by KIKE CALVO
A fascinating new study explores how anthropologists and psychologists disagreed over one deceptively simple question: can human emotion truly be measured?
Long before artificial intelligence began scanning faces for emotion, attention, or intent, scientists were already debating whether human expression could ever be fully translated into measurable data.
A newly released paper by Heewon Kim titled:
Trained eyes, coded faces: Visual instruments and expertise in the sciences of facial and bodily expression
examines how mid-twentieth-century anthropologists and psychologists developed radically different ways of understanding facial and bodily movement—and how those differences still shape modern technologies today.
Published in the journal History of the Human Sciences, the paper traces intellectual debates between figures such as:
The result is not only a history of science.
It is also a history of how humans increasingly became visually measurable.
The Human Face as Scientific Territory
Today, facial analysis technologies are everywhere.
Phones recognize faces.
Security systems track expressions.
AI tools claim to detect:
- stress
- attention
- happiness
- deception
- fatigue
simply by analyzing facial movement.
But Kim’s paper reminds us that these technologies emerged from decades of disagreement over what faces actually mean.
Can emotion be objectively measured?
Or is every expression inseparable from culture, interpretation, and context?
Two Different Ways of Seeing Humans
One of the paper’s most compelling arguments is that anthropologists and psychologists approached visual analysis very differently.
For anthropologists such as Margaret Mead and Ray Birdwhistell, understanding human behavior required trained cultural interpretation.
Expressions could not simply be isolated from social meaning.
A smile, gesture, or posture might communicate entirely different things depending on:
- culture
- environment
- learned behavior
- context
Visual technologies like film and photography were useful—but they could never fully replace human interpretation.
Ray Birdwhistell and the Language of the Body
The paper places special attention on Ray Birdwhistell and his pioneering work in:
kinesics
—the study of body movement as communication.
Birdwhistell believed bodily gestures functioned similarly to language systems.
But importantly, he resisted simplistic universal readings of facial expression.
To him, human movement was culturally coded.
Meaning depended on the observer’s expertise and understanding.
The Rise of Facial Coding
Meanwhile, psychologists such as Paul Ekman and Silvan Tomkins increasingly pursued measurable, universal emotional expressions.
Using photography, film analysis, and eventually computational systems, they attempted to transform facial movement into standardized units.
The face gradually became:
- measurable
- codifiable
- machine-readable
This shift would eventually influence modern AI emotion-recognition systems still used today.
“Code” Meant Two Very Different Things
One of the most fascinating insights in Kim’s paper revolves around the idea of:
code
Anthropologists viewed codes as:
- socially learned meanings
- cultural systems
- interpretive frameworks
Psychologists and emerging computational systems increasingly understood codes as:
- discrete units
- measurable data
- standardized information
This distinction fundamentally shaped how scientists approached human behavior.
And it continues to influence AI systems today.
Why This Matters in the Age of Artificial Intelligence
Kim’s historical analysis feels remarkably contemporary.
Modern facial recognition and emotion-detection technologies often assume that physical facial movements correspond directly to internal emotional states.
But the anthropological critique remains deeply relevant:
👉 Can algorithms truly understand irony, context, performance, culture, or ambiguity?
Or are they simply measuring muscular movement while missing meaning itself?
The paper suggests these unresolved historical tensions remain embedded within today’s technologies.
The Power of Visual Instruments
Another major theme of the paper is that visual technologies themselves shape scientific understanding.
Film and photography were not neutral recording tools.
They influenced what researchers believed they were observing.
The same question applies today to AI systems.
Technologies do not merely document human behavior.
They also define what counts as meaningful behavior.
Faces as Data
Perhaps the most unsettling implication of the paper is how gradually human faces became transformed into datasets.
What once required human interpretation increasingly became:
- numerical information
- coded expressions
- machine-readable systems
In today’s world, billions of faces circulate through automated infrastructures every day.
And many people remain unaware that these systems emerged from decades-old debates about perception, expertise, and interpretation.
A Traveler’s Reflection
Travel reveals how culturally shaped human expression truly is.
A gesture interpreted one way in one country may mean something entirely different elsewhere.
This is why visual anthropology remains so important.
Human behavior cannot always be separated from:
- history
- language
- place
- social norms
Context changes meaning.
A Luminous Perspective
At Luminous Photo Expeditions, we are deeply interested in how humans observe, interpret, and visually document the world.
Photography and visual technologies are never entirely objective.
They shape:
- memory
- identity
- perception
- power
Kim’s paper offers a powerful reminder that the act of “reading” faces has always involved tension between:
- technology and interpretation
- measurement and meaning
- data and lived experience
And those tensions are becoming increasingly important in the age of artificial intelligence.
Final Reflection
Perhaps the central question raised by this research remains unresolved:
👉 Can human emotion ever truly be translated into code?
Or does something fundamentally human always resist measurement?
As AI systems increasingly attempt to analyze faces automatically, the debates explored in this paper feel less historical—
and far more immediate.
Learn More
This article is inspired by:
Kim, Heewon.
Trained eyes, coded faces: Visual instruments and expertise in the sciences of facial and bodily expression.
History of the Human Sciences (2026).
DOI: 10.1177/09526951251398489
The original paper explores the historical development of facial coding, visual expertise, and the scientific interpretation of human expression from the 1950s to the 1970s.
