They crossed their arms, so they’re defensive. They broke eye contact, so they’re lying. They leaned in, so they’re interested. You’ve absorbed a hundred rules like these, and most of them are wrong.
Reading body language is a real skill with real limits, and the gap between what the research supports and what the internet claims is enormous. This article separates the two: which nonverbal signals actually carry information, which popular rules collapsed under scrutiny, and how to read people in a way that survives contact with evidence.
Two Body Language Myths to Drop First
Almost every popular guide to body language rests on claims the research does not support. Two are load-bearing.
Myth 1: “93% of communication is nonverbal”
You’ve seen the 7-38-55 rule: 7% words, 38% tone, 55% body language. It traces to two small 1967 studies by Albert Mehrabian, and it has been stretched far past what they showed.
Those studies examined a narrow case: how people resolve a contradiction between a spoken word and the tone or facial expression delivering it, using single words, tiny samples, and only female participants. They were never about communication in general. Mehrabian himself has repeatedly said the ratio should not be generalized this way.
The practical damage is real. If you believe words carry 7% of meaning, you will overweight posture and underweight what someone actually told you.
Myth 2: You can spot a liar by watching them
This one is worth internalizing, because it is expensive to get wrong.
Charles Bond and Bella DePaulo synthesized 206 studies covering nearly 24,500 observers judging whether someone was lying. The result: average accuracy was about 54%, barely better than a coin flip. Observers did better on truths (around 61%) than on lies (around 47%), meaning people are actively worse than chance at identifying an actual lie.
Professionals with training did not do meaningfully better. There is no reliable Pinocchio’s nose: no gesture, gaze pattern, or fidget that dependably separates deception from honesty. Anxiety looks like guilt. So does being questioned while innocent.
If you take one thing from this article, take this: you cannot detect lies from body language, and confidence that you can is the dangerous part.
What the Face Actually Tells You
The idea that six basic emotions produce universal facial expressions has anchored body language advice for decades. It is now seriously contested.
A major 2019 review in Psychological Science in the Public Interest by Lisa Feldman Barrett and colleagues examined the evidence for inferring emotion from facial movements across anger, disgust, fear, happiness, sadness and surprise. Their conclusion was that facial movements do not reliably map onto specific emotion categories. People scowl when angry sometimes, but also when concentrating, confused, or reading small print. The same emotion produces different faces across people and situations.
This is a live scientific debate rather than a settled overturning, and researchers in the basic-emotion tradition dispute the review’s conclusions. But the disagreement is itself the point: the confident claim that a given expression equals a given feeling is not something the field agrees on.
What follows practically: treat a facial expression as a question, not an answer. A furrowed brow means something is happening. It does not tell you what.
Reading Body Language: What Actually Carries Information
Stripping out the myths leaves less than the pop-psychology canon promises, but the body language signals that remain are genuinely useful.
1. Deviation from a person’s baseline
This is the single most valuable principle, and it is why generic sign-lists fail. Some people gesture constantly; some are naturally still. Some avoid eye contact from temperament, culture, or neurodivergence.
The signal is not the behavior. It is the change. Someone animated who goes still, or reserved who becomes restless, has told you something. This requires knowing the person, which is exactly why reading strangers is so unreliable.
2. Clusters, never single cues
Crossed arms alone means nothing; people cross their arms because the room is cold or the chair has no armrests. Crossed arms plus turning away plus shortened answers plus a change from five minutes ago is a pattern worth noticing.
3. Timing relative to speech
Genuine gestures tend to arrive with or slightly before the words they accompany. A noticeable lag between a statement and its expression is worth attention, not as proof of anything, but as a prompt to ask a question.
4. Mimicry as a rapport signal
Tanya Chartrand and John Bargh documented the chameleon effect: people unconsciously copy the postures and mannerisms of those they interact with, and this mimicry is associated with liking and rapport. It is one of the better-supported nonverbal findings. Note that it indicates rapport generally, not romantic interest specifically.
5. Distance and orientation
How close someone stands and which way they angle carries information, especially when it changes mid-conversation. Someone who turns away or steps back has adjusted something. Cultural norms for personal space vary widely, so the meaningful unit is again the shift, not the absolute.
What About Power Posing?
Worth addressing, since it is the most famous body language claim of the last fifteen years.
The original 2010 finding suggested expansive postures raised testosterone, lowered cortisol, and increased risk tolerance. A larger 2015 replication by Eva Ranehill and colleagues found no effect on hormones or risk tolerance, though self-reported feelings of power did replicate.
Dana Carney, lead author of the original study, has since stated publicly that she no longer believes the effect is real. Amy Cuddy continues to defend a narrower version. The honest summary: the hormonal claims did not hold up; the “you may feel somewhat more confident” part is modest and survives. Standing tall is fine. Expecting it to change your biochemistry is not supported.
How to Actually Read People Better
The body language techniques that work are less cinematic than the ones that don’t.
Establish a baseline before interpreting anything. In a first meeting you have no baseline, so your read is close to guessing. Spend early minutes observing normal behavior rather than decoding it.
Use nonverbals to generate questions, not conclusions. The correct response to noticing discomfort is asking “how are you finding this?”, not silently concluding they are hiding something. This single reframe converts an unreliable instrument into a useful one.
Ask, then listen to the words. Given how weak nonverbal inference is, the highest-information move is usually a direct question. This connects to the broader skill of how emotionally intelligent people think: they check interpretations rather than trusting them.
Label your own states precisely. Accuracy about others is built on accuracy about yourself, which is why emotional granularity underpins this whole skill.
Account for your own bias. You will find what you expect to find. Once you decide someone is hostile, ambiguous behavior starts reading as hostility, which is the mechanism behind the cognitive biases running your judgment. First impressions form fast and then filter everything after, worth understanding through the neuroscience of first impressions.
Be wary of confident readings of you. People who claim to read you accurately and then tell you what you “really” feel are often doing something else, a pattern that overlaps with everyday manipulation tactics.
Frequently Asked Questions
Does crossing your arms mean you’re defensive?
Not reliably. It is comfortable, it is warm, and chairs without armrests encourage it. Only meaningful as part of a cluster and as a change from that person’s baseline.
Does avoiding eye contact mean someone is lying?
No. This is among the most persistent and least supported beliefs about deception. Gaze avoidance varies with culture, anxiety, autism, shyness, and simple thinking. Liars sometimes make more eye contact, precisely because they know the stereotype.
Can you learn to read microexpressions?
Commercial training programs make strong claims that independent evidence does not clearly support, and the underlying assumption that expressions map neatly onto emotions is itself contested. Treat these programs skeptically.
Is body language different across cultures?
Substantially, for gestures, personal space, eye contact and touch. This is a major reason cross-cultural nonverbal reading goes wrong, and another argument for asking rather than assuming.
What is the most reliable body language signal?
Change from that individual’s baseline, observed as a cluster over time. No universal cue outperforms simply knowing how a specific person normally behaves.
The Takeaway
Honest body language reading looks nothing like the version sold in listicles. There is no gesture dictionary, no lie-detection ability, and no clean map from expression to emotion.
What exists is subtler and more demanding: knowing someone’s baseline, noticing clusters and changes, holding interpretations loosely, and using what you observe to ask better questions. That is less satisfying than a decoder ring, and considerably more accurate.
The people who genuinely read others well are not running a lookup table. They are paying attention, staying curious, and checking rather than concluding.
nn
For the same caution applied to one specific question people ask constantly, see signs someone secretly likes you — where the gap between confident folk wisdom and the actual evidence is unusually wide.
This article summarizes published psychological research and is educational rather than diagnostic. Where the science is actively contested, that is noted in the text.
