How memory finds things
You put a thing in a folder because you believe that later, when you want it back, you will remember the name you chose today: the mood you were in, the project it belonged to, the word that seemed obvious at the time. That belief is almost always wrong. The folder survives; the reason for it does not.
What survives is stranger and more useful. You remember that the piece was long, that it argued against something you believed, that you read it on a train, that it mentioned bees. Recall does not walk down a tree of names. It arrives sideways, from whatever happened to be nearby. Ask anyone to find a half-remembered article and watch what they actually do: they do not open folders, they describe it.
The shape of the problem
A hierarchy asks every item to live in exactly one place, which is a strong claim about a thing you have read once. An essay about urban planning that is really about how institutions forget is filed under cities, and then it is gone, because the day you want it you are thinking about institutions.
The alternative is older than the computer. Bush’s memex was built on trails rather than drawers: a path between documents, laid down by the person reading and followed later by the same person, or someone else. What it kept was the connection between two things, which is the part a folder throws away.
What this means for a library you actually keep
So the question is whether anything has to be filed at all: whether a library can arrange itself by what its contents mean, continuously, so that a half-remembered description is a good enough address. That is a much lower bar than remembering a name, and it is the one your memory was already clearing.
