Case Study 13-2: What Paul Ekman Found — and What Was Built on Top of It

A Research and Perspective Case

Introduction

In 2009, a television drama premiered on the Fox network that ran for three seasons and drew millions of viewers. Lie to Me was built around the premise that a scientist could read lies from facial microexpressions with extraordinary precision — could see, in a 1/25-second flash of a face, what a person was truly thinking. The show was based, its creators said, on the real research of psychologist Paul Ekman.

Ekman, who had spent forty years studying facial expressions, was a technical adviser on the series. He later expressed some ambivalence about what the show had done with his work.

That ambivalence is instructive. Because what Ekman actually found is genuinely important — foundational to how we understand emotional expression and relevant to anyone interested in nonverbal communication. And what popular culture built on top of those findings tells us something important about how research travels from laboratories to practical application — and what gets distorted in the journey.

What Ekman Actually Found: The Foundational Research (1960s–1990s)

The Cross-Cultural Universality Studies

In the late 1960s, Paul Ekman and his colleague Wallace Friesen set out to test what was then a contested question: Are facial expressions of emotion universal across cultures, or are they culturally learned?

The dominant position in mid-century social science, associated with figures like Margaret Mead and Gregory Bateson, was that emotional expression was culturally variable — that the face was a kind of language, and that its vocabulary was learned, not innate. If this were true, you would expect people from different cultures to use and read expressions differently, much as speakers of different languages use different words.

Ekman and Friesen designed studies to test this. They showed photographs of facial expressions to participants in countries including the United States, Brazil, Chile, Argentina, and Japan — and found significant cross-cultural agreement in how expressions were labeled. But critics argued that this proved only that mass media exposure (especially American films and television) had spread Western emotional display norms globally.

To address this objection, Ekman conducted his most famous study: he traveled to Papua New Guinea and worked with members of the Fore culture — an isolated group in the highlands who had had minimal contact with Western media, television, or photographs. He showed them photographs of Western faces and described emotional scenarios (e.g., "Your child has died" — show me the face of this person). The Fore participants matched expressions to scenarios at rates significantly above chance. And their own posed expressions were recognizable to Western participants.

The conclusion Ekman drew: at least six basic emotional expressions are universal across cultures — innate rather than learned. These were happiness, sadness, fear, anger, disgust, and surprise. He later added contempt as a seventh.

This was a significant finding. It suggested that core emotional expressions are, to a meaningful degree, biological — part of our evolutionary inheritance — rather than purely cultural constructions.

The FACS System

Parallel to the cross-cultural work, Ekman and Friesen developed the Facial Action Coding System (FACS) — a comprehensive taxonomy of all observable facial muscle movements. They identified 44 action units (AUs): specific movements of specific muscle groups that combine to produce recognizable expressions.

FACS was not a theory of deception. It was an anatomical map. It allowed researchers to describe facial movements with precision and repeatability. An expression that colloquially might be called "a sad face" could be broken down into specific AUs — the inner corner of the eyebrows raised (AU1), the lip corners pulled down (AU15), the chin raised and wrinkled (AU17) — allowing researchers to agree on exactly what they were studying.

FACS became the gold standard in facial expression research. It is used today in psychology, computer vision, and animation (Pixar animators used FACS principles to design emotionally legible characters). The system itself is Ekman's most scientifically durable contribution.

Microexpressions

In the 1970s, Ekman and Friesen analyzed film footage of psychiatric patients and identified a phenomenon they called "microexpressions": very brief, involuntary facial expressions that appeared when a person was attempting to suppress or conceal an emotional state.

In one famous instance, they analyzed footage of a psychiatric patient named Mary who was attempting to hide her suicidal thoughts from clinical staff. By slowing the footage, Ekman identified a flash of intense distress — lasting less than a quarter of a second — that appeared on her face immediately before a normal-looking expression reasserted itself. The microexpression had not been visible in real time.

This was the core microexpression insight: that suppressed emotional states leak through the face in brief, often involuntary flashes. Ekman's subsequent research found that microexpressions typically fall into the categories of the seven basic emotions identified in his universal expression research.

He developed training tools — the Micro Expression Training Tool (METT) and the Subtle Expression Training Tool (SETT) — and found that with training, people could learn to detect microexpressions at above-chance rates, and that trained readers performed better than untrained readers in identifying concealed emotional states.

What Became Controversial: The Deception Claim

Here is where the research and its popular application diverge in ways that matter.

Ekman's original microexpression work was about suppressed emotional states — not specifically about deception. A microexpression reveals that a person is feeling something they are trying not to show. It does not, by itself, tell you why they are suppressing it, what they plan to do about it, or whether they are lying.

But the popular application of his work — including Lie to Me, popular training programs, and government security screening programs — collapsed that distinction. The claim that emerged was: microexpressions can detect lies.

This is a meaningful overreach.

The Problem with "Detecting Deception"

Detecting an emotional state and detecting deception are categorically different problems.

A person showing signs of suppressed fear during an interrogation may be suppressing fear because they are lying — or because they are terrified of being falsely accused, because they have an anxiety disorder, because the authority differential in the room activates a stress response regardless of guilt, or because they have a traumatic history with interrogation-style interactions.

An emotional state does not explain itself. It tells you what a person is feeling, not why, and not what it means about their verbal statements.

Ekman himself was aware of this distinction and wrote about it. He noted that what a microexpression reveals is emotional state, not intent or truthfulness. However, the applied programs — particularly the Transportation Security Administration's SPOT (Screening of Passengers by Observation Technique) program, which trained security officers to use behavioral indicators including Ekman-derived microexpression signals to identify potential threats — did not always honor this distinction.

The Evidence on Lie Detection

A substantial body of research has examined whether humans — trained or untrained — can detect deception above chance rates. The meta-analytic findings are not encouraging for strong claims.

Charles Bond and Bella DePaulo's 2006 meta-analysis of deception detection research found that people detect lies at approximately 54% accuracy — barely above the 50% baseline of random chance. Trained professionals (customs officers, police, FBI agents, federal judges) performed no significantly better than untrained undergraduates. In several studies, higher confidence in judgments was inversely correlated with accuracy.

Ekman-based microexpression training showed some improvements in detection accuracy — but in controlled experimental conditions, and with modest effect sizes. The question of whether those improvements translated to real-world detection across varied populations, with the time pressure and contextual noise of actual security or interrogation settings, was less well-supported.

A 2003 report by the National Academies of Sciences examining polygraph testing — a technology premised on reading physiological arousal as an indicator of deception — raised related concerns about the gap between laboratory findings and real-world application. Similar critiques applied to behavioral detection programs.

The Lie to Me Problem

The television program Lie to Me depicted its protagonist reading lies with near-perfect accuracy from brief behavioral cues, in real time, under complex conditions. This is fiction — but fiction that millions of people took as a representation of what the underlying science supports.

The effects of this popularization are worth considering:

  • Law enforcement and security professionals enrolled in training programs premised on detection accuracy that the research does not support.
  • People in everyday contexts began to apply "microexpression reading" with overconfidence — treating a flash of what looks like suppressed emotion as evidence of lying rather than as a signal to ask another question.
  • The genuine findings of Ekman's research — that emotional expression is substantially universal, that suppressed emotions leak in brief facial signals — were reframed as a lie detection technology.

Ekman himself, in a 2019 interview, noted that he was not responsible for how his findings had been commercialized and that some applications went beyond what the research supports. He maintained that trained observers can improve their detection of suppressed emotional states, while also acknowledging that the deception-detection claims made in applied settings often exceeded the evidence.

What the Research Actually Supports

It is useful to be precise about what the science behind microexpressions and emotional expression actually warrants, separated from its popular application.

Supported: - Core facial expressions for a set of basic emotions show significant cross-cultural recognition, suggesting some degree of biological universality - Emotional states produce observable facial signals, some very brief, that occur before or despite suppression efforts - Trained observers can learn to detect these brief signals at rates above untrained baseline - The FACS system provides a rigorous, replicable framework for describing facial muscle movement - Awareness of incongruence — moments when someone's face briefly shows something inconsistent with their words — is a meaningful signal to slow down, ask more questions, or probe more gently

Not well-supported: - That microexpression reading is reliable enough to identify liars with meaningful accuracy in real-world conditions - That emotional arousal signals detected by trained observers can be confidently attributed to deception rather than other causes (fear of accusation, social anxiety, trauma response, cognitive load) - That the accuracy advantage of trained over untrained observers translates robustly to real-world, time-pressured, high-stakes detection settings - That any single behavioral indicator is reliably diagnostic of deception, concealment, or harmful intent

The Important Limitation: Emotional State Is Not Motivation

This limitation bears its own section because it matters practically for how you use this chapter's material.

Knowing what someone is feeling does not tell you why they are feeling it. A person who is visibly anxious is anxious — but whether they are anxious because they are lying, because they feel falsely accused, because authority figures activate their stress response, because they have social anxiety, because they are conflict-averse (like Marcus), or because something else entirely is happening is not available from the emotional signal alone.

This is not a minor caveat. It is the central epistemological constraint on the use of emotional expression reading.

In conflict communication, this translates to a practical discipline: use emotional signals as data to inform your next move, not to draw conclusions about intent. If you notice that someone's face flashed what looked like contempt when you made a particular point, that is relevant information — it may mean they disagree more strongly than they're saying, or they're masking frustration, or they interpreted your statement as condescending. It is a signal to probe, to check in, to ask a question. It is not evidence of deception, malice, or bad faith.

The correct response to a microexpression is typically a question, not an accusation.

The Research Legacy

Set aside the commercialization and the television drama. What remains is genuinely important.

Ekman's cross-cultural research shifted the scientific consensus on emotional expression from a purely culturalist position (all expression is culturally learned) to a more integrated view in which some expressions are biologically grounded, while display rules and contextual interpretation remain culturally shaped. This integration — not a simple "all universal" or "all cultural" — is where the current evidence sits.

His FACS system gave researchers a precise language for describing faces. This is no small contribution: science advances partly on the basis of shared measurement systems, and FACS created one for facial expression that has endured for fifty years and generated thousands of studies.

His microexpression work introduced a phenomenon that is real and measurable: emotional states do leak through faces in brief, involuntary ways. Suppression is imperfect. The body tells truth the mind would prefer to manage.

What that truth is, and what it means, requires interpretation — careful, provisional, contextually informed interpretation. The body can tell you someone is feeling something they haven't said. Figuring out what to do with that is the human work that no algorithm has solved.

Questions for Discussion

  1. Why do you think Ekman's research was so readily transformed into a lie-detection technology in popular culture? What does this transformation tell us about how we want nonverbal communication science to work — about what we wish were true?

  2. The chapter argues that microexpressions should prompt questions, not conclusions. Can you think of a scenario in which even that more cautious application could cause harm? What safeguards would make microexpression awareness more reliably useful?

  3. Research suggests that trained professionals are no better at detecting lies than untrained undergraduates, yet confidence in one's ability to detect deception often increases with training. What does this suggest about the relationship between expertise and accuracy in judgment? How does it apply beyond lie detection?

  4. Ekman's cross-cultural research found above-chance recognition of basic emotions across cultures, but subsequent research found lower rates than originally reported. How should we update our practical approach to reading emotions across cultural contexts, given this more uncertain evidence base?

  5. How does the "detecting emotional state vs. detecting deception" distinction apply to everyday conflict situations — not security or interrogation contexts, but ordinary difficult conversations between people who know each other?

Selected Sources

Ekman, P., & Friesen, W. V. (1969). The repertoire of nonverbal behavior: Categories, origins, usage, and coding. Semiotica, 1, 49–98.

Ekman, P., Sorenson, E. R., & Friesen, W. V. (1969). Pan-cultural elements in facial displays of emotion. Science, 164(3875), 86–88.

Ekman, P. (1992). An argument for basic emotions. Cognition and Emotion, 6(3–4), 169–200.

Ekman, P. (2003). Emotions Revealed: Recognizing Faces and Feelings to Improve Communication and Emotional Life. Times Books.

Bond, C. F., & DePaulo, B. M. (2006). Accuracy of deception judgments. Personality and Social Psychology Review, 10(3), 214–234.

Barrett, L. F., et al. (2019). Emotional expressions reconsidered: Challenges to inferring emotion from human facial movements. Psychological Science in the Public Interest, 20(1), 1–68.

National Research Council. (2003). The Polygraph and Lie Detection. The National Academies Press.

Porter, S., & ten Brinke, L. (2008). Reading between the lies: Identifying concealed and falsified emotions in universal facial expressions. Psychological Science, 19(5), 508–514.