The claim “90% of communication is body language” is not supported as a general scientific finding.
Likely origin: It is usually a distorted version of psychologist Albert Mehrabian’s late-1960s experiments on communicating feelings and attitudes. Participants judged messages in which a speaker’s words, tone, and facial expression were inconsistent—for example, a positive word delivered with a negative tone. Mehrabian summarized the relative influence on perceived liking as roughly 7% verbal, 38% vocal, and 55% facial. Combining vocal and facial cues produced the famous “93% nonverbal” figure. This was not a survey, and it did not measure communication in general.
Back-of-the-envelope plausibility check: If 90% of meaning were body language, a text message, email, book, transcript, or phone call would preserve only about 10% of the intended information. A 1,000-word technical document would effectively convey the equivalent of just 100 words. Translation, radio, and screen readers would barely work. In reality, people can communicate detailed facts, instructions, arguments, and stories without seeing the speaker at all. That alone rules out a universal 90% rule.
What the research probably measured: The experiments examined how people infer a speaker’s emotional attitude or degree of liking when verbal and nonverbal signals conflict. Under those narrow conditions, tone and facial expression may carry more weight than the literal word. The studies used small samples, limited stimuli, and isolated words or brief recordings—not ordinary conversation or the transfer of factual information. “Body language” is also inaccurate here because the 38% component concerned vocal delivery, not bodily behavior.
How the number gets misused: It is often presented as if words contribute only 7–10% of all meaning, or as proof that gestures and posture matter nine times more than content. That leap ignores context. Nonverbal cues can strongly affect perceived