The Paper Road

Paper was made in China by the second century of the common era, and probably earlier. It reached the Islamic world in the eighth century, by tradition through craftsmen taken at the battle of Talas in 751, though the story is neater than the evidence. Paper mills were operating in Baghdad by the end of that century. It reached Europe through Islamic Spain and Sicily, with mills in Iberia by the twelfth century and in Italy by the thirteenth.

That is roughly eleven hundred years from invention to general European use, for a technology with no moving parts.

The consequence is easy to state and hard to overstate. Before paper, European writing was on parchment. Parchment is animal skin, prepared by a long process, and a single large book could require the hides of an entire flock. The cost of the substrate dominated the cost of the book.

The Baghdad translation movement is not separable from this. Rendering the Greek scientific corpus into Arabic meant producing an enormous quantity of text, and the arrival of paper in the same century is not a coincidence. A translation programme at that scale on parchment would have been economically impossible.

The same logic applies to the European sequence, in the order that is usually reversed. Printing is credited with the explosion of European text in the fifteenth century. Printing required cheap paper. The press without a paper industry is a machine with nothing affordable to print on, and paper mills preceded Gutenberg in Europe by two centuries.

This is the least discussed variable in the whole history of transmitted knowledge, and it deserves to be the first one checked. What survives from any period is constrained by what the period could afford to write on and by whether that material lasts.

The archive has met the second half of that already. The pre-colonial Philippines wrote on bamboo and palm leaf, which rot. Egypt's papyrus survives because Egypt is dry. Mesopotamian clay survives because clay is nearly indestructible, which is why we have more everyday documentation from Sumer than from classical Athens.

So the density of the historical record is substantially a map of substrate durability crossed with local humidity, and every civilisational comparison drawn from that record inherits the distortion.

There is a live version of the same problem. Digital storage has an extremely short demonstrated lifespan compared to parchment or clay, and it depends on continuously maintained infrastructure and readable formats. A cuneiform tablet needs a person who can read Akkadian. A file needs the format, the software, the operating system, the hardware and the electricity.

We have written more in thirty years than in all previous history, on the least durable medium ever used for the purpose. Which of those two facts turns out to matter more is not yet decided.