Two Names on a Stone

Stand on the corner of Rochussenstraat and 's-Gravendijkwal. Look at the brick. High up, under a concrete sill, a pale stone tablet is set into the wall. The capitals are cut deep. You can still read them from the pavement.

I read them most days. De eerste steen gelegd. The first stone, laid, by Gerrit van Duffelen, nine years old, and Cornelis Pelt, six years old, on 3 June 1899¹. A factory and a house. Its foundation stone. Cut with the names of a nine-year-old and a six-year-old, the sons of the firm's two founders².

I always wondered who they were. The tablet is still there, a diagonal crack across its right side and the letters sharp. It is the one piece of the 1899 building you can still read from the pavement, and I live behind it³. What it belongs to, the founding permit and the works it authorised, is the deep dive on the founding document.

Gerrit, the nine-year-old, became a painter. The first document that calls him kunstschilder, a painter by profession, is his sister's Rotterdam marriage act of 9 March 1911, where he signs the register at twenty-one in his own hand¹. His file is the dossier on Gerrit van Duffelen.

Cornelis 'Cees' Pelt, the six-year-old, went to an office. His 1918 Rotterdam marriage act lists his trade as kantoorbediende, a clerk². His line does not end in the archive. It reaches a family living today, set out in the Pelt descendants.

Some of the people I wrote to wrote back. Beverley and Murk especially, descendants of the two boys on the stone. The archive gave me names and dates. It could not give me faces, or the things these two families had kept. They did. Beverley shared picture after picture of Gerrit's paintings, an album she sent from Canada. Murk sent a photograph of his family, and the one surviving likeness of Cornelis Pelt himself. They answered my messages, told me what they knew, and let a stranger put it here. Thank you.

Some months ago I bumped into a picture of my house on the internet, from its years as a music venue, and I began the search from those two names and their ages on the stone^1 2. A neighbour had once told me the building used to sell carpets. That was all anyone knew. Carpets, and two names in the wall.

Around the turn of the millennium the corner was a stage. WORM was formed in 2000 out of the merger of the Dodorama venue and shop, a film workshop and Popifilm³, and its agenda pages billed events at Rochussenstraat 169 across 2002 and 2003⁴. Those venue years survive mostly in archived web pages, gathered in the WORM web archive. The photograph that started me off is the one above, reproduced here by permission of Kunstinstituut Melly.

Busy job. Family. No time for archives. But the question would not sit still. So I did the only thing I had time for. I experimented.

I work with AI. One evening I pointed it at my own house. We did not build a small thing. We built a solution: AI agents that go into the public archives and squeeze the history out of them, running on a research method strict enough that no claim gets through without a source. The agents worked while I slept. Permits, marriage acts, address books, old newspapers. Every occupant, firm and picture followed back to where it came from, and checked. Night after night, for weeks. The days went to work and family. The house got the nights.

We have held the ground-floor dwelling since 1 July 2019⁵, and I am the one who went looking. This whole site is what came back.

Think of it as a book with an open archive behind it. Four eras follow the building's own changes of state: the Furniture Works, the Furnishing House, Dodorama and WORM, and Thirty-Three Homes. Read them in order and the corner changes trade under you.

Beside the story runs a right-hand column of deep dives. A person, a firm, a place, a register, a question. Each sits next to the narrative and never breaks into it. Behind both stand the Archive, where every source, entity, register and open question can be looked up, and the Methodology, which explains how each statement was made, and where the work was done by machine and where that stopped.

One rule holds throughout. A claim carries its status wherever it appears: confirmed, probable or a lead, the label travels with the sentence, and the evidence sits a click away. The Dutch and archival terms are explained as they arise in a plain glossary, and the whole span is laid end to end in the timeline.

I should be honest about how this was made. Almost all of it is automated, and I steered and decided. It is not perfect. It is not complete. The more you dig, the more you find, and I could dig for years. I chose to release it anyway, a working version now rather than a perfect one never. There are errors in here. I know it, and I will fix them over time, at night, when I can. Acknowledgments thanks the archives, the families who opened their collections, the people of the venue years, and the tools and AI systems the research leaned on. Contacts is a footer page for corrections and additions, because the history of one building is never quite finished.

The book itself starts before the house. Part I opens on the polder ground and the estate that decided what could be built on this corner, and works up to the day the permit was granted.