaeus (and, more fully, John Chrysostom) make little use of Acts; Barrett attributes Acts’ neglect in the second century to its apparent irrelevance until the Marcionite controversy. Tyson, by contrast, believes that it was not simply irrelevant until Marcion but became relevant to that controversy because it was written directly to meet the Marcionite challenge. Reconstructing dates based purely on reconstructed external situations is usually hazardous, although understandably often hazarded; as noted above, few have been persuaded by the anti-Marcionite position, and (I believe) with substantive reason. (It is interesting that some patristic writers, such as Tertullian, seem to have almost avoided Acts in their conflicts with Marcionites.) Because I have mentioned the Marcionite case above, I will focus here on Pervo’s arguments based on patristic evidence.
The earliest argued piece of patristic evidence that may attest knowledge of Acts comes from Poly. Phil. 1.2:
Phil
Acts 2:24
ὃν ἤγειρεν ὁ θεός
ὃν ὁ θεὸς ἀνέστησεν
λύσας τὰς ὠδῖνας τοῦ ᾅδου
λύσας τὰς ὠδῖνας τοῦ θανάτου
The language is unusual and not found in this form before Polycarp except in Acts, another early Christian document. Unless one supposes that Luke and Polycarp draw on a common tradition here (which is possible but, against the principle of Occam’s razor, introduces an unnecessary level of complexity), we have a likely allusion to the language of Acts in the early second century. Pervo accepts this as a possible but not certain allusion; because he will date Acts about 115, an allusion in Polycarp would not necessarily undercut his proposed date. Beyond these connections, at a very early date 1 Clem. 17.1 and Polycarp Phil. 6.3 apparently betray knowledge of Acts 7:52, which also mentions prophets foretelling Jesus’s first “coming” (using ἔλευσις, a term that is, apart from minor textual variants, a NT hapax). It would be difficult to demand patristic sources earlier than these, since these are among the earliest.
Perhaps also relevant, because the longer ending of Mark betrays familiarity with Luke’s Gospel, it seems fair to look for allusions to ideas in Acts as well. Of these, several are probable or possible (cf. esp. Mark 16:20 with Acts 14:3; perhaps Mark 16:18a with Acts 28:3–6; perhaps Mark 16:19 with Acts 1:11). Some date this ending to the period 100–140 C.E., which would provide quite early evidence for allusions to Acts. Certainly the longer ending of Mark circulated by the late second century, used in Tatian’s harmony of the Gospels (ca. 170) and attributed by Irenaeus to Mark (ca. 180). The terminus a quo for this addendum to Mark is, however, uncertain, making the allusions less decisive than certainly early sources such as Clement and Polycarp.
Some scholars, however, who favor a second-century date are more impressed by lack of evidence, which they interpret as negative evidence. Pervo argues, for example, that Eusebius cites Papias regarding only Mark and Matthew, not Luke or John. Given our small corpus of Papian material, however, this could well be coincidence and is surely an argument from silence. Pervo contends that Papias either did not know Acts or disagreed with it, “for he twice contradicts it.” The “contradictions” do not, however, amount to much, by the standards of the Apostolic Fathers. First, Papias frg. 3 differs from Acts and Matthew regarding Judas’s death. Yet the early fathers certainly knew Matthew, whom Papias’s account also contradicts; Papias simply had more traditions from which to choose than the two preserved in the NT.
The second “contradiction” proves even less compelling. Papias frg. 11.3 claims that the Jews killed both James and John the evangelist—fitting Mark 10:35–40, Pervo reasons, but not Acts 12:2. Yet that Jewish people authored the persecution fits Acts 12:2 more clearly than it fits Mark 10:35–40, and Papias, who is merely being summarized, does not specify that the two died together. Arguing from Papias’s failure to cite Acts is thus an argument from silence, since Papias was heir to multiple traditions of the apostles and preferred oral to written sources (frg. 3.4; 7.2). Moreover, Papias’s knowledge of Justus Barsabbas (Acts 1:23; Euseb. H.E. 3.39.9–10) probably either reflects Acts or, if derived from Philip’s daughters, as Eusebius says, provides independent support for tradition reported in Acts.
Pervo also points out that Hegesippus (ca. 110–80) reflects no knowledge of Acts though he is interested in the early bishops of locations (for which Eusebius cites him) and was in Rome about 160–170. Against making so much of the silence here, we have primarily Eusebius’s citations of Hegesippus, and Eusebius need not rely on Hegesippus to provide information that he already has in Acts! Even for some time afterward, church fathers were more interested in the Gospels, especially Matthew, and Paul’s letters than in Acts. This interest indicates a preference for use, however, rather than identifying exhaustive knowledge of relevant works.
Pervo might be correct about Justin’s failure to cite Acts, but the evidence here is ambiguous. Haenchen finds possible echoes of Acts in Justin 1 Apol. 39.3 (recalling the idea, though not the words, of Acts 4:13), 49.5 (alluding to Acts 13:48), and 50.12 (using Luke 23:49; 24:25, 44–45 and wording from Acts 1:8). Acts is not Justin’s favorite book, but he may know of its content, directly or indirectly.
Pervo concludes that Irenaeus, writing about 180, “remains as the earliest certain witness to the existence of Acts.” Pervo recognizes that Luke’s Gospel is attested as early as Marcion and hence was in use by 110 or 120, but he separates Acts from Luke and thinks it unclear that Marcion knew Acts. The vast majority of scholars today view Luke-Acts as a unity and would view evidence for the Gospel’s date as an indicator of a near date for its second volume (although Acts can be later).
But is dating a document no earlier than its first clear attestation in later sources a fair criterion, in any case? Should we date Luke’s Gospel decades before Acts and Matthew, the patristic church’s favorite Gospel, before Mark? Taking the latest possible date of a document as its most probable actual date would radically reshift the dates assigned to many of our ancient sources. Irenaeus hardly adopted precisely these four Gospels randomly, especially given his emphasis on church tradition; and is it an accident that he chose the four Gospels more reflective of first-century Judean traditions than our other extant gospels (the “apocryphal” gospels and gnostic sayings treatises)?
In my view, the correspondences with the external Roman world of the first century (such as the correct titles of officials such as Gallio or Festus in the correct years), especially in the Pauline portions of Acts, would be highly unlikely for a second-century author unless Luke had available some fairly accurate written sources (more than I have argued as likely). Certainly a novelist (in Pervo’s view) would not have gone to the trouble of locating these details. Second-century apocryphal acts lack the connections we have in Luke’s Acts, a historical monograph; the genre differs.
Moreover, patristic use (and Irenaeus’s confidence of unanimity among the “orthodox”) makes it difficult to date any of the four Gospels as late as the mid-second century (the Fourth Gospel, the best candidate, cannot be so dated because a manuscript fragment dated earlier than this period exists). But Acts certainly comes from the same author, at least, as Luke’s Gospel, whatever one’s view of their unity.
More concretely, Pervo cites Acts’ dependence on Paul’s letters, but as we have noted, most scholars reject this dependence (and above has given reasons for favoring this rejection). He also cites dependence on Josephus, which is also addressed above. Less concretely, he cites anachronisms that he believes require a later date for Acts, for example, Peter supposedly speaking of Jerusalem’s residents and telling them of a field called something in their language. Most readers, however, recognize these words as simply the narrator’s digression (see comment on ).
Pervo compares Acts with writings from “the first third of the second century.” Yet one also can compare Luke with first-century sources, though extant first-century sources are fewer in number, and explain second-century parallels as the common growth of a Christian language stemming from earlier usage. Moreover, Pervo’s date for Acts would place it fairly close to Pliny the Younger’s letter to Trajan (Ep. 10.96), which portrays a state of affairs quite different from the one to which Acts responds.
Because I believe that the arguments are compelling in favor of the author having been one of Paul’s traveling companions (probably for a brief time as early as ca. 48 C.E.), I would restrict any date estimate I would offer to within the probable lifetime of such a companion. Such a limitation does not necessarily preclude any first-century options, since we know neither the author’s age when he first traveled with Paul nor the extent of his longevity.
The date of Acts is uncertain, but my best guesses, for reasons that follow, are in the early 70s, with dates in the 80s and 60s still plausible, and a date in the 90s not impossible. The arguments limiting the range between 70 and 90 (the majority view) seem to me stronger than the alternatives; because I also favor the tradition of Lukan authorship and because I find a strong apologetic for Paul engaging a range of concrete accusations about his involvement in specific local riots (see , ), accusations that would seem less relevant in later years, I am inclined to suggest a date in the early 70s rather than a later time. Luke the apologist would hardly invent a history of local riots surrounding Paul, yet unable to deny them, he must explain them, at a time when the local memories of such riots remained alive and Paul’s legacy remained contested (see more fully , ; , ). Yet though I believe this argument strongly favors a date fairly soon after Paul’s ministry, to an extent any suggested date is merely an educated guess, and as I have noted one need not date Acts so early to maintain eyewitness authorship. If Luke was about thirty (he could have been younger or older; if one accepts the tradition that he was a physician, he was probably at least twenty) when he traveled with Paul about 50 C.E. (Acts 16:10), he would be in his early fifties by the early 70s and in his early sixties by the early 80s (or in his seventies in the 90s if one dates Acts so late). If the author was as young as twenty ca. 50 C.E., he would be only seventy years old ca. 100 C.E. Paul’s memory remains fresh, Diaspora churches are establishing their identity distinct from local synagogues, and the troubles of the later Domitianic era have not yet arisen.
That Luke names Theophilus, a person of status, and provides an optimistic legal apologetic might suggest that his audience is facing neither Nero’s persecution (albeit perhaps relevant only in Rome) nor the local persecutions that occurred in some cities under Domitian. Even under Nero (cf. Tac. Ann. 15.44), it is not clear that the Corinthian Christians, though Corinth might emulate Rome, experienced direct persecution. By contrast, Christians in Philippi may well have (Phil 1:28–30); in any case, Christians in Roman colonies would certainly have known of persecution in Rome and have had reason for concern. If Philippi is a major center of Luke’s audience, this might suggest a date after Nero, though his policies could appear more rational to distant colonies than in the capital (by the time of his demise, Nero was discredited at least among the Roman aristocracy). Further, as best one can reconstruct the implied audience from the data available, Theophilus may belong to a church that knows and appreciates the Pauline mission, but he and many others may have become believers since Paul’s own work there.
All such dating, however, is at best educated guesswork, given the inadequacy of our data; we cannot even reconstruct the period on the basis of the setting without also knowing the location. Nevertheless, given my arguments for authorship, a first-century date, at least, should be regarded as secure. This commentary assumes a post-70 date (within, however, living memory of Paul). If this premise is mistaken, it will bring into question some of my interpretive judgments at key points (such as Paul’s speech in Acts 22:3–21 as constituting a final public invitation to the people of Jerusalem to turn from their anti-Roman course before Jerusalem’s destruction). Such a misjudgment would, however, affect only a relatively small portion (less than 1 percent) of the commentary proper, and I still think this dating likelier than not.
For older studies of authorship, date, and provenance, see Mattill and Mattill, Bibliography, 147–52, §§1940–2028.
If Acts merely pretends to be by a companion of Paul, the dramatic date of the narrator’s world could differ from the actual date of composition (as probably in the case of Tacitus’s Dialogus de oratoribus; see Kennedy, “Survey of Rhetoric,” 33); but as I argue elsewhere, this is unlikely.
Fitzmyer, Luke, 1:53–57 (dating Luke to ca. 80–85 and Acts afterward). (Pervo, Dating Acts, 24, meanwhile, dates Acts long after Luke, while allowing the same author.)
Brownson, “Introduction to Anabasis,” 234–35, allowing that it might have been written shortly after his return to Greece (394 B.C.E.) but not published till ca. 370 B.C.E. (since his Hellenica in 380 B.C.E. must refer to another’s work on the same subject).
Luke may have proved one of the compelling Diaspora Christian voices in his support if the dispute was serious enough to threaten it (cf. 2 Tim 1:8, 15–16; earlier, Phil 1:14–17).
See the extensive survey of views in Pervo, Dating Acts, 359–63 (following Hemer on some of them), which lists roughly thirty-one in the 60s, forty-eight in the 70s–80s, twenty in the 90s, and eleven for ca. 100 or any decade thereafter. Hemer focused on earlier sources and Pervo on later ones, though later dates remain fewer even in Pervo’s list. More conservative scholars with early dates could be added, but undoubtedly, the list of those with later dates could also be amplified. I use Pervo’s survey merely to illustrate the range of views, with roughly 10 percent after 100, almost twice that number (18 percent) for the 90s, almost three times that number (28 percent for the 60s) and more than four times that number (nearly 44 percent) for the 70s–80s. Tyson, “Dates,” and idem, Marcion, ix, uses three categories: early (60s); middle (70–100); and late (ca. 110–20); with most scholars, I fall in his “intermediate” range. Others prefer the three categories of early (60s), middle (70–90), and late (95–100, based on use of Josephus; e.g., Gangel, Acts, 2).
If one argues instead for dependence on Paul’s collected letters, one must ask the source of Luke’s other early sources, many unusual deviations from the reports of those letters and silence about those letters (and key elements in them), and (if one thinks of a full collection) why the strong divergence from the itinerary in the Pastorals’ traditions. Most works after Paul’s letters do not hew to Pauline chronology and so forth the way that Acts does. See discussion in , , of this introduction.
See Keener, “Apologetic.” From my examination of Luke’s apologetic arguments in ibid., I would argue that the memories of civic disorder surrounding Paul’s ministry are far too concrete and localized to fit a setting half a century after Paul’s death (pace the milieu that Pervo, Mystery, 7, envisions for Acts).
This has been argued perhaps most thoroughly and convincingly recently by Mittelstaedt, Historiker (who dates the bulk of the work to Paul’s Caesarean custody in 57–59 C.E.).
Harnack, Acts, 293 (although he ultimately maintains the likelihood of a later date, 296–97).
Goodenough (“Perspective of Acts,” 58) dates it to the early 60s, when controversies about Paul remained alive (cf. Phil 1:16–18), though he ascribes little historical value to much of Acts (51–59).
Supporting the early date, see, e.g., C. Williams, Acts, 18–19; Clark, Acts, 390; Munck, Acts, liv; Mattill, “Date and Purpose”; Blaiklock, Acts, 16; Bruce, Acts1, 11, 481; Finegan, Apostles, 6; Arrington, Acts, xxxiv–xxxvi (tentatively); Hemer, Acts in History, 365–410; Robinson, Trust, 72–74; idem, Redating, 90; Le Cornu, Acts, xxviii–xxix; Gangel, Acts, 2; Bock, Acts, 25–27 (tentatively); Freedman and MacAdam, “Witness”; Mauck, Trial, 41–45; Delebecque, Actes (for Luke’s first draft, essentially our Alexandrian text); particularly extensively, Mittelstaedt, Historiker (for composition before Paul’s death, 219–21); cf. Moberly, “Planned and Shaped” (mostly completed before the story ended).
See Euseb. H.E. 2.22.1–8 (Martin, Acts, 319).
Fitzmyer, Acts, 52. For the appropriateness of Mark ending at 16:8, see Boomershine and Bartholomew, “Techniques”; Hooker, Message, 119; discussion at Keener, Matthew, 702n295. Abrupt endings were common, whether for literary works (e.g., Ps.-Philo, Biblical Antiquities; Thucyd. 8.109.1; Val. Max. 9.15.ext. 2; Lucan C.W. 10.542–46; Plut. Fame Ath. 8, Mor. 351B; Men. Rhet. 1.3, 367.8) or for speeches (Isaeus Pyrr. 80); see further discussion at Acts 28:30–31. Farmer, Verses, however, argues from internal (79–103) and external (3–75) evidence that the discussion for the inclusion of Mark 16:9–20 should remain open.
Dion. Hal. Thuc. 12 complains that Thucydides promised in his introduction to recount the entire war (he actually makes the promise in Thucyd. 5.26).
See Whittaker, “Introduction,” xii–xv.
See arguments against the early date in Witherington, Acts, 807. Most of these counterarguments are reasonable except the claim that Luke 1:1–4 and Acts 1:1 suggest “some time and distance” from the events; they could suggest distance from Jesus’s ministry without suggesting distance from Paul’s. In any case, for whatever it is worth, most scholars do place Acts after Paul’s death (e.g., Conzelmann, “Luke’s Place,” 299).
See Witherington, Acts, 61; cf. Trompf, “Death of Paul” (emphasizing Luke’s divine-retribution theme). Whatever the reasons (possibly his own failing health or death), Livy’s failure to narrate Augustus’s death does not mean that Augustus did not predecease Livy, who outlived him by three years (Laistner, Historians, 80).
Witherington, Acts, 807.
A view noted in Bruce, Acts3, 12–13n6.
So ibid., 13–14 (doubting the relevance of 1 Tim 1:3).
The last volume in Thucydides’s work has been thought unfinished, though this is debated (Wade-Gery, “Thucydides”).
This line of argument is most commonly associated with Robinson, Redating; cf. also Mattill, “Date and Purpose,” 341–48; Igenoza, “Hypothesis” (noting that Africans, in view of their traditional context, find successful predictive prophecy less troubling than Westerners).
Pervo, Dating Acts, 337.
See Longenecker, Acts, 31–34; several of these arguments also appear in Bruce, Acts1, 11; for the favorable Roman setting, see esp. Bruce, Commentary, 22–23; for issues current in view of Paul’s approaching trial, see Vine, “Purpose and Date.” Bruce, Acts1, 11, claims that Jewish-Gentile relations, a focus of Acts, were less of an issue after 70 C.E., but aside from the need of Acts to record foundational history, the Johannine literature shows that the issue continued to be a live one well into the Domitianic era.
Kistemaker, Acts, 23, following Parker, “Treatise,” doubts that Acts could end with such optimism after Nero’s persecution. But Achaia probably escaped this persecution, and we cannot say whether Macedonia had martyrs at that time; Luke’s “optimism” may also relate to his apologetic.
In a later period, see esp. the Sardis synagogue (e.g., Hanfmann, “Campaign”; Mitten, “Sardis,” 61–62; Bonz, “Approaches”; Botermann, “Synagoge”; esp. Seager and Kraabel, “Synagogue”; Trebilco, Communities, 37–54).
The lampstands, symbols of the churches (Rev 1:20), traditionally symbolized Judaism (CIJ 2:117, §890; 2:128, §910; 2:131, §918; 2:137, §932; 2:142, §943; 2:147, §956; 2:149, §961; 2:165, §980; 2:234–35, §1197; 2:235, §1198); in the Diaspora (especially Asia), see CIJ 2:12, §743; 2:32, §771; 2:38–39, §781; 2:40, §783; 2:42, §785; 2:43, §787; 2:43–44, §788; 2:45, §790; 2:46, §792; 2:47, §794; 2:50, §798; 2:52, §800; 2:53, §801; 2:94–95, §855; 2:100, §862; 2:108, §873); for scores of other references, see Keener, John, 133–34.
If Luke ever visited Philippi again, he would at least know Paul’s letter to the Philippians, and given his likely company in Caesarea and Rome, he was, presumably, with Paul when he wrote it. But he lacked reason to show much interest in Paul’s letters, which become paramount only after the memory of Paul and his direct teaching to churches had faded.
Goodenough, “Perspective of Acts,” 58.
Manson’s view is summarized in C. Williams, Acts, 13–14; cf. perhaps similarly Marshall, Acts, 48 (“towards AD 70”). Usually, however, scholars date Acts either before 64 or after 70 (e.g., Kent, Jerusalem to Rome, 15–16).
E.g., Keener, Revelation, 36–38; Cuss, Cult; Price, “Man and God.”
Witherington, Acts, 793, following Walaskay, Came to Rome, 64.
Windisch, “Tradition,” 310.
A number of “conservative” scholars (including Witherington, Acts, 61; D. Williams, Acts, 12–13; Bruce, Acts3, in contrast to his earlier work) date Acts after 70. I note this to respond in advance to the suspicion that dating a book a decade later than some others might bring into question the orthodoxy of one’s argument (cf., e.g., Doriani, “Review,” 35). I am hoping that those who date the book just a decade earlier (or later) will not ignore my entire commentary on that account!
Brosend, “Means,” 358 (using Papias frg. 3.15). On Peter’s martyrdom, see, e.g., John 21:18–19; 1 Clem. 5.4; later, Tert. Scorp. 15; Euseb. H.E. 2.25.5–8; 3.1. Even were we to think that Mark writes under Peter’s tutelage, it would have likely been toward the end of Peter’s life in Rome (1 Pet 5:13; cf. Col 4:10; Phlm 24). Of course, not all scholars accept Markan priority (Farmer, Synoptic Problem; cf. Longstaff, Conflation, 218; Murray, “Conflator”; for proto-Matthew, see Lowe and Flusser, “Theory”), though not all proposed solutions for the “minor agreements” between Matthew and Luke challenge that hypothesis (see, e.g., Friedrichsen, “Agreements”). I discuss these issues more fully in Keener, Matthew, 8–10 (favoring Markan priority).
For something like Lukan priority, see Lindsey, Jesus, 84; Young, Parables, 129–63.
E.g., Windisch, “Tradition,” 310; van der Waal, “Temple in Luke”; Goulder, Type and History, 112–15; Tiede, Prophecy, passim; Gaventa, Acts, 51; Langner, Hechos, 35, 40–41.
See, e.g., Test. Mos. 6:8–9; Jesus son of Ananias in Jos. War 6.300–309; further sources and discussion in Keener, Matthew, 561. Cf. the difficulty of knowing how Tacitus treats the Jewish War in view of this section of his work being lost (Feldherr, “Tacitus’ Jews,” 316).
Cf. Dodd, “Fall”; Robinson, Redating, 13–30; Mattill, “Date and Purpose,” 341–48; see now esp. Mittelstaedt, Historiker. Aune, Prophecy, 396–97n39 (to p. 176), accepts a post-70 date for Luke but remarks at how little Jesus’s prophecies have been transformed to suit the new setting. Robinson, Redating, 28, argues that even Jerusalem’s being surrounded by armies (apparently Lukan redaction) is not a prophecy after the event.
Cf. Theissen, Gospels, 203–34, esp. 220–21, 230–32, for a date for Q in the 40s; pace Mack, Myth, 84. Whether or not the correct date is the 40s, I believe that Q predates Mark (explaining several of Matthew’s and Luke’s agreements against Mark as simple dependence on Q where Mark abbreviated it).
Nor does Luke make the issue of the second coming explicit in Luke 21:7 (though paraphrasing Mark 13:4, in a context anticipating both) as Matthew does in Matt 24:3.
C. Williams, Acts, 15. Already in 1876, Abbott, Acts, 14, dated Acts before 70 because it does not mention Jerusalem’s destruction. For data on the war of 66–70, see, e.g., Cornfeld, Josephus; Mendels, “Revolt” (and bibliography there).
Of course, Luke could presuppose his readers’ knowledge of the event from Luke 21, even if not from extratextual knowledge.
Perhaps the fall of Jerusalem even provided Luke an additional incentive to reiterate Paul’s appeal to a different path of relations with Gentiles, the rejection of which had now led to tragedy.
Cf. Taylor, “Jerusalem Temple,” who contrasts post-70 desires to rebuild the temple with Luke’s summons to seek the restoration of the divine presence through turning to Christ.
Bruce, Acts3, 16 (see also 17). The description in Luke 19:41–42 fits both OT prophecies and what one expects generally at cities’ fall (Aune, Prophecy, 176, citing B. Reicke and others), but Aune also thinks (176) that Luke applied it to Jerusalem’s destruction. Luke 13:35 may also be relevant (Matt 23:38, at least, clearly refers to the temple, in context; Luke is less explicit about the interpretation, but it is probable).
Cf. Rome as Babylon (conqueror of Israel) in 1 Pet 5:13; Rev 14:8; Sib. Or. 5.143, 159–61 (though it refers back to literal Babylon in 5.434–46); perhaps Bar 1:1; 2 Bar. 67:7; 77:12–79; 80:4; 4 Ezra 3:1, 28, 31; 4Q163 6–7 II, 4–5; frg. 8–10. Many in the Roman Empire believed that literal Babylon was now ruined (Pliny E. N.H. 6.30.121–22), though others differed (Philost. Vit. Apoll. 1.25), and “Babylon” could still be employed figuratively for Parthia (Lucan C.W. 1.10). Among the rabbis, cf. the four kingdoms (e.g., Sipre Deut. 317.4.2; 320.2.3; cf. Dan 2:37–45; 7:3–8; 2 Bar. 39:7; Sib. Or. 8.6–11).
It may also stand behind Luke 19:45–46 and 23:45 (see, e.g., comment in Keener, John, 524–27; idem, Matthew, 686–87; Kingsbury, Christology, 132; Matera, Kingship, 139).
C. Williams, Acts, 14–15. See also discussion in Bouwman, “‘Livre’ et date.”
For a consensus date in the 70s–80s, see, e.g., Zwiep, Ascension, 168; Puskas and Crump, Introduction, 149; cf. Chance, Acts, 5. Powell, Acts, 37, suggests that “the great majority of scholars date Acts somewhere around A.D. 80–85, with the admission that this cannot be established with precision.”
Harnack, Acts, 290–97 (sometime after 78 C.E.; he does allow [296–97], however, as early as the beginning of the 70s).
E.g., Faw, Acts, 19; Ehrhardt, Acts, 4; Polhill, Acts, 27–31; Jervell, Apostelgeschichte, 86; and scholars listed below.
Fitzmyer, Acts, 54–55 (adding in favor of his specific date in the 80s only the observation that many opt for 80–85, and he finds no reason to reject that). A date in the 80s is most common (e.g., Juel, Promise, 7; Sterling, Historiography, 329–30), sometimes alongside an insistence on trustworthy sources for the 40s and 50s (Martin, Foundations, 66).
Witherington, Acts, 165–72. Bruce offers the same suggested range (Acts3, 18). Hengel, Peter, 91, suggests 75 to 85.
Dunn, Acts, xi. Wayne Meeks and others think that Luke fictitiously raises Paul’s status to fit a later generation, but Paul’s own letters show sufficient contact with the elite to question this premise (e.g., Rom 16:23; see Gill, “Élites,” 105–18, esp. 109–12, 117–18).
Noting that he fails to distinguish as clearly as his successors, especially Matthew, between the temple’s desolation and the Son of Man’s return. A dating of Mark to ca. 64 (Bruce, “Date of Mark,” 78) or around the outbreak of the war (Anderson, Mark, 26) allows sufficient time, though we cannot know Mark’s date with certainty. Because Mark seems to expect the Romans to erect an image of the divine emperor on the site of the temple, Collins, Mark, 14, dates the work before 70 (albeit just barely before 70).
Hengel and Schwemer, Between Damascus and Antioch, 10, adding, “At the time Luke wrote, it was still well known who Peter, John, James, Stephen, Paul, Philip, Barnabas or even Mark had really been. Around 110 or 120 this was no longer the case in the same way.”
Theissen, Gospels, 280.
Haenchen, Acts, 86 (this is his date for preparing Acts, not necessarily its publication or circulation).
E.g., Pervo, “Dating Acts”; idem, Dating Acts, 149–99 (esp. 152–60, on Theudas and Judas); idem, Mystery, 4 (a “very strong probability”); idem, Acts, 12. Against this position, see, e.g., Fitzmyer, Acts, 52; Barrett, Acts, 2:xliii; Ehrhardt, Acts, 2–4; also Conzelmann, Acts, xxxiii (though he dates Acts to 80–100).
Gnuse, “Vita apologetica,” noting their shared experiences of education, shipwreck, and dream-revelation. Yet Paul’s own letters do imply education and clearly speak of shipwrecks (2 Cor 11:25) and revelations (2 Cor 12:1–7)!
Any such details would count for honor in an apologetic biography insofar as they were true. Thus, e.g., Paul (Phil 3:5) and Josephus are both Pharisees; but only Josephus is a priest (Paul was from a different tribe; Rom 11:1; Phil 3:5). By 90, Paul’s legacy was probably also more secure, requiring less apologetic (1 Clem. 5.5; 47.1; Ign. Eph. 12.2; Rom. 4.3; Poly. Phil. 3.2; 9.1; 11.2–3; Papias 28.1).
With, e.g., Wandrey, “Iosephus.”
Pervo concedes that a mistaken citation of Theudas, who acted more than a decade after the speech’s “dramatic date,” could have occurred as early as the 60s (Dating Acts, 310).
With C. Williams, Acts, 19–22; Bruce, Acts1, 24; Packer, Acts, 13, 48. Luke often conflicts with Josephus where they address the same topic (Bruce, Acts1, 25; Hemer, Acts in History, 95; Gempf, “Speaking,” 288).
On probable Christian suffering in at least some locations during Domitian’s reign, see, e.g., Söding, “Widerspruch”; Jones, Empire, 323–24; Hemer, Letters, 7–11; Keener, John, 178–79; cf. Dio Cass. 67.14.1–2; Euseb. H.E. 3.17–18. Contrasting Luke’s vision of the empire with John’s, cf. Kim, Caesar, 180–90.
Bornkamm, Paul, xx.
D. Williams, Acts, 12, notes against a late date that Luke’s Christianity remains spontaneous and Spirit-led, not very institutional. This may be right, but it is not clear that the church had become stratified everywhere by the 90s; Johannine literature betrays no sign of such stratification (though some think that Diotrephes in 3 John 9 claimed a hierarchical role; see the summary of this position in Brown, Epistles, 733–34, although Brown disagrees with it), in contrast to Ignatius a few decades later.
Aged people could write memoirs with accuracy; see, e.g., Seneca the Elder (esp. Controv. 1. pref. 2–4; 9. pref. 1); for other elders with sharp minds, see, e.g., Cic. Brut. 64.229; Val. Max. 8.7.1; 8.7.ext. 5, 9–13; Philost. Vit. soph. 1.9.494; 2.21.604; Eunapius Lives 482 (also Keener, John, 102–3, for numerous examples). My most regular correspondent, and one of the sharpest, at the time of this book’s editing is in her nineties.
E.g., Townsend, “Date,” 58; Pervo, “Dating Acts”; Price, Son of Man, 33; Parsons, Acts, 3, 17 (ca. 110 C.E.); Walker, “Portrayal” (anti-Marcionite); for a post-Marcionite date, Tyson, “Date of Acts” (though on some approaches to Marcion this thesis might date Acts close to a period when Irenaeus, writing perhaps three decades after Marcion, was defending the four Gospels!).
O’Neill, Theology, 175. Justin opposed Marcionite teaching, although this part of his polemical work is now lost; see Efroymson, “Connection,” 105; Chadwick, “Defence,” 290; cf. Stylianopoulos, Justin, 20–32. Justin allows for more continuity with the Torah than in Barnabas or Marcion (see Osborn, Justin, 5, 160–61), but by prioritizing some parts of the law over others (Stylianopoulos, Justin, 51–52, 58–68, 74) and presenting Christ as a new law (Dial. 11).
Knox, “Reflections,” 112. Because he believes that Marcion’s date is unknown, Knox requires the finished product of Luke-Acts to be dated no later than 125 C.E. (“Acts and Corpus,” 286). Yet the “formidable” nineteenth-century history of interpretation that Knox (followed by Tyson and others) cites in support of his approach to Marcion and Luke-Acts may depend on Knox’s misreading of the sources (see Roth, “Marcion’s Gospel”).
See the summary in Bruce, Acts3, 18–19n21. The nearly mid-second-century position is rare enough for Powell, Acts, 36, to attribute it to “an occasional scholar.” For one careful yet firm response to the thesis that Luke expands Marcion’s Gospel, see Hays, “Marcion.” But although the second-century view remains a distinctly minority position, it has been growing in some scholarly circles especially since recent arguments by Pervo and Tyson.
In contrast to Roman historians, Luke had no annals and perhaps, in Acts, no prior written sources from which to draw.
Barrett, Acts, 30–48, esp. 48; cf. also lxix. C. Williams, Acts, 2, finds allusions in 2 Tim 3:11 and the gnostic Gospel of Truth; Conzelmann (Acts, xxvii–xxxii), by contrast, finds attestation in the Apostolic Fathers but not in the Pastorals. Most patristic allusions (e.g., Justin 1 Apol. 39.3; 50.12; 2 Apol. 10) need not, however, establish a first-century date for Acts; Luke uses Mark (and acquaintance with Paul) without a fifty-year gap. (But cf. earlier 1 Clem. 2.1, unless it independently attests the agraphon in Acts 20:35.)
Theissen and Merz, Guide, 32. Against those who argue that Luke depended on the completed version of Matthew (as opposed to a proto-Matthew much like other scholars’ Q), material like the tensions between the infancy narratives and the two authors’ divergent treatments of Judas’s death (Matt 27:3–10; Acts 1:18–19) make such dependence well-nigh impossible.
With Witherington, Acts, 61–62 (also arguing from the “we” passages that Luke sees his own time as a continuation of Acts); cf. Wallace and Williams, Acts, 11 (contending for a date before Paul’s letter collection, on which Ignatius drew).
Pervo, Dating Acts; Tyson, Marcion; cf. Pervo, “Dating Acts”; Tyson, “Dates.”
Tyson, Marcion, ix; the survey of views, 1–23 (though focusing on those who support a second-century dating).
Views also change over time; Pervo, Dating Acts, 2, is aware that an antithesis against a previous consensus tends to readjust the resulting center or synthesis.
Jones, “Rejoinder,” thinks that the Jewish Christian source of Ps.-Clem. Rec. 1.27–71 comments on—and seeks to supplant—the imagery in Acts. Our extant material, however, appears later than Irenaeus. For a full survey of possible references or allusions to Acts in the first six centuries, see Bovon, “Reception”; most relevant here are his observations regarding Irenaeus’s sources (68), the mid-second-century Epistula apostolorum (69), the Muratorian Canon and Western text (70), and some other possibilities that some have suggested (71–72).
As noted by Pervo, Dating Acts, 331.
Tyson, Marcion, 50–78 (see esp. 73, 76–78); idem, “Wrestling” (emphasizing the pro-law Paul of Acts). Although one can read Acts this way, it is not the only way, and it makes Luke’s accuracy on much earlier matters of testable detail remarkable. He suggests that a pre-Marcionite Gospel of Luke began with Luke 3:1; Marcion’s version of this Gospel followed; then the author of Acts created the present canonical Luke (adding 1:1–2:52 and other material) and later wrote Acts, publishing the work as an entirety ca. 120–25 C.E. (Tyson, Marcion, 119–20). Without positing a definitive date, Gregory, “Reception of Acts,” 64, observes that the theological affinities discerned might mean not that Acts was written in the early second century but that it was read then.
From an index of Scripture citations in Tertullian, I estimate that several of Tertullian’s works cite the Gospel of Luke about twice as often, and Matthew about three times as often, as Acts. Yet Tertullian cites Luke closer to sixteen times his Acts citations in his anti-Marcionite work.
Tyson, Marcion, xi, 76–77, is explicit that his views echo those of Knox, who was his mentor.
The conjunction of the verb ἔλυσας and the noun ὠδῖνας in Job 39:2 LXX is hardly comparable, and the correspondence here is close.
Pervo, Dating Acts, 17–20.
Though a date so early should render Tyson’s Marcionite approach untenable, Tyson could counter that this part of Acts belonged to a pre-Marcionite version. Occam’s razor, however, prefers fewer necessary conjectures, the multiplication of which could be used to defend most hypotheses.
Charles E. Hill pointed this likelihood out to me in personal correspondence (Jan. 18, 2010). There is also a possible but not necessary conceptual allusion in Ep. Diogn. 9.1 to the idea in Acts 14:16; 17:30 (Hill, Polycarp, 148), which Hill believes may well be the work of Polycarp (ibid., 97–165). Because the longer ending of Mark betrays familiarity with Luke’s Gospel (cf. Mark 16:9 with Luke 8:2; Mark 16:12–13 with Luke 24:13–35), it may well allude to ideas in Acts as well (cf. esp. Mark 16:20 with Acts 14:3; perhaps Mark 16:18a with Acts 28:3–6; perhaps Mark 16:19 with Acts 1:11; the tongues of Mark 16:17 may be more general—See comment on and McDonnell and Montague, Initiation, 314). The terminus a quo for this addendum to Mark is, however, uncertain, making the allusions less helpful than definitely early sources.
Cf. Mark 16:9 with Luke 8:2; Mark 16:12–13 with Luke 24:13–35. For other connections, see, e.g., France, Mark, 686.
France, Mark, 687, suggests that Mark 16:17–18 alludes to signs in Acts and 16:20 summarizes Acts. The tongues of Mark 16:17 may reflect knowledge of more general practice (See comment on Acts 2:4 and early Christian tradition in McDonnell and Montague, Initiation, 314).
Bratcher and Nida, Handbook on Mark, 506, noting that it may have been known by Justin (1 Apol. 45) and dating Tatian’s Diatessaron to 140 C.E. Scholars usually date the Diatessaron later, however.
Anderson, Mark, 358.
Pervo, Dating Acts, 20.
Ibid.
Ibid., 378n52. Too recent for engagement here, MacDonald, “Papias,” suggests Luke’s dependence on Papias.
Pervo, Dating Acts, 378n52. My edition (Lightfoot, Harmer, and Holmes) identifies it as frg. 5.1.
Pervo, Dating Acts, 20. Others date his sojourn in Rome slightly earlier.
See, e.g., Dibelius, Studies in Acts, 88–89.
Pervo, Dating Acts, 20–22.
Haenchen, Acts, 8. Pervo doubts the allusion to Acts 1:8 (“power” appears in Luke 24:49; but “receiving power” does sound as if it draws on Acts 1:8; moreover, unless one accepts Pervo’s separation of Luke from Acts [Pervo, Dating Acts, 24], Justin’s knowledge of the Gospel would likely presuppose the existence, sometime not many years after it, of Acts). For uncertain but not impossible allusions to Acts (3:14; cf. 26:29), see Justin Dial. 8; 16. Justin seems more interested in Jesus’s sayings (see Bellinzoni, Sayings), which he may quote from memory (see Kaye, Apology, 104–8). (He might also know a harmony [Osborn, Justin, 126–28], testimonia [115–17; Parkes, Conflict, 99; Chadwick, “Defence,” 282; Barnard, Justin Martyr, 74], postcanonical material derived from the Synoptics [Bellinzoni, Sayings, 139; cf. 21, 140], Barnabas [Williams, Justin Martyr, xxi–xxii], and some source shared with the Pseudo-Clementines [Kline, “Sayings”]). He has much less interest in Paul (though some find Paulinisms [Stylianopoulos, Justin, 96] and contacts with 2 Thessalonians [Barnard, Justin Martyr, 63]). Even Justin’s complete silence about Acts would not, however, argue that Acts did not exist. For an analogy: apart from the logos (similar to John’s, Justin Dial. 128; Apol. 2.13; Holte, “Logos Spermatikos”; Barnard, Justin Martyr, 85–100), which Justin could have developed independently (Bauer, Orthodoxy, 156; Barnard, “Study,” 160; against dependence even on Philo, see Barnard, “Logos Theology,” 137; Osborn, Justin, 73–74; Chadwick, “Defence,” 296–97), Justin alludes to the Fourth Gospel rarely (Chadwick, “Defence,” 283; cf. Bauer, Orthodoxy, 208–12), probably only to John 3:3 (Apol. 1.61; Osborn, Justin, 137, attributes this instead to a baptismal liturgy, but Barnard, Justin Martyr, 60–62, thinks that the cumulative weight of parallels suggests allusions). Nevertheless, earlier gnostic use and 𝔓52 make clear that the Fourth Gospel was already in circulation.
Pervo, Dating Acts, 23. For Irenaeus as the first clear citation, see Gregory, Reception (who finds the first certain use of Luke’s Gospel in the mid-second century, and Acts a bit after that), but he is not arguing that Acts was written this late.
Pervo, Dating Acts, 25.
Ibid., 24. Could it be the pro-Pauline, pro-Gentile stance of Acts that made much of Luke’s Gospel acceptable to Marcion? Given Pervo’s date of ca. 115, however, his view would not necessarily suffer by allowing Marcion’s knowledge of Acts.
In a public response to discussion after his paper (Gregory, “Irenaeus”), Gregory emphasized that we lack 85 percent of works that one writer names as existing then and also that Greco-Roman writers often are quoted in extant works only later. Gregory, “Reception of Acts,” 64–65, warns against arguing from silence based on when works are first explicitly attested. One might note that most scholars believe that a fragment of John’s Gospel demonstrates its circulation probably earlier than its attestation in Justin. Meanwhile, Tyson, who refuses to date Acts before its explicit citation in extant sources, infers Luke’s actual use of Josephus and Paul’s letters without explicit citations of them.
Because I have argued that Luke’s sources in Acts are primarily oral, his accuracy on such points cannot be compared with that of, e.g., Tacitus, who writes long after many events described by him yet has annals and other written sources from which to draw.
On 𝔓52, see, e.g., Roberts, Fragment; Metzger, “Papyri,” 40. The fragment is virtually identical to the passage in John’s Gospel (Dibelius, Jesus, 13).
A unity contested on one level (that of purpose more than of authorship) by Parsons and Pervo, Rethinking.
Pervo, Dating Acts, 51–147; idem, “Dating Acts”; cf. also several others, e.g., Walker, “Acts and Corpus Revisited.”
Pervo, Dating Acts, 149–99. Despite his commendable detail, I do not find his appeal to dependence on Josephus convincing.
Ibid., 309–17.
Ibid., 310.
Ibid., 201–57 (quote, 257). Pervo uses other comparisons with the Apostolic Fathers (259–308) to establish a date most likely (307) ca. 110–120 (hence ca. 115, p. 346). He argues that social and legal issues resemble those found in second-century apologists (327–28), e.g., “a willingness to ridicule polytheistic practices” (327)—yet such ridicule occurs commonly in earlier Jewish sources, including Isa 46 (see fuller comment on Acts 14:15–17)! On Acts and the apologists, see also Pervo, “Apologists.”
Certainly, Luke’s Gospel fits better with Matthew and Mark than with later gnostic “sayings gospels” or even the apocryphal gospels. Parallels with Acts are so many as to suggest to me Luke’s substantial knowledge about Paul (though Pervo would argue that most of this derives from Paul’s letters).
Cf. also, e.g., Stagg, Acts, 22 (shortly after the Judean-Roman war); Polhill, Acts, 27–31. For Luke responding to still-concrete memories of riots associated with Paul, see Keener, “Apologetic.”
Then again, because apprenticeship ended younger, it is not surprising to know of one physician perhaps aged seventeen (Lewis, Life, 153–54, citing P.Giss. 43 = SB 10630; for physicians less than twenty in Asia Minor, he cites Phoenix 36 [1982]: 271). Some of Paul’s other putative traveling companions were younger than Paul, probably adolescents at the beginning of their work with him (John Mark and Timothy). Further, even if the author of Acts is identified with Luke the physician in Colossians, we cannot be certain that the medical training preceded his initial voyage with Paul (Acts 16:10–12).
Given average post-childhood longevity and a median guess of an age of twenty-five to thirty when he first traveled with Paul, we might prefer a date before 85 C.E., but only as an estimate.
These did not affect every locality: among the seven central churches of Asia, persecution appears only in Rev 2:9–10, 13; 3:9; see Keener, Revelation, 37–39, esp. 39.
Later, in the period of the Apostolic Fathers, the Corinthian church apparently continued to struggle with prosperity (cf., e.g., 1 Clem. 13.1; 38.2; 59.3); while this temptation also existed for some in Philippi (Poly. Phil. 2.2; 4.1, 3; 5.2; 6.1; 10.2; 11.1–2), they may also risk persecution (Poly. Phil. 8.2; 9.1–2; 12.3). The persecution in Philippi seems less obvious than in Paul’s day, however (Phil 1:28–30; 1 Thess 2:2).
Although one could read the narrative thus simply in view of Luke 21, the point would be much stronger for a Diaspora audience after 70.