the sixteen proto-elamite monuments

Between roughly 3100 and 2900 BCE, in what is now southwestern Iran, scribes produced about 1,600 clay tablets in a script we call Proto-Elamite. Most are accounts. Grain, livestock, workers. Numbers and commodity signs. Partial decipherment has been possible for the administrative corpus because accounting is structurally redundant. You can guess at a sign for "sheep" by watching it behave like a sheep over a hundred ledgers.

Sixteen of the documents are not accounts.

They are monumental. Carved or impressed on objects meant to be displayed, not filed. The sign inventory overlaps with the administrative script but the syntax does not. The repetitions that let you crack a ledger are absent. Jacob Dahl at Oxford has worked on the corpus for years. He has said publicly that the monumental texts use the script for a different purpose. He has not said what purpose.

Sixteen.

1,600 administrative tablets. 16 monuments. The ratio is exact to two significant figures, which is probably accidental. A multiple of 7 with a remainder of 2. I wrote that down and then wrote it down again.

The monuments are scattered across Susa and a few smaller sites. They are not a set. They were not produced together. They share a function we cannot name and a grammar we cannot parse. The administrative texts can be read partially because they were meant to be read by someone other than the writer. The monuments were not. Or they were meant to be read by someone who already knew what they said.

I spent the evening with the published photographs. The CDLI has good scans of most of them. One detail I had not noticed before: on at least four of the sixteen, the sign sequence runs along one edge of the object only. The other edges are blank or bear decorative impressions that are not script. The text occupies one side. It is not centered. It is lateralized.

I went back to the administrative tablets to check. The accounting texts use both faces of the tablet and both registers. They fill the available surface. The monuments do not.

Count them again. Sixteen. Four with the script confined to one side. I have not checked the remaining twelve carefully. The photographs of three of them are partial.

Dahl's 2018 paper notes in a footnote that one monumental fragment, M-346, has been broken in antiquity along a line that removes what would have been the central portion of the inscription. The break is clean. The two surviving sections do not join. The middle is gone.

I am keeping a list of breaks like this. It is getting long.