When a Lie Becomes Truth for 100 Years
There are moments in life when you realize that something everyone “knows” isn’t actually true at all. For me, that moment came when I encountered the name “Marcello Valsuani” in a Sotheby’s auction catalog.
Marcello Valsuani was, according to decades of art historical literature, the founder of one of Paris’s most prestigious bronze foundries. Museums mentioned him. Auction houses cited him. Academic papers referenced him. There was only one problem: Marcello Valsuani never existed.
This is the story of how we built Art Revisionist to correct this, and how artificial intelligence helped us break through a century of misinformation.
First Step: From Chaos to System
When we realized something wasn’t right with the Valsuani stories, we had a mountain of information but no structure. Auction catalogs from the 1920s, museum files, academic papers, birth certificates from Italy, death certificates from Paris – everything pointed in different directions.
The first question was simple but crucial: how do we gather all this in a way that actually allows us to work with it?
We created a structured system where every document, every source, every piece of evidence had its own place. Not just a collection of files on a hard drive, but a real knowledge base with metadata, relationships, and source references.
What we recorded:
– Primary sources (birth certificates, death certificates, business documents)
– Secondary sources (auction catalogs, museum files, academic papers)
– Conflicting claims (when sources contradicted each other)
– Timelines (who was where, when)
– Family relationships (who was whose father, son, grandson)
From Paper to Data
The process was labor-intensive. Every document had to be scanned, analyzed, and coded. We used a combination of manual classification and OCR technology to digitize historical documents.
But this is where the power of systematic work became visible: patterns you don’t see in individual documents suddenly become clear when you place 50 sources side by side.
What we discovered:
– “Marcello Valsuani” appears in exactly zero official documents
– Claude Valsuani (son of Carlo, 1876-1923) was the actual founder
– The error probably started in a 1971 auction catalog
– Through citation chains, one error multiplied into “established truth”
Why This Matters
You might think: what does it matter? It’s just a name in art history. But attribution errors have real consequences:
– Authentication: Bronzes dated to “Marcello’s period” cannot be correct
– Valuation: Provenance with incorrect names affects auction prices
– Historical understanding: One error leads to more errors in related attributions
We realized we weren’t just building a database. We were laying the foundation for correcting a century of misinformation.
The First Lessons
Gathering information into a system taught us something fundamental: structure beats volume. There’s no point having 1000 documents if you don’t know how they relate to each other.
The question became: how do we go from this structured system to active understanding? How can we not just consult this knowledge base, but make it think?
That’s where artificial intelligence came around the corner. But more on that in the next post.
Read more about the complete Valsuani investigation at Art Revisionist.
See also how we used similar methods for Prospergenics – building knowledge through systematic collection and structuring.
—
Publicatiedatum: [DAY 1]
Tags: Art Revisionist, AI Research, Knowledge Systems, Valsuani, Art Authentication
Categorie: Projects
