Nonverbal Communication

Abstract

Nonverbal communication is the exchange of information through channels other than words: facial expression, gaze, gesture, posture, interpersonal distance, touch, and vocal tone. Charles Darwin opened its scientific study by arguing that emotional expressions are evolved and partly universal, a claim Paul Ekman and Wallace Friesen revived with cross-cultural evidence and a taxonomy sorted by origin, usage, and coding. Later traditions mapped the communicative use of space, the varying reliability of each channel, the thin slices of behavior that let observers predict outcomes from seconds of exposure, and the weak, much-overstated cues to deception. What remains contested is how much any channel actually conveys, since the signal proves informative for some judgments and poor for others. This article and three interactive demonstrations present nonverbal communication as a set of channels, a problem of accuracy, and an integral part of language.

Keywords: nonverbal communication, facial expression, proxemics, interpersonal accuracy, emotional expression

People broadcast information without words almost continuously, and observers read it just as continuously: the set of a mouth, the length of a glance, how close a stranger stands. Nonverbal communication is the study of that traffic — the channels through which it flows, what each one reliably carries, and how a receiver turns a stream of movement and spacing into an inference about another mind. The field is old enough to predate experimental psychology itself, beginning with Charles Darwin's claim that the expression of emotion is an evolved endowment shared across cultures and with other animals (Darwin, 1872). It is also broad: it gathers the study of the face, the hands, the eyes, posture, the use of space, touch, and the vocal qualities that ride alongside speech into a single question about how bodies inform one another. The sections that follow lay out the channels, ask how much each actually communicates, and trace the modern shift from treating these signals as leakage of hidden feeling toward treating them as part of language.

What Nonverbal Communication Is

Nonverbal communication is the whole of human communication that is not carried by the words themselves. It is the superordinate category under which the study of the face, of body motion, of the voice, and of the use of space all fall, and in the Medical Subject Headings vocabulary it sits directly above those narrower topics. Its intellectual charter is Darwin's *The Expression of the Emotions in Man and Animals*, which argued that expressions such as the smile, the sneer, and the startle are not arbitrary cultural signs but evolved responses, serviceable in our ancestors and recognizable across human groups and in other species (Darwin, 1872). The claim lay largely dormant until Paul Ekman and Wallace Friesen revived it with evidence that members of a preliterate culture with little exposure to the West nonetheless matched the same facial configurations to the same emotion stories that Westerners did, suggesting a core of cross-cultural constancy in facial expression (Ekman & Friesen, 1971).

Alongside the evolutionary thread runs a descriptive one. Ekman and Friesen proposed that nonverbal behavior can be sorted along three dimensions — its *origin* (reflexive, learned, or cultural), its *usage* (whether it is tied to speech, to emotion, or to the body), and its *coding* (whether its form is arbitrary or resembles what it depicts) — yielding categories such as emblems, illustrators, affect displays, regulators, and adaptors (Ekman & Friesen, 1969). The anthropologist Ray Birdwhistell pushed the descriptive program furthest for body motion specifically, treating gesture and posture as patterned and learnable the way language is (Birdwhistell, 1970). Across both threads the founding insight is the same: nonverbal behavior is neither a transparent window onto feeling nor random noise, but a structured, partly conventional, researchable system. The first demonstration lets the reader sort example behaviors into the channels through which nonverbal information travels.

Select a channel to see what it conveys, an example, and how reliably it can be read.

Note. The channels operate at once and are only partly independent. No single one is a high-bandwidth conduit of hidden truth; each is informative for some judgments and poor for others.

Types of Nonverbal Communication

In the Medical Subject Headings vocabulary, nonverbal communication is the parent descriptor for a family of narrower topics, each naming a channel or a specialized form. This is a classification for indexing the literature, not a theory of the mind — MeSH groups terms so that papers can be retrieved, and its parent-and-child links need not line up with any psychological taxonomy, such as the encoding and decoding stages discussed below. The schemes are orthogonal: a single raised eyebrow is at once a *facial expression* in the MeSH sense and, functionally, an affect display, a conversational regulator, or an emblem depending on its use. The table lists the narrower MeSH descriptors filed beneath nonverbal communication; those that have an article on this site are linked.

Table 1

Narrower MeSH Descriptors Under Nonverbal Communication

Note. These are the descriptors MeSH files directly beneath Nonverbal Communication. Lipreading does not yet have its own article on this site. Proxemics and gaze, though central to the subject, are indexed elsewhere in the MeSH tree rather than as children of this descriptor.

The Channels and What They Carry

The channels of nonverbal communication differ sharply in what they convey. The face is the richest source of emotional information, and the strongest evidence for its partial universality is that the same configurations are read as the same emotions across cultures with very different exposure to one another, consistent with Darwin's evolutionary account (Ekman & Friesen, 1971; Darwin, 1872). Gaze does double duty, signaling attention and intimacy while also regulating the flow of conversation. Gesture, treated in detail under kinesics, binds to speech so tightly that the two are often analyzed as one system. The voice carries a parallel stream of paralinguistic information — pitch, timing, loudness — that colors the words it transports.

The channel that is easiest to overlook is space itself. Edward T. Hall coined the term *proxemics* for the study of how people use interpersonal distance as a communicative medium, arguing in *The Hidden Dimension* that humans carry an invisible structure of zones around the body — an intimate zone for close contact, a personal zone for friends, a social zone for acquaintances and business, and a public zone for formal or impersonal encounters — and that moving across these boundaries is itself a message, read differently across cultures (Hall, 1966). Distance interacts with the other channels: Michael Argyle and Janet Dean showed that intimacy is held near a comfortable equilibrium by trading off closeness against mutual gaze, so that people forced together in a crowd compensate by looking away (Argyle & Dean, 1965). The second demonstration lets the reader move a figure through Hall's four proxemic zones and see where each boundary falls.

Move the two people together or apart. The demo reports which of Edward T. Hall’s four zones the separation falls into and how that zone is typically used.

Note. The boundaries are Hall’s estimates for middle-class North American adults; comfortable distances differ markedly across cultures, which is part of why crossing them unexpectedly reads as a message. Illustrative values from The Hidden Dimension.

Encoding, Decoding, and Accuracy

A nonverbal signal has two sides: a sender who *encodes* a state into behavior and a receiver who *decodes* behavior into an inference. Miles Patterson's sequential functional model organized the field around this exchange, arguing that nonverbal behavior serves identifiable functions — providing information, regulating interaction, expressing intimacy, exercising social control, and serving the practical demands of a task — and that the meaning of a given behavior depends on which function is operative for sender and receiver at that moment (Patterson, 1982). The same lowered gaze can express submission, manage a turn at talk, or simply accompany concentration; decoding requires inferring the function, not reading a fixed code.

How accurate is that decoding? Two findings frame the answer. On one hand, observers are remarkably good at extracting valid information from very little. Nalini Ambady and Robert Rosenthal's meta-analysis of *thin slices* showed that judgments based on brief excerpts of expressive behavior — often under half a minute, sometimes a few seconds — predict consequential outcomes such as teaching evaluations and clinical judgments about as well as far longer observation (Ambady & Rosenthal, 1992). On the other hand, accuracy is neither uniform nor high across the board. Judith Hall and colleagues, surveying the field, describe nonverbal communication as a set of partly independent channels of context-dependent reliability — accurate for some states and judgments, poor for others — rather than a uniform high-bandwidth conduit of hidden truth (Hall et al., 2019). The nonverbal channel is informative, but selectively so, and the clearest case of its limits is deception.

Reading Deception

The folk belief that liars betray themselves through nervous nonverbal tells is one of the most studied claims in the field, and the evidence against its strong form is decisive. Bella DePaulo and colleagues' comprehensive meta-analysis of cues to deception pooled well over a hundred behaviors across scores of studies and found that most putative tells — gaze aversion, fidgeting, posture shifts — bear little or no reliable relation to lying; the few cues that do discriminate are weak and easily swamped by individual differences and situational nerves (DePaulo et al., 2003). Because the cues are faint, even an observer who attends to the right ones gains little: a weakly diagnostic behavior shifts the odds only slightly, which is why laypeople and professionals alike detect deception at rates barely above chance. The worked example makes this reasoning quantitative, and the third demonstration lets the reader see how a cue's diagnosticity and the base rate of lying combine to determine how much a single tell is worth.

Set how often a cue — say, a break in eye contact — appears when people lie and when they tell the truth, and how common lying is. The demo computes the probability that a message showing the cue is actually a lie.

Note. With the default weak cue (60% of lies, 40% of truths) and a 20% base rate, seeing the tell raises suspicion only from 20% to about 27%. Make the cue realistically weaker — as DePaulo and colleagues found real cues to be — and the number barely moves off the base rate. That is why reading body language to catch lies works so poorly.

Table 2

Research Traditions in Nonverbal Communication

Note. The traditions are complementary rather than competing. A single conversational episode can be analyzed for the evolved origins of its expressions, the channels carrying them, the accuracy of their reception, and their integration with the words.

Worked Example

Why does attending to a nonverbal tell help so little in detecting a lie? The reason is Bayesian, and a simple calculation makes it concrete. Suppose a conversation in which one message in five is deceptive, so the base rate of lying is *P*(lie) = 0.20. Suppose an observer watches for a specific cue — say, a break in eye contact — and that this cue appears in 60% of deceptive messages but also in 40% of truthful ones: *P*(cue | lie) = 0.60 and *P*(cue | truth) = 0.40. The cue is more common among liars, so seeing it should raise the observer's suspicion. By how much?

Bayes's rule gives the updated probability that a message is a lie, given that the cue appeared:

*P*(lie | cue) = [*P*(cue | lie) × *P*(lie)] / [*P*(cue | lie) × *P*(lie) + *P*(cue | truth) × *P*(truth)].

Substituting, the numerator is 0.60 × 0.20 = 0.12, and the denominator is 0.12 + (0.40 × 0.80) = 0.12 + 0.32 = 0.44. So *P*(lie | cue) = 0.12 / 0.44 = 0.27. The cue has moved the observer's estimate from 20% to only about 27% — a message showing the dreaded tell is still probably truthful. Now make the cue more realistic, because DePaulo's meta-analysis found real cues far weaker than folklore assumes: let it appear in 55% of lies and 45% of truths. The numerator becomes 0.55 × 0.20 = 0.11, the denominator 0.11 + (0.45 × 0.80) = 0.11 + 0.36 = 0.47, and *P*(lie | cue) = 0.11 / 0.47 = 0.23 — barely above the 20% base rate. The arithmetic captures the core lesson of the deception literature: when a cue is only weakly diagnostic and lying is uncommon, even a correctly observed tell changes the odds very little, which is why nonverbal lie detection performs so poorly (DePaulo et al., 2003). The third demonstration implements exactly this relation.

Discussion

The study of nonverbal communication has moved, over a century and a half, from a search for a universal code to a careful accounting of what each channel reliably conveys. Darwin's evolutionary thesis and Ekman's cross-cultural evidence establish that some expressions, especially facial ones, have a shared biological core; Hall's proxemics and Birdwhistell's kinesics establish that much else is learned and culturally patterned. Between the universal and the conventional lies the field's central empirical question: given a particular behavior in a particular context, how much does it actually tell an observer? The modern answer, from the thin-slices literature to the meta-analyses of accuracy, is that the nonverbal channel is genuinely informative but unevenly so — excellent for rapid first impressions that predict outcomes, mediocre to poor for the kind of concealed-state detection that folklore prizes (Ambady & Rosenthal, 1992; Hall et al., 2019).

What the mature field resists is the two opposite oversimplifications the subject invites. One is the romantic view that the body involuntarily broadcasts the truth, so that a trained reader can detect concealed feeling like print; the weakness of deception cues refutes it (DePaulo et al., 2003). The other is the deflationary view that nonverbal behavior is mere accompaniment to the real business of words; the demonstrated information in gesture and the predictive power of thin slices refute that. Between them, nonverbal communication has become the study of how a receiver, under uncertainty, infers another mind from the body that houses it.

Current Directions

Two strands are especially active. The first reframes the subject as *multimodal language*. Judith Holler and Stephen Levinson argue that face-to-face communication is processed as a single integrated signal in which visual bodily cues and speech are bound together and even speed comprehension, so that gesture, gaze, and the fast machinery of turn-taking are not add-ons to language but part of its core design (Holler & Levinson, 2019). On this view the historical separation of 'verbal' from 'nonverbal' is itself an artifact; what the brain receives in conversation is one multimodal stream. The companion question — whether producing gesture serves the speaker's own thought — is pursued in the gesture-for-conceptualization hypothesis of Sotaro Kita and colleagues, on which representational gestures help a speaker package spatial and motor information for speaking and problem-solving (Kita et al., 2017), and it rests on evidence that gesture carries real information to listeners in the first place (Hostetter, 2011).

The second strand returns to emotion and the body with the tools of affective neuroscience. Eliska Prochazkova and Mariska Kret proposed a neurocognitive model of emotional contagion through mimicry, in which observing another's expression triggers a rapid, partly automatic imitation that feeds back into the observer's own affective and physiological state, synchronizing two people and grounding the perception of emotion in simulation rather than inference alone (Prochazkova & Kret, 2017). Both strands move nonverbal communication away from cataloguing signals and toward explaining how bodies and brains couple during real interaction.

Common Misconceptions

Communication is 7% words, 38% tone, and 55% body language.

The '7-38-55 rule' misstates two narrow experiments by Mehrabian and Wiener on how people resolve an inconsistent message about feelings, where the tone and face outweighed the words. The figures were never a general law of communication, and applying them to ordinary informative speech is unsupported (Mehrabian & Wiener, 1967).

Liars betray themselves through nervous body language.

A comprehensive meta-analysis found that most supposed tells — gaze aversion, fidgeting, posture shifts — bear little reliable relation to lying, and the few genuine cues are weak. Nonverbal lie detection accordingly runs barely above chance (DePaulo et al., 2003).

Body language is a universal code that means the same everywhere.

Only a core of emotional facial expression shows strong cross-cultural constancy (Ekman & Friesen, 1971). Much else — gesture emblems, comfortable interpersonal distance, display rules — is learned and varies markedly between cultures (Hall, 1966).

Glossary

Affect Display.

A facial or bodily expression of emotion, one of Ekman and Friesen's five categories of nonverbal behavior.

Channel.

A distinct medium of nonverbal information — the face, gaze, gesture, posture, interpersonal distance, or voice — each with its own signals and reliability.

Decoding.

The receiver's side of nonverbal exchange: inferring a sender's state or intention from observed behavior.

Display Rule.

A learned, culturally specific convention governing when and how an emotion may be shown, which modulates the universal expressive core.

Emblem.

A gesture with a direct verbal translation, such as a thumbs-up, that can substitute for words; typically culturally specific.

Emotional Contagion.

The tendency to catch another's emotional state, on influential accounts through rapid mimicry of expression that feeds back into one's own affect.

Encoding.

The sender's side of nonverbal exchange: expressing an internal state as observable behavior.

Equilibrium Theory.

Argyle and Dean's account in which intimacy is held near a set point by trading off eye contact, proximity, smiling, and topic; also called affiliative-conflict theory.

Facial Expression.

The configuration of facial muscles that signals an emotional or social state; the most studied nonverbal channel.

Gaze.

The direction and duration of looking, especially mutual eye contact, which signals attention and intimacy and helps regulate turn-taking.

Kinesics.

The systematic study of body motion — gesture, posture, and movement — as a channel of communication.

Mimicry.

The rapid, often automatic imitation of another's expression, posture, or movement, implicated in rapport and emotional contagion.

Nonverbal Communication.

The exchange of information through channels other than words, including the face, gaze, gesture, posture, space, touch, and vocal tone.

Nonverbal Leakage.

The notion that concealed states escape through hard-to-control channels; influential historically but a weak and unreliable basis for lie detection.

Paralanguage.

The vocal qualities that accompany speech — pitch, loudness, rhythm, and timing — carrying information alongside the words; also called vocalics.

Proxemics.

Edward T. Hall's study of interpersonal distance as a communicative system, organized into intimate, personal, social, and public zones.

Thin Slice.

A brief excerpt of expressive behavior, often seconds long, from which observers form judgments that predict consequential outcomes.

Key Researchers

Nalini Ambady

(1959-2013). Social psychologist whose thin-slices program showed that brief samples of expressive behavior predict consequential interpersonal outcomes with surprising accuracy.
Wikipedia - Wikidata

Ray L. Birdwhistell

(1918-1994). American anthropologist who coined the term *kinesics* and founded the systematic study of body-motion communication, modeling it on descriptive linguistics.
Wikipedia - Wikidata

Bella M. DePaulo

Social psychologist at the University of California, Santa Barbara, whose comprehensive meta-analysis of cues to deception showed that nonverbal tells are weak and lie detection is barely above chance.
ORCID - Google Scholar - Faculty Page

Paul Ekman

(1934-2025). American psychologist who, with Wallace Friesen, revived Darwin's universality thesis with cross-cultural evidence and built the origin-usage-coding taxonomy of nonverbal behavior.
Wikipedia - Wikidata - Google Scholar

Edward T. Hall

(1914-2009). American anthropologist who founded proxemics, the study of interpersonal distance as a communicative system, in *The Hidden Dimension*.
Wikipedia - Wikidata

Judith A. Hall

University Distinguished Professor Emerita at Northeastern University; her work on nonverbal communication and interpersonal accuracy includes the field's modern synthesis of what nonverbal cues reliably convey.
Google Scholar - Faculty Page

Mariska E. Kret

Professor of cognitive psychology at Leiden University; her research develops a neurocognitive model of emotional contagion through expressive mimicry in humans and other primates.
ORCID - Wikidata - Google Scholar - Faculty Page

Albert Mehrabian

(b. 1939). Professor emeritus at the University of California, Los Angeles, whose early experiments on inconsistent affective messages produced the widely misquoted figures for the relative weight of the nonverbal channels.
Wikipedia - Wikidata - Faculty Page

Frequently Asked Questions

What is nonverbal communication?

Nonverbal communication is the exchange of information through channels other than words: facial expression, gaze, gesture, posture, interpersonal distance, touch, and vocal tone. It is the broad category under which the study of the face, body motion, and the use of space all fall, and observers read it more or less continuously during interaction (Hall et al., 2019).

Are facial expressions of emotion universal?

A core of them appears to be. Darwin argued that emotional expression is evolved and shared across cultures (Darwin, 1872), and Ekman and Friesen found that members of a preliterate culture matched the same facial configurations to the same emotions as Westerners did (Ekman & Friesen, 1971). Learned display rules then modulate when and how expressions are shown.

Is body language really 93% of communication?

No. The so-called 7-38-55 rule misapplies two narrow experiments by Mehrabian and Wiener on how people resolve a message in which words and tone conflict about feelings; it was never a general law and does not apply to ordinary informative speech (Mehrabian & Wiener, 1967).

Can body language reveal that someone is lying?

Barely. A comprehensive meta-analysis found that most supposed nonverbal tells have little reliable relation to deception, and the few real cues are weak, so lie detection from behavior runs only slightly above chance (DePaulo et al., 2003).

What is proxemics?

Proxemics, named by Edward T. Hall, is the study of how people use interpersonal distance as a communicative system. Hall described intimate, personal, social, and public zones around the body, with boundaries and meanings that vary across cultures (Hall, 1966).

How can observers judge people accurately from just a few seconds?

Through what Ambady and Rosenthal called thin slices: brief excerpts of expressive behavior carry enough valid information that snap judgments predict outcomes such as teaching evaluations about as well as far longer observation does (Ambady & Rosenthal, 1992).

What functions does nonverbal behavior serve?

Patterson's sequential functional model identifies several: providing information, regulating interaction, expressing intimacy, exercising social control, and serving task demands. The meaning of a behavior depends on which function is operative at the moment (Patterson, 1982).

Is nonverbal behavior separate from language?

Increasingly researchers think not. On the multimodal-language view, gesture, gaze, and voice are processed together with speech as a single integrated signal that can even speed comprehension, rather than as a separate nonverbal track (Holler & Levinson, 2019).

References