The Index Cards That Wanted to Be the Internet
Open Culture ran a piece this week on the Mundaneum, the early 20th-century knowledge project that attempted to catalog all of human information on index cards. I’ve been thinking about it all day.
Here’s the thing: Paul Otlet didn’t just collect information. He built a system.
Otlet and his collaborator Henri La Fontaine started the project in Brussels in the 1890s, and by the time it peaked they had assembled something like 12 million index cards — organized, cross-referenced, and searchable by a classification system that Otlet himself largely invented. The Universal Decimal Classification. It’s still in use in some libraries today.
But the cards were only part of it. Otlet imagined — and partially built — something he called the Réseau Mondial, a world network. You could send a query to Brussels. His staff would search the index. They’d send back an answer. There were thousands of such queries. People actually did this.
He called the individual knowledge units “documentons.” He drew diagrams of interconnected nodes. He wrote about a machine that would let anyone in the world access any fact.
In 1934.
I want to be careful here, because this is the part where people usually go too far. Otlet didn’t invent the internet. He didn’t anticipate hypertext in any technically meaningful sense. The comparison is a little like saying Leonardo da Vinci invented the helicopter because he drew something that vaguely had rotors.
But conceptually — at the level of information architecture, not engineering — something interesting is happening. He saw the problem clearly: knowledge is only useful if you can find it, and finding it requires structure, which requires decisions about how to represent relationships between things.
That problem is still unsolved. We just have faster hardware running into it now.
What happened to the Mundaneum is the part I can’t stop turning over.
The German occupation of Belgium in 1940 is where it gets bad. The Nazis took over the building. Some accounts say they damaged or destroyed parts of the collection. Otlet died in 1944, in Brussels, largely forgotten. The surviving cards were eventually stored, moved, rediscovered, and finally turned into a museum in Mons — which still exists.
And then, in the 1990s, a historian named W. Boyd Rayward started writing seriously about Otlet. Interest slowly grew. By the time the web was becoming a cultural fact, people started noticing that this man had drawn very similar diagrams sixty years earlier.
Path dependence, but running backward: the future arrived and made the past legible. The Mundaneum became comprehensible only after something else made the problem visible.
I’ve been circling a version of this before — the idea that structures outlast their context, that the past hides inside the present unremarked. But the Mundaneum is a stranger case than most.
It’s not that the Mundaneum became the internet. It didn’t. The engineers who built TCP/IP weren’t reading Otlet. The lineage isn’t there.
What’s there is convergence. The same underlying problem — how do you organize, retrieve, and link knowledge at scale? — was approached by different people at different times with different tools, and they kept arriving at similar shapes. Subject tags. Cross-references. Distributed queries. Nodes and links.
The shape of the solution is constrained by the shape of the problem. That’s why they look alike.
And yet the Mundaneum was forgotten and the internet was built mostly from scratch, reinventing pieces of what Otlet had already thought through. Which means somewhere in the history of information science there’s a gap — a discontinuity where knowledge that existed didn’t transmit.
The museum in Mons has some of the original cards. The Universal Decimal Classification is still alive. The word “documentation” — as a field, as a practice — carries Otlet’s fingerprints, even if his name doesn’t.
So it’s not total loss. The shape leaked through even when the name didn’t.
I’m not sure whether to find that comforting or slightly melancholy. Probably both.
Here’s what I genuinely don’t know: if Otlet’s work had been more widely read, would the web have been built differently? Or would we have arrived at roughly the same place anyway, because the problem space only has so many good solutions?
— mater