Zohran Mamdani! Look at that face. Look at the eyes, the nose, the beard. A Muslim? A Middle Eastern type! He must be a terrorist—or at least a supporter of terrorism.
Absurd? Of course.
Mamdani is Muslim, but he is not even Middle Eastern: he was born in Uganda to Indian parents. More importantly, neither terrorism nor support for terrorism can be diagnosed from someone’s eyes, ears, nose, jaw, skin color, beard, ancestry, or overall facial appearance.
If you want to judge Mamdani’s politics, examine his statements, associations, policies, votes, and actions. Those constitute evidence. His facial anatomy does not.
Yet humanity spent centuries attempting to do something remarkably similar.
The practice became known as physiognomy: the supposed ability to determine a person’s character, intelligence, temperament, morality, and even criminal tendencies from physical appearance, particularly the face.
At different points in history, the eyes supposedly revealed honesty or deceit. Ears indicated intelligence or stupidity. Noses revealed courage, weakness, or lust. Lips exposed sensuality. Chins revealed determination. Foreheads could supposedly betray criminal tendencies.
It sounds ridiculous today.
For centuries, however, educated people took such ideas seriously.
Physiognomy has no valid scientific basis as a system for reading individual character from facial anatomy. It is pseudoscience.
An ancient temptation
Attempts to connect appearance with character are extremely old.
Physiognomic ideas appeared in antiquity, including in a work traditionally attributed to Aristotle, although modern scholarship does not generally regard Aristotle himself as its author.
One recurring method was comparison with animals. If an animal possessed a particular physical characteristic and was believed to possess a particular temperament, a human resembling that animal could supposedly possess a similar temperament.
The reasoning was essentially analogical.
If lions were courageous and a person’s appearance resembled a lion, perhaps the person was courageous too.
Eyes attracted enormous attention. Their size, shape, position, movement, and color could supposedly reveal temperament.
Ears mattered too. Different traditions associated their size and form with intelligence, dullness, cunning, and other characteristics. In some physiognomic traditions, small ears were associated with undesirable qualities including dullness, cunning, or thievery.
Then there was the nose.
Different nose shapes supposedly indicated courage, timidity, nobility, sensuality, strength, or weakness.
The forehead, eyebrows, lips, teeth, chin, jaw, complexion, hair, posture, and proportions of the head could all become clues to the invisible personality supposedly hiding underneath.
There was no modern controlled psychological research behind these systems.
People observed faces, attached character judgments to them, remembered apparent successes, and constructed elaborate theories from their impressions.
Confirmation bias could do the rest.
Lavater turns the face into a personality map
Physiognomy became enormously influential in eighteenth-century Europe, particularly through the Swiss pastor and writer Johann Kaspar Lavater.
Lavater attempted to systematize physiognomy in extraordinary detail.
Almost every part of the human face could supposedly communicate something about the person behind it.
Consider the eyes.
Black eyes were associated with energy. Brown or hazel eyes could supposedly indicate vivacity and affability. Characteristics of the eyelids could suggest cunning.
The nose offered another personality test.
An aquiline nose could supposedly indicate an imperious temperament and strong passions. Small nostrils could indicate timidity.
Then came the mouth.
Thick, fleshy lips were associated with sensuality and laziness. Tightly closed lips could supposedly indicate avarice. Even the distance between the nose and mouth could acquire psychological significance: a large distance was associated with a supposed lack of prudence.
And then there was the chin.
A projecting chin could signify energy. A pointed chin could indicate craftiness. A receding chin could suggest a lack of firmness. A small chin could supposedly indicate timidity.
Virtually everything became evidence.
The face could reveal genius, benevolence, obstinacy, arrogance, melancholy, dishonesty, stupidity, strength, weakness, and countless other characteristics.
It looked systematic.
There were categories, observations, illustrations, and detailed descriptions.
But creating a complicated classification system does not make the underlying idea true.
You can measure nonsense with extraordinary precision and it remains nonsense.
From strange noses to the criminal face
The history becomes considerably darker when physiognomic thinking enters criminology.
In the nineteenth century, Italian physician and criminologist Cesare Lombroso developed his theory of the “born criminal.”
Lombroso believed that some criminals represented evolutionary throwbacks and that their supposed criminal nature manifested itself physically.
The jaw mattered. So did the forehead.
Facial asymmetry could become suspicious. Prominent or unusually shaped ears were among the characteristics he examined. Other supposed signs involved the skull and various bodily measurements.
Different kinds of criminals were even described as possessing different characteristic appearances. Thieves, murderers, sexual offenders, and other categories could supposedly display characteristic anatomical features.
This transformed an amusing pseudoscience into something much more dangerous.
A person no longer needed to commit a crime before becoming suspicious.
His face could convict him first.
Once people believe morality can be read from anatomy, prejudice can acquire the appearance of scientific legitimacy.
The criminal does not merely commit crimes.
He supposedly looks like a criminal.
Look at his eyes!
The basic instinct never disappeared.
People still say:
“He has evil eyes.”
“She looks dishonest.”
“He looks like a psychopath.”
“That guy looks dangerous.”
Sometimes these judgments may accidentally be correct. That makes them particularly seductive.
Return to Mamdani.
Someone already hostile toward Muslims could look at his photograph and believe that his face confirms everything he thinks he knows about Muslims.
Another observer might look at exactly the same face and perceive intelligence, confidence, friendliness, or compassion.
Nothing about the anatomy has changed.
The observer has.
Knowledge of someone’s religion, political affiliation, nationality, criminal record, or reputation can even alter how we retrospectively perceive his face.
Once we know someone committed a horrific crime, his photograph may suddenly seem sinister.
But imagine seeing exactly the same photograph without knowing anything about him.
Would the evil still be visible?
Often, what we think we see in a face is partly information supplied by our own brain.
Why physiognomy feels convincing
Human beings really do form extraordinarily rapid impressions from faces.
Within moments, another person can appear trustworthy, dominant, submissive, intelligent, aggressive, friendly, competent, attractive, dangerous, or harmless.
The impression itself is psychologically real.
That does not mean its interpretation is accurate.
A naturally lowered eyebrow position can make a neutral person appear angry. Large eyes may contribute to perceptions of youthfulness or innocence. Facial structure can influence perceived dominance.
Attractiveness produces another famous problem.
The halo effect can cause observers to attribute unrelated positive characteristics to attractive people. Someone who looks attractive may consequently be perceived as more intelligent, competent, trustworthy, or socially desirable without sufficient evidence.
The reverse can happen as well.
A face that violates cultural expectations can generate negative impressions that have nothing to do with the person’s actual behavior.
Humans therefore really do “read” faces.
The crucial question is whether we are reading the other person’s character or merely our own psychological response to his appearance.
Confirmation bias does the rest
Suppose you meet someone and immediately think he looks dishonest.
Two months later, he lies to you.
“See? I knew it from the beginning.”
Your confidence in your ability to read faces increases.
But what happens when you meet another hundred people who look similarly dishonest and behave perfectly honestly?
Probably nothing.
You do not maintain a spreadsheet recording every failed physiognomic prediction.
The memorable hit survives. The countless misses disappear.
This is confirmation bias, and it can make completely unreliable methods feel extraordinarily accurate.
Fortune-telling, astrology, stereotypes, and physiognomy can all benefit from similar psychological mechanisms.
Race made the idea much more dangerous
Physiognomic thinking also became entangled with historical attempts to classify human populations and attach intellectual, behavioral, or moral characteristics to physical differences.
This is where seemingly trivial speculation about noses and foreheads becomes socially dangerous.
Physical appearance can become a shortcut from perceived group membership to assumed character.
The imaginary argument about Mamdani demonstrates the mechanism:
He looks Muslim.
Muslims have committed terrorist attacks.
Therefore, he looks like a terrorist.
Every step is problematic.
There is no single Muslim facial anatomy. Islam is a religion followed by people from enormously diverse populations. And even if someone’s religion were known with certainty, group membership would not establish an individual’s political beliefs, morality, or propensity for violence.
A Muslim can be a terrorist, a pacifist, a surgeon, a socialist, a conservative, a secularist, a billionaire, or almost anything else a human being can be.
The same applies to Christians, Jews, Hindus, atheists, Arabs, Europeans, Africans, Asians, and every other enormous human category.
An individual’s actions require individual evidence.
Faces can contain information—but that is not physiognomy
Rejecting physiognomy does not require pretending that a human face contains absolutely no information.
Obviously it does.
Faces provide probabilistic information about approximate age. Some aspects of biological development are visible. Facial expressions can provide information about someone’s current emotional state, although expressions themselves require context.
Health can affect appearance.
So can hormones, nutrition, body composition, genetics, injury, disease, aging, sleep, alcohol consumption, smoking, and countless other variables.
Behavior can affect appearance too.
Someone who frequently smiles may develop characteristic habitual expressions. Long-term lifestyle differences can affect skin, teeth, body composition, and other visible characteristics.
This does not vindicate physiognomy.
There is an enormous difference between saying “chronic smoking can affect someone’s appearance” and saying “this nose reveals dishonesty.”
One is an empirically testable biological relationship.
The other is a personality horoscope written onto someone’s face.
Modern physiognomy has algorithms
Artificial intelligence has revived the temptation in a much more sophisticated form.
Imagine giving an algorithm millions of photographs together with psychological measurements.
Perhaps the machine could discover patterns that Lavater and Lombroso could never see.
Researchers have indeed investigated whether algorithms can predict psychological or behavioral variables from facial photographs at rates above chance.
But interpreting such results is extraordinarily difficult.
A photograph contains much more than bone structure.
It contains hairstyle, grooming, makeup, facial expression, body weight, lighting, camera quality, head position, clothing, background, image selection, and potentially socioeconomic and demographic information.
Even the choice of photograph is behavioral information.
A person choosing a formal professional portrait is making a different decision from someone uploading a drunken party photograph.
An algorithm that successfully distinguishes groups may therefore be exploiting dozens of indirect clues.
It has not necessarily discovered personality encoded in facial anatomy.
This creates a dangerous possibility: eighteenth-century physiognomy equipped with twenty-first-century computing power.
The computer’s answer can look scientific simply because a computer produced it.
But are there any real correlations?
Here the answer becomes more interesting.
Physiognomy is pseudoscience.
But the stronger proposition that no aspect of human facial morphology could ever have any statistical relationship whatsoever with behavior would also be difficult to justify.
Humans are biological organisms. The face and brain develop within the same body and can be influenced by overlapping genetic, developmental, hormonal, environmental, and health-related factors.
Small statistical relationships are therefore possible.
One of the most studied examples is the facial width-to-height ratio, usually abbreviated fWHR: roughly speaking, facial width relative to the height of the upper face.
Some research has reported associations between greater fWHR and aggressive, threatening, or dominant behavior, particularly in men.
A meta-analysis found a small positive association between men’s fWHR and aggression, approximately r = .11.
That is interesting.
It is also an excellent demonstration of why modern science is not physiognomy.
An association of r = .11 is small. There is enormous overlap between individuals. Many men with relatively wide faces will not be particularly aggressive, while many aggressive men will not possess relatively wide faces.
Moreover, the broader literature is mixed. Some large studies have found little evidence that fWHR meaningfully predicts antisocial tendencies, and results can depend on populations, methods, measurements, sex, body size, and other variables.
“Some studies detect a small population-level correlation” is scientifically interesting.
“Wide-faced men are aggressive” is not a scientifically justified conclusion.
What about sociopaths and psychopaths?
This requires particular caution because popular culture loves the idea that psychopaths somehow look different.
People talk about “psychopathic eyes” or claim that they can recognize a sociopath from a photograph.
There is no scientifically established psychopath face.
Psychopathy refers to a constellation of interpersonal, affective, and behavioral traits that can include callousness, manipulativeness, shallow affect, lack of empathy, impulsivity, and persistent antisocial behavior.
“Sociopath” is a looser popular term rather than a distinct facial or medical category that can be diagnosed from appearance.
Researchers have investigated relationships between facial morphology, perceived aggression, dominance, antisocial tendencies, and related characteristics. Some small statistical associations have been reported.
But none allows us to inspect someone’s eyes, ears, jaw, forehead, or facial proportions and diagnose psychopathy.
Even the small fWHR findings concern statistical tendencies such as aggression or dominance—not a facial diagnostic test for psychopathy.
That distinction matters enormously.
A weak correlation between one measurement and one behavioral tendency does not reveal someone’s entire personality.
The tiny grain of truth does not rescue physiognomy
This is perhaps the most interesting part of the story.
Physiognomy is wrong, but it is not necessary to claim that every imaginable correlation between appearance and behavior must equal exactly zero.
Pseudosciences often survive by attaching enormous conclusions to tiny pieces of reality.
Suppose researchers discover a weak correlation between one facial measurement and aggression.
The scientific conclusion might be:
There appears to be a small statistical association under particular conditions, although its mechanism, robustness, confounders, and predictive usefulness require further investigation.
The physiognomic translation becomes:
Scientists have discovered the face of an aggressive man.
Those statements are worlds apart.
Genes, hormones, nutrition, health, environment, body size, social experience, development, and behavior can influence human beings in overlapping ways.
Some physical and psychological variables will inevitably correlate.
But correlation is not destiny.
And a population-level correlation is not a reliable method for judging the individual standing in front of you.
Your face is not your character
Return once more to Zohran Mamdani.
Look at his eyes, ears, nose. Look at his jaw.
What do they tell us about whether he supports terrorism?
Essentially nothing.
His statements might tell us something; his actions might tell us something; his associations might tell us something. And his political record might tell us something.
Evidence matters.
His facial anatomy does not.
The same principle applies to everyone else.
The old physiognomists promised something extraordinarily seductive: that we could look at another human being and see intelligence, criminality, courage, sexuality, honesty, aggression, weakness, or evil written directly onto the body.
They assigned character to eyes, intelligence to ears, temperament to noses, sensuality to lips, determination to chins, and criminality to jaws and skulls.
Modern science offers something considerably less exciting: complicated people, noisy measurements, modest correlations, enormous individual variation, confounding variables, probabilities, and uncertainty.
You cannot look into someone’s eyes and discover whether he is a sociopath. And you cannot measure his ears and determine whether he will steal. You cannot inspect his nose and discover his courage.
And you cannot look at a Muslim politician’s face and discover whether he supports terrorism.
The old physiognomists promised that the human face could expose the human soul.
It cannot.
Leave a Reply