History may be one thing after another but a literary history is the author's attempt to convey their own vision. The authorial representation should therefore guide scholarly interpretation of the text.
Steve Mason's dissertation on the Pharisees in Josephus is a masterclass in narratological interpretation. Previous scholars had taken a single line of Josephus's autobiography (an odd phrasing which seemed to declare Pharisaic affiliation) as evidence for doubting all the passages in Josephus's narrative work which expressed criticism of Pharisees. Since that dissertation, scholarship in general has recognized Mason's superior exegesis of the one difficult line and henceforth overturned the previous dogma that Josephus identified as a Pharisee. Also since that dissertation, Josephus scholars have come around to the larger issue raised by Mason's methodology. Unfortunately, NT scholars have not. If you haven't sussed quite yet what that larger issue might be, please bear with me a bit longer.
Frank Ankersmit has argued that a narrative text cannot be comprehended in the same way linguists exegete a single propositional statement. Let us examine the difference.
When reading one sentence, we decode individual words while looking for grammatical cues and patterns of syntax. Piecing those things together is how one comprehends the proposition. The task is to build meaning from the bottom up. If a statement mentions "Elvis Presley" we might wonder whether that refers to Ed Sullivan Elvis or military service Elvis or Blue Hawaii Elvis or Las Vegas Elvis or the Elvis purportedly haunting Graceland today. As exegetes of a single reference, we must deduce which Elvis the author means us to recognize, and our initial deduction may be confirmed or corrected a few words later, or perhaps a few lines later, further down in the text. This is not necessarily the case with a large collection of sentences.
The first distinction to recognize is that a narrative passage cannot "refer" in the same way as an individual word or a propositional statement. If you mention the name of my dog, you "refer" to one identifiable being. It is "picked out uniquely" by the label I use. Even if I may have given my dog a common name, we can easily clarify with further specification. "I mean the cocker spaniel Taffy who lived at Bill's childhood home until he was 13." The ideal one-to-one correspondence is at least achievable, even if our collective vocabulary (or "cultural repertoire") does not contain a single term to label each object on Earth. With enough added words, we can still pick out which dog I mean to describe.
Narrative cannot do that. It simply cannot. A published story about the battle of Waterloo can only "refer" to its subject in general (what Ankersmit idiosyncratically calls "aboutness"). Of course I can cite a year to specify which Treaty of Paris or Ocean's Eleven I might have in mind, but that is not the point. The one and only battle of Waterloo can be identified with a single place and time but it cannot be identified by a single set of words. That is, we cannot select one verbal discourse as the proper description of that singular battle. It was unique, but words cannot designate it uniquely. If that were possible, then we should expect to find only one proper description of the battle, only one proper biography of Napoleon, only one acceptable version of any given event in Earth's history. Obviously, this is not the case. The one-to-one correspondence between words and things cannot be extended to words and events. This necessarily shifts our hermeneutic away from objective decoding of a linguistic construction and towards the subjective interpretation of an aesthetic construction.
Narrative is not reference. Narrative is representation.
The construction of meaning from narrative, therefore, cannot be treated exclusively as a bottom-up endeavor, in the same way we construct meaning from the propositional statement. Unlike the grammatical and syntactical bits of a sentence, the individual sentences within a larger narrative text create a whole which is not merely the sum of its parts. Thus, instead of piecing together the meaning of each text by tackling one sentence at a time, one must rather prioritize a determination about the overall meaning of an authorial representation IN ORDER TO have any chance at properly understanding each single statement within it. Although this prescription may seem paradoxical, because of course we must read the text one statement at a time, the point is that our hermeneutic spiral must build and build and built UNTIL we have reached the point where we can look back on each part and see how they all fit in the whole.
Impressively, that is precisely how Mason engaged the corpus of Josephus's writings. Where other scholars had camped out on one troublesome sentence and erased several passages which appeared to contradict it, Mason argued that we should rather prioritize understanding Josephus's whole body of work before turning again to that difficult bit in his autobiography. When taken in context, Mason demonstrated conclusively, there was a different way to interpret the one bit. Josephus had not joined the Pharisees. He had merely studied them for a while. The single ambiguous statement was best understood in the context of Josephus's larger narratological construction.
I have argued the same thing about Luke and Quirinius: whatever else we think about that difficult bit of text, it cannot be taken as evidence that Luke thought Jesus was born in the year 6 CE. Rather, the larger story Luke tells obviously fits in a world where Judea and Galilee were being administered in a unified way. Rather, a reconstruction of Luke's esthetic vision must be prioritized. The author's view of the story he means to tell should be our first reconstruction, after which we can use that reconstruction as context for attempting to understand what Luke was trying to say about that odd proconsul Quirinius.
I have argued the same thing about Matthew and Archelaus. Instead of getting stuck on one word ("basileuei") and declaring that to be inaccurate (because Archelaus was an ethnarch and not a king), we should rather go back several verses and prioritize Matthew's narratological context. Even if Jesus living in Egypt is pure fiction, Matthew's angel wakes Joseph precisely at Herod's death so that Matthew's readers can recognize the timeframe. Archelaus was indeed, just then, playing the king. He was literally "kinging" in Judea, in his father's position. Recognizing this temporal context lends dramatic irony and praises God's foresight, because Galilee was not yet safe from the dangerous princeling, but it would soon be once Augustus slapped down the upstart.
I have argued the same thing for the book of Acts. Instead of treating each episode as more or less random, we must first reconstruct the authorial viewpoint. First, the author inserts himself into the story, aligns himself not just with the gentile mission and Paul but also with the narrative climax in Caesarea. The author, as character, appears to remain for two years with the Caesarean church as his base. That the apprehensions they feel for Jersualem have lingered is not anti-Jewish but anti-mother-church. Within that context, it cannot be coincidental that Luke's outstanding point of view characters for the early chapters - Stephen and Philip and Cornelius and Barnabas, the ones Luke goes out of his way to introduce and to follow - also happen to align with the gentile mission, with Antioch, and with Caesarea. Philip and Cornelius in particular signal an authorial focalization. These are not random characters with marginal stories Luke felt bound to include. They are some of his favorites. In terms of comprehending narrative to be representational, that authorial construction should be prioritized as the context for everything else in the text.
As I said at the top of this post, history itself may be one thing after another but a literary history is the author's attempt to convey their own vision. One cannot merely analyze bits of narrative content and then think about them as truth claims, or potentially true. The exegetical analysis of any passage must root itself in a larger awareness of the "history" as a literary construction, an authorial representation. The individual truth claims are not merely tainted by authorial bias; they are painted with brushstrokes of meaning. As Josephus's overall view of the Pharisees elucidates a proper exegesis of each statement he makes about them, so should we apply Luke's overall disposition (against the heavy-handed Christians in Jersualem) to our exegetical analysis of each line and episode in the book.
Narrative history is not simply a story, to be taken or left. It is not simply "narrative" because that's what we call this stuff. Narrative history is an authorial representation. Before we judge the potential veracity of a narrated claim, we must first be careful to understand the meaning of that claim, NOT with a bottom-up semiotic and linguistic approach, BUT with a top-down approach of narratology and literature.
The authorial representation should guide scholarly interpretation of the text. After which, THEN AND ONLY THEN, historical critical judgment should absolutely proceed to have its field day by assessing the evidential value (or lack thereof) in each claim and action depicted. But those claims must be judged with a full awareness of their nature as brush strokes, rather than log entries.
At any rate... I said all that to say this.
If any NT scholar happens to be thinking about doing a project called "Acts as History" I dearly hope they pay attention to these important distinctions.
Anon...