Le Corbusier published the Modulor in 1948 and a second volume in 1955. The system begins from a figure—a man with arm raised, approximately 2.26 meters in total height—and derives from this figure a cascade of measurements that the designer can use to dimension anything from a door handle to a city block. The measurements are organized around the golden ratio and the Fibonacci sequence, producing proportions that Le Corbusier believed would feel naturally harmonious to human perception because they were derived from human proportion.
The system was applied most visibly at the Unité d'Habitation in Marseille, where the apartment dimensions, the height of the corridors, and the dimensions of the communal spaces are all generated by the Modulor. People who live in the Unité frequently report that the spaces feel right in a way they cannot entirely articulate, which is taken as evidence that the Modulor was doing something real. The system was also adopted, in various degrees of fidelity, by the mass housing estates built across Europe and America on Corbusian principles—with results that ranged from the Unité's qualified success to the catastrophic failures of the tower-in-the-park estates.
The particular masquerading as the universal. Le Corbusier's Modulor figure was, by his own account, chosen partly for aesthetic reasons—a tall, handsome man who looked like the hero of an English novel. The supposedly universal measure of man was a specific man, and the 'universality' encoded one body's proportions as the standard against which all other bodies would be measured and found wanting. This is the defining move of every biased AI system: a standard derived from a particular—a training dataset skewed toward certain populations, a benchmark calibrated on certain performance patterns—is presented as the objective measure, and deviation from it is coded as deficiency rather than as evidence that the standard was never universal.
Standardization as liberation and as mold. The Modulor was not wrong to seek a standard. A built environment needs shared dimensions; a doorway must have a height, a stair a rise. Used flexibly, as a proportional guide rather than a procrustean law, the Modulor produced spaces of genuine harmony. The error was not standardization itself but the mistake of treating one standard as universally correct—of hard-coding the assumptions about one kind of body into a system that would be applied to all bodies. A standard that cannot accommodate the people it omits is not a universal. It is a particular that has been given the institutional authority to define deviation as error.
The prescriptive undertow. A measure of man, deployed at scale, does not merely fit the world to human beings; it begins to fit human beings to itself. If every door, chair, and counter is built for the Modulor man, the world subtly instructs everyone to be that man, and those who cannot are made to feel the building was not meant for them. AI standardization carries the same prescriptive undertow. A model of the creditworthy applicant, applied across a society, does not just predict who will repay; it shapes who gets the chance to. A language model trained on text skewed toward certain cultural registers does not merely describe communication; it begins to define what legitimate communication looks like, and to penalize modes of expression that fall outside its training distribution.