What a face was supposed to say
From the Renaissance to the twentieth century, people tried to read character in a face. The enterprise failed, and the way it failed is instructive.
There is a long Western tradition that sought to read a person in their face. It produced substantial books, it had serious readers, and it failed.
That failure is worth recounting, because it points precisely to where the mistake lay.
The three great attempts
Giambattista della Porta published De humana physiognomonia libri IIII in 1586, a lavishly illustrated work that set human faces alongside animal heads and claimed to deduce character from them. The book was a European success.
Johann Caspar Lavater, two centuries later, published his essays on physiognomy — translated into French as early as the eighteenth century and available online today. He was read everywhere, debated by the greatest minds of his time, and his method was taken for a nascent science.
The twentieth century tried again, differently: the biotypologies of Kretschmer and Sheldon, which claimed to link body shapes to temperaments. They were discussed seriously — the Société d’anthropologie de Paris was still devoting studies to them in the 1970s — and then abandoned.
In France, Louis Corman founded morphopsychology in 1937, and its French society still exists today. It is the only one of these traditions to have survived as a living practice.
Why it failed
For two distinct reasons, and they must be kept apart.
The first is moral, and it is serious: these systems served to justify hierarchies between people, and to lend scholarly authority to prejudice. The legacy of physiognomy is indefensible, and it is enough to disqualify these authors as authorities.
The second is methodological. They read by eye. Each reader saw what they expected to see, each conclusion confirmed the starting hypothesis, and two physiognomists looking at the same face did not say the same thing.
What remains valid, and it is narrow
A single observation, very old and independent of everything else: the face expresses something of how the body is functioning.
This is true and verifiable. Fatigue shows there, tension shows there, a difficult period shows there. Every beauty tradition has observed this, and it is what makes the face useful as a sign.
What does not hold is the leap: moving from “the face expresses a state” to “the shape of a nose indicates a character.”
What this failure teaches
That the problem was not the wish to classify. It was classifying from what one sees.
A reading by eye inherits all the flaws of the eye: it depends on the lighting, on the reader’s expectations, on the mood of the day, and on what the subject chooses to show. This is exactly why physiognomy never stabilised and why the old typologies contradicted one another from one practitioner to the next.
That was the only point to correct. Not to give up on distinguishing between people — but to stop doing so from a datum that varies and lends itself to interpretation.
A reading that starts from a fixed datum, one that no one declares and that the gaze does not alter, escapes that particular reproach — and it is the only reproach that, technically, mattered.