In the summer of 1966, a researcher named Paul Ekman traveled to Papua New Guinea to visit an isolated group of people who had had virtually no contact with Western culture — the South Fore. He brought photographs of facial expressions and asked the Fore to match them to emotional scenarios. Fear, sadness, anger, happiness, disgust, and surprise were all recognized accurately, in the same configurations that Western participants had identified. The finding was significant: if people who had never encountered Western media recognized the same emotional expressions, those expressions were likely not learned but universal. Ekman had found what he believed were the basic building blocks of the human emotional face.
The research that followed over the next five decades built an entire field around the readable face — microexpressions, leakage cues, the markers of deception, the involuntary betrayals of concealed emotion. By the early 2000s, Ekman's work had become the basis of a television series (Lie to Me), extensive law enforcement training programs, and an entire popular genre of body language guides promising to reveal the hidden code of human behavior. The implicit promise was compelling: with enough knowledge of the right cues, you could see through the social performance to the truth beneath.
The evidence tells a more complicated story. Reading people accurately is a genuine cognitive skill, but it operates through different mechanisms than the popular account suggests, is subject to systematic biases and limitations that popular accounts rarely acknowledge, and depends far more on contextual understanding and baseline comparison than on the identification of universal signal cues. Understanding what the research actually shows — what can be read from behavior and what cannot — is both more intellectually honest and more practically useful than the simplified version.
"We are far more influenced by nonverbal communication than we realize, but that does not mean we are accurate in reading it." — Paul Ekman
Key Definitions
Microexpression: A brief, involuntary facial expression, lasting between one-twenty-fifth and one-fifth of a second, that occurs when a person attempts to suppress or conceal an emotional response. Associated with Paul Ekman's research program at the University of California San Francisco.
Leakage: The phenomenon in which suppressed emotional states or deceptive intentions 'leak' through behavioral channels that are harder to control consciously — voice pitch, subtle facial movements, physiological changes — even when the person is actively managing their presentation.
| Signal Channel | Reliability for Deception Detection | Reliability for Rapport | Notes |
|---|---|---|---|
| Verbal content | Low | Moderate | Easily controlled; people craft words |
| Vocal tone and pace | Moderate | Moderate | Harder to control under stress |
| Facial expression | Low-Moderate | High | Micro-expressions visible but training required |
| Body posture and movement | Moderate | High | Mirroring and leaning are reliable rapport cues |
| Physiological arousal (flushing, sweat) | Moderate | Low | Indicates arousal; does not identify its cause |
Thin-Slicing: Nalini Ambady's term for the ability to make accurate inferences about a person or situation from very brief behavioral exposures. Popularized by Malcolm Gladwell in 'Blink' (2005) with some oversimplification of the underlying research.
Baseline Behavior: An individual's characteristic behavioral patterns under normal, unstressed conditions. The reference point against which meaningful deviations can be assessed.
Prosodic Cues: Elements of speech other than the literal words — pitch, tempo, rhythm, pausing, emphasis, and vocal quality — that carry emotional and intentional information and are harder to consciously control than word choice.
Paul Ekman and the Science of Facial Expression
Ekman's Papua New Guinea research of the late 1960s established the cross-cultural universality of basic emotional expressions and launched a research program that continued for fifty years. His taxonomy of basic emotions and their facial configurations became, for a time, one of the most cited frameworks in psychology.
The FACS — Facial Action Coding System — which Ekman developed with Wallace Friesen in the 1970s, provided a comprehensive anatomical catalog of facial muscle movements (action units) and their combinations. It allowed for the systematic description and comparison of facial expressions in research contexts and remains a widely used tool in affective computing, clinical research, and cognitive neuroscience.
Microexpressions, which Ekman identified and studied extensively, occur when full emotional expression is suppressed but the genuine expression briefly escapes control — typically in less than a second. Ekman claimed that trained observers could learn to detect these brief expressions and thereby gain access to concealed emotional states. His training programs were adopted by numerous law enforcement and security agencies.
What the Replication Evidence Shows
The scientific reception of Ekman's specific claims has grown more critical over time, particularly as large-scale meta-analyses became possible. The core finding — that basic emotional expressions are recognizable across cultures — remains reasonably robust, though subsequent research by Lisa Feldman Barrett and colleagues has raised important questions about whether emotional expressions are as discrete and universal as Ekman's original model suggested.
The applied claim — that microexpression reading can reliably identify deception — has fared less well. Studies consistently find that lie detection accuracy rates, even for trained observers, cluster only modestly above chance. The conceptual reason is important: microexpressions indicate suppressed emotion, and suppressed emotion during an interaction does not reliably indicate deception. A truthful person may suppress fear of not being believed. A liar may suppress nothing if they have rehearsed sufficiently. Emotion and honesty are not the same variable.
Thin-Slicing: What Brief Exposures Can and Cannot Reveal
Nalini Ambady's research on thin-slicing produced some of the most intriguing findings in interpersonal psychology. Her studies found that observers could make accurate predictions about teaching effectiveness from 30-second silent video clips, surgical malpractice claims from 40-second voice samples, clinical supervisors' ratings from 20-second video clips, and personality traits from brief behavioral observations.
The accuracy of thin-slicing predictions, across Ambady's research, was typically modest but significantly above chance — which is notable given how little information was available. The implication is that humans have genuine pattern-recognition abilities that extract meaningful signals from behavioral observations, even very brief ones.
Malcolm Gladwell's 2005 book 'Blink' popularized these findings with the argument that rapid, unconscious cognition — the 'blink' of a decision — is often more accurate than deliberate analysis. The book was widely read and is intellectually engaging, but the research community generally considers it to have overstated the case. Gladwell's account underemphasizes Ambady's own findings about the conditions under which thin-slicing fails, the systematic biases it is subject to, and the important role that expertise plays in calibrating the unconscious patterns being matched.
What Thin-Slicing Actually Predicts Well
The research suggests that thin-slicing accuracy is highest for characteristics that are genuinely and consistently expressed through observable behavior. Extraversion is reliably read from brief exposures because extraversion is defined, in part, by behavioral expression — warmth, energy, talkativeness, approach behavior. It makes conceptual sense that thin-slicing works for it.
Thin-slicing is less accurate for characteristics that are not reliably expressed through observable behavior. Honesty, intelligence, and competence in domain-specific tasks are all relatively poorly predicted from brief behavioral observation. The subjective sense of confidence in thin-slicing judgments, unfortunately, does not track with accuracy in these domains — people feel certain about inferences that the evidence does not support.
Leakage Cues: The Channels That Betray
The concept of behavioral leakage — originally developed by Ekman and Friesen in a 1969 paper — describes the differential controllability of different behavioral channels. The face is, in Ekman and Friesen's framework, the most controlled channel because it is the one we most deliberately monitor and manage. Voice is somewhat less controlled. Body movement and physiological responses are less controlled still.
This differential controllability has practical implications: if a person is attempting to manage their presentation, the most reliable signals of their actual state will emerge in the channels they are least attending to. A person who has composed their facial expression carefully may reveal stress through vocal characteristics — slightly elevated pitch, faster speech rate, increased disfluencies (ums and ahs). They may reveal it through postural changes, leg movement, or physiological indicators like flushing or perspiration.
Prosodic cues are particularly information-rich and consistently underestimated. Research by Klaus Scherer and colleagues has extensively documented the relationship between vocal characteristics and emotional states. Voice pitch tends to rise with anxiety. Speech rate tends to increase with urgency and decrease with sadness. Temporal disruptions — pauses, stutters, repairs — increase under cognitive load, which deception requires. These cues are more reliable than gesture-based interpretations precisely because they are harder to consciously control.
The Baseline Problem: Why Universal Signals Are Insufficient
One of the most important and most frequently missed insights in the behavioral observation literature is the primacy of baseline. The meaning of any behavioral observation depends on its relationship to the individual's baseline behavior, not on its relationship to a population norm.
Joe Navarro, who spent twenty-five years as an FBI special agent and has written extensively on nonverbal communication, makes this point central to his practical framework. In his view, the first step in reading any person is establishing their baseline: how do they normally move, speak, make eye contact, gesture, and position their body when they are comfortable and unstressed? Only deviations from that individual baseline carry diagnostic information.
The research on this point is consistent. A person who normally has limited eye contact is not displaying a significant signal when they have limited eye contact in a high-stakes conversation. A person who characteristically fidgets in almost all situations is not displaying anxiety-specific fidgeting when they fidget during questioning. Popular body language guides that assign universal meanings to specific behaviors fail on this dimension: they provide a baseline derived from population averages rather than from the specific individual being observed.
The Role of Context
Context compounds the baseline problem. The same behavioral cue can have entirely different causes depending on the situational context in which it appears. Reduced eye contact may indicate: cultural norms (direct eye contact with authority figures is considered disrespectful in many cultures), cognitive absorption (the person is thinking carefully), shyness (baseline characteristic), concealment (the person wishes to avoid scrutiny), or simple distraction. Without contextual information, assigning a specific meaning is unjustified.
Research by Bella DePaulo and colleagues on behavioral cues to deception found that the behavioral indicators most commonly believed to signal lying — gaze aversion, increased body movement, speech disfluencies — showed inconsistent and often opposite relationships with actual deception across studies. Liars, in controlled experiments, often show reduced rather than increased movement because they are managing their presentation. The intuitive model of deception behavior is systematically wrong in important ways.
The Limits of Lie Detection
The most practically important finding in the behavioral observation literature, given how confidently people believe they can detect deception, is the systematic evidence that they cannot.
Charles Bond and Bella DePaulo's 2006 meta-analysis, examining 206 studies and over 24,000 participants, found a mean lie detection accuracy of 54 percent — barely above the 50 percent that random guessing would produce. This held across different populations, different experimental paradigms, and different types of lies. High-stakes lies, where motivation to conceal was presumably higher, were not significantly easier to detect than low-stakes lies.
Critically, the meta-analysis included professional groups often presumed to have specialized detection skills. Police officers, judges, customs officials, psychiatrists, and Secret Service agents all performed at roughly the same accuracy level as undergraduate students with no training. The one group that showed slightly above-average accuracy in some studies — Secret Service agents — performed no better than other professionals in other studies, suggesting the small effect was not robust.
This is not a pessimistic finding for interpersonal intelligence; it is a calibrating one. The ability to read people accurately is real, but it does not reside primarily in the detection of deception from behavioral cues. It resides in the patient accumulation of baseline knowledge about specific individuals, in the careful use of contextual information, and in the maintenance of calibrated uncertainty rather than false confidence.
Practical Interpersonal Intelligence
What the research suggests about actually reading people well is considerably more demanding and less glamorous than the microexpression-detection model implies.
Effective interpersonal intelligence is built primarily through attention to the specific individual over time. The most reliable source of information about what a person's behavior means is your accumulated experience of that specific person in different contexts. This is baseline-building in practice: not a one-time assessment but a continuous updating of the model you hold of someone's characteristic patterns.
High-quality listening contributes more than observation. Research on expert social intelligence — the kind that therapists, diplomats, and skilled negotiators exercise — consistently finds that it is built more on careful listening to the content and structure of speech than on behavioral observation. What people say, how they organize their narrative, what they emphasize and what they omit, and how their stated positions relate to their past stated positions are all more reliable signals than body language.
Prosodic attention — attending to how something is said rather than only what is said — supplements the verbal content usefully. Voice characteristics are harder to control than words, and discrepancies between verbal content and vocal tone are among the more reliable behavioral signals available.
Maintaining calibrated uncertainty rather than false confidence is perhaps the most practically valuable disposition. The person who knows they are uncertain reads people more accurately over time than the person who is confident in their false impressions. Treating behavioral observations as hypotheses to be tested rather than conclusions to be acted on is the epistemic stance that interpersonal accuracy requires.
Practical Takeaways
Establish baseline before drawing conclusions. Before interpreting any behavioral cue as meaningful, ask yourself: is this a deviation from how this person normally behaves? Without baseline, behavioral observation is noise.
Weight vocal cues more heavily than gestural cues for detecting genuine emotional states. Prosodic characteristics are harder to control than gesture and more consistently related to underlying states.
Accept that lie detection accuracy is near chance. If you believe you can reliably detect lying from behavioral cues, the research strongly suggests you are overconfident. This is useful information — it should prompt greater reliance on evidence, context, and explicit communication and less reliance on intuitive deception detection.
Invest in listening quality. Attending carefully to what people say, how they frame it, and what they consistently emphasize is a more reliable path to interpersonal understanding than behavioral observation.
Use observations as hypotheses, not conclusions. When a behavioral observation is interesting, treat it as a question: what might explain this? Consider multiple explanations. Collect more information. Resist the temptation to act on a single observation as though its meaning were unambiguous.
References
- Ekman, P., & Friesen, W. V. (1969). Nonverbal leakage and clues to deception. Psychiatry, 32(1), 88-106.
- Ekman, P. (1978). Facial Action Coding System: A Technique for the Measurement of Facial Movement. Consulting Psychologists Press.
- Ambady, N., & Rosenthal, R. (1992). Thin slices of expressive behavior as predictors of interpersonal consequences: A meta-analysis. Psychological Bulletin, 111(2), 256-274.
- Bond, C. F., Jr., & DePaulo, B. M. (2006). Accuracy of deception judgments. Personality and Social Psychology Review, 10(3), 214-234.
- Gladwell, M. (2005). Blink: The Power of Thinking Without Thinking. Little, Brown.
- Navarro, J., & Karlins, M. (2008). What Every Body Is Saying: An Ex-FBI Agent's Guide to Speed-Reading People. HarperCollins.
- Barrett, L. F. (2017). How Emotions Are Made: The Secret Life of the Brain. Houghton Mifflin Harcourt.
- DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, K., & Cooper, H. (2003). Cues to deception. Psychological Bulletin, 129(1), 74-118.
- Scherer, K. R. (2003). Vocal communication of emotion: A review of research paradigms. Speech Communication, 40(1-2), 227-256.
- Ekman, P. (2003). Emotions Revealed: Recognizing Faces and Feelings to Improve Communication and Emotional Life. Times Books.
- Hall, J. A., & Schmid Mast, M. (2008). Are women always more interpersonally sensitive than men? Impact of goals and content domain. Personality and Social Psychology Bulletin, 34(1), 144-155.
- Zuckerman, M., DePaulo, B. M., & Rosenthal, R. (1981). Verbal and nonverbal communication of deception. Advances in Experimental Social Psychology, 14, 1-59.
Frequently Asked Questions
What are microexpressions and how reliable are they?
Microexpressions are brief, involuntary facial expressions lasting between one-twenty-fifth and one-fifth of a second, occurring when a person attempts to conceal or suppress an emotional reaction. Paul Ekman, a psychologist who spent decades at the University of California San Francisco and then at his own research group, conducted extensive research establishing that a set of basic emotions — happiness, sadness, anger, fear, disgust, contempt, and surprise — are expressed through universal facial configurations across cultures. Microexpressions occur because the suppression of emotional expression is only partially successful; the genuine emotion leaks through briefly. However, the reliability of microexpression reading in real-world lie detection is substantially lower than popular accounts suggest. Controlled studies consistently find that untrained observers detect microexpressions at chance levels, and even trained observers show accuracy rates only modestly above chance in naturalistic conditions.
What is thin-slicing and is it accurate?
Thin-slicing refers to the ability to make accurate judgments about a person, situation, or pattern from very brief exposures — 'thin slices' of experience. The term was popularized by Malcolm Gladwell in his 2005 book 'Blink,' drawing on research by psychologist Nalini Ambady. Ambady's research found, for example, that 30-second silent video clips of teachers were sufficient to produce evaluations of teaching effectiveness that correlated strongly with end-of-semester student evaluations. Thin-slicing accuracy varies considerably by domain. It tends to be higher for characteristics that are genuinely expressed through observable behavior and appearance — extraversion, for instance — and lower for characteristics that are not, such as honesty or intelligence. Gladwell's popular account overstated the accuracy of first impressions and underemphasized the research showing that thin-slicing is subject to significant systematic biases.
How good are people at detecting lies?
The research on lie detection is sobering for anyone who believes they can read deception through body language. A comprehensive meta-analysis by Charles Bond and Bella DePaulo, published in 2006 and analyzing 206 studies with over 24,000 participants, found that the average lie detection accuracy rate is approximately 54 percent — barely above the 50 percent that would result from random guessing. Professional groups often assumed to be skilled lie detectors — police, judges, customs officers, therapists — perform no better than untrained civilians in controlled studies. The popular belief that nervousness, gaze aversion, and other behavioral cues reliably indicate deception is not supported by the evidence. Most of these cues are associated with other states as well and are context-dependent in ways that make them unreliable signals of dishonesty specifically.
What is baseline behavior and why does it matter for reading people?
Baseline behavior refers to an individual's characteristic patterns of expression, posture, gesture, and speech under normal, unstressed conditions. It is the behavioral foundation against which deviations can be meaningfully interpreted. The concept is central to FBI behavioral analysis training and is emphasized by practitioners like Joe Navarro, a former FBI special agent who has written extensively on nonverbal communication. The critical insight is that nonverbal cues are only meaningful when interpreted relative to an individual's baseline, not relative to a universal standard. What looks like nervousness in one person is their normal demeanor. A person who typically avoids eye contact is not displaying an unusual signal when they avoid eye contact in a high-stakes situation. Without establishing baseline, behavioral observations are systematically misinterpreted.
What is the context effect in reading people?
The context effect describes the powerful influence that situational context has on the interpretation of behavioral cues. The same behavior can have entirely different meanings depending on the circumstances in which it occurs. Crossed arms, for example, may indicate defensiveness in a confrontational conversation or simply cold temperature in a chilly room. Reduced eye contact may indicate deception, shyness, cultural norms about eye contact with authority figures, or absorption in thought. Research by Paul Ekman and colleagues emphasizes that no single nonverbal cue has a fixed meaning independent of context — a point often lost in popular body language guides that assign definitive meanings to specific gestures. Effective interpersonal reading requires integrating behavioral cues with contextual information: the setting, the relationship, the stakes of the interaction, and the individual's established baseline.