Non-verbal communication, oriented

Researchers define non-verbal communication as messages sent without words, through behavior, appearance and setting. The field sorts it into channels: kinesics (body motion, including gesture, posture and facial expression), oculesics (eye behavior), haptics (touch), proxemics (space and distance), vocalics (the voice apart from the words), chronemics (time), and appearance and artifacts (body features, clothing and objects). Behavior becomes communication when someone reads it as a message, intended or not, and the channel is always on: falling quiet is not the same as sending nothing.

Words can assert a checkable fact. A smile with prolonged gaze and a close approach asserts nothing; observers read it for what it says about the relationship, most centrally dominance and affection, and that reading is a guess. Most non-verbal behavior has no vocabulary or grammar, with emblems such as the thumbs-up as the exception. Much of it runs outside awareness, and body language is often assumed to reveal more truth than words, though one view of faces calls the idea that they leak hidden feeling unverified.

The common mistake is treating a cue like a dictionary entry. Say a friend scowls when you ask about the weather. The scowl could mean the weather is bad, the question seems silly, she is preoccupied, or she is still upset about yesterday. Her posture or gaze might settle it, or might not, and setting, culture and personality all shift what a pattern means.

Rule: treat a single cue as a question rather than an answer, and look for other evidence in its context before you settle on a reading.

Non-verbal research shares the usual threats to reliable science, such as low statistical power and publication bias, in which positive, novel results reach print more often than null ones. A significant result is not necessarily a large one, since significance also depends on sample size, and one failed replication does not settle a question either.

The 93 percent rule, usually given as 7 percent words, 38 percent tone and 55 percent body language, traces to Mehrabian's 1967 studies. People judged a stranger's attitude, in one study from the single word "maybe", in the other with tone and word content set against each other. Both measured judgments of attitude, meaning feeling or liking, and the percentages describe how people weigh channels that conflict; retellings stretched them into 90 percent of all information. The defensible version: non-verbal cues weigh heavily in judging feelings and relationships, with no fixed share of meaning.

Power posing followed the same arc. The original study reported that two minutes of expansive posing raised testosterone, lowered cortisol and increased risk taking, and a talk on it went viral. Larger studies, and meta-analyses that corrected for publication bias, found no reliable effect on hormones or risk taking. What survives is a small rise in self-reported feelings of power, which may partly reflect participants giving the answers they think the study expects. In one large experiment, observers rated people in expansive poses as more competent; what a pose does to the poser is the disputed part.

The trap is treating print or popularity as evidence, and journal articles still restate the 7-38-55 split. Say a workshop tells you to hold a power pose before an interview so your hormones shift. Studies that found such effects tended to be small, and large preregistered ones mostly found none.

Rule: trace a body-language claim to its original study, check what it measured, and give more weight to large preregistered replications than to a viral first result.

Ekman and Friesen's scheme sorts body movements by what they do. Emblems carry an agreed meaning inside a group, like a thumbs-up, so they can replace words. Illustrators depict or stress what is said, such as hands held apart to show a fish's size. Regulators manage the flow of talk, like an open palm offering someone the turn. Adaptors, such as scratching or fidgeting, are thought to carry no deliberate message, and affect displays convey an emotional state.

Set against speech, a cue can repeat it (saying "left" and pointing left), complement it (saying "here" and pointing left, so only the hand gives the direction), accent a syllable with a quick beat, regulate turns, substitute for words, or contradict them. The same motion can do several of these jobs, so its function is read from the words around it.

Gesture also helps the person making it. Speech and gesture appear to be planned together, and people gesture more when a word is hard to retrieve, though some studies disagree. Listeners often understand better when gestures accompany the words, and a child whose gestures express a strategy the words do not is often ready to learn that problem.

The trap is assuming a gesture means the same everywhere, or that fidgeting means nerves. The A-OK circle signals approval in English-speaking settings but is obscene in Brazil. Observers judge liars by their fidgeting, yet a meta-analysis found liars make fewer movements than truth tellers, perhaps because lying adds mental demand. Say you are presenting to a client abroad: check what a gesture means there before you let it stand in for words.

Rule: read a gesture against the words it accompanies, and never assume an emblem means the same thing in every place.

The Facial Action Coding System breaks visible facial movement into numbered action units, each tied to its muscles; the zygomaticus major, for example, pulls the lip corners up in a smile. The system itself names no emotion. Reasoning backward from a face to the feeling behind it, called reverse inference, can mislead: actors posing emotional scenarios varied their faces within each emotion, and only fear, happiness and surprise came out with moderate reliability. In a meta-analysis, once clarity and congruence were accounted for, faces were no more influential than body postures.

The face may also feed back into feeling. People holding a pen in their teeth, which engages a smiling muscle, rated cartoons as funnier than people holding it in their lips. A multi-lab replication with 1,894 participants found no such effect, yet a meta-analysis of 138 studies found a small, variable one.

A Duchenne smile adds an eye-narrowing cheek raise to the lip-corner pull, and observers read it as more felt, yet people can produce that eye action on purpose. One theory sorts smiles by job: reward smiles share enjoyment, affiliative smiles signal friendliness, and dominance smiles assert status. Display rules, learned norms about which feelings to show, when and to whom, lead people to dampen, intensify, neutralize or mask an expression.

The trap is treating a smile as a readout of happiness. Studies put accuracy at telling genuine from posed smiles around 55 to 70 percent, and in one study observers named the feeling behind a masking smile only 20 to 50 percent of the time. Say a server smiles warmly all evening. That smile may share enjoyment, signal friendliness or meet the job's demands, and managing only the outward display, called surface acting, is linked with emotional exhaustion.

Rule: describe what the face did before you guess what it means, and let the situation, not the smile alone, carry the inference.

Basic emotion theory, popularized mostly by Paul Ekman, holds that certain evolved emotions, including at least happiness, sadness, fear, anger and disgust, each come with universal signals. Early support came from cross-cultural studies matching posed faces to emotion words. Critics answer that choosing from a short list can inflate apparent agreement, and that exporting English emotion words assumes each has an exact local match. The theory of constructed emotion holds instead that emotions are built from past experience, bodily state and the situation, so a category need not come with a specific face.

Both sides have evidence. A review of 21 studies found facial expression production largely universal, with culturally variable nuances. In a test with Hadza hunter-gatherers, agreement with US participants was closest when pleasantness separated the faces; when neither pleasantness nor arousal could, only 58 percent of Hadza participants beat chance, against 94 percent of US participants.

Culture also decides where a message lives. Hall described low-context communication as carrying most of the information in the explicit message, and high-context communication as leaving most of it in the setting and in what people already share. Research links collectivism with high-context, indirect styles and individualism with low-context, direct ones, and every culture uses both, though one tends to dominate. People recognize emotion more accurately when the person expressing it comes from their own cultural group.

The trap runs two ways: assuming a face means the same everywhere, and turning a cultural average into a rule about one person. National character stereotypes have not matched measured personality traits, and cultural averages hide wide individual variation. Say a new colleague from abroad rarely meets your eyes. The national pattern cannot tell you whether that reflects respect, heavy thought or something else.

Rule: expect regularity in broad strokes and variation in detail, and never read a person through their culture's average.

Gaze shows where attention is directed, and it also regulates talk, expresses intimacy and exerts social control. Mutual gaze is arousing, and conversation holds it only in brief stretches; museum visitors preferred an actor's direct gaze for about 3.3 seconds. Speakers tend to look at the listener as they finish a turn. Pupils widen with arousal largely beyond voluntary control, but dilation also comes with pain or exercise, so it signals arousal, not its cause.

Immediacy behaviors, such as eye contact, facing someone directly, leaning forward and a relaxed posture, signal closeness and work together as a system. Orientation defines who is inside a group conversation: in Kendon's F-formation, members arrange themselves around a shared space each can easily reach. Expansive postures go with holding power and constricted ones with lacking it, a correlation separate from any claim that posing changes the body.

The trap is reading one cue as a verdict. Averted eyes come with heavy thinking, planning speech, very personal questions and, in Japanese culture, respect. The Neuro-Linguistic Programming claim that looking up and to the right marks a lie found no support in three tests. Crossed arms alone are an imprecise guide to feeling, and movement level fails too: moderate anxiety can bring more fidgeting, while heavy pressure may freeze someone into rigid stillness. Say a job candidate looks away while answering a hard question. That fits heavy thinking as well as anything else, and the glance alone cannot say which cause is at work.

Rule: read gaze and posture as patterns in context, never as a single tell.

Edward Hall proposed four distance zones, intimate up to about 45 cm, personal to about 1.2 m, social to about 3.65 m and public beyond that, drawn from observation alone. Spacing is negotiated in the moment: people let friends closer than strangers, and in a paper-and-pencil task across 42 countries, average preferred distance ranged from 77 to 140 cm. When someone comes too close, people step back or add distance another way, by looking away, falling silent or turning aside. Equilibrium accounts treat closeness as balanced across channels, so standing closer can be offset by less eye contact, though friends and partners sometimes return a close approach instead.

Space also works as territory. Altman separated primary territories, one person's exclusive domain, from secondary ones tied to personal connections and public ones open to all, and people mark a space as theirs with decor and objects. Density is a physical feature of a setting, while crowding is the felt sense of having less privacy than you want; in housing studies, perceived crowding was more strongly linked to aggression than density was, a correlation rather than a cause. Seating matters too: children seated near each other in grade 3 to 5 classrooms were more likely to form new friendships, though teachers assigned the seats.

Touch carries meaning as well. People can recognize emotions such as anger, gratitude and sympathy from touch alone, and observers saw whoever started a touch as more dominant. Cultures are often labeled high-contact or low-contact, but in a UK and Japan comparison closer bonds allowed more touch in both countries, while UK participants found social touch more pleasant. One study concluded that without closeness and consent, even comforting touch can feel distressing.

The trap is trusting a tidy rule. Servers who briefly touched diners were reported to get bigger tips, yet several controlled lab tests of touch and generosity found no effect. Say a manager pats every employee on the shoulder to seem warm. Touch sits on a fragile line between reassurance and control, and one proposal holds that it helps only when it conveys care and matches the intimacy the receiver prefers.

Rule: treat distance, territory and touch as negotiated, and let the other person's response, not a chart of zones, tell you what is welcome.

The pitch movement at the end of an utterance helps mark a question, and a rising ending is often heard as less confident. Louder and faster speech is usually heard as more confident. Listeners respond faster to a word that follows a pause, whether an um or a silence.

The voice carries emotion through brief sounds such as laughs and cries as well as through the tone of speech, and in two studies listeners identified emotion better from the sounds. Listeners agreed on impressions such as dominance from just 25 milliseconds of a vowel, but agreement among listeners is not accuracy. Lower male pitch was heard as more dominant.

Time is a channel too. Gaps between conversational turns are short across languages, around 200 milliseconds in one study, so speakers seem to plan replies before the other person finishes. In corpus studies, gaps of 700 milliseconds or more after a question go with answers that resist what was invited, and in experiments gaps of 600 or more led listeners to that inference. Hall contrasted monochronic time, one task after another, with polychronic time, where activities and relationships run together. Waiting is tied to power and status, and two people interviewed about waiting for prison health care described feeling second-class. People also differ in how past, present and future frames guide them, though a widely used questionnaire for this has replicated unevenly across countries.

The trap is judging the person by the voice. Accented speakers are often rated lower in hiring, and whether an accent lowers belief in what is said is contested. Say a caller pauses for a second before agreeing to your plan. The pause is a reason for a follow-up question, not a conclusion.

Rule: use pitch, pace and pauses on purpose, and hold voice-based impressions of other people loosely, since quick agreement is not accuracy.

One glimpse of a face is enough for people to form impressions such as trustworthiness, and those impressions have predicted online lending and even death penalty decisions, yet their accuracy is limited at best. Short excerpts of behavior, called thin slices, can support some judgments: in one study, personality judgments beat chance after just 5 seconds for every trait tested except agreeableness, and going from 60 to 300 seconds added little. Attractiveness strongly shapes ratings of social competence and intelligence but only weakly shapes ratings of moral integrity. Clothing is read for role, group and status, and a visitor's chair facing a supervisor across a desk, often set lower, signals more power than seating side by side.

Biases then shape the reading. A halo effect spreads one positive trait to a person's other traits, and a horn effect spreads a negative one. Information from the start of an interaction can bias the whole impression. Observers tend to explain other people's behavior by character rather than circumstance. And research on self-fulfilling prophecies holds that a teacher who wrongly expects less of a student may treat that student in ways that draw the expected poor performance.

The trap is mistaking a fast, shared impression for a true one. Accounting managers rating photographed candidates favored the more attractive ones. Say a reviewer calls an applicant unprofessional because of a tattoo. That is a judgment about appearance, and in one survey private acceptance of tattooed teachers exceeded the acceptance people attributed to others. Researchers recommend checking the reason for a behavior instead of assuming it.

Rule: notice when an impression rests on looks or a first moment, and find out the reason behind a behavior before you judge it.

Researchers treat rapport as a quality of the pair, built from mutual attention, an exchange of positive cues and, they argue most importantly, coordinated non-verbal behavior. Synchrony, partners adapting to each other's rhythms, often arises on its own and does not require copying. People tend to mimic those they like and like those who mimic them, though a preregistered virtual-reality test found avatar mimicry did not raise rapport. Follow-up questions and validation predicted connection in deep talk with strangers, while nods and uh-huhs are weaker evidence of listening.

Text strips out face, voice and body, so writers carry tone with stretched letters, capitals and emoji, and the same emoji can mean different things to different age groups. On a typical video call the camera sits above the on-screen eyes, so someone trying to make eye contact seems to look lower, and constant self-view is a suspected source of video fatigue. In one messaging experiment, a partner who replied at once seemed more accessible than one who replied after seven minutes.

The trap is deliberate mirroring. Mimicry is tied to liking mainly when it goes unnoticed, and being caught copying someone can backfire. Say you copy a client's posture to build rapport. If the client notices, it can cost the liking it was meant to win.

Rule: build rapport by attending and responding rather than copying, and on a screen put into words what the missing cues would have carried.

A meta-analysis of 206 documents and 24,483 judges found people sort lies from truths about 54 percent of the time, where guessing gives 50. Law enforcement personnel did only marginally better, and confidence was almost unrelated to accuracy. Gaze aversion, the most widely believed cue, does not separate liars from truth tellers. What a person says carries more signal than how they move: liars tend to report fewer details that an investigator could check. Cognitive-load interviewing tries to make lying harder than telling the truth, with unexpected questions about minor details, surprise drawings or recall in reverse order, though a 2018 review says the methods need validation and rehearsal may undermine them.

A micro-expression is a very brief, usually faint facial movement said to escape when someone conceals a feeling. In one study only about 22 percent of participants showed any, and they were no more frequent during deception. Training raises scores on the trained recognition task, with weak evidence of transfer. One mock-crime study found brief negative expressions separating liars from truth tellers, offered as first evidence of a possible cue, so these expressions are real but fall well short of a reliable detector.

The trap is mistaking a manner for honesty. Levine argues that many people's demeanor does not match their honesty, and an honest autistic person, for example, may seem deceptive. In the Othello error, an honest person who is not believed grows distressed and so looks the way a liar might. A US airport program trained officers to watch passengers against a checklist of behaviors, and a government review concluded that published research did not support such indicators for identifying deception. Say an interviewer sees a suspect look away and tense up. Those cues fit fear of being accused as well as lying.

Rule: never read a lie from demeanor; when honesty matters, ask about details that can be checked.

Decoding skill is measured with performance tests: you watch short clips, say what the person feels, and are scored against a key. How good people say they are shows little or no relation to those scores. Training mostly improves the specific skill trained, with little sign that gains spread to other cues. People also differ widely in how readable they are; in groups of strangers, more facially expressive members were rated warmer and sat more centrally in liking networks.

Expression and reading vary between people. Autistic and non-autistic adults produce and interpret cues differently, and the double empathy problem describes misunderstanding that runs both ways, so researchers warn against judging neurodivergent behavior on neurotypical terms. Culture, personality and trauma can make expression more restrained, and automated emotion systems may misread restrained or unusual expression as flat, deceptive or insincere.

The trap is turning a reading into a verdict about someone's honesty or worth. Police manuals have listed poor eye contact and restless movement as signs of lying, airport screeners were taught similar cues, and jurors are told to notice witnesses' behavior; those at risk include nervous truth tellers. Say a quiet colleague rarely smiles in meetings. That describes how she expresses herself; it does not tell you whether she is engaged, and asking her could surface the real reason.

Rule: treat every reading of another person as a hypothesis, confirm it with other evidence, and never use cues to judge honesty or worth.