Half the external evidence in this archive comes from royal inscriptions. Reading them as if they were reports is the single most common error in this field, and it is made by defenders and critics alike.
They were not written to inform anyone
An Assyrian annal was composed to be deposited in the foundations of a building, or carved on a monument in a courtyard. Its audience was the gods, the king’s successors, and whoever might one day dig up the wall. It was not a dispatch and it was not an archive entry.
This matters for a practical reason. A text with that purpose has no obligation to be complete, and no incentive to record a campaign that failed. It is not lying in the way a modern forgery lies; it is doing exactly what the genre requires, which is to present a reign as the gods would wish to see it.
The conventions are visible, and that is what makes them usable
Royal inscriptions repeat themselves. The same formulae for tribute, for siege, for the destruction of a rebel city recur across reigns and across empires, often word for word. That repetition is what lets a reader tell the difference between a formula and a fact.
When Sennacherib says he shut Hezekiah in Jerusalem like a bird in a cage, the image is memorable and it is also standard. When the same text gives thirty talents of gold, a specific number that matches a figure in an entirely different record, the specificity is doing something the formula is not.
So the working rule is not that royal inscriptions are unreliable. It is that the parts which are conventional and the parts which are particular carry different weight, and a page that quotes one as if it were the other is misleading its reader.
Silence is the trap
The most frequent misuse is the argument from omission: an Assyrian text does not claim to have taken Jerusalem, therefore Assyria admits it failed.
That reads a modern expectation of full disclosure into a document designed for a wall. A campaign ending in tribute rather than conquest is reported as tribute, which in this genre is a success. Nothing is being conceded. The text stops where the genre stops.
The reverse trap is just as common. Because royal inscriptions omit failures, some treatments conclude they are worthless. But an inscription that names a foreign king, gives his kingdom, and records a transaction with him is establishing that king’s existence whatever it says about the outcome — and that is often the claim actually under examination.
What a royal inscription can settle, and what it cannot
- Can settle: that a named ruler existed, ruled the state named, and was in contact with the empire in a datable year.
- Can settle: the political framework — who was subject to whom, who rebelled, who paid.
- Usually cannot settle: numbers. Tribute figures, army sizes and captive counts are the parts of the genre most open to inflation.
- Cannot settle: outcomes the text has an interest in. A withdrawal will not be reported as a withdrawal.
- Cannot settle: anything about the motives, beliefs or internal politics of the other side.
The Moabite exception, and why it is instructive
The Mesha Stele is a royal inscription of exactly this kind, produced by a king of Moab rather than an empire. It claims that Israel perished forever, which is false — the kingdom continued for another century and a half.
That single sentence is worth more to a careful reader than a dozen cautionary paragraphs. Here is a royal inscription making a claim we can check, and failing it, in the ordinary way of the genre. It is the clearest available demonstration that the same text can be sound on the political framework and worthless on the rhetoric, in adjacent lines.
Related: the dossiers on Ahab at Qarqar, Jehu on the Black Obelisk and Mesha of Moab, and the method note on what external attestation means.