How Writing Was Invented (Twice, By Accountants)

Writing is the technology that made every other technology cumulative. It was not invented by artists or priests. As far as the evidence goes, it was invented by people who needed to know how much grain someone owed them.

The first texts are receipts

The earliest known writing is cuneiform, developed in Mesopotamia — roughly modern Iraq — around 3200 BC. It was pressed into wet clay with a wedge-shaped reed, which is where the name comes from. The oldest tablets are not stories or laws. They are inventories: quantities of barley, counts of livestock, lists of workers and their rations.

Clay turned out to be a spectacular accident of preservation. Papyrus rots, but a clay tablet in a burning building simply bakes harder. A great deal of what we know about the ancient Near East survives because someone's archive caught fire.

Invented more than once

Writing appears to have been independently invented a small number of times: in Mesopotamia, in China, and in Mesoamerica, with Egypt arriving early enough that scholars still argue about influence. Everything else — every script you can name — is borrowed, adapted, or descended from one of those.

The alphabet, by contrast, seems to be a one-off. A single system devised in the Semitic-speaking world around 1800 BC, adapted by the Phoenicians, passed to the Greeks (who added vowels), and from there to the Romans and onward. Latin, Greek, Cyrillic, Hebrew, Arabic and the scripts of India all trace back to that one invention. The letters you are reading are a direct descendant.

The Rosetta Stone was a tax notice

Egyptian hieroglyphs went unread for roughly 1,400 years. The key was a granite slab found in 1799 carrying the same decree in three scripts — hieroglyphic, demotic and Greek — which gave scholars a known language to work from.

The content is gloriously mundane: a priestly decree concerning royal honours and tax exemptions. The most famous object in the history of decipherment is, in substance, bureaucracy. Jean-François Champollion completed the decipherment in 1822, having begun teaching himself ancient languages as a boy specifically to do it.

The scripts still waiting

Not every script has been read. Linear A, used in Bronze Age Crete, remains undeciphered — while Linear B, which looks similar, was cracked in 1952 and turned out to be an early form of Greek, and also mostly inventories. The Indus Valley script is likewise unread, partly because no bilingual Rosetta-Stone equivalent has ever been found.

There are, in other words, civilisations whose writing survives and whose voices do not. Which is a reminder that literacy is not just about having the marks — it is about someone, somewhere, still knowing what they meant.