Before asking whether a passage is true, ask what kind of thing it is trying to be. Most arguments about the Bible are arguments between two people reading the same page as two different genres.
Nobody reads a poem the way they read a receipt. We do this automatically with modern writing — a headline, a horoscope, a court transcript and a novel arrive with their genre attached, and we adjust before the first sentence. Ancient writing arrives without those signals. The cues that told a first reader what they were holding are gone: the scroll it was on, the setting it was read in, the conventions it was playing against.
So the question has to be asked deliberately, and it has to be asked before the question of accuracy. A text can only fail at what it was attempting. Grading a parable for historicity is a category error, and so is treating a royal annal as though it were disinterested reporting.
The kinds this site keeps meeting
Annal and chronicle. Year-by-year records kept by an administration: campaigns, accessions, building works. The Babylonian chronicle tablets are the type. They are terse, dated and comparatively unemotional, which makes them the most useful external material this site has — but “unemotional” is not the same as neutral. They record what the administration kept track of.
Royal inscription. A king’s own account of his reign, cut in stone or clay and often buried in a foundation. Sennacherib’s prisms are the type. The genre has fixed conventions: victories are total, tribute is submission, and defeats are not mentioned. An inscription that stops short of claiming a city was taken is following its own rules, not making a confession.
Narrative history. Kings, Samuel, Chronicles, Acts. Written to explain as well as to record, arranged around a thesis about why events went as they did. Ancient historiography did not share the modern convention that speeches are quoted rather than composed; a first-century reader expected an author to supply what a speaker would plausibly have said. That is not fabrication by the standards of the form, but it does change what a speech can be used as evidence for.
Prophetic oracle. Poetry, usually, addressed to a situation and delivered in the first person of the deity. Oracles are conditional far more often than they are read as being, and they are frequently collected and arranged long after delivery. Dating the collection and dating the oracle are two different problems, and confusing them is the commonest mistake in this field.
Apocalypse. A revelation given to a seer through a mediator, heavy with symbolic beasts and numbers, typically written in a crisis and typically set in the past. Daniel and Revelation are the biblical examples, and there are many non-biblical ones, which is a considerable help: the conventions of the genre are documented, and can be checked against texts whose dates are known independently.
Wisdom. Observations about how life generally goes. Proverbs is a collection of them, and reading a proverb as a promise is the error the form invites. Job and Ecclesiastes are wisdom arguing with itself — which is evidence that the tradition knew its own generalisations did not always hold.
Law. Ancient legal collections were not statute books in the modern sense. They read more like collected rulings and teaching examples, and several ancient Near Eastern collections survive to compare against. Where biblical law differs from its neighbours, the difference is where the interest lies.
Letter. Occasional writing, addressed to a specific situation, and — crucially — we usually have only one side. Reading a reply without the question is like hearing half a phone call. Paul’s letters answer things we have to reconstruct from the answers.
Parable and fable. Fiction told to make a point, and understood as such by its first hearers. Nobody in the first century asked for the prodigal son’s name.
Why it decides the verdict
The five-grade scale on this site is applied to claims about events, and only to claims about events. Genre is what determines whether a passage is making one.
A dossier can ask whether Nineveh fell, because Nahum’s oracle points at a public event that other people also recorded. A dossier cannot ask whether the good Samaritan really stopped, because the parable is not claiming he did. If a page on this site ever grades a parable, a hymn or a proverb, it has made a mistake, and the corrections page is where that would be recorded.
Where genre is itself the argument
The hardest cases are the ones where the genre is disputed, and those are the interesting pages. Are the opening chapters of Genesis a cosmology, a polemic against neighbouring cosmologies, a liturgy, or an account of events? Is Jonah a prophetic book or a satire about a prophet? Is Job a case history or a thought experiment? Each of those questions decides in advance what would count as evidence — and readers who disagree about the answer usually find they were never really arguing about the evidence at all.
Where a page on this site takes a view on genre, it says so, and says what the view costs. Related: how we date a text, and what external attestation means.