Every dossier on this site opens with a dating range and an authority for each end of it. This note explains what those numbers are, and what they are not.
Why dating carries the whole argument
If a text can be shown to predate the event it describes, its correspondence with that event is evidence of something. If it cannot, the correspondence is compatible with the text having been written afterwards — and a description written after the fact is not a prediction, however precise it is. That is the entire hinge. It is why a dossier that skipped straight to “look how exactly this matches” would be worthless, and why we make ourselves state the dating first, before anyone has been impressed by anything.
A range, not a date
Ancient texts are almost never datable to a year. What can usually be established is a window: an earliest plausible point of composition and a latest one. We publish both limits, and we publish the reasoning behind each limit separately, because the two ends are frequently held up by completely different kinds of argument — and are frequently disputed by different people.
A dossier cannot be published here with a range but no authority for it. “Scholars date this to the seventh century” is not a dating; it is a rumour about a dating.
What sets the limits
- Manuscript evidence. The hardest kind of upper limit: a text cannot have been written later than the oldest copy of it that survives. Palaeography and, where available, physical dating give that copy a range of its own — which is why a manuscript date is itself a range, and why stacking one range on another needs care.
- Internal references. A text that names a king, a city or an institution cannot easily predate that thing’s existence, and often cannot postdate its disappearance. This is powerful and also the easiest place to reason in a circle, since the identification of the reference is sometimes the very thing in dispute.
- Language. Vocabulary, spelling conventions and grammatical forms shift over time and can place a text in a broad linguistic period. It is a blunt instrument: it usually rules out centuries rather than pinning down decades, and later editing can leave a genuinely old text looking younger.
- Citation by others. If another datable work quotes the text, the text is older than that work. Clean when it applies, and it rarely applies to the cases people most want dated.
- Redaction history. Many biblical books show signs of having been composed in stages. “When was this written?” can then have more than one true answer, and a responsible dossier says which layer it is talking about.
Where the disagreement usually is
For the texts this site examines most often, the dispute is rarely about the evidence itself. It is about what is allowed to count. A reader who holds that a text can describe an event before it happens will read an apparent prediction as evidence for an early date; a reader who does not will read the same passage as evidence for a late one. Both are reasoning consistently from a prior commitment, and the passage cannot settle between them.
When that is the situation, the honest verdict is grade 3 — disputed, and we say plainly that the answer depends on the dating and that the dating depends on the assumption. Presenting a contested date as settled, in either direction, is the single most common failure in popular writing on this subject.
What we do not do
We do not narrow a published range to make a case tidier. We do not report the range from one school as though it were the consensus. And where a well-supported minority dating exists, it goes in the case against, stated at its strongest — not summarised into weakness.