Key Takeaways
Measure facial behavior before treating a face as an emotional verdict
Study the face itself. This edited collection shifts attention from whether observers can recognize posed expressions to what people’s faces actually do during spontaneous behavior. Researchers measure facial movement as an outcome or correlate, then ask how it relates to experience, interaction, development, health, or personality. The studies range from infant taste responses to facial behavior during psychotherapy and coronary disease interviews.
Observation is not interpretation. A visible movement can be recorded without deciding what a person feels. That distinction makes findings comparable across fields, but it does not make the face a transparent window into inner life. Coding is laborious, requires training, and must be interpreted in context.
What is striking is the book’s change in the unit of evidence: instead of asking what a face means to a viewer, it asks what measurable actions occur and when. That shift invites cumulative research, much as standardized observation does in biology. It also restrains overconfidence. A facial movement can correlate with emotion without uniquely identifying it, and a silent face may conceal, regulate, or simply fail to communicate an experience.
Code muscle movements anatomically, then infer emotion with restraint
Describe before labeling. The Facial Action Coding System, or FACS, divides visible facial movement into 44 action units, plus categories for head and eye movement. An action unit describes a visually distinguishable muscle action; it does not name an emotion. A single muscle can contribute to different visible actions, so the system is based on appearance changes rather than a simple one-muscle, one-code map.
Choose measures deliberately. FACS is comprehensive, allowing researchers to discover patterns not specified in advance, but takes substantial training and time. Selective systems such as EMFACS are faster because they code only emotion-relevant actions. Facial electromyography can detect activity invisible to the eye, but electrodes are intrusive and signals from nearby muscles can overlap.
The distinction between measurement and inference is the book’s methodological backbone. It resembles a basic principle of clinical assessment: record the sign separately from the diagnosis so that new interpretations remain possible. The trade-off is real. Comprehensive coding costs time, while selective coding can miss an unexpected pattern. Neither approach eliminates judgment; the key is to make clear which part of the result is observed movement and which part is an interpretation.
Use timing and control to separate reflex from emotion
Startle is unusually fixed. In one study, a sudden blank-pistol shot produced a rapid, brief response involving eye closure and movements of the neck, head, and trunk. The main actions began within 200 milliseconds, and the response usually ended in under a second. Knowing when the shot would occur reduced its intensity but did not eliminate it.
Control provides another clue. Participants asked to suppress the startle could not fully do so, while simulated startles began much later and often included the wrong facial action. Observers distinguished real from simulated startles only about 60% of the time when timing cues were removed. The authors therefore judged startle more like a reflex than an emotion, while emphasizing that the comparison was tentative.
The study offers a useful diagnostic idea: resemblance is not identity. A reflex and an emotion may both change the face, yet differ in latency, duration, and susceptibility to voluntary control. This is a stronger test than relying on the event that triggered the response. Still, the evidence is specific to startle. Emotional episodes are not all alike, and their timing or controllability cannot be inferred from one reflex.
Track feelings and expressions at the moment they change
Compare moment to moment. In a study of 20 women watching emotionally evocative films, researchers used cued review: participants replayed each film and marked moments when they remembered an emotion changing. Facial expressions and reports of experience coincided more often than chance, with overlap around 46% to 47%, compared with an estimated chance level of 30%.
Intensity sharpened the match. At film moments when many participants showed an emotional response, facial expression and reported emotion more often agreed on the emotion involved. The study also found stronger correspondence for more intense responses. Cued review helps locate remembered changes, but it remains retrospective, so memory and labeling can affect the report.
The important move is treating an emotional episode as a changing process rather than a single score taken afterward. That approach fits the temporal character of emotion and prevents a long interval from blending distinct reactions together. Yet the method cannot fully separate real-time experience from later reconstruction. Its best use is as a bridge between continuous behavioral records and carefully timed reports, not as a direct recording of consciousness.
Distinguish enjoyment smiles from social smiles by their mechanics
A smile is not one signal. An enjoyment smile combines lip-corner pulling with contraction around the eyes, often called the Duchenne smile. Research in this collection also finds that enjoyment smiles tend to have more regular timing than other smiles. In a deception study, 31 student nurses who maintained a lie showed fewer enjoyment smiles and more smiles mixed with traces of negative emotion than when they honestly described pleasant feelings.
The group result is not a lie detector. Facial patterns differed across conditions, but many individuals could not be classified from these measures. Other studies found that observers could distinguish enjoyment from nonenjoyment smiles more accurately when comparing two smiles from the same person, especially when the eye action was clear.
The distinction matters because a smile can communicate politeness, concealment, enjoyment, or several things at once. Reading all smiles as happiness collapses those functions and invites false conclusions. The findings support attention to muscle pattern and timing, but they do not justify treating a single facial cue as proof of deception. In everyday interaction, smiles are strategic and socially situated, so their meaning depends on the accompanying behavior and the context.
Check the analysis design before trusting weak expression correlations
The calculation changes the answer. In a study of 61 women responding to humorous material, correlations between judged funniness and facial behavior ranged from .28 to .96 depending on how the same data were analyzed. Within-person designs generally produced stronger correlations than comparisons between people. Aggregating responses across jokes also changed the estimates.
People differ in expressiveness. Some subjects smiled less than their funniness ratings would predict; others smiled at jokes they rated only mildly amusing. Cheerful mood was associated with a lower threshold for smiling, helping explain why comparisons across people can look weak. The study cautions that low correlations do not automatically mean experience and expression are unrelated.
This is a pointed lesson about statistical design: different correlations answer different questions. Within-person analysis asks whether a person smiles more at funnier stimuli; between-person analysis asks whether people who smile more also rate humor as funnier. The study shows why those answers can diverge. It also raises a caution: methodological flexibility should clarify the question, not become a search for the most flattering coefficient.
Read embarrassment as an unfolding act of social repair
The sequence carries meaning. In a study of people reporting embarrassment or amusement after an awkward facial-action task, embarrassment was associated with gaze down, controlled smiling, head movement away, and sometimes face touching. The typical pattern began quickly and lasted about five seconds. Amusement more often began with a smile and included fewer signs of gaze and head avoidance.
Observers could distinguish prototypes. Across judgment studies, observers identified embarrassment more accurately than chance and separated its display from amusement, shame, anger, disgust, and enjoyment. The findings fit an appeasement account: visible embarrassment can signal awareness of a social breach and a wish to repair relations. But its exact form may vary with the situation and culture.
Embarrassment is especially revealing because its meaning lies partly in how actions unfold. Looking down, managing a smile, and turning away can communicate social awareness without a verbal apology. This links facial behavior to the regulation of relationships, not just the expression of private feeling. The account should remain bounded, however: the observed sequence came from particular eliciting situations, and observers’ judgments are not direct proof of the sender’s intention.
Treat pain expressions as evidence, not as a pain meter
A small facial pattern recurs. Across electric shock, cold, pressure, and ischemic pain, four actions were repeatedly associated with pain: brow lowering, tightening around the eyes, upper-lip raising, and eye closure. Combining these actions produced a useful index for distinguishing pain from baseline behavior.
The face is not a calibrated scale. Facial expression showed only a limited relationship to self-reported pain, tolerance, and threshold. Patients could suppress much of their genuine expression, and their deliberately faked pain often resembled the real display, sometimes with greater intensity. Facial measures can therefore add evidence, especially when speech is limited, but they cannot independently establish pain severity or its cause.
The pain chapters make a strong case for using observable behavior when verbal report is unavailable, while warning against turning expression into a numerical substitute for experience. This distinction matters in clinical settings: the face can contribute evidence without serving as a definitive diagnosis. The overlap between faked and genuine pain also shows why judgments should combine facial behavior with history, context, and other indicators.
Use infants’ faces to test what sensation and culture shape
Newborns already discriminate tastes. Infants only hours old showed different facial responses to sweet, sour, salty, and bitter solutions, including hedonic signs such as acceptance of sweet and rejection of unpleasant tastes. The study used a detailed infant-appropriate coding system rather than relying on observers’ broad impressions.
Culture and age still matter. Japanese and American infants showed broad similarities in response to arm restraint, but differed in how quickly some distress expressions appeared at five months. By twelve months, those latency differences were smaller. The findings support early common capacities without implying that development, social experience, or display practices have no influence.
Infant research offers a way to separate early sensory capacities from later social learning, but it does not make that separation simple. A newborn’s facial movement can signal discrimination without proving a fully developed adult emotion. The cross-cultural findings are valuable precisely because they show both commonality and variation. Strong claims about universality need matched situations, multiple response measures, and age-sensitive comparisons.
Link facial patterns to health, but validate every clinical use
Applied findings open new questions. The collection examines facial behavior in depression, schizophrenia, suicide-related assessment, psychotherapy, adolescent adjustment, coronary disease, and smoking craving. In men with coronary artery disease, facial anger and nonenjoyment smiles during a clinical interview were associated with transient myocardial ischemia. In smoking studies, brief waits were accompanied by positive facial responses, while longer waits brought more negative expressions.
Automation can widen access. Early computer systems agreed with manual coding at roughly 80% to 90% for selected actions, but their range was limited. These tools may reduce coding labor, yet performance on a narrow set of movements is not clinical validation. Facial patterns should support, not replace, broader assessment.
The book’s applied studies show the face functioning as a new measurement channel across psychological and physical health. Their value lies less in offering instant diagnoses than in generating testable links between behavior and outcomes. The same caution applies to automation: speed and agreement on selected movements do not establish fairness, generalizability, or clinical usefulness. Those questions require validation in the populations and settings where decisions are actually made.
Analysis
The book’s central contribution is methodological as much as psychological. It turns the face from an illustration of emotion into an observable stream of behavior that can be measured, compared, and related to other systems. FACS supplies a shared vocabulary for that work, while the studies demonstrate why the vocabulary matters: a reflex, a smile, embarrassment, pain, and psychiatric symptoms can look superficially similar to an observer but differ in timing, muscular composition, context, and relation to experience.
The anthology also exposes a persistent tension. Facial actions are measurable, but their meanings are not self-evident. The same action can serve emotional, conversational, defensive, or social functions, and a facial signal can diverge from felt experience. The strongest chapters therefore combine measurement with independent evidence, such as self-report, interaction context, physiological change, or observer judgments. The weaker inference would be to translate a movement directly into a diagnosis or motive.
A modern reading should treat the face as one channel in a multimodal system. Speech, posture, physiology, culture, and the observer’s expectations all help shape what a display means. Automated analysis can increase scale and temporal precision, but it also risks turning imperfect labels into apparently objective scores. Training data, population diversity, consent, and the distinction between describing movement and inferring emotion remain essential safeguards.
The lasting value of this collection is its insistence that careful observation can create better questions. It does not deliver a universal decoder for human feeling. Instead, it shows how to build cumulative evidence, discover patterns that posed photographs conceal, and test claims that would otherwise remain speculation. Its most durable lesson is disciplined curiosity: measure precisely, interpret cautiously, and never confuse a visible face with the whole person.
Review Summary
What the Face Reveals receives mostly positive reviews, with readers praising its scientific approach to understanding facial expressions and deception. Some find it an invaluable resource for improving their ability to read people. The book is described as detailed and research-based, appealing to those interested in the science behind facial expressions. However, a few readers note its technical nature, making it challenging to read at times. Overall, it's considered a comprehensive work on facial expressions and their implications.
Glossary
Facial Action Coding System (FACS)
Comprehensive facial movement codingAn anatomically based system for describing visually distinguishable facial movements. FACS codes action units, their intensity, timing, and combinations, without requiring the coder to infer an emotion. Researchers can later interpret those coded movements in relation to emotion, interaction, or other variables.
Action Unit
A visible facial movementAn Action Unit, or AU, is a minimally distinguishable facial action defined by the appearance changes produced by one muscle or a combination of muscles. AUs are the basic descriptive components used in FACS. They do not, by themselves, name an emotion.
EMFACS
Emotion-focused FACS scoringA selective application of FACS that codes only facial actions judged relevant to emotion, rather than every visible facial movement. EMFACS is faster than comprehensive FACS, but it can miss actions or configurations that were not included in its predefined emotion-related categories.
Duchenne smile
Smile with eye-muscle actionA smile involving both the zygomatic major, which pulls the lip corners upward, and the outer portion of the orbicularis oculi, which raises the cheeks and changes the skin around the eyes. The collection links this pattern especially closely to enjoyment, while distinguishing it from other smiles.
Display rules
Culturally learned expression normsDisplay rules are learned cultural expectations about when, where, and toward whom emotions may be shown, muted, exaggerated, or masked. They can shape observable facial behavior without necessarily changing the emotion being experienced, as in the reported comparison of Japanese and American students viewing stressful films.
Cued-review
Replay-assisted emotion reportingA procedure in which participants review a recording of an emotional stimulus and indicate moments when they remember their feelings changing. It helps align reports with particular moments in a sequence, but remains retrospective and may be affected by memory and emotion labeling.
prototypical affective microsequences
Brief recurring affective patternsShort sequences of facial behavior and interaction that recur during exchanges between a client and therapist. The book uses this concept to describe how affect is regulated between partners, including sequences in which one person’s behavior invites a particular emotional response from the other.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.