Invented by Accountants

The earliest writing anybody has found is a list of how much barley somebody owed.

The Uruk tablets, from around 3300 BCE, are receipts, inventories and ration lists: quantities of grain, beer, oil, textiles, sheep and human labour, with the occasional official's name attached. Somewhere around nine-tenths of the surviving early corpus is administration of one kind or another. Literature (hymns, laments, the first stretches of what becomes Gilgamesh) arrives some centuries later, once the system exists and somebody thinks to use it for something else.

This inverts the way most people imagine the invention. Writing was not created to preserve thought. It was created to keep track of things, and the capacity to record a poem is a by-product of a technology built to record how much barley went where.

The best part is how gradual it was, and the reconstruction belongs to Denise Schmandt-Besserat.

Across the Near East, for four thousand years before writing, people used small shaped clay tokens: cones, spheres, discs, cylinders, each standing for a quantity of some commodity. Three cones and a disc might mean three measures of grain and one jar of oil. To record an obligation between two parties, the tokens were sealed inside a hollow clay ball and the outside was stamped with the parties' seals, so neither could alter the count.

Which creates an obvious problem. A sealed envelope cannot be read without breaking it, and breaking it destroys the evidence.

The solution was to press the tokens into the outside of the ball before sealing them in, so anybody could see the contents without opening it. And once the impressions on the outside carried the whole of the information, the tokens inside were redundant. The ball flattens into a tablet. The impressions become signs.

Writing emerged as the solution to a filing problem, over millennia, with no inventor.

But an accounting notation is not yet a script, and the step that separates them is the one that matters. A system of pictures can only record what can be pictured. It can write "sheep" and "five" and "granary". It cannot write a name, or a verb, or a negation, or the word "if".

The escape is called the rebus principle, and in English it would be drawing a picture of a bee and a leaf to write the word "belief". The sign stops standing for the thing and starts standing for the sound, at which point the script can write anything the language can say, including the names of people, which is very often what the first phonetic writing in any system is actually doing.

Every independently invented script that developed into a full writing system made this move: Sumerian, Egyptian, Chinese, Maya. The ones that never made it stayed as notations for counting things, useful and limited.

Which brings up a habit worth resisting. It is common to see the alphabet described as the endpoint of a progression: pictures, then syllables, then letters, each better than the last. This is a nineteenth-century ranking exercise rather than a finding. Chinese has been in continuous use for over three thousand years, carries one of the world's great literatures, and works perfectly well for a modern industrial society. Japanese deliberately mixes two syllabaries with thousands of characters and shows no interest in simplifying.

What actually matters is fit. A language with a small inventory of possible syllables is well served by a syllabary. A language with heavy consonant clusters is not, and does better with an alphabet. The script follows the language, not a ladder.

The last thing to understand is why, having been invented, writing stayed in so few hands for so long.

Cuneiform required control of several hundred signs, many of which could be read as a word, as a syllable or as a silent classifier depending on context. Learning it took years in a scribal school, and Mesopotamian and Egyptian school texts are refreshingly frank about what the qualification was worth; a scribe does not carry baskets, is not beaten, is not conscripted. A profession whose position rests on a barrier has no incentive to lower it.

So the wall came down from outside. Simpler scripts, and the cheap materials to write them on, spread through traders rather than through palaces: the consonantal alphabets of the Levantine coast were developed by people who needed to write quickly for commercial purposes, not by bureaucracies with all the time and trained labour in the world. That is the pattern to carry forward. The institution that owns a technology rarely simplifies it. Somebody outside, who needs it to be cheap, does.