Appearance
10.3 — Body Language, Honestly
You have heard that 93 per cent of communication is non-verbal — 55 per cent body language, 38 per cent tone, 7 per cent words.
It comes from two small studies by Albert Mehrabian in 1967, and it does not mean what it is used to mean.
What he actually studied: single spoken words, delivered in tones that either matched or contradicted their meaning, with participants judging the speaker's feeling towards them. The question was: when the channels conflict, which one do people believe?
The finding was about resolving contradiction between channels, in judgements of liking, using single words. It says nothing about communication in general.
Mehrabian himself repeatedly stated that the formula was being misapplied, and pointed out the obvious: if words were seven per cent of communication, you could follow a lecture in a language you do not speak.
The real proportion depends entirely on what is being communicated. For directions to the station, words are approximately all of it. For whether someone is upset with you, tone and face carry a great deal.
The number is wrong. Keep the underlying point, which is that when the channels disagree, people believe the non-verbal one — and that is genuinely useful.
What the body actually carries
Sorted honestly.
Reliable
Arousal level. Fidgeting, stillness, restlessness, speed of movement. You can tell activated from calm, and this is the most reliable body signal there is.
Orientation and attention. Which way someone's feet, torso and eyes point tells you where their attention is. Someone whose feet are pointed at the door is, in a real sense, already leaving.
Distance. How close people stand and how they respond to your approach carries real information about comfort — with the enormous caveat that comfortable distance varies substantially by culture, by relationship and by gender.
Synchrony. When two people are engaged, their movements and postures tend to align — leaning at the same time, matching rhythm. This is one of the better-supported non-verbal findings, and desynchrony is a genuine signal that something is off.
And changes from baseline. Someone who normally gestures a great deal going still, or someone normally still becoming restless. Change is the signal; the absolute posture is not.
Weak or unreliable
Crossed arms as defensiveness. The most quoted item in every popular book. It is also what you do when you are cold, when there is nowhere to put your arms, when the chair has no armrests, and when it is comfortable. Weak signal at best, and it produces confident wrong conclusions constantly.
Touching the face or nose as deception. No reliable relationship (10.6).
Specific gestures with fixed meanings. Almost all of the "leg cross means X, steepled fingers mean Y" material is invented, and gestures vary enormously across cultures.
And any single cue read in isolation. The general rule: no isolated body signal has enough validity to conclude anything.
Power posing, and what happened to it
Worth its own section because it is the clearest case of a popular non-verbal claim being tested and failing.
The 2010 claim: holding an expansive posture for two minutes increases testosterone, decreases cortisol, increases risk-taking and increases felt power.
What replication found: the hormonal changes did not appear. The risk-taking did not appear. A large replication with several times the original sample found only a small subjective feeling of power, which nobody disputed.
One of the original authors publicly stated that she no longer believes the effect is real.
What survives: posture affects how you feel somewhat, and how others read you. That is not nothing. The specific claims that made it famous were not supported.
Mirroring
What is real: people who are engaged with each other spontaneously match posture and movement, and there is evidence that being mirrored increases liking.
What is oversold: deliberately copying someone's posture as a rapport technique. The effect in studies is modest, and being caught doing it is far worse than not doing it — it reads as manipulative because it is.
The useful version is not a technique. It is a diagnostic. Notice whether synchrony is happening. If you are engaged and they are not moving with you at all, something is off, and that is worth knowing.
The baseline principle
This is the single most important idea in reading bodies, and it makes almost everything above usable.
A behaviour means nothing without knowing what this person does normally.
Someone who fidgets constantly, fidgeting: nothing. Someone who is always still, fidgeting: something.
Someone who never makes eye contact, not making eye contact: nothing.
Someone who talks quickly, talking quickly: nothing. Someone measured, suddenly speaking fast: something.
So the procedure is always the same: establish the baseline first, then watch for deviation. And even then, a deviation tells you that something changed — not what changed. Stress, discomfort, excitement, needing the toilet, remembering something they forgot to do. The signal is "look here", not "here is the answer".
Cultural and neurological variation
Distance, touch, gesture, eye contact and expressiveness vary substantially across cultures, and reading someone from a different background against your own norms produces systematic errors — usually in the direction of concluding that reserved people are cold and expressive people are aggressive.
And there is neurological variation. Autistic people frequently have different body language — less eye contact, different gesture, atypical facial expression, stimming movements — and it does not mean what it would mean in a non-autistic person. Reading it as disinterest, dishonesty or rudeness is a common and damaging error, and it happens in interviews, in classrooms and in courtrooms.
Also: people with anxiety look nervous, which reads as guilty. People in pain look irritable. People who are tired look uninterested. In every case the reading is about the state, not about the person or the truth of what they are saying.
The one thing worth training
Not detection. Your own signals.
You cannot reliably read other people's bodies. Other people are reading yours, with the same low accuracy and the same high confidence — and unlike their internal states, yours is under your control.
The three that matter, and they are not complicated.
Orientation. Point your body at the person. Put the phone away and turn.
Stillness. Restlessness reads as impatience or anxiety regardless of what you feel.
And matching the register. Not mirroring as a technique — matching the seriousness of the moment. Someone telling you something painful while you are relaxed and cheerful experiences that as not being taken seriously, and that is the most common non-verbal failure in ordinary life.
What to do with this
Drop the 93 per cent figure and correct it when you hear it. It is one of the most consequential misquotations in popular psychology, because it licenses people to ignore what was actually said in favour of what they think they detected.
Establish a baseline before reading anything. No baseline, no reading.
Treat any single cue as a prompt to ask, never as a conclusion. "You've gone quiet — everything all right?" This converts a guess into data and takes four seconds.
Assume cultural and neurological variation by default rather than as an exception.
And spend the effort on your own signals rather than on detection. Turn towards people, be still, and match the register. That is a larger return than any amount of trying to read theirs.
Next: 10.4 covers the channels that actually carry the most information and get the least attention — the voice, the specific words someone chooses, and what writing reveals.