6. Approaching Acts as a Historical Source
6
Granted that Acts is ancient historiography, what kind of ancient historiography is it? The question can be answered in different ways, depending on one’s objective: from the perspective of form, Acts is (as argued above, following many others) a historical monograph; from the perspective of purpose and function, it is (as also suggested above, again following others) apologetic historiography and the foundation story of a people or (more precisely) a movement. Here we consider the degree of Luke’s fidelity to historiographic ideals.
As already noted, most ancient historiography included both an interest in information and an interest in pleasing rhetorical presentation. This chapter investigates Luke’s use of information, examining a question of great interest to modern historians of antiquity (and of significance to ancient readers of histories as well): to what extent can we trust Luke’s reliability regarding the historical claims that he offers?
I pose this question of reliability here in terms of the standards of ancient historiography, not modern historiography. As already argued, ancient historians expected substantial accuracy regarding events, but they did allow the construction of scenes supporting such events, sometimes based on inferences (especially regarding speech and discourse material). Yet, as also noted, there was a range of practices among ancient historians; not all proved equally reliable by either ancient or modern standards. Establishing a general range of reliability on major points for ancient histories in general does not settle where Acts lies on this range.
To better answer this more particular question, we must first survey Luke’s use of sources and then turn to some specific evidence. Writing according to the constraints of ancient rather than modern historiography, Luke exercises flexibility even in his own summaries of the events he reports; he does not expect his readers to flinch when he gives accounts that diverge in some details (cf., e.g., Luke 23:50–53 with Acts 13:28–29; Luke 24:40–51 with Acts 1:3–9). Taking this flexibility for granted, I instead focus in this chapter on points of likely correspondence with external history.
To some extent, however, this chapter merely introduces some issues that must also be addressed later. The next chapter will take up the genuine major test case for Luke’s reliability, namely, comparison with Paul’s letters. I believe that this test case will argue strongly for Luke’s use of substantial historical information (and that most readers will find this evidence for Luke’s use of such data persuasive).
After these discussions we will turn to the questions of speeches and signs reports in ancient historiography and in Acts (chs. – below). Because of the special problems associated with these issues, the present discussion on reliability addresses primarily nondiscourse, nonsupernatural narrative in Acts. To separate speeches and signs from other historical narrative is not to prejudge the case but to recognize that, in view of ancient (especially for speeches) and modern (especially for signs) questions, more space is required to address these issues separately.
No one believes that nothing happened in history (or even in our day-to-day lives) other than what we can prove or provide evidence for. When we discuss historical “probability,” we acknowledge that some events that our methods deem probable did not happen and that others that our methods deem improbable did happen; our methods entail much educated guesswork. Nevertheless, an educated guess is apt to be correct more often than an uneducated one. When we speak of historical probability, then, we speak in terms of the principles of historiography that tend to give us more accurate answers. Normal historical methodology suggests that Paul appeared before Gallio but that no one in the Sanhedrin scene of Acts 5 should have mentioned Theudas. This methodology cannot guarantee such conclusions, but insofar as we ask questions based on extant evidence, we can offer conclusions that are probable on the basis of such evidence. This is what scholars normally mean by “historical probability.” The degree to which our conclusions are tentative in a given case depends on the quality, quantity, and representative character of our extant evidence.
Although we often lack direct access to firsthand ancient sources, we can look to the ancient historical writers who still had access to such sources, then test them by means of other writers to try to determine the degree of their fidelity to those sources. Comparing different ancient historians such as Herodian and Dio Cassius turns up discrepancies but also confirms that both were employing substantial historical data. Such a comparison will also reveal that such writers did not always choose to cover the same ground; thus, for example, there are many omissions in Herodian, but hints of the information suggest that he did not lack the information itself.
Although ancient historians did not always have access to the best sources for earlier eras, their treatment of more recent history was generally more dependable, as suggested above (). The Roman historian Tacitus, for instance, recorded much of the history of first-century Rome, often using imperial annals; he is widely regarded as one of the most reliable sources for the history of this period. When Tacitus wrote biography, he maintained the same standard he had upheld in writing Roman history: although his Agricola, a biography of his father-in-law, has a particular agenda (to praise his father-in-law while condemning the depravity of the Flavian era), it is certainly as historically reliable as Tacitus’s resources allowed him to make it. Indeed, Tacitus, as Agricola’s son-in-law, also had firsthand acquaintance with the data he reported.
Other historians reporting contemporary or recent events were also substantially reliable, although one must consider how critically each writer used his sources and how freely he adapted them. Suetonius’s biographies of the twelve Caesars provide critical information to modern historians of antiquity; they may be less reliable than Tacitus, but where Suetonius errs, it is by depending too uncritically on his sources, not by fabricating material. Collecting so much material was no mean feat or small service to subsequent historians. Other historians and biographers, such as Livy and Plutarch, took much more freedom to moralize and spice up their narratives. But as noted above, even Plutarch plainly believes that it is historical data he is using to make his moral points, and his record frequently parallels other historical sources. Historians did make errors, but they could expect their successors to expose their errors when discovered (Diod. Sic. 4.56.7–8).
Josephus may provide an example of a freer historian. Josephus’s history and autobiography are dominated by his apologetic Tendenz; J. D. Crossan wryly but insightfully remarks of Josephus’s War, “Nobody from the highest aristocracy on either side is guilty of anything.” Even many of his adaptations of biblical accounts emphasize points pleasing to his Roman sponsors and Gentile audience. He also interprets his sources for his Hellenistic audience in other ways. After promising to add nothing to Moses’s laws (Ant. 4.196; cf. 1.17), he finds among them a specific prohibition against theft from pagan temples (4.207), a prohibition against women’s testimony (4.219), and the requirement of seven judges per city (4.214). Numerous studies have traced Josephus’s adaptation of biblical accounts, but whereas the degree of adaptation varies from one narrative to another, it remains balanced by his general fidelity to the basic biblical account. He seems to have viewed his “translation” task as including interpretation and adaptation for his audience; Plato, Cicero, and others understood the translation task similarly.
Few agree with Josephus’s population estimates (see discussion at Acts 2:41); probably all that he had available in most cases were guesses. His speeches are likewise viewed as suspect (see discussion in below). Nevertheless, archaeology confirms that he usually gets right even many minor details unaffected by this Tendenz (although his accuracy has exceptions). Inscriptions likewise sometimes confirm his accuracy on disputed details, sometimes against other historians.
Josephus retells the same event in different ways in different books, yet this practice suggests not that the event never happened but that he presents it from a different perspective. While adding details and perspectives, he even retains the stories of David’s sin with Bathsheba (Ant. 7.130–31) and Uriah’s murder (7.131–46), though—perhaps with an eye toward anti-Judaic polemic such as Apion’s sources—he omits the episode of the golden calf (3.95–99). Although Josephus was not striving for modern standards of historical accuracy, E. P. Sanders concludes that “wherever he can be tested, he can be seen to have been a pretty fair historian.” (A fuller comparison with Josephus is offered here below.)
Although we have established that historians intended the substance of their works to be historically reliable and that they accomplished this intention much more commonly with recent sources than with older ones, historians varied in their degree of reliability. Evidence examined here so far suggests that Luke writes a historical monograph; that he therefore reports events that he believes to be true (whatever the accuracy of particular details); and that given the recent nature of the events he reports (and especially the more recent ones), he is likelier than not to be largely correct in the events that he claims. But how far should we press this general principle?
For example, it is certain that Paul was publicly punished at Philippi (1 Thess 2:2; Phil 1:30), likely that he was jailed there, and reasonably likely that this happened in response to an exorcism as Luke claims. But to what degree should we press Luke on the details (such as particular conversations in that narrative or even the meal with the jailer)? Where on the range of reliability should we expect him to fall—in other words, how probable should we expect his reliability to be in any given case, and where does the burden of proof lie, and how heavily? And what sort of sources does Luke have for the period before Paul’s ministry? From a strictly historical standpoint, such questions can be resolved historically (to the extent that a strict historical method can resolve them) only by working through the text passage by passage with such questions in mind. But we can start by laying out general observations below concerning historians’ use of sources, including Luke’s use of them.
Ancients expected historians to utilize sources, whether written (most commonly), oral (for recent or local events), or their own firsthand experiences (where they were eyewitnesses). Although Luke is less apt to name his sources than most extant ancient historians (who tended to be elite historians), he clearly had sources available (e.g., Luke 1:1–2; Acts 16:10).
Omniscient narrators, common in epic poetry, were not common to historians. Historians, in fact, often felt constrained to indicate some of their sources, especially if they depended on written sources and wrote for upper-class readers who might have access to these various works. Accuracy of sources, unimportant in some genres, was known to be a concern of historians (Pliny Ep. 9.33.1). One could establish one’s point better by naming various earlier sources supporting it (Suet. Jul. 9.3). A historical writer who does not include everything that has been written might need to explain that he had in fact read almost everything but did not judge it all suitable for inclusion (Dio Cass. 1.1.1–2).
That ancient historians, biographers, and anthologists depended on earlier sources is not in question; both biblical and Greco-Roman traditions frequently cite them. (Other means of detecting them, such as different textures within a narrative, are usually less dependable.) They often cite varying accounts, even when preferring one over another. Arrian prefers over other sources his two earliest ones, which often agree, and he chooses between them when they diverge; when sources diverge too much, he frankly complains that the exact truth is unrecoverable. Likewise Philostratus notes a point where his sources diverge and his research provided no definitive resolution; such a complaint hardly implies that ancient historians regularly practiced free invention. At one point Plutarch names five sources for a “majority” position and nine for a minority one, plus an extant letter attributed to the person about whom he writes, but he then adds that the minor divergence does not affect our view of his hero’s character (the main point for him; Alex. 46.1–2). Valerius Maximus, a more popular and less careful writer than some others, rarely cites his sources (and often confuses his data), but he mentions them occasionally when they diverge (e.g., Val. Max. 5.7.ext. 1; 6.8.3). Occasionally historians also found ways to harmonize traditions (Diod. Sic. 4.4.1–5). Sometimes they appear to have smoothed out contradictions in their sources in their own rewriting.
At the same time, we dare not assume that historians have such sources only when they cite them; as a matter of rhetorical elegance, they avoided documentation where they found it unnecessary. Historians usually mentioned these sources only when they conflicted or the author disagreed or was unsure about their reliability or about which sources were best. Such conflicts arose more often when they were treating events of the more remote past than when they were treating the present. Luke (like the other Gospel writers) does not identify specific sources (except perhaps “we” in Acts 16:10–28:16), probably in part because he discusses events of a recent generation on which sources have not yet diverged greatly. A historian might also more generally refer readers interested in more detail to “other historians” (e.g., Vell. Paterc. 2.48.5), not unlike Luke’s oblique reference to other authors in Luke 1:1. The Gospel writers’ reticence to name sources might also follow some Jewish conventions on this point; in some early Jewish works we can identify the sources only because they are extant. The more popular audience that the Gospel writers anticipated may be a more important factor in their failure to name sources; popular works of various genres were less likely to cite sources, even when they clearly depended on them. Earlier scholarly exaggerations of the contrasts between elite and popular literature aside, the Gospels do not reflect an elite audience (though Luke’s audience appears socially higher than the others’).
Including material missing in earlier extant sources does not always betray fabrication. A writer providing information missing in some earlier historians sometimes was drawing from sources unavailable to the other historians, whether the sources were written, oral, or both. Moreover, even writers who preserved their sources redacted them, even in the case of sacred cultural texts and philosophic works. Some scholars argue that such changes were more common at the written than at the oral level, perhaps because the former reflected literary urban culture and the latter more traditional society.
Polybius complains that Timaeus failed to make appropriate use of earlier historians’ works although living in Athens provided him access to them (Polyb. 12.25d.1); this complaint reinforces the picture that many sources were available and that good historians were expected to consult them when possible. For Polybius, real history requires a study of documents, although this must take a third place behind visiting the locations in question and reviewing their historical context (12.25e.1, 25i.2).
We should not underestimate the research sources available to ancient writers, especially if (as in Luke 1:1) they mention the existence of such works. Clearly, an abundance of contemporary sources existed then that are no longer extant (cf. the many contemporary histories of Nero noted in Jos. Ant. 20.154); for example, Pliny the Elder, explaining that he could not survey everything (N.H. pref. 18), notes that he surveyed about two thousand volumes (though using especially a hundred) and supplemented them with other data (pref. 17). Few of his sources remain extant, but we can appreciate Pliny’s preservation of much of the content.
Tacitus omits reporting most of Seneca’s dying words, recorded by the latter’s secretaries, because in Tacitus’s day they remained too well-known (Ann. 15.63). He elsewhere cites “historians of that era” (Ann. 5.9) as sources for events a century before his time. He normally follows annals and earlier histories (sometimes specified only when they became subjects of history, e.g., Ann. 4.34–35), but also consulted personal memoirs from perhaps half a century earlier (Ann. 4.53). Suetonius’s sources include notes that he took from official “libraries and archives,” and while he proved less critically discerning about his various sources than Plutarch, modern historians appreciate “his hesitation to impose his own judgments” on his material. Suetonius apparently made some local inquiries for his work (Vesp. 1.4), and sometimes could establish his point by naming various earlier sources supporting it (Jul. 9.3).
One may compare sources already available to Tacitus, Suetonius, and Plutarch when writing about the short-lived Roman emperor Otho, who perished roughly half a century earlier. One source that they sometimes shared might be the no-longer-extant work of Fabius Rusticus (cf. Tac. Ann. 13.20.2; 14.2; 15.61). Plutarch consulted witnesses, including an officer who described to him what he saw while Plutarch was touring the site with him (Plut. Otho 14.1). One of Suetonius’s sources concerning Otho is explicit: his own father Suetonius Laetus was a tribune serving under Otho, and shared with him information about Otho’s character and actions (Otho 10.1). By ancient standards, these writers, like Luke, were writing a fairly brief time after the events, within living memory of eyewitnesses. Indeed, on the majority view of the dating of Acts, Luke writes closer to most of his events than these historians wrote to the time of Otho.
Luke’s original prologue (Luke 1:1–4) probably introduces the entire two-volume work, although not necessarily every statement covers every section of his work equally. (This commentary investigates some aspects of his prologue in further detail at Acts 1:1.) At least for his Gospel, Luke claims the availability of many written documents covering the same ground he covers (Luke 1:1). Some of these may have constituted sources for him to investigate; certainly, they were not all written at once, and those he mentioned wrote before him, although we cannot say how early the first narratives were. Did Luke more frequently find such sources trustworthy or suspect?
Ancient historians often cited inadequacies in earlier writers as reasons for writing, although some explicitly disclaimed such reasons. Some rhetorical critics also offered this claim of unique superiority for select historians before their time. Many scholars think that Luke thus finds fault with the work of his predecessors; ἐίιχειρέω appears elsewhere in Acts for futile or mistaken efforts (Acts 9:29; 19:13), and an adjective cognate to ἀσφάλεια in Luke 1:4 addresses distinguishing truth among “competing claims” (Acts 21:34; 22:30; 25:26).
Luke’s language, however, is far less harsh than that of writers who genuinely criticized their predecessors, and it certainly need not imply that he is claiming that they lacked eyewitness sources that he has. Some ancient writers noted that they simply had a fresh perspective to add to the work of their predecessors. Another writer could claim to prefer even his predecessors’ style to his own but to regard his own selectivity as an improvement, or could note that he has nothing new to add in terms of information but offers it in a different way.
Luke departs from the work of his predecessors not primarily in terms of accuracy of information but instead in terms of rhetorical superiority: he promises to write καθεξῆς, that is, with the proper sequencing of events in terms of cause and effect, necessary to an accurate account. Smaller accounts within the larger one (e.g., 1:21–22; 26:4–5) and historical retrospectives in speeches serve to reinforce this purpose. The term need not specify chronological order but refers rather to logical and persuasive order; for a history, however, logical order would include as much chronology as possible. Because Luke’s sources, whose sequence he usually follows meticulously in his Gospel, control his arrangement there but are probably not entirely chronological, Luke’s purposeful sequencing of a continuous narrative is probably more effective chronologically in Acts. In contrast to some other genres, history should be written sequentially (Pliny Ep. 1.1.1), though with the flexibility to allow for completing narratives begun at different times (cf. the problem in Acts 11:30; 12:25).
Other elements of Luke’s prologue may point to either rhetorical or factual excellence. Balch argues that some historians (including Polybius, Dionysius of Halicarnassus, and, in his view, Luke) felt that their readers would be dissatisfied with an epitome containing almost exclusively deeds (e.g., 2 Maccabees or Mark); they wanted to hear the reason for events, and so the historians provided narratives more ἀκριβῶς, fully, hence including speeches that explained events. He contends that ἀκριβῶς in Luke 1:3 modifies γράψαι and means to write fully rather than to follow accurately.
Although this reading of 1:3 is possible (it might better fit the contrast with earlier writers in 1:1–2, and it may accurately depict the situation with speeches [see below]), Luke’s own use of ἀκριβῶς elsewhere could suggest accuracy instead of, or in addition to, completeness; accuracy was also part of the term’s nuance (e.g., Diod. Sic. 1.4.4; 1.6.2). In most Lukan texts, one could interpret the term either way (Acts 18:26; 22:3; 23:15, 20; 24:22; 26:5), but Acts 18:25 seems to imply primarily “accuracy” (possibly implying the same for 18:26, but 18:26 may play on the other nuance of the term), and the texts about Pharisaic scrupulousness (22:3; 26:5; cf. 18:25–26) probably specify “accuracy,” in view of similar descriptions of Pharisaic learning in Josephus (Ant. 17.41; 19.332; War 1.110; 2.162; Life 191). The phrase can suit an “exact” account elsewhere.
Luke’s rhetorical superiority to earlier Gospels of his day (though not to elite historians) is likely: as a comparison with our other extant Gospels may confirm, his Gospel is far closer to the general vein of extant Greco-Roman historiography. Nevertheless, his research and arrangement did not lead him to depart substantially from the facts of his predecessors; he saw his purpose as confirmatory (Luke 1:4). Their proximity in basic substance may reflect the relatively recent nature of the events they report.
Luke has more information available than he recounts, and sometimes this information is already known to his ideal audience (see, e.g., Luke 1:4). Thus he sometimes recounts a matter as if his audience already knows it, when in fact he has not mentioned it (e.g., Acts 9:39; 17:6). The large majority of scholars hold that Luke used Mark in writing his Gospel, and most also hold that he shared a common source with Matthew; some even suggest a common source or sources with John. Some have also used Semitisms to suggest Semitic sources behind some of his work (esp. Luke 1–2), although these may simply display Luke’s skill at writing “biblical” Greek. Luke is well informed about the first-century C.E. history of Judea’s rulers (e.g., Luke 3:1; Acts 23:24; 24:27; 25:13), including plausible events not recorded elsewhere (e.g., Luke 13:1). This suggests that he found records or memories of Judean events available; neither he nor Josephus (who was not alive during the period reflected in the earlier accounts in his War) composed histories from thin air.
Luke may avoid naming specific sources in part because he writes on a more popular level than most extant works, which tend to reflect elite circles. Elite circles were more concerned about plagiarism, though it seems to have been fairly common. Pliny the Elder notes that he begins his first volume by listing his sources—even though most of them had not done their predecessors the same favor (N.H. pref. 21). “For you must know,” he warns, “that when collating authorities I have found that the most professedly reliable and modern writers have copied the old authors word for word, without acknowledgement” (pref. 22); how petty, he charges, to prefer theft to repaying one’s loan (pref. 23). While Luke does not identify his sources, however, both Luke 1:1 and the Gospel’s internal evidence make clear that he used sources. If Luke has reworked them, often adapting their wording and making them his own, most ancient circles would not count this as “plagiarism.”
But if Luke avoids naming sources for literary reasons, his ethical justification may be simply that he, like his predecessors, depends on recent and communal memory (“accomplished among us,” Luke 1:1). The earliest Christian biographers and historians were not bound to written works the way that historians and biographers narrating the distant past were. (Thus, for example, as we have mentioned, Tacitus does not need to cite sources in his biography of his father-in-law, Agricola, the way Plutarch does for Alexander of Macedon.) The Gospel reflects the shared, recent memory of the community, of things “accomplished among us” (1:1), passed down orally from many leaders “to us” (1:2), and already widely known (1:4).
Luke does use written sources (cf. 1:1) in his Gospel, surely Mark and likely Q (and very possibly others as well), but these sources themselves depended on the shared tradition of the community of which they had become an important part. The matter appears even more clear-cut in Acts: the sources here are probably entirely oral (reports heard and interviews) or based on firsthand acquaintance (see the fuller discussion on authorship and esp. comment at Acts 16:10), which made for history writing in the best of the Hellenistic historiographic tradition.
Even in elite circles, if a writer alluded to well-known works that the hearers might recognize, the thought was not plagiarism but literary sophistication; if he could assume his audience’s knowledge, he did not need to state the source of his quotes. Luke’s audience knew some gospel tradition, which seems to have been “common property” of the early churches (cf. Luke 1:4). If Luke’s preferred written sources, supplemented with oral ones (cf. 1:2–3), were recognizable and common property, we might also not expect him to feel the need to identify them. (Naming sources is not in itself a question of genre, since elite mythographers also can name their sources. The practice of naming sources when it was relevant to do so simply illustrates that prose writers about the past typically built on what preceded them instead of freely inventing in the way dramatists or novelists far more often did.)
Scholars devote volumes to the “Synoptic Problem” and to how (in the most common view) Luke has used Mark and the material he shares with Matthew (commonly called Q). I have entered this fray elsewhere and here simply summarize my agreement (to a reasonable degree of probability) with the general consensus of scholarship on this question. I have also argued that, as with other ancient Mediterranean literature reporting on past events (e.g., Paus. 1.23.2), we should allow for a considerable range of other sources, written and especially (for events this recent) oral, which are no longer extant.
If one works through the Lukan texts from a synopsis of the Gospels, one will note both that he exercised considerable freedom in adaptation (such as the construction of symposium scenes) and that much of the Gospel is tightly bound to its sources in depicting events, sayings, and even many of the details of these events. It seems reasonable to assume that if Luke, in most cases, follows his sources closely where we can test him, he follows a similar approach with his sources that are no longer extant.
That Luke used sources in Acts, as in the Gospel, is likely. The general uniformity of style in Acts no more counts against the use of sources there than the stylistic similarities between Acts and the Third Gospel count against sources in the latter. Likewise, Luke’s usually fairly consistent style cannot count against his use of sources any more than Josephus’s rewriting biblical narratives in his own style should lead us to think that he did not follow biblical accounts. Historians typically rewrote sources in their own style.
Most often, however, Luke’s sources for Acts were probably oral reports. If Luke’s sources were mostly oral (except for the travel journal in the “we” sections), we should expect even less stylistic indication of sources than in Josephus. Whereas many had written accounts of Jesus (Luke 1:1; of which pre-Lukan sources at least Mark remains extant today), Luke’s second volume was probably breaking new ground that apparently failed to be understood even in the mid-second century. Because we lack the sort of overlapping documents Matthew and Mark provide for the Gospel, we also have no hard evidence for reconstructing sources in Acts, and this reduces most arguments about them to speculation. Trying to reconstruct the sources of Acts is akin to trying to reconstruct Mark if all that we had available was Luke. Few source-critical reconstructions agree, and whatever sources Luke may have had he rewrites in his own style. Thus most scholars today are content to admit that Luke has sources in Acts but that we cannot reconstruct them. As Beverly Gaventa reminds us, “Whatever Luke’s sources may have been, they are no longer available.”
Despite our inability to reconstruct them, Luke’s use of oral sources remains probable (see the discussion below). In this case, what should we expect of such oral sources? Many studies of human memory approach memory as reconstruction more than copy, with a recollection of components rather than of wholes. Such reconstruction has constraints; the reconstruction approach to memory explains inaccuracies in eyewitness reports, but constraints explain “the relative accuracy and the broad element of stability.” Interpretive elements color both one’s initial experience and the recollections; although these do not necessarily distort the content, the result is “more like a painting than a photograph.” Among elements of experience most apt to support the preservation of the sorts of memories common among Luke’s likely informants are distinctive and consequential events and frequent retelling. Narrative structures used as a grid to interpret these components of memory may be borrowed from narrative structures in the larger culture. ( below addresses memory more fully.)
The kinds of oral reports that Luke might have had available (beyond the “we” narratives, where, I argue, despite the skepticism of many, the evidence is far more convincing than not that Luke was present; see the much fuller discussion at Acts 16:10) are not hard to conjecture. If the author spent months traveling with Paul, his memory of Paul’s occasional anecdotes would provide a sufficient oral source for many of the reports in Acts (Acts 9:1–30; 11:25–30; 13–20); Philip (whom Luke could have interviewed or listened to at 21:8–10) could provide material about the Hellenists (Acts 6–8) and about ministry in Caesarea (8:40; 10:1–48); Jerusalem sources would cover most of the rest of the book, and Luke apparently had up to two years for any interviews with Judeans in Judea, though probably only a minimum of that in Jerusalem proper (see 21:15; 24:27; 27:1). If the author was, as many scholars think, a Gentile (see the discussion on authorship in ), Luke’s direct access to Jerusalem sources may have been somewhat limited (cf. 21:18), and Galilean sources would have proved even less available; but he would likely have had abundant access to accounts of those who had settled in Caesarea (cf. 21:8) and perhaps some familiarity with the coastal cities (cf. 8:40; 9:32–43). Luke had less concrete material for the earlier chapters and hence fleshed them out “with summaries, speeches, and vivid descriptions.” Once his narrative reaches Acts 13, Luke appears to possess more “substantial and reliable” data, an observation consistent with the view that Paul could have been an oral source. The reports seem most complete in the “we” narratives, an observation consistent with the view that the author may have been present on the occasions mentioned.
How much would Luke have adapted his sources? Certainly Luke, like other Greco-Roman writers, feels free to vary details, since he does so even when retelling the same event in his finished narrative work, yet in such cases he preserves the basic core of the account. In the Gospel he preserves the basic substance of his sources where we can compare them. Luke’s editing of Mark is most often stylistic, eliminating historical presents and less appropriate imperfects; some scholars argue that he might be even more conservative with Q. (Mark presumably employed historical presents for vividness, but some other writers employed them only sparingly, and some ancient critics complained of inconsistency in verb tenses.) Luke also edits to adapt details of the tradition for his Diaspora audience; for example, in a parable he transforms a Palestinian wadi into a river, and in a narrative he transforms men digging through a thatched Palestinian roof into men removing roof tiles, as in the northern Mediterranean region. Luke’s practice on this point fits what we expect for rhetorical adaptation of narratives (outlined, e.g., in Aelius Theon’s Progymnasmata).
Martin Dibelius observes that Luke was “more bound by his material in the Gospel,” so that there he acted as editor but in Acts “as an author.” This view underestimates Luke’s authorial role in shaping his Gospel but may nevertheless reflect a genuine difference between the two works. Because no predecessors had (so far as we know) established a pattern for a Christian work such as Acts, Luke may have exercised somewhat greater freedom in selecting and retelling his oral sources than he could have done in a Gospel. (This observation warns us against arguing from silence that Luke, in contrast to other ancient historians, must have invented material wherever we cannot reconstruct his sources.)
At the same time, where we can test him, Luke does not indulge this freedom recklessly; even the three fairly free retellings of Paul’s conversion in Acts (9:1–18; 22:5–16, 21; 26:12–18) preserve the basic substance. Luke remains a historian, not a novelist or tragic poet. Indeed, the reports in Acts may be viewed as even closer to firsthand sources than those in the Gospel.
Although it is not implausible that Luke’s Palestinian informants for earlier parts of Acts might account for his Semitisms, such supposed linguistic evidence has not proved widely convincing. Some scholars have found Semitic sources in the early sections of Acts (often as far as Acts 15) as in the Gospel (Luke 1–2). Today scholars more often attribute this apparent characteristic to Luke’s semitizing or archaizing style in these passages rather than to discrete sources that he employed. Luke removes Semitisms from Q less often than he edits Mark’s grammar; often he elsewhere adds Semitic constructions, perhaps by imitating the LXX. Some think that the LXX reflected or generated a sort of “Jewish Greek.” Others find such constructions in Koine, possibly reflecting the bilingual Eastern milieu in which Koine evolved.
The currently dominant opinion is that many of these characteristics reflect Koine, even if knowledge of a Semitic language (or imitation of the LXX) may have generated some “peculiarities” in individual writers. Where Luke’s style is especially Semitic, however, it may reflect biblicizing idiom from the more literally rendered (hence semitizing) parts of the LXX, as distinct from his usual style elsewhere. Luke’s editing of Mark shows that he controls the style in his narrative; he replaces loanwords in Mark and “restyles Mark’s simple paratactic sentence structure into participial and relative clauses that are more in keeping with the higher literary language.”
Luke is, as is widely acknowledged, quite capable of varying his style, a skill appreciated in his rhetorically astute milieu. Such variations in style have led to questions about more than one work’s authorship, so it is not surprising that it would lead some scholars to suspect non-Lukan sources behind Luke’s narrative. If Luke composed his own semitizing passages, he was also capable of more stylish, classical rhetoric, as in Luke 1:1–4. Even the most skilled rhetoricians employed different styles in different settings; adapting one’s style to the setting in the narrative was regarded as rhetorically appropriate. It is Luke’s careful modulation of his style that renders doubtful most source theories and his retention of a “we” from a source not written by him. (See the discussion of the “we” material especially at Acts 16:10.) This conclusion does not rule out the possibility that informants shaped the use of some Semitisms early in Acts, but it reduces the likelihood of identifying them.
How reliable were Luke’s sources? I have argued elsewhere, based on a broad range of parallels from Mediterranean antiquity, that the gospel traditions would have been preserved mostly reliably for the generation or two (usually estimated at ca. thirty-five to sixty-five years) necessary to reach our written Gospels. This would have been even more the case with the more recent events toward the end of Acts (by our dating, at most twenty years), though these may have been of less thoroughgoing interest to the churches. It also appears that Luke undertook to examine extant sources for his gospel narrative firsthand in a manner no longer available to us, by interviewing some survivors closest to the events described.
For the Greeks, the very term used for research or investigation, ἱστορία (historia), left “no doubt possible about what was early considered the defining characteristic of the genre. . . . The method . . . consisted basically of the interrogation of witnesses and other informed parties” and then weaving their responses into a cohesive narrative. Even if some writers failed to travel to all the places their narratives covered, travel was apparently a familiar component of historical research. Herodotus initiated this emphasis on research (Hdt. 1.1), traveling widely; Thucydides, who cross-examined his sources, assumed this approach as the standard (Thucyd. 1.22.2; 5.26). Diodorus Siculus claims to have visited the sites of his history in Asia and Europe, complaining that even some of the best historians err when they do not visit the sites in question (Diod. Sic. 1.4.1). Appian (Hist. rom. pref. 12) claims to have checked out his reports by traveling to Carthage, Spain, Sicily, Macedonia, and elsewhere. Likewise, the later historian Herodian insists that he accepted nothing secondhand without tracking down all the facts (Hdn. 1.1.3). Although most of Philostratus’s sophistic subjects were long deceased, he interviewed some who still lived, even on multiple occasions (Vit. soph. 2.23.606). According to later tradition, even mythographers traveled from one city to the next gathering their various oral traditions (Philost. Hrk. 48.11).
Polybius avers that investigation is “the most important part” of writing history (Polyb. 12.4c.3). His proposed method for conducting investigation, given the limitations of space and time, was to interview people, critically evaluate reports, and accept the most reliable sources (12.4c.4–5). He severely criticizes Timaeus for neglecting travel to the locations about which he writes (12.25e.1), complaining that written sources alone are not sufficient (12.25e.7) or even the most critical part of historical study (12.25i.2). Although interviews were impossible when one was dealing with the distant past, writers preferred them when living witnesses remained available. Greek historians often traveled to the locations of events and consulted those who were considered reliable oral sources. Condemning writers who sought to make guesses sound plausible, Polybius notes that in his research he also came across documentary evidence. Probability is one helpful test, but visiting a location and interviewing witnesses there is much better (12.9.2).
Greek historians had their weaknesses, but primary research was one of their strengths. Klaus Meister observes,
Although Greek historians usually referred to only one or two predecessors and quoted them uncritically (an “unscholarly” practice according to modern understanding), their primary research was often superior to that of modern historians: they relied to a large degree on “autopsy” and their own experiences, collected and examined the oral transmission, questioned eyewitnesses and sources, and visited the scenes of events in order to gather their information on the spot.
In practice, not all historians in this period traveled. Romans focused on Roman history, most of which was available locally through armies and legates sending word back to the senate. Thus Roman historians sometimes simply collected information without field research; they nevertheless expected accuracy in their writing. Was Luke more like a typical Greek or like a typical Roman historian on the matter of travel? If, as many contend (including myself; see discussion at Acts 16:10), the author of Acts also authored the “we” source, Acts stands much closer to the Hellenistic mold. Luke would not have traveled as widely as Polybius, but he would be acquainted firsthand with material and locations in his more detailed “we” material, including Philippi, Jerusalem, and especially Caesarea. Given Luke’s Greek style and his focus on events in the Greek East, those who contend that Luke was primarily a “Hellenistic historian” rather than a Roman one are undoubtedly correct, in any case.
Luke informs his audience that he “was thoroughly familiar with” the matters about which he writes (Luke 1:3). The verb παρακολουθέω does not specifically denote investigation or research, “although both these activities might be required in the process”; but it does indicate thorough acquaintance by one means or another. In view of other ancient prologues using the same language, Luke’s use of παρηκολουθηκότι (1:3) suggests that he has a thorough familiarity with reports (λόγων, 1:4) and that he is able to evaluate their accuracy. This was a “stock term from discussions of poetics and possibly also from rhetorical handbooks,” used to affirm one’s “impeccable credentials” for writing and advertise the writer’s reliability.
How extensive was Luke’s firsthand acquaintance and/or his investigation? As already noted, Roman historians typically consulted records; Greek historians in the tradition of Polybius traveled and consulted with witnesses. Luke’s travels allowed him to confirm many of the oral traditions circulating (1:2), traditions already known to his audience (1:4). Presumably, for the Gospel, this would have at least included interviewing Jerusalem followers of Jesus who could confirm and augment the stories circulating among Diaspora Christians. Luke’s appeal to αὐτόίται (“eyewitnesses”) fits the appreciation for research in Hellenistic historiography and has parallels in some histories.
The specific term αὐτοψία (mentioned by Meister, above) is a rare one, but the convention of appeal to such sources is much more widespread. Although the term may not be common, the emphasis on visual observation that it implied was; both Greek and Roman historians emphasized it. The ideal for ancient historians was to have seen events themselves; since this was not always possible, they often had to depend on oral sources. When they needed to defend their work, historians typically showed their reliance on oral sources and autopsy, from Polybius through the first century. Tacitus sought information from witnesses where possible (and otherwise had access to imperial annals for numerous points). When the term is used, it normally indicates “those with personal/first-hand experience: those who know the facts at first hand.” Luke emphasizes eyewitness attestation not only in his preface, but commonly enough to reinforce the preface’s claim.
That Luke’s research serves a primarily confirmatory purpose (Luke 1:4) suggests that Luke in the end does not arrive at conclusions far distant from his sources. Ancient rhetoricians often appealed to common knowledge to make a point; although this information may have sometimes been gossip, it seems unlikely that it was normally simply a rhetorical deception. Appeals to common knowledge suggest that the knowledge was in fact widespread and hence that what Luke reports in his Gospel (and, to a lesser extent, in Acts) was already in wide circulation at the time of his writing, probably within the lifetime of some who had known Jesus’s public ministry.
Many stories that Luke includes in his Gospel were probably widely disseminated, and other information may have also been broadly disseminated among many churches. The early churches throughout the empire were already informally networked long before Luke wrote and certainly long before the more explicit network of bishops that we recognize by the early second century. In Mediterranean antiquity in general, travelers regularly carried news from one location to another, residing especially with members of their own social group, such as Jews or Christians or resident aliens from their homeland (see also “Travel and Geography in Acts” in , esp. , below). In the cosmopolitan cities of the eastern Mediterranean world, such as Alexandria, Antioch, Jerusalem, Ephesus, and Corinth, one regularly found people who had moved there from various kingdoms. When one learned of someone traveling near a place where friends resided, one might prepare and send a letter.
Even during Paul’s years of planting churches, urban Christians traveled (1 Cor 16:10, 12, 17; Phil 2:30; 4:18) carrying letters (Rom 16:1–2; Phil 2:25), relocated to other places (Rom 16:3, 5; perhaps 16:6–15), or sent greetings to other churches (Rom 16:21–23; 1 Cor 16:19; Phil 4:22; Col 4:10–15). In the first century, many churches knew what was happening with churches in other cities (Rom 1:8; 1 Cor 11:16; 14:33; 1 Thess 1:7–9), and even shared letters (Col 4:16). Missionaries could speak to some churches about other churches (Rom 15:26; 2 Cor 8:1–5; 9:2–4; Phil 4:16; 1 Thess 2:14–16) and send personal news by other workers (Eph 6:21–22; Col 4:7–9). Although different locations might have their own struggles or theological emphases, the idea that the early, geographically distinct Christian communities were theologically and socially isolated from one another is a fiction created by modern scholarship on the basis of inferences incompatible with the hard evidence.
Even scholars who, such as Haenchen, doubt that the author of Acts was well acquainted with Paul can suggest that Luke could have either traveled to major Pauline centers (e.g., Philippi, Corinth, Ephesus, Antioch) or depended on informants who did so. Yet given the level of detail of Luke’s knowledge of Paul at many points (albeit in Lukan style), we would expect an even stronger acquaintance than this. Luke knows some centers’ histories (e.g., Philippi) in greater detail than others (e.g., Corinth, Antioch) but is, of course, most detailed in the “we” narratives (which include, e.g., Philippi but not Corinth or Antioch). See the discussion at Acts 16:10 for the likelihood that Luke did not simply travel after Paul but traveled with him at some points. (See also below, against the traditional claim that Luke could not be acquainted with the real Paul, a claim often based on a misleading modern Protestant interpretation of Paul.) In any case, the quality of Luke’s research seems evident from the volume of information in Acts that can today be confirmed as accurate from other sources, especially Paul’s letters; his contradictions of other historical sources are relatively few, certainly, by widely accepted ancient standards.
As a Hellenistically educated monotheist steeped in Palestinian Judaism but writing history for a Diaspora audience, Josephus may invite comparison to Luke. The comparison is less complete than one might initially suppose; since both adapt broader Greco-Roman writing conventions (which spanned a range of forms) and since Luke addresses a less sophisticated, more popular audience than does Josephus, one cannot simply reason from Josephus’s methods to Luke’s. Luke’s historiographic method may still be closer to that of Thucydides and Polybius, as some have argued: certainly they did exert continuing influence, and Thucydides was widely read. Or Luke may have gone his own way, in eclectic fashion. (Luke’s preface, at least, owes little to any specifically Jewish literary conventions, and his use of Scripture was too common in the early Christian movement to require other specifically Jewish literary models.)
Nevertheless, Josephus provides one among several possible analogies for Luke’s historical objectives, one we cannot afford to neglect. This is especially true because of Josephus’s analogous apologetic interests. Such shared interests need not require shared historical methodology, given the vast rhetorical distance between the two writers, but at the least they suggest the possibility of shared rhetorical techniques and invite further exploration here.
Clearly Josephus both preserves and adapts prior material. Josephus does not uncritically affirm all his sources; he can note that some make a particular claim (Ant. 19.60), and refuse to decide himself the cause of some events, diplomatically leaving final decisions to the reader’s discretion (19.61, 108); he can also question a source’s bias (e.g., 16.183–84). Yet we can best test Josephus’s reliability where he is reproducing accounts to which we have ready access, most often biblical accounts.
As indicated above, Josephus, like some other Jewish writers, dramatizes, adds speeches, omits what appears counterproductive, and inserts his own apologetic slant. Sometimes Josephus apparently “corrects” or adjusts biblical accounts on the basis of other biblical passages. This practice may not represent his method everywhere; he may strive for greater rhetorical sophistication in his magnum opus, the Antiquities (where we have most of his overlap with Scripture), than in his earlier War. But even here Josephus tends to adapt events in his sources, not create them.
Even when Josephus adds extrabiblical events to biblical accounts (such as Moses’s exploits as an Egyptian prince), our sources often confirm that he is following earlier extrabiblical traditions, not composing from his imagination. For postbiblical as well as biblical events, Josephus also clearly depends on sources, such as the work of Nicolaus of Damascus (e.g., Ant. 12.127), although he critiques what he sees as Nicolaus’s bias (14.9; 16.183–84).
Like other historians whose works have survived, Josephus affirms his intention to write truth in contrast to historians who he thinks are unconcerned with it (20.156–57, 260, 262, 266). This claim recalls the standard that historians demanded but does not resolve the extent to which Josephus fulfills it. (He specifically condemns a rival historian for lack of accuracy [Life 336] and compares him to a forger of contracts [Life 337–39]. In this instance, however, for all we know, Josephus might be the one lying.) He condemns those who wrote about events without any firsthand knowledge, contrasting these writers with his own knowledge of the Judean-Roman war (Ag. Ap. 1.45–49). Not only was he an eyewitness of much that he claims (e.g., 1.48–49); he claims to have even taken notes (1.49), receiving reports also from deserters from the city (1.50).
Yet whether because of conflicting sources, careless composition, or neglecting to explain information that would resolve some of the tensions, Josephus sometimes contradicts himself, even in some of his most rhetorically refined material and closest to his own lifetime. (Sometimes the apparent mistakes might instead betray deliberate literary connections, but this approach does not explain all of them.) At one point (Ant. 20.16) he appears to claim not only that Herod, brother of Agrippa I, removed a high priest but that the authority to remove high priests continued among his descendants, whereas elsewhere Josephus is clear that Agrippa II (Agrippa’s son, not Herod’s) ultimately assumed this role. Josephus sometimes forgets to include information; although promising to recount more about Helena’s and Izates’s benefactions during Jerusalem’s famine (20.53), he never returns to the subject (at least not in this extant volume).
How does Josephus usually fare in his reports of the postbiblical period? Scholars on Josephus range from skepticism to appreciation, generally depending on the aspects of his historiography on which they focus their attention. On a number of points he fails to impress—for example, in his speeches (which below will address). Most historians today also argue that Josephus’s population estimates are unreliable and that he is sometimes mistaken on distances. Since it is unlikely that Josephus actually counted people or measured distances, such matters do not otherwise affect the substance of what he reports about events. He occasionally makes mistakes on even his most public information, although this appears to be the exception rather than the rule.
Josephus’s biases are, despite his claims of straightforwardness, difficult to miss. He is probably more accurate in his general narration of events than in his summaries; he does not fabricate events, but he puts his slant on them. Indeed, his overarching summaries and apologetic perspectives sometimes conflict with the information he himself reports. He presents the Judean-Roman war as virtually an accident in which incompetent governors and a few Jewish “bandits” forced Judea into unwilling conflict with Rome. He likewise defends Titus; after all, he was writing under the patronage of the Flavian dynasty. His perspective is that of an aristocrat favoring Jewish aristocrats; in some of his work he favors Pharisees, one of whom he had allegedly been; he may also exaggerate Jewish privileges in the empire (albeit based on genuine precedents) to further his apologetic on his people’s behalf; and he presents Jewish sects in terms of Greek philosophic ideals. Josephus’s biases do not, however, significantly hamper our use of his work for historical reconstruction. “These biases are to be expected” and are usually easy enough to recognize. “With proper allowance made for his special interests and recognition that he was sometimes misinformed, the reader will find Josephus on the whole reliable.”
If we leave aside examples of bias (including in summaries), speeches, and difficult estimates, however, Josephus proves generally accurate on matters of historical detail. As noted above, whereas archaeology has challenged some of Josephus’s claims, it has vindicated him on many detailed points. Apart from distance estimates, Josephus proves generally reliable on geographic matters, and archaeology confirms most of his observations about Jerusalem, which he knew well: here, despite some errors, he is often accurate in even quite minute details, such as the measurements of columns. One of Herod’s three towers remains, 66 feet square and 66 feet high—just as Josephus claims (War 5.166).
Outside Jerusalem, the Caesarea Ancient Harbor Excavation Project has largely confirmed Josephus’s description of Caesarea’s harbor, despite some omissions in his description. Josephus appears to confuse the directions of the theater and amphitheater from Caesarea’s harbor but supplies many correct details.
As one Israeli archaeologist emphasizes regarding the physical data (where Josephus fares better than when we critique his fairly blatant biases or rhetorical distortions):
Before archaeological excavations, it was the vogue among historians—very serious historians—to argue that Josephus in many places relates sheer nonsense, that he is not historical, that he exaggerates, and so forth. But the more we dig in Jerusalem and at Masada and at Herodium and in [Herodian] Jericho, the greater respect we—both archaeologists and historians—have for the accuracy of Josephus. He is one of the greatest historians. Of course he had his own prejudices. But show me any historian without them. Josephus is accurate not only for his own period but for previous periods as well, for example, the Hellenistic period.
Despite reservations, scholars normally accept Josephus’s main outline of events as accurate. As Mosley points out, Josephus not only claims accuracy (itself no guarantee of it) but presented copies of his work to Vespasian, Titus, Herod of Chalcis, and Agrippa—and so dared not seriously misrepresent the events known also to them (Ag. Ap. 1.50–51) (particularly not to their detriment!).
How does Luke compare with Josephus? Luke’s adaptation of Mark generally sticks close to Mark’s stories and (with very rare exceptions) even his sequence. If this is representative of his use of written sources, Luke may stay closer to his sources than Josephus stays to the Bible in his Antiquities. How each adapts oral sources (probably more to the point in Acts) is harder to discern, but the use of written sources is the best guide to the question we have, and so at the very least we should not think Luke less attached to his sources than is Josephus. As argued above, Luke is not primarily a “rhetorical” historian; by contrast, Josephus in his Antiquities (more than in his War), following the model of Dionysius and Isocrates, often is. If we frequently depend on Josephus, despite his Tendenz and literary freedom, for historical information where he is our primary source, we ought not hesitate to grant the same courtesy to Luke.
Chronology was important to history and, where possible, to some other genres. Military historians such as Polybius and Thucydides could provide careful chronology because they or their sources had annals or notes available; the same was not true when one’s sources depended on oral recollections. Thus, for example, Seneca the Elder promises to recall events of the past so long as he need not do so in sequence; Papias reports that Mark recounted Peter’s reports but not in order. Biographies (such as the Gospels) were even less constrained by chronology than history was; having surveyed Augustus’s life, Suetonius promises, “I shall now take up its various phases one by one, not in chronological order, but by classes, to make the account clearer and more intelligible.” Thus Suetonius would summarize either the positive or the negative minority of an emperor’s acts before focusing on the good or evil that characterized him. Only if one specified chronology (and in the process contradicted other historical sources) did one open oneself for criticism. Sometimes historiography also required one to backtrack or advance into the future, to keep to a train of thought or a geographical region.
Luke’s Gospel follows the sequence of Mark except in a few places (most obviously the programmatic scene of Luke 4:16–30; cf. Mark 6:1–5) necessary for his own literary structure; but Mark’s Gospel itself need not be chronologically arranged, and there can be no guarantee of chronology for Luke’s other sources in his Gospel. The nature of events over the greater length of time covered in Acts allows for better estimates of sequencing there, and Luke’s fairly consistent following of Mark’s sequence may suggest that his chronology in Acts is fairly reliable. Yet even in Acts, he must sometimes restructure the chronology to retain a cohesive narrative (cf., e.g., Acts 11:30; 12:25).
When a historian was using episodic narrative “woven . . . around the impact of a personality,” similar to biography, a topical arrangement was appropriate. Even in history more generally, chronology could be negotiable. Dionysius of Halicarnassus complains that Thucydides’s chronological style breaks up his flow of thought and obscures his narrative whereas Herodotus more naturally follows the flow of events. Debates about the chronology of Acts sometimes center on Paul’s activity in Tarsus and Antioch, of which Luke tells us little; but Paul’s letters correspond closely to Luke’s itinerary for him in the places where Luke’s narrative is detailed (see below). We may not know how precise Luke’s chronology is in Acts, but were it less than fully precise, this would not count against Luke as a historical writer.
Scholars are sometimes troubled by Luke’s omissions of matters stated in Paul’s letters. Given the tension between Luke’s space and the length of the period he covers, however, numerous omissions are not surprising, and Luke is a rhetorically strategic writer who uses only what serves his purpose (which may help explain the work’s preservation). Thus, for example, if he wishes to report positive precedents from Roman officials, he may have cause to omit the reasons for which Claudius expelled Rome’s Jewish community: if the dominant interpretation of Suetonius is correct, Christians may have been involved in the dissension in Rome (see comment on Acts 18:2). Acts may likewise omit some early Christian controversies no longer relevant to Luke’s own audience; the rhetorical principle of τάχος allowed Luke to omit some details to reserve “more space for the truly momentous controversies that changed the shape of the early church.”
We can hardly deny an ancient writer’s status as a historian because of deliberate omissions, whether due to his focus (a necessary constraint in all history) or, less acceptable today, his bias. Historians often did omit some events and did not regard it as a criticism when someone else testified to the accuracy of what they reported but offered to supply additional information. Sometimes hints of information suggest that a historian who omitted it did not lack the information itself. For that matter, ancient historiographers demanded selectivity and sometimes made this characteristic a major distinction between “history” and “chronicles.” Sound historical method thus did not require the inclusion of all available information but merely prohibited the creation of events.
By modern standards, ancient writers can be accused of bias in their selectivity. Thus a modern historian who recognizes that most of what Xenophon’s Hellenica reports is accurate nevertheless considers him inaccurate for what he fails to report. So fond is Xenophon of Sparta that he omits its failures and the greatness of its rival, Thebes. But while ancient readers were willing to charge some authors with bias motivated by malice (e.g., Plut. Mal. Hdt.), standard rhetorical practice did allow them to select what they considered most essential for their purposes.
Thus one ancient rhetorical critic warned good historians not to focus too much on what was unimportant or to neglect what was important. Tacitus frankly admits that he does not treat all senate business, but only that of moral value for his audience (Ann. 3.65). The historian Polybius complains that some writers focus too much of their narrative on matters that others have treated sufficiently or on matters that would not be helpful to their readers (Polyb. 15.36.10). That Luke omits some themes to dwell on others is not surprising. Polybius insists that he must omit some matters about Roman customs (6.11.4–6), pleading with his critics that they should evaluate writers by what they recount, and assume that omissions are due to ignorance only when what they recount is found to contain errors (6.11.7–8).
Hengel observes that Luke omits what is unnecessary and sometimes “abbreviates some events so much that they become almost incomprehensible,” elaborating and combining as needed. Likewise, he argues that Luke knew Palestine’s geography well in some places (especially Jerusalem, Caesarea, and the coastal plain he would have traveled in Acts 21–24) but poorly in others (namely, Samaria and Galilee)—though like any outsider entering the land would. Getting geography as right as possible was important; some ancient historians could critique other historians harshly for erring in their geography. But Tacitus, whose interest was rhetoric, politics, and morals, also is obscure and even inaccurate in his geography of Britain. Luke’s geographic obscurity at points thus hardly disqualifies his work from the genre of history.
All this can also be found in the secular historians of Greek and Roman antiquity. On the other hand, one can hardly accuse him of simply having invented events. . . . He is not just an “edifying writer,” but a historian and theologian who needs to be taken seriously. His account always remains within the limits of what was considered reliable by the standards of antiquity. That means that the author’s assurance in Luke 1:3 is more than mere convention; it contains a real theological and historical programme, though this cannot be measured by the standards of a modern critical historian.
By the standards of modern historiography, the real problem is not that “Luke simply invented whatever he needed,” a premise for which we have no clear supporting evidence. Rather, the primary historiographic problem is “that he often passed over in silence things he thought were either unpleasant or secondary.” But it is important that we avoid anachronistically judging Luke by modern criteria; by the historiographic standards of his day such as those found in Dionysius of Halicarnassus and Lucian—and the criteria by which we evaluate the usefulness of ancient historians for reconstructing history—Luke was a good historian. As Loveday Alexander notes, Luke’s adaptation of Mark falls within the range of adaptation acceptable in ancient works; a first-century audience “would have found the reality of Luke’s work well within the range of the expectations aroused by the preface.” Likewise, W. C. van Unnik notes that Luke follows most of the rules that generally characterize ancient historiography, such as lively movement and appropriate speeches.
Arguing that Acts is a historical monograph allows us to make some statements about historical intention but does not clarify the nature of Luke’s historical endeavor in the way that modern historians of antiquity would like. Ancient historiography was a broad overarching genre providing a wide range of historical works; not every historian proved equally fit for the task or even understood the nature of the task in precisely the same terms. Some modern interpreters grow impatient with historical questions, but historical questions are also literary questions: if the genre of Acts is history, what did this genre entail, and how well did Luke conform to the general expectations that his implicit genre initially invited?
Historically, a great deal more is at stake than students of Acts itself sometimes recognize. Although the stakes should not dictate the outcome of the exploration (this approach would invoke a functionally utilitarian epistemology that could distort the historiographic enterprise as modern historians ideally practice it), they should be taken seriously by those who dismiss available sources too readily because (for equally utilitarian reasons) they do not think they will need them. Scholars often note that Acts provides some of the most complete evidence available for first-century eastern Mediterranean life from a nonaristocratic perspective. For some eastern Mediterranean cities in the mid-first century, Acts “provides the sole literary insight” from antiquity. Classicists recognize the various weaknesses of their sources, but they nevertheless find them more practical than modern alternative reconstructions offered without any extant evidence.
Although Luke-Acts is merely valuable for reconstructing some features of ancient Mediterranean life, it is indispensable for understanding the life of early Christianity. Luke may have his agendas, but without Acts we lack sufficient information to say much, beyond modern speculation about the first three and a half seminal decades of the early church as the Christian movement exploded across the empire sufficiently to provoke even the emperor’s attention. Its unique role in supplying much information about this period makes Acts indispensable for historians interested in the early church.
Those who reconstruct early Christian history on the basis of a very selective reading of Acts, such as the Tübingen school, are likely to read it in light of their own biases (in that case, Hegelian dialectic); those who exclude Acts altogether may be forced to reconstruct Christian origins almost by arguing from silence after they have eliminated the clearest concrete sources. (This practice becomes obvious, for example, in the case of many wildly divergent interpretations of Paul, often according to scholarly fashions dominant at the time, often evident among those who rule out any use of Acts and depend entirely on Paul’s letters.)
More skeptical scholars sometimes dismiss the historical value of Acts on the basis of historical implausibilities that they believe they find there, but plausibility structures are often socially constructed entities. Classicists, on the whole, seem far more ready to make use of Acts as a historical document than NT scholars are. This may be because NT scholars study Acts in more minute detail, noticing narrative inconsistencies and rhetorical adaptation in details, areas where Luke, like other ancient historians, could exercise more literary freedom than in the substance of events. But it seems noteworthy that those who work closely with other ancient historical productions tend to find Acts a useful source (even if some would be skeptical of supernatural claims in both Luke and other historians). Such external controls put Acts in context and support its utility for historical reconstruction and have generally pointed classical scholars toward respect for Acts. New Testament scholars raise questions on matters of detail; classicists appreciate Acts as a source on a par with comparable ancient sources.
Thus, more than a century ago, in the heyday of NT criticism’s historical skepticism shaped by Tübingen’s academic orthodoxy, William Ramsay advocated a second-century date for Acts and viewed it as an imaginative work. It was, however, his archaeological investigations of specific localities that led him to a first-century date and trust in Luke as a historian. In 1921 Alfred Wikenhauser compared Acts with secular sources and found few divergences; later the respected specialist in Roman law A. N. Sherwin-White found that Acts fit the civic structures and laws of the early empire, showing a writer well informed on details in the way that a later writer could not have been.
Our respective perspectives are also shaped in part by our perception of our pedagogic and scholarly task. New Testament professors often emphasize Acts’ “unreliability,” using modern history as the foil and seeking to wean modern students from their traditional assumptions that fail to distinguish events themselves from the narration of them. By contrast, those who use ancient historiography as the foil generally respect Acts much more highly because Acts fares well by the standards of ancient historiography. As noted above, Luke wrote in the genre of ancient historiography, not in the as-yet nonexistent genre of its modern successor. Although debunking anachronistic genre assumptions of students may be a noble task, the very comparison with modern historiography, if unqualified, runs the serious risk of playing into these anachronistic assumptions.
How important is internal consistency as a historical criterion? From a literary perspective, we should not press our demands for consistency beyond what could work for other historical sources; even apart from some legitimate insights that deconstruction provides concerning the limitations of any narrative, some of what appears inconsistent is often a deliberate portrayal of weaknesses or narrative tensions. In historical works, some apparent inconsistencies stem from incomplete reporting or the blending of equally incomplete sources.
But granted such limitations, ancients themselves recognized that historical accounts should fit the other historical data available. Thus Polybius challenges a claim of Fabius as implausible, reasoning from undisputed historical evidence. He subsequently allows that Fabius may be right sometimes, but demands that critics evaluate his claims by examining the facts. Even the “purest” of historical reports may contain some events that seem implausible, but they are likely to fare much better (and are expected to fare much better) than carelessly fabricated or novelistic reports. Yet provided we allow for the sort of inconsistencies (real or apparent) and reports of supernatural phenomena (or claims of such phenomena) that characterize most ancient historians, plausibility remains a useful criterion for historical inquiry. We should be able to test Acts in one of the ways that Polybius and others suggest we evaluate historians, namely, by examining the plausibility of accounts.
This commentary will pursue the question in specific passages, but we should note here that plausibility proves not necessarily an account’s accuracy but Luke’s ability to write cohesive historical narrative; this is, first of all, more a literary than a historical question. Indeed, a poor writer with good sources could produce an implausible narrative yet preserve much accurate information, just as a skilled writer could make fiction sound plausible (a skill that orators employed both in law courts and in encomia). Nevertheless, in general, it counts more in favor of a narrative’s accuracy if a writer fairly consistently provides internal cohesiveness and plausibility than if he or she generally writes implausibly. We would expect Luke to meet this criterion more readily where he writes at greater length and in more detail (especially in “we” material) than where he offers primarily summary reports enlivened with speeches.
On the whole, scholars seem more appreciative than not of Acts as a legitimate source for historical reconstruction, but this appreciation appears far from unanimous. Some scholars are fairly skeptical regarding the historical reliability of Acts, especially voices from the Acts Seminar, successor to the notorious, self-selected Jesus Seminar. Many studies today welcome discussions of “history” and Acts (hence Acts’ historical context) but reject questions of its historical reliability as representing an older era’s approach.
Although the current emphasis on historical context is paramount for interpreting Acts, rejection of reliability questions altogether seems to me inconsistent with Acts’ genre: if Luke writes as an ancient historian, he makes claims about history. We cannot hold him to the standards of peer-reviewed modern historiography, but neither can we isolate his narrative world from the earlier real world he emphatically claims (Luke 1:3–4) that it emulates. He is not writing about mythical characters but about Peter, Paul, Agrippa I, Gallio, Felix, Festus, Agrippa II, and so forth.
It is therefore legitimate to investigate the degree of correspondence between Luke’s portrait of the events and what can be reconstructed of the events (where they can be reconstructed at all)—that is, to ask where on the broader continuum of ancient historiography Luke fits. Granted, we cannot reconstruct historical details with certainty, but many literary interpretations of the works likewise remain debated. That a task cannot be completed perfectly is not necessarily reason to abandon it; otherwise we would have to abandon inquiry about ancient history altogether.
Because Luke’s narrative, where we can compare it with external data, corresponds with the real world far more often than not, scholars who focus on these data are impressed with his carefulness. In contrast to some earlier generations, many scholars today have grown in their appreciation of Luke’s historiography. Hengel thus opines that “Luke is no less trustworthy than other historians of antiquity.” Michael D. Goulder, comparing numerous details in Acts with Paul’s letters (see below on this subject), argues that Acts is a work of ancient history and points out, “One may get other matters wrong, and one may tell the story with a spin on it, but one cannot be repeatedly accurate unless one is well informed.” Or as Luke Timothy Johnson puts it:
Taking into consideration his use of the one source we can check [i.e., Mark], his general accuracy in matters we can otherwise confirm from archaeological or documentary sources, and the overall agreement between his account of Paul and that in the letters, we can conclude that Luke is accurate in what he tells us, certainly by the standards of Hellenistic historiography (Hengel 1979). The phrase “in what he tells us,” however, is critical. Luke writes selectively.
C. K. Barrett notes that we lack certainty at many points but that the story must have happened in much the way Luke tells it, to yield the outcome that history reveals that it did.
Many scholars thus argue for Luke’s general dependability as an ancient historian, judged by the criteria that we use for other ancient (rather than modern) historians. They conclude this even if most of these scholars also allow that Luke (like other historians) reported some historical inaccuracies (most often found in Luke 2:2 and Acts 5:36–37). Given Luke’s freedom to rewrite details in passages such as Luke 24:39–52 (esp. 24:47–52) and Acts 1:3–12 (or even the slight differences among Paul’s conversion accounts), Luke clearly does not work from a modern perspective and would hardly be as disturbed by questions about the rarer proposed exceptions (e.g., Acts 5:36–37) as some of his subsequent readers have been.
To judge Luke’s integrity by modern assumptions of how he should have written, rather than by his objectives and standards as a Christian Hellenistic historian, disrespects Hellenistic historians in general, violating our own historical canons against anachronism. All extant ancient historians contradict others at some points.
Some scholars defend Luke even at Luke 2:2 and Acts 5:36–37, but if evidence allows us to accept Luke’s accuracy in at least most cases, this is sufficient to keep the burden of proof where it belongs: except where one has compelling reason to believe otherwise in particular instances, Luke’s reports about events should normally be respected. Provided we also apply other tests at our disposal in relevant cases, we expect Luke to usually reflect events in his sources if he does so where we can test him. Apart from the more extreme postmodernists who are skeptical about most ancient history, detractors against this position apply a skepticism that they would not apply in normal historiography. (When scholars treat what has become Scripture as a special case, they may reflect a canonically determined bias, even if in their case the bias is against the material.)
No one regards the speeches as verbatim, but few scholars would regard Luke’s speech summaries as less historical than the fleshed-out speeches of other historians of his time. (Indeed, he might fare generally better than Josephus; see below). I believe that the evidence probably argues against the usual “conservative” chronology of Acts 15 and Gal 2, but this is a problem for one scholarly position, not for the historical nature of Acts. In the Gospel, the divergences from Mark and possibly from material shared with Matthew are all within the expected literary freedom of ancient writers.
The radical skepticism of the old Tübingen school retains few advocates today; indeed, on the whole, extreme critical approaches have proved too subjective to yield consensus. (Even Haenchen’s skepticism is too severe, applying a standard of suspicion under which even Polybius and Thucydides would fare badly. Contrary to the views of some modern critics, the NT writers knew more about their subjects than we modern critics can at our remove.) Although radical views such as those associated with nineteenth-century Tübingen have not vanished, scholarship today is divided especially between those who argue for a high degree of Acts’ historical reliability and those who allow that it is at some times reliable but that this is not Luke’s focus. Most scholars today would probably concur with Fitzmyer: the genre and external corroborations of many points in Acts do “not guarantee, of course, the historicity of every Lucan statement or episode, but” they do reveal “that what is recounted in Acts is substantially more trustworthy from a historical point of view than not.”
By many modern standards, Luke may have taken significant literary liberties to produce a coherent narrative rather than a documented chronicle. Even for his own day, Luke provides less “documentation” of his sources than the generally longer and more rhetorically sophisticated histories of his day. But Luke is accurate where it counts for popular historical writing of his day: he probably participated in some of the events he narrates; relevant for more of his narrative, he interviewed eyewitnesses where possible and wrote of relatively recent events. So long as we recognize and take into account that Acts is a single-volume, popular-level apologetic history, we should give Luke the same courtesy of trust for historical events that we accord other competent ancient historians discussing recent events.
When all has been examined, a number of scholars rank Luke among the more accurate historians, such as Polybius, Thucydides, and Tacitus, though others would view this verdict as excessive. We must note again that Luke’s rhetorical techniques reflect the fact that he writes for a somewhat less demanding, more popular audience than such historians (but, on the other side of it, less interested in rhetorical flourishes and other rhetorical expectations). Yet even most of the scholars who are more skeptical accept that Acts has historical value.
Despite some scholars’ reservations, it seems most logical to presume that Luke’s standards of reliability where we cannot test him are roughly equivalent to his standards where we can. This is the same procedure we would apply to establish (more narrowly than genre alone permits) the relative burden of proof for the factual accuracy of any ancient historian; writers could not know what information would remain extant, even in the unlikely event that they wrote only for future generations; thus the burden of proof rests heavily on interpreters who expect the degree of reliability to decline precipitously where we cannot test it. Of course, while it is not our focus here, the test also cuts the other way: the evidence will suggest that Luke has considerable information, increasingly detailed in later chapters, but he also takes the ancient historian’s storytelling liberties to narrate the story in a cohesive, interesting, and edifying way.
We can test the historical accuracy of Acts by noting its relatively conservative use of sources and the likely reliability of these sources (as noted above). We can also test its reliability by comparing it with other sources that cover the same material and then determining which sources most consistently represent the likely consensus of ancient evidence. Paul’s letters, as a primary source, provide the most significant test case for Acts, as Q and Mark do for Luke’s reports about Jesus in the Gospel; here, however, the issues are significant enough to merit a separate chapter on Paul and Acts (ch. 7 below). We must also compare Acts with its extratextual setting wherever the opportunity arises. I shall survey such correspondences here, but this more detailed task belongs especially to the commentary that follows. First I mention some of the other evidence, though I will treat most of it in detail only under the requisite passages.
Leaving aside places where we cannot test the work, external evidence corroborates Luke’s accounts at many points. One area of interest here is Luke’s accuracy in titles for local officials, which changed from one locale to the next and sometimes within a locale from one decade to the next. Whereas Agrippa I and II are rightly “kings,” Antipas is a tetrarch (in contrast to Mark 6:14; cf. the correction also in Matt 14:1); Philippi’s magistrates are praetors with rod-bearing lictors; Thessalonica has its “politarchs”; Asia has its “Asiarchs”; and “the chief municipal officer of Ephesus is the ‘town clerk.’” Long ago Cadbury observed that just as one can evaluate a person’s
knowledge of modern Oxford and Cambridge by his ability to name correctly the presiding officer of each college whether as Master, Principal, Provost, Warden, rector, President or Dean, or one’s knowledge of old Germany’s political constitution when more than three hundred territorial sovereignties existed within the Holy Roman Empire ruled by kings, electors, and princes, counts and counts palatine, dukes and grand dukes, landgraves, margraves and all the rest, so one may test Luke’s knowledge of municipal institutions in Aegean cities. His language fully meets the test.
Charles Talbert notes that congruence with the period depicted does not by itself establish historicity, but he also notes that the “points at which the contemporary color of Acts can be challenged . . . are few and insignificant compared to the overwhelming congruence between Acts and its time and place.” Likewise, he contends that the sequence of events in Acts fits external history and Paul’s letters fairly well (except Gal 2:1–10). External sources confirm numerous points.
Colin Hemer provides a massive list of such correspondences, including (for a small sampling of his examples):
Not all of these examples are equally compelling, but the cumulative weight of the best of them should be sufficient to dispel any notions that Luke is writing a novel rather than ancient historiography or that Luke is a careless historian unconcerned with detail. Hemer also notes numerous internal correspondences that suggest consistent information on which Luke never capitalizes for his narrative—for example, “we” leaving off and resuming at Philippi (16:17; 20:5–6) and again in Jerusalem (21:18; 27:1) or the assumption of Paul’s personal possession of funds (21:24; 24:26; 28:30).
Some scholars have simply dismissed such historical correspondences as (in the case of Hemer, though not Cadbury or Talbert) the work of a “conservative” scholar (academic shorthand in some circles for a work that does not come to the appropriate conclusions and hence may be dismissed as closed-minded). This is, however, an ad hominem label rather than an argument. Hemer provides abundant data, more often than not data difficult to contest. Such correspondences do not prove that everything Luke wrote is factual, but they do indicate that he had access to solid information; he may have adapted it, but he was not ignorant of it. And serious historical interest would limit the degree of adaptation in a way that a primarily novelistic approach (or very careless historiography) would not.
As already noted, it was an ancient historical work, rather than an ancient novel (despite some displays of local color), that would be typically confirmed on incidental matters by external data. Such a description certainly suits Acts. The key test is Paul’s letters (see below), but other extrinsic data provide helpful tests as well.
Although we should not expect any two ancient historians to agree on every detail, we find for Acts numerous confirmations of substance that fit the mainstream of historical but not of novelistic works. For example, Josephus and Luke concur that Agrippa died after accepting public adulation in Caesarea (Acts 12:19–23). Moreover, some objections to Luke’s reliability are ill founded. Whereas Haenchen complains that Gentile soldiers were not stationed in Judea during Agrippa’s reign, Josephus explicitly tells us otherwise (Ant. 19.357; though we cannot be sure the Cornelius episode happened during Agrippa’s reign, in any case). Following are examples in support of Luke’s usual accuracy (some overlapping with those offered by others above).
Although accurate local color does not prove reliability, obviously it is consistent with it, especially when Luke would have had to travel or depend on informants to obtain detailed information (i.e., where it was not simply common knowledge). As Richard Wallace and Wynne Williams, both classicists, observe, “It is the accuracy over quite obscure details which is striking.”
Luke gets correct the local color in Ephesus; for example, Ephesus did sometimes publicly defend the Artemis cult, Asia Minor’s cities had many riots and much unrest in this period, and economic concern for the Artemis temple was at a peak in precisely these decades. In addition, Luke’s portrayal of Ephesians calling Artemis both ἡ θεά and ἡ θεός fits local Ephesian usage (rare outside Ephesus). Even what we can reconstruct of the progress of the riot in his account (admittedly hypothetical, given uncertainties over some of our archaeological data there) appears to fit the local topography of Ephesus.
Far more impressively, Luke correctly describes details of the Anatolian interior (such as geography; key sites, occurring often on known roads and in the right sequence; and the greater than usual linkage of Zeus and Hermes), yet these are not (and appear before any) “we” narratives and Luke makes no claim to have traveled to these locations. Nor had most of his Aegean contemporaries; it is unlikely that Luke relies on a (non-Pauline) source for his geography here. Today, as another commentator notes, “even skeptical critics” are impressed with Luke’s accuracy on “geographical, social and political minutiae,” but “reference works that would make such research possible did not exist in the first century.” The relative inaccessibility of such information suggests that, to acquire it, either Luke would need very reliable information from someone who traveled or he himself would have had to travel in the period described—with or without Paul. Yet he does not write himself into narratives in the Anatolian interior and elsewhere (as someone seeking to validate a fiction might). (This commentary treats most of these correspondences in the comments on the specific passages where they appear, but some scholars have documented such information at much greater length.) Likewise, Jerusalem in Acts becomes increasingly dominated by nationalism in precisely the decades for which Josephus reports its growth, yet Luke neither draws attention to this shift (as he might if deliberately trying to highlight his realism) nor obscures this reality (as a novelist pursuing free composition undoubtedly would).
Other elements may contribute to an impression of historical accuracy. Detail by itself does not; it appears in poetry and drama as well as history, and in poetry and drama it is a sign of invention. But Luke provides far more detail in the “we” narratives, where he was present (or, in some views, had an eyewitness source), than he does elsewhere, a pattern consistent with what we would expect of genuine reminiscences. Likewise, he provides more historical detail where his sources are better (more recent events known to Paul) than in the early chapters of the book. This conforms to what we would expect of a fairly conservative historical method, not the relatively free composition of a novelist.
Luke’s narrative arena in Acts contains real geography (often known to his audience, especially in the Aegean region) in quite recent history, in contrast to novels. Such settings demanded more accuracy than the distant times or exotic locations sometimes featured in other kinds of works. When Luke speaks of Paul’s conflicts in synagogues of specific locales, or the behavior of local authorities, or the founding of local churches, he reports matters that may well be preserved in local memories at the time of his writing. Local churches could dispute his assertions; synagogues could treat what they heard of his reports in the way they responded to and generated other polemic. Luke could not afford to get his basic facts wrong if he wished a wide readership, especially in the regions on which his narrative focuses. And it seems likely that just as Luke is disposed to quote biblical texts accurately, he would also wish to communicate accurately the history of his community. Although the former belonged to the community’s common repository of information, it appears that some of the latter did as well (Luke 1:4).
Gospel scholarship has discovered over time that the criterion of dissimilarity works effectively only in a positive way (affirming material), not a negative one (excluding it). This is because we should not expect the historical Jesus to have differed on every point with either his environment (whose language and imagery we would in fact expect him to use) or his followers (who, as disciples, would in fact be expected to respect, preserve, and propagate much of his teaching). The criterion of Palestinian environment thus effectively cancels out that of dissimilarity fairly often. In the case of Acts, we cannot exclude a claim’s accuracy merely because it fits what we know of early Christian beliefs; we dare not assume that what was useful to Luke (whose Christian faith was undoubtedly nurtured in the setting of early Christianity) could not have also happened in early Christianity. Historians often selected historical material that strengthened their case, but no one supposes that whatever fits their case must therefore be fabricated. Thus, as with the Gospels, the criterion proves more useful in its positive form. Some apparently non-Lukan material, or material embarrassing to later Christians (such as Paul’s split with Barnabas), is unlikely to have been invented.
Although Luke’s sources for Paul’s life, especially in the “we” material, are more thorough, many scholars believe that Luke also preserves genuine traditions about the early Jerusalem church. He portrays some of the information as being well known in Judea (Luke 24:18; Acts 10:37; 26:26); although we could accuse him of fabricating this portrayal, such fabrication would have been riskier in his own day when presumably some people remained—especially scattered refugees from Judea—who could have discounted such claims (cf. Luke 1:2, 4). More concretely, his Pentecost narrative seems to presuppose, regarding the festival, some Jewish Christian theologizing on which he himself fails to capitalize; perhaps the Christology and more clearly the imminent end expectation portrayed in Acts 3:12–26 may also suggest some knowledge of ideas that circulated in an earlier period of the Jerusalem church, ideas that Luke would not, apart from his historian’s mantle, feel as much need to emphasize.
Luke’s portrayal of Pharisaic tolerance (e.g., 5:34–39) fits much of what we know of first-century Pharisees outside the gospel tradition. Likewise, we would probably not expect Luke to invent the complaint of the widows in 6:1 and then work to soften the charge. Nor would he likely invent the account of Philip’s ministry to the African court official, an account that seems in tension with the church’s official account of Cornelius (which Luke also approves). His information regarding the official fits ancient historical accounts of Meroë better than it fits ancient fictional ones.
The Hellenists’ mission in Antioch in 11:19 is incidental to Luke’s focus on major characters such as Barnabas and Paul (who arrive only afterward) and hence probably reflects pre-Lukan information too widely known to ignore. Luke’s tradition about the origin of the title “Christians” in Antioch (11:26) is also likely, since Luke himself prefers different titles for Jesus’s followers (and “Christians” does not appear complimentary when employed in this period). Although a global famine (11:28) may be Lukan hyperbole (a not uncommon ancient device), many data corroborate widespread famine conditions during much of Claudius’s reign (though this reign was admittedly probably still often remembered in Luke’s day).
Luke’s report of Peter’s escape, often considered legendary, fits both the timing of Agrippa’s brief reign and archaeological data for Jerusalem’s topography (see comment ad loc.). Luke’s report of James’s death in 12:2 appears incidental to his focus on Peter’s story and is historically likely. Because the house of John Mark’s mother is spoken of as well known, Luke probably is drawing on authentic local tradition for it. Luke’s narrative seems to presuppose knowledge of John Mark (12:12) and James the Just (12:17), suggesting that Theophilus had heard somewhat more than (or at least a fuller version of) the gospel tradition itself (Luke 1:4). That some degree of leadership would pass to James the Lord’s brother (Acts 12:17; 15:13–22; 21:18) would fit Middle Eastern kinship expectations (although Luke does not specify the sibling connection, which appears in Gal 1:19; 2:12).
Agrippa’s death in Caesarea agrees, on the essentials, with Josephus’s account, and Agrippa’s personality in Acts (see esp. comment on Acts 12:1, 3) fits the more negative side of his personality in Josephus. Although each source mentions and emphasizes points the other omits and hence neither is likely dependent on the other, the overlap is considerable:
Ant
Acts 12:19–23
Agrippa was in Caesarea (19.343)
Agrippa was in Caesarea (12:19)
Setting of games in the theater in honor of Caesar; no mention of the embassy (19.343–44)
βῆμα
δῆμος
*
Mentions Agrippa’s glorious robe as a cause for praise (19.344)
Mentions his royal apparel, though without details that would explain why this mention is important (12:21)
No mention of Agrippa’s speech before he is struck, but a rhetorically apt one is composed for him afterward (19.347)
Agrippa is speaking when he is praised (12:21)
†
Flatterers acclaim Agrippa as divine (12:22)
Agrippa struck just afterward (19.346–48)
Agrippa struck just afterward (12:23)
Because he did not rebuke the acclamation (19.346–47)
Because he did not defer the glory to God (12:23)
‡
§
He died (19.350)
He died (12:23)
*A theater was a natural meeting place for the δῆμος, as would be games.
†They also plead for mercy (“Be propitious to us,” Jos. Ant.19.345 [LCL, 9:378]), which might fit Luke’s context of at least the embassy (Acts 12:20).
‡Eusebius treats Josephus’s account here as corroborating Acts (H.E. 2.10.10; Witherington, Acts, 390n112).
§That the duration is unspecified may fit the nature of Luke’s tradition or interests (cf. Acts 9:23).
Much of Paul’s ministry in Paphos fits external evidence. One Lucius Sergius Paullus is “the only senator attested from this generation of the family” and hence, the classicist Stephen Mitchell concludes, surely the one Paul met in Cyprus. Acts 13 and the inscription in question both fit together well with the way senatorial careers usually proceeded in this era and could indicate the same individual. Given the rareness of the name and the normal pattern of a senatorial career, Mitchell is also convinced that this would be the same Sergius Paullus who became consul suffectus in 70 C.E. (CIL 6.253). Various rulers depended on magicians or similar sorts of advisers, and Sergius Paullus, who was a Roman from southern Anatolia, grew up surrounded by Eastern culture and perhaps was more open to it than most Roman officials would have been; for what it is worth, evidence also suggests that there may have been other Jewish magicians in Cyprus. Perhaps most significantly, the Sergii Paulli, though Italian by descent, “were native to Pisidian Antioch,” which fits historically the next location of the team’s ministry even though Luke never clarifies the connection.
That Luke mentions only Perga (Πέργη, normally transliterated outside the NT as Pergē) at Acts 13:13 and only Attalia on the departure of Paul and Barnabas at 14:25 fits historical reality, since Perga was the appropriate goal if they wished to reach the Via Sebaste for travel into the inland highlands whereas Attalia was the more appropriate goal for a departure port. Most of Luke’s portrait of events in Pisidian Antioch fits what we know of the town, as does Paul’s interest in evangelizing there (for details, see commentary loc. cit.). Even some non-Lukan midrash in Paul’s speech there appears to be pre-Lukan (also comporting well with Pauline language there), although we do not know Luke’s source (unless it was, if anyone, probably Paul himself and probably orally), and it is tenuous to automatically infer that something being “pre-Lukan” necessarily makes it what Paul must have said on an occasion.
Moreover, in view of Paul’s ministry in prominent locations such as Corinth, who would invent ministry in such out-of-the-way towns as Iconium and Lystra? As mentioned above, the pairing of Zeus and Hermes is well attested at Lystra, and it seems unlikely that Luke acquired such information by visiting this site personally; other elements of local color are also dominant (see comment ad loc.). Luke’s compressed account there also suggests traces of prior information (such as mention of “the gates,” without clarifying what these gates are, 14:13).
James’s speech in 15:13–21, like some other speeches to Jewish audiences in Acts, seems to imply Jewish exegetical arguments that Luke himself only summarizes and nowhere develops (for a more concrete argument for the Jerusalem Council itself, see below). The conflict between Paul and Barnabas (15:37–39), as something Luke probably would not have invented, meets the criterion of embarrassment; so, most likely, does Luke’s implied portrayal of Mark. This embarrassment leads to another: the new mission initially appears to wander aimlessly now that Paul lacks Barnabas’s help and has pressed beyond the previous sphere of ministry (16:6–9). It is not the sort of report Luke would likely invent about his hero.
The known timing of both Gallio’s governorship and particularly explicit anti-Jewish sentiments (though the latter could have also occurred at other times) fits Luke’s description in Acts 18 well. Paul’s letters indicate his friendship with Prisca and Aquila; Luke’s additional specification that Aquila hails from Pontus (which neither is important to his narrative nor reflects a particularly high-status location in this period) makes little sense except as historical tradition. Given the relatively low status of artisans, Luke is unlikely to have invented such activity for Paul, Priscilla, and Aquila (18:3) in a document dedicated to a person of status (Luke 1:3; Acts 1:1).
The riot in Ephesus also likely indicates historical tradition. Luke plays the events of this section for all they are worth dramatically, but he would not likely invent them. To invent a riot associated with Paul’s activity would weaken the apologetic case Luke so carefully cultivates in much of his work against the charges that Paul stirs riots (See comment on Acts 24:5); he is at such pains to explain Paul’s innocence (here and on other occasions) that creating the problem to begin with makes little sense. Given his apologetic considerations, it is far more likely that Luke toned down the repercussions of the incident than that he created it. That Luke could not deny Paul’s involvement suggests that there were people who recalled that the unrest there related to his missionary activity (cf. Acts 21:27–28). That Paul later avoids Ephesus (although Luke provides a less suspicious reason) suggests that Paul experienced more decisive hostility there than in most other locations. Luke also appears to assume knowledge of a wider body of information than he reports, mentioning Alexander the Jew and Paul’s relation to the Asiarchs as if they were known (19:31, 33–34).
As mentioned above, Luke gets the correct local color for Ephesus. More specifically, local elements of the account reflecting the spirit of Ephesus include the following:
Cumulatively, such details have persuaded many scholars that Luke displays considerable accurate information, at least about Ephesus. As the Oxford Classical Dictionary points out, Luke here “gives a vivid picture of the Artemisium’s religious and economic importance for the Roman city.” Historian Fergus Millar observes, concerning Luke’s report of this incident, “No text illustrates better the city life of the Greek East, its passionate local loyalties, its potential violence precariously held in check by the city officials, and the overshadowing presence of the Roman governor.” Thus many scholars, though stressing Luke’s style, affirm that he depends on a report of an authentic event. Nevertheless, taken by itself, local color could reflect mere knowledge of Ephesus rather than of the events reported; what carries more weight is the testimony of Paul’s letters (see below).
The careful itinerary in 20:13–15 and 21:1–3 and their contexts, belonging to “we” material, serves no overt theological purpose, yet as one early twentieth-century scholar remarked, “the narrative portion of the twentieth chapter of Acts is full and correct enough for a guide-book.” Moreover, as Barrett notes, Paul’s passing Ephesus and meeting the elders in Miletus poses the sort of problem that Luke (who was present on the voyage, 20:15) would not “voluntarily concoct.”
Likewise, the Tyrian prophecies in Luke’s “we” narrative (21:4) meet the criterion of embarrassment: given the mistrust of failed prophecies or of those that contradicted other sources of knowledge, Luke presumably would not have invented a prophecy worded as if prohibiting Paul from doing what he in fact did (potentially allowing less sympathetic readers to infer that, against Luke’s character portrayal of Paul, Paul’s disobedience to prophecy led him into his trouble with Rome). Although we might assume that narrators would clean up such tensions, such narratives probably reflected the ambiguities of genuine charismatic experience (1 Cor 13:9–12; 14:29; 1 Thess 5:20–22).
In one letter written from Corinth, Paul expected trouble in Judea (Rom 15:31), and we next meet him in Roman custody (e.g., Phil 1:13; Phlm 1, 9). Romans normally did not arrest people without either catching them in a subversive act or receiving accusations from locals. Yet local accusers probably would not have gently handed Paul over to Romans on the charge of violating the temple (which, our evidence suggests, at least a sufficient number of them would have been prepared to avenge more directly); more clearly, Luke would not have missed the opportunity to parallel Paul and Jesus if the Sanhedrin arrested Paul and handed him over to the governor (cf. Acts 24:6). A mob scene such as Luke describes in his “we” narrative is therefore plausible. While it makes for good adventure, it also supports the charges of stirring riots (24:5) that Luke is at pains to refute in his own narratives by blaming Paul’s enemies. Luke would hardly invent a situation that could be construed as supporting the very charges for which he frames his apologetic narrative.
The one offense for which Jerusalemites could automatically kill an offender, even (Josephus claims) a Roman citizen, was desecration of the temple; this is precisely the most plausible charge if we grant that Paul was nearly lynched in the temple (for stronger corroborative support from Paul’s letters, see below). That the report “came up” (ἀνέβη) to the chiliarch in charge of the garrison (21:31), that the soldiers “ran down” (κατέδραμεν, 21:32), and the later mention of “stairs” (ἀναβαθμοί, 21:35, 40)—all these details fit the topography of the temple. Once Rome had intervened and taken Paul into its custody, however, he could be condemned only after a proper investigation. The portrayal of subsequent political influence and legal corruption in Paul’s custody arrangements is certainly believable; ancient reports abound with these factors (See comment on Acts 24:26), and legal “justice” functions this way against the politically weak in many parts of the world today. Of course, these are matters of plausibility rather than of external corroboration.
Though appearing in conversation, even the relative citizenship and status claims of Paul and Claudius Lysias fit the period in question better than Luke’s own. Dio Cassius indicates that Claudius bestowed citizenship freely and almost indiscriminately (Dio Cass. 60.17.5). His successor, Nero, cracked down on cases of corruption in the previous administration’s imperial favors (e.g., Tac. Ann. 14.50), suggesting that Luke’s portrayal of Lysias’s acquiring citizenship fits the narrow window of time during Claudius’s rule (41–54 C.E.). The tribune’s name, “Claudius Lysias” (Acts 23:26), supports this picture. “Lysias” is a Greek name, suggesting that he was born in the East. The name “Claudius” reveals that he became a citizen between 41 and 54 C.E.; new citizens normally took the name of the emperor under whom they were enfranchised. That Luke nowhere draws the connection between this incident and the tribune’s name (reported in a separate location in Acts) and that the tribune’s name would be preserved in the court’s official documents suggest that Luke’s information here is historically dependable. The verbal exchange that Luke reports (for which Paul could have been the source) might even reflect the shift during Claudius’s reign: early in that reign, Lysias had to pay a high price for citizenship; thus he could wonder whether Paul was one of the newer citizens who had obtained the franchise cheaply toward the end of Claudius’s reign (Dio Cass. 60.17.6).
Luke would have hardly gone to the trouble of securing precise names for the governors of Judea, in the right sequence, for his period if he were writing a novel; the roles (and characterization) of Felix and Festus, on the other hand, fit well a work of history. Even Felix’s question of Paul’s province (Acts 23:34) and his immediately subsequent decision to decide the case himself (23:35) fit Paul’s lifetime better than a period after 70 C.E. and especially substantially afterward. Cilicia Pedias, like Judea, was in precisely this period part of the larger Roman province of Syria-Cilicia (cf. Gal 1:21). Cilicia became a separate province again only under Vespasian in 72 C.E.; it was thus separate again when Luke was writing, on the post-70 dating assumed in this commentary. Because Paul belonged to the very administrative sphere in which Felix worked, transferring the case was less appropriate. Like many other points in Acts, this detail reflects a situation much more relevant to Paul’s day than to Luke’s, with Luke having nothing to gain even in credibility by so presenting it (since much of his audience would not have known this detail). Thus Luke likely reflects accurate historical information about a real conversation at this point.
The portrait of Paul’s final trials also comports well with external data. Acts approximates so closely the procedure of Roman law, indeed, that many Roman historians have treated it as the clearest illustration of how governors in the provinces conducted such trials. A. N. Sherwin-White, the one Roman historian who has written on the matter at length and with particular expertise, has adequately demonstrated the legal accuracy of many features of Luke’s accounts, sufficiently challenging the unwarranted skepticism of Haenchen and others. Haenchen is characteristically skeptical about the substance of Paul’s defense speeches, not distinguishing them from other kinds of speeches in the book, but they fit what one expects for records of official proceedings, which would have been available to the defendant and were normally summaries of roughly the length Luke provides.
The emperor to whom Paul appeals in Acts 25:10–12 is Nero, yet to Paul’s detriment at some later hearing, Nero turned out to be vicious against Christians in 64 C.E. In 54–59, however, Nero remained under the moderating influence of Seneca the philosopher and Burrus the praetorian prefect. This distinction fits the historical context of Paul’s time better than the perspective dominant when Luke was writing, by which time Nero’s repression of Christians would have dominated Christian thinking about Nero. Would Luke have invented an appeal that could have made Paul appear foolish or lacking in prophetic insight?
Even apart from this portrayal, there is no compelling reason to doubt Paul’s appeal. Roman citizens would have been sent to Rome for trial more frequently than other persons; thus later the Christians whom Pliny detained and who were Roman citizens he sent to Rome for trial (Pliny Ep. 10.96.4) instead of simply executing them as he did the others (10.96.3). One could counter that Festus could have sent Paul to Rome even without the appeal that Luke here describes, but such an approach would fail to explain the coincidence between Paul’s destination in Rome while in Roman custody and his explicit earlier hope to visit Rome after Jerusalem (Rom 15:23–28). This “coincidence” more readily supports the thesis that Paul may have had something to do with his Roman destination (albeit with the procurator’s cooperation), just as depicted in Luke’s narrative (Acts 25:10–11).
A cover letter concerning Paul’s case would surely have been sent to Rome with him, from which Luke might infer the sort of favorable verdict he reports in 26:30–32 (although not the exact words). Had Agrippa thought Paul guilty, Festus surely would have handled the matter in a more politically advantageous manner instead of sending him on to Rome and increasing the imperial committee’s workload. Paul’s citizenship explains why he is also not handled in a manner more politically expedient vis-à-vis his high-status local accusers. On the usual first-century datings of Acts, Agrippa II remained alive and respected at the time of initial publication, so that Luke would hardly have dared to misrepresent Agrippa’s interest (which even in Acts is more a matter of neutrality than commitment; See comment on Acts 26:28).
Scholars have often noted the detailed correspondence of weather and topography patterns between Paul’s voyage to Rome and what is known from subsequent mariners’ testing (27:1–28:14). These connections are developed in greater detail in the commentary; here it is sufficient to say that the account must reflect the observations of an eyewitness, not unbridled imagination. Minimalists could excise from this account the participation of Paul, but one should note that they can delete this element precisely because it is the only element not available for external testing. Yet it seems intrinsically likelier that Luke recounts a genuine experience on this voyage (which Paul certainly undertook, traveling from the east to Rome) than that he discovered another genuine storm account that coincidentally shared the same itinerary taken by Paul and the other members of his “we.” (It would not have been Paul’s first storm, 2 Cor 11:25.)
Where we can test Luke most fully, however, is where his testimony overlaps with Paul’s letters (ch. 7 below explores these connections at much greater length; the exploration is even more detailed in the commentary on the relevant passages).
The twentieth century’s most renowned historian of Greco-Roman antiquity, Eduard Meyer, opined that Luke was a great historian and that Acts, “in spite of its more restricted content, bears the same character as those of the greatest historians, of a Polybius, a Livy, and many others.” Some interpreters have criticized Meyer and other historians of antiquity for depending on Acts, complaining that they write outside their specialization. But the questions of how first-century readers would have understood, and how historians of antiquity best evaluate, Acts’ historical claims are in fact questions not of theology but of ancient historiography, especially when Luke’s overarching perspectives are no more pronounced than those of other ancient historians. Whose specialization is better to offer such analyses than specialists in Greco-Roman historiography? Indeed, as one historian notes, where the details can be checked, Luke is, on the whole, “marked by carefulness” far more than Josephus is. Josephus provides us more historical information concerning the period in general, but where we can check them (especially by using Paul’s letters), Luke is proportionately more often correct in the limited details that he chooses to report.
Historians read their sources critically but regard some sources as more useful than others. As historical sources go, Acts is an extremely valuable one, offering serious evidence of dependence on considerable eyewitness testimony. Ancient historians could strive merely for verisimilitude where their details were incomplete, and so one arguing for substantial historical testimony in Acts need not suppose that Luke has tradition for every detail. But the contrary supposition, that Luke simply invents a novel about the early church and Paul, runs counter to the numerous correspondences between Paul’s letters and Luke’s portrayal (see below). These external correspondences with the Pauline portion of Acts, indeed, appear more often, in proportion to Acts’ size, than external correspondences for most first-century historians or biographers and their subjects (though this is partly because most biographic subjects left no epistolary corpus). Having written detailed commentaries on both Matthew and John, I am far more impressed with Luke’s historical interest and acumen, which are in keeping with Luke’s use of this genre.
These observations do not of course make him a modern historian who reports taped words verbatim or who may neglect some kinds of rhetorical, moralistic, and literary concerns. But it does mean that we should be readier to lend the benefit of the doubt to Luke than to the sort of modern hypotheses of Christian origins that dismiss our only concrete historical source and then argue from the silence that remains. Such hypotheses are all too common, yet they prove speculative, subjective, and impossible to verify or falsify; occasionally they gain respect through repeated citation, but otherwise they merely compete with other equally speculative scholarly proposals. Writing when eyewitnesses could still be consulted, two millennia closer to the events he reports than we are, Luke remains, for reconstructing Christian origins, a source that we neglect only to our historiographic peril. Whatever limitations we might attribute to him, Luke knows more about his subject than we do.
Scholarship on these issues sometimes divides into partisan camps instead of examining each issue on its own terms. Often (although not always) this approach has prejudged conclusions and led to ignoring the work of scholars in different camps (not to mention reviews fixated more on these categories than on the broader contents of the books being reviewed). For example, some Acts scholars have a tendency to ignore Ramsay’s archaeological work as unduly “conservative,” whereas Anatolian archaeologists depend heavily on this research, which subsequent study has more often than not confirmed. Sometimes a work is judged “critical” on the basis of conclusions adhering to scholarly consensus rather than of the breadth or depth of research. Worse, it is sometimes judged “critical” (regardless of scholarly consensus) only if it agrees with the critic’s measure of historical skepticism (as if no amount of evidence or analysis could lead to a different conclusion)—a rigid, doctrinaire approach that represents its own kind of “fundamentalism.”
On the range of scholarship publishing for a wide range of readers, my conclusions on matters of Luke’s historiography will be viewed as fairly “conservative” (though on many or most points fairly mainstream): I contend that Acts is a historical monograph, affirm genuine eyewitness presence in the “we” narratives, and find frequent external corroboration of Luke’s claims. Taking into account the conventions of ancient historiography, I have high respect for Luke’s reliability. I believe that this is where the strongest evidence points, and I make no apology for this claim.
At the same time, the evidence does not always support a partisan line for some common conservative conclusions: for example, against many (though not all) conservative scholars, I do believe that the evidence favors identifying the Jerusalem Council with Gal 2:1–10 rather than with an earlier famine visit (I prefer the “late” date for Galatians, though the difference is not great here: the late date is less than a decade later than the “early” date). Likewise, ancient historiography often handled speeches and details differently than we do (see below); although Luke may have been more faithful to his sources than were some of his contemporaries, it is neither necessary nor feasible to suppose that Luke followed modern rather than ancient historiographic conventions, conventions that did not yet exist. As even most conservative biblical scholars today recognize, Luke and other ancient historians had theological, moral, political, and literary/rhetorical agendas in how they shaped their accounts. Ancient historians allowed such factors, especially the literary concern for a complete narration, far more play than modern historians would.
With some other fairly conservative commentators (such as Bruce and Witherington) but in contrast to a number of others, I think that our current form of Acts was published after 70 C.E. (though exactitude in dating in either direction is admittedly impossible, given the limited data from which we draw our inferences; see ). For a particular issue in a matter of detail, I cannot imagine (as some scholars have argued) that Josephus and Luke each had in mind a separate revolutionary leader named Theudas. In sum, hearing Luke’s voice in its first-century setting cannot always be fitted into predetermined conclusions; impartial historical investigation will not always yield the results expected by traditional schools of thought. Nevertheless, many scholars of varying persuasions value the fair quest of truth, even when it requires some adjustments in their paradigms, and most do not uncritically adopt all ideas of one school. I have done my best to be honest with where I think the evidence points.