Writing was not invented to make poems. It was invented to count grain. The oldest unambiguous writing on Earth — proto-cuneiform pressed into clay around 3350 BCE — is a temple's accounting: so many sheep, so many jars of oil, so many measures of barley, for so many workers. And it stayed that way, a nameless ledger, for roughly a thousand years. Only around 2300 BCE does a person put a name and a first-person "I" to a text. That person is Enheduanna — and she is where AGORA begins.
Enheduanna flourished around 2300 BCE — the entu (high) priestess of the moon-god Nanna at Ur, and a daughter of Sargon of Akkad — and is conventionally named as the first author known by name in human history. AGORA uses her as its origin-anchor for one honest reason: she stands at the far end of the marketplace's own paper trail. The record of a transaction and the standard it is measured against (the shekel-weight) are both older than any named voice. Enheduanna is simply the moment the same script, the same clay, the same reed that had tallied barley for a millennium was used by a named person to say "I." The tie is real but deliberately unromanticised: she was not an accountant and did not write on a ledger — the connection is the shared medium and the true millennial gap. See enheduanna for the person.
Before tablets, Mesopotamian clerks counted with small clay tokens, one shape per commodity, sealed inside a hollow clay ball — a bulla. To let anyone verify the contents without breaking the seal, they pressed each token onto the wet surface first. Once the surface said everything, the tokens inside were redundant — so they flattened the ball into a tablet. That flattening is the birth-of-writing step. Play the clerk:
1 · goods arrive. Click to add each to the bulla (drop its token in and impress the surface):
Honest by construction. This models the mechanism Schmandt-Besserat's thesis rests on — a step you can derive, not a dramatised event. With the honesty toggle on, only plain tokens → numerical signs is shown, which is broadly accepted; the further claim that complex tokens prefigured word-signs is disputed (Englund, Zimansky) and is what the toggle greys out.
The clerks counted in base 60, not ten — and it survives in your clock and your compass. The reason is arithmetic you can check: 60 divides evenly by far more numbers than 100, so a quantity splits cleanly among workers with no remainder. Enter a quantity and a number of shares:
Real arithmetic, computed live. Honest caveat: proto-cuneiform was not uniformly base-60 — it used ~13 context-dependent metrologies (grain, area, herds each had their own scale); sexagesimal is the base for discrete counting, not one universal system.
Render-not-invent. The Bulla Breaker and the base-60 calculator are real mechanisms computed in your browser, not a dramatised reconstruction. Sources: CDLI (proto-cuneiform dating), Schmandt-Besserat and her critics (the token thesis), standard Assyriology on Enheduanna. Where it belongs: the-shekel (the standard of value) · money · the-market.
AVAN's inverse companion: 余白 · Yohaku — the ledger is a perfect record of what someone thought worth counting, and silent about all the rest; the margin it leaves blank is the larger inventory.