Kinesics: What the Face Signals, and What It Does Not

Body language has an academic name and a genuinely useful taxonomy. It also carries a claim that the evidence does not support.

This article keeps the taxonomy, because it holds up. It replaces the universality claim with what the measurements actually show, which is more interesting than either side of the argument usually puts forward.

The short version is that facial expressions are real signals that are read above chance, and that they are also ambiguous in specific, measurable ways.

The Taxonomy Is Real

Nonverbal communication was broken into fields so each could be measured. Six categories cover most of it, and the boundaries between them are conventional rather than natural.

Kinesics covers body movement: gesture, posture, and the face. Oculesics covers the eyes. Haptics covers touch.

Proxemics covers space, and it is the one category with numbers attached. Chronemics covers time use. Paralanguage covers how the voice carries meaning alongside the words.

For space, Edward Hall set out four distance zones that he observed commonly among North Americans.

Intimate distance runs from touching out to about eighteen inches, and is reserved for people you are close to. Personal distance covers eighteen inches to four feet, which is where friends talk.

Social distance runs from four to twelve feet, and public distance from twelve feet out to twenty-five or more. At public distance a speaker becomes formal.

That framework is a useful description of a pattern. What it is not is a law, and Hall himself was careful that distance also depends on the relationship, the topic, and the room you are in.

Where touch and time sit

Haptics and chronemics are the two categories the original post named and then left alone, so they are worth a line each.

Touch is the first form of communication a human receives, and it carries relationship information more directly than gesture does.

The same action changes meaning with the relationship behind it. A hand on a shoulder reads differently from a hand on a shoulder, depending on who is doing it and why.

How people use time works the same way. Punctuality is read as reliability in some settings and as something else entirely in others.

Neither is universal, and neither is fake. They are conventions that carry real information, which is exactly what the other four categories are.

The Claim That Does Not Hold

Here is the sentence this article was written to fix. Expressions of the face are "universally understood. Regardless of location or culture, people interpret expressions in the same way."

Barrett and colleagues call that assumption the common view, and their review of the evidence concludes that it rests on flawed assumptions and incorrect interpretations of research.

Their argument is not that faces are unreadable. They concede the opposite. People do sometimes smile when happy, frown when sad, and scowl when angry, more often than chance alone would produce.

What they deny is the word diagnostic. A signal that is genuinely produced and genuinely read above chance is not the same as a signal specific enough to identify one emotion from one other.

They document that variation along two axes at once, and both cut against reading a face as a readout.

They also note something the universal claim leaves out completely. The same configuration often communicates something that is not an emotional state at all. A scowl is as likely to be concentration, effort, or discomfort as anger.

Paul Ekman, who argued in the 1960s and 1970s that a small set of facial expressions of emotion is recognised worldwide. That position remains one half of a live argument rather than the settled fact it is often presented as. Paul Ekman Group, LLC, CC BY-SA 3.0, via Wikimedia Commons

Where the argument started

The universality claim is old, and its origin is usually skipped.

Charles Darwin argued in 1872 that human facial expressions of emotion are inherited rather than learned, and that an infant's grimace and an adult's share a root.

The plates in that book were not photographs of people feeling things. Guillaume Duchenne stimulated the facial muscles electrically and photographed what the contractions produced.

A staged composite from 1872: a photograph of a crying infant placed beside a diagram of the face muscles that produce the expression of grief. The muscles were stimulated electrically, so the picture records what the face can do, not what anyone felt. Oscar Gustave Rejlander, public domain, via Wikimedia Commons

That distinction matters and it is easy to lose. A photograph of a muscle being pulled is evidence about anatomy. It is not evidence about what a signal means to whoever sees it.

Ekman's cross-cultural work in the 1960s and 1970s took that inheritance argument seriously and tested it. That is why the basic-emotions position is a serious scientific claim and not merely a popular one.

Why the Categories Blur

The most useful finding here is anatomical, and it explains the ambiguity rather than just reporting it.

The face is built from about 42 muscles that can be controlled largely independently of one another. Researchers call the controllable units Action Units.

Nobody moves all of them at once. An expression is a pattern that unfolds in time, and different muscles join in at different moments.

The instrument that made this measurable

Naming 42 muscles is not itself a finding. What turned the anatomy into evidence was a coding scheme built to record it.

The Facial Action Coding System breaks the face into Action Units: the smallest movements a system can tell apart, whether they are made alone or in combination.

That matters because it replaces a label with a measurement. Instead of scoring a face as angry, a coder records which specific units are active and in what order.

Two people can then be compared on the same recorded units instead of on an impression. That was a genuine advance, and it is the reason the later debate could be stated precisely enough to have.

It also moved the disagreement. Once disagreement is expressed as specific units rather than as competing intuitions about an expression, it becomes testable.

The Glasgow work pushed the same idea into time. Its platform could drive individual units on a three-dimensional face, which let them ask not just which units switch on but when each one joins in.

One of Duchenne's plates recording the muscle configuration associated with fear. Duchenne photographed the face while stimulating the muscles, which is how individual movements could be attributed to specific anatomy. Guillaume Duchenne, public domain, via Wikimedia Commons

Jack and colleagues found that the early part of the signal is shared between categories that are supposed to be distinct.

Fear and surprise both open the eyes wide. Anger and disgust both produce a wrinkled nose. Early in the sequence a viewer cannot reliably separate those two pairs.

The categories only separate later, as additional muscles join in and the signal becomes more specific over time.

Why the early signals blur

There is an evolutionary reason for the overlap, and it is a good one.

The early signals are the cheap ones. They require the fewest muscles and the least time, which makes them the fastest to produce and the fastest for another person to act on.

A wide-open eye and a wrinkled nose do the same job early. They both mean something is wrong, and they both get an information transfer out of the way before anything more specific has time to appear.

The Glasgow team also pointed to a physiological benefit. A wrinkled nose limits the inhalation of harmful particles, and widened eyes take in more visual information.

Both pay off early, so both are selected for. That is a good reason for the signals to be shared.

It also means a fast, coarse signal and a slow, fine one are running at the same time, and they are not always telling the same story.

What the Observer Brings

The second half of the story is about the person doing the reading, not the face doing the signaling.

Jack and colleagues found cultural specificity in both which muscle groups people use and in the timing of those muscles. The pattern is not simply copied across cultures, and it is not simply identical either.

The Glasgow work reported something sharper: observers from East Asian populations place more weight on the eye region and less on the mouth region than Western observers do.

Same face, different weighting. The signal reaches both observers, but they are not treating the same feature as decisive, which is a different thing from receiving a different message.

That is the difference between universality and reliability. A signal can be genuinely shared across humans without every observer extracting the same thing from it.

An illustration from the 1872 book, from the argument that human facial expressions of emotion are inherited rather than learned. That argument is the ancestor of the modern debate, not the last word in it. Charles Darwin, public domain, via Wikimedia Commons

Why this is not merely academic

The common view is embedded in things that make decisions about people, and Barrett and colleagues document where.

They report that defendants perceived as untrustworthy receive harsher sentences, and that the effect is stronger when the defendant looks angry.

An incorrect inference about a defendant's emotional state can cost someone their children and their freedom. A scowl read as rage is not a small mistake.

Agents of the FBI and the Transportation Security Administration were trained to detect intent from facial configuration. A product shipped under the claim that its detected emotions are understood to be communicated universally.

None of this means faces carry no information. It means treating a face as a readout of an inner state is a decision with error, and one that gets made about other people without their say.

What the Categories Are Good For

None of this makes the taxonomy useless, and it is worth saying what the categories do earn their keep on.

They give you vocabulary for things you would otherwise describe vaguely. Kinesics, oculesics, haptics and proxemics are four different things, and knowing they differ stops you treating a spacing problem as a sincerity problem.

They keep analysis separate from interpretation. That a touch occurred is an observation. That it meant affection is a reading, and the gap between them is where most mistakes live.

They make the numbers available. Hall's distance zones give you something concrete to check against. Knowing that personal space runs to about four feet beats an intuition that it felt close.

And they keep the cultural question in view, which is the one the original post closed off by declaring everything universal.

What they do not earn is the right to skip the context. A reading that ignores who is looking, where they are from, and what else is happening is not more objective. It is just less careful.

Key Terms in Plain English

These are the words in this article that would send you to a dictionary. Each one is given here the way it is actually used above.

Term	What it means here
Kinesics	Body movement and gesture, including the face.
Oculesics	The eyes, and what gaze direction signals.
Haptics	Touch, and what it communicates about relationship.
Proxemics	Use of space, including how close people stand.
Chronemics	Use of time, including punctuality.
Paralanguage	Vocal qualities that carry meaning beside the words.
Action Unit	One independently controllable movement of a face muscle.
The common view	The belief that a face reliably reveals an inner emotional state.
Diagnostic	Specific enough to identify one state from all others.
Above chance	Occurring more often than random pairing would give.
Universality	The claim that people read the same signals the same way everywhere.

The words in this article that would otherwise send you to a dictionary.

Key Takeaways

• The taxonomy is sound. Kinesics, oculesics, haptics and proxemics are four different things with different evidence behind them.
• Hall set four distance zones: intimate to 18 inches, personal to 4 feet, social to 12 feet, public to 25.
• Facial expressions are produced and read above chance. That much is not in dispute.
• They are not diagnostic. Variation runs across cultures, across situations, and between people in the same situation.
• A scowl as often signals concentration or effort as anger.
• About 42 face muscles act independently, and an expression unfolds over time rather than firing at once.
• Fear and surprise share wide-open eyes early; anger and disgust share the wrinkled nose.
• The early signals blur because coarse danger signals are cheaper and faster than specific ones.
• East Asian observers weight eye signals more heavily than Western observers do.
• This is why the claim matters: defendants read as angry get harsher sentences.

Frequently Asked Questions

So can you read someone's emotions from their face?

Partly, and with real error. Faces do signal, and people do read them above chance. But the mapping is not specific enough to call it a readout, and the error is not random noise.

Was Darwin wrong?

His 1872 argument was that expressions are inherited rather than learned, and it seeded the whole debate. What followed did not settle in his favor or against him, which is why the argument is still live.

Are the Duchenne photographs evidence of emotion?

They are evidence of anatomy. The muscles were stimulated electrically and photographed working, which tells you what the face can do. It does not tell you what a signal means to a viewer.

Is haptics different from kinesics?

Yes. Kinesics is movement and gesture; haptics is touch, and it carries relationship information more directly than gesture does. The same touch between two people can mean very different things.

Why use Hall's distances if they are only descriptive?

Because a description with numbers beats an intuition without. Knowing personal space runs to about four feet gives you something you can actually check, and Hall treated the zones as patterns rather than rules.

What should I do with a smile from a stranger?

Notice it, and do not treat it as data. The honest reading is that you saw a smile and you do not yet know why, which is a different and more useful sentence than a conclusion.

Key references
• Barrett et al., expressions reconsidered
• Jack et al., evolving hierarchy of signals
• Jack et al., not culturally universal
• Boise State, types of nonverbal communication

Images: Wikimedia Commons, with the author and license named in each caption. Figures and dates follow the four sources above.