Funicular Models

In the late 19th century, Gaudí hung thousands of chains from the ceiling of his Sagrada Família studio, each weighted with small metal bags. He spent years adjusting weights and positions, photographing the hanging forms, then inverting the photographs: the curves created by gravity hanging downward, when flipped, became perfect arch geometries under compression. This peculiar apparatus became one of the most brilliant “computers” in the history of architecture—the funicular model.

A reconstruction of this model is now displayed in a corner of the Sagrada Família’s underground museum and is one of the visit’s must-see exhibits. The original was destroyed during the Civil War in 1936, but later architects spent years rebuilding it from high-resolution photographs and written records. A mirror is positioned above the hanging model today, allowing visitors to look up and see directly how inverted gravity forms perfect arches.

Gaudí chose this method because he had grown dissatisfied with the mainstream architectural reliance on pure mathematical calculation. He believed gravity had already “solved” the optimal geometry of structural mechanics; the architect’s job was not to invent new formulas but to borrow the answers that nature had already provided. The catenary—the curve formed by a hanging chain under its own weight—is mathematically described by the hyperbolic cosine function. But Gaudí never needed to know this equation; he used physics to simulate physics.

Gaudí’s funicular models had a profound influence on modern structural engineering. When he first presented this working method in the early 20th century, many peers mocked it as “primitive.” Yet today, finite element analysis (FEA) software operates on exactly the same mechanical principles—just with digital simulation replacing Gaudí’s chains and sandbags.

When the Sagrada Família architectural team began digitally modeling Gaudí’s funicular forms in the early 21st century, they were astonished to discover that virtually all of the arch geometries he “hand-calculated” with chains and weights over a century ago were structurally correct. The margin of error was minimal, far surpassing any paper-based calculation method available in his time. Those little sandbags stood in for supercomputers that wouldn’t exist for another hundred years.