Students of language all too often have to sigh and turn away when they hear yet another expert claim that 75% – or is it 85%? Or 93%? – of communication is via body language. Where does this zombie stat come from, and why won’t it go away? I’m delighted that Gonen Dori-Hacohen has taken on the job of tracking it down, and it makes for some fascinating reading.

“Although some have suggested that as much as 93% of conversational meaning is communicated nonverbally (Mehrabian, 1968), more conservative estimates indicate that nonverbal behaviors account for 60 to 65% of the meaning conveyed in an interpersonal exchange (Birdwhistell, 1970; Burgoon, 1994). That is, even conservative estimates ascribe nearly twice as much meaning-making power to nonverbal communication as to verbal—and it is not difficult to understand why, given the number of nonverbal channels and the range of nonverbal behaviors to which people have access.” (Guerrero & Floyd, 2006, p. 2)

The claim that 60 to 93 percent of nonverbal communication is repeated and accepted. And it is nonsense. Not because much of current communication is mediated without bodies. It’s nonsense because imagine seeing me stating it in a TED talk, but in a language, you do not know. Will you understand 90% of the meanings? Or even 65%? You might understand some elements, like my stance, but that is it.

Indeed, Elizabeth Stokoe says that Max Atkinson had used this translation argument to get Mehrabian (the person wrongly cited as 1968 above) to admit that his argument was widely overblown (Stokoe, 2018). Communication is verbal and nonverbal, and reducing it to competition among different channels is wrongheaded, especially claiming that nonverbal is more important to meaning.

This blog is another attempt to explain how we got these claims. Mehrabian was the easy target, so we will get to him at the end, but what about Birdwhistell and his “conservative estimate”? It is probably a joke that has gone wrong.

Tracing the myth

Textbooks and handbooks spread this misinformation. They cite each other, and it took me three generations of books to get to the original research. Let us return to the textbook above (in turn cited as a source by Jones, 2024, a textbook my department wants to adopt). Nonverbal communication gives twice the “meaning” that verbal communication gives. This textbook cites Albert Mehrabian, Ray Birdwhistell and Judee Burgoon. Let’s start with Burgoon (1994), which is a chapter in a handbook (that keeps being reprinted) on interpersonal communication. Burgoon writes: “More reasonable estimate comes from Birdwhistell (1955), who claimed that 60 to 65 percent of the meaning in a social situation is communicated nonverbally. Although he offered no empirical evidence…” (Burgoon, 1994, p. 234). First, it is Birdwhistell’s 1970 and not 1955. Second, if Birtwhistell provided no evidence, why cite him? Burgoon herself cites Philpott’s (1983) MA thesis, a meta-analysis of nonverbal communication experiments, to support Birdwhistell’s claim. However, Birdwhistell and the MA thesis cannot align, mainly because, as we will soon see, the MA thesis quotes Birdwhistell without understanding what he meant.

We need not delve far into Phillpott’s (1983) MA thesis, but we do find him reoporting that Birdwhistell “estimates that ‘no more than 30 to 35 percent’ (1970, p. 157) of meaning is based on verbal information.” (p. 6). He continues: “[a]lthough Birdwhistell’s … figure is presented as no more than a ‘guess,’ it has come to be treated as gospel” (p. 6-7). Thus, like Borgoon, Phillpott acknowledges that Birdwhistell’s number is a “guess,” but it is accepted as a fact. In the remainder of the thesis, Phillpott combines the prior analyses, taking research that was disproven and the one that disproves it to achieve the numbers Borgoon cites. However, the MA thesis makes Birdwhistell a principal of the claim that 65% of communication is nonverbal.

Birdwhistell: Kinesics, Communication, and a Joke?

Ray Birdwhistell was influential in introducing the study of body language to various scientific disciplines. In his influential “Kinesics and Context” (1970) he was: “… trying to demonstrate the necessary interdependence of the kinesic and linguistic;” (p. 17). Interdependence: so, you would think, not separable.

However, when discussing gestures, Birdwhistell writes: “Our present guess is that in pseudostatistics probably no more than 30 to 35 per cent of conversation or interaction is carried by the words.” (1970, P. 157-158). The use of “guess,” the “pseudo” to refer to statistics, followed by a “probably” all suggest non-seriousness. Relatedly, “carrying a conversation” does not equal “meanings” in it. Considering the extreme attention Birdwhistell gave to context and his stress on the interdependence of all communication channels, this sentence probably mocks the “x percent of meaning being verbal or nonverbal” research. Nonetheless, this half-joking sentence is taken as Birdwhistell’s legacy and then used as the “more conservative” finding by people who probably have not read another word Birdwhistell wrote.

Mehrabian: Experimenting with variance

“[A]s much as 93% of meaning in any interaction is attributable to nonverbal communication. Albert Mehrabian asserts that this 93% of meaning can be broken into three parts (Figure 5.2).⁴ Mehrabian’s work is widely reported and accepted.” Fn 4. Mehrabian, A. (1971). Silent messages. Wadsworth.(Quote from Wrench et al. 2020, pp. 159-160).

To be fair, Mehrabian (1971) never claimed the statistic in first sentence. He was interested in “inconsistent communication” (Mehrabian, 1971 p. 42) where “we may express something verbally while our facial expressions [16], posture [111], tone of voice [73, 150, 161], or gestures [39] say the opposite.” (p. 142) From these situations, the myth starts:

“we can say that a person’s nonverbal behavior has more bearing than his words on communicating feelings or attitudes to others. The equation we just presented is a generalization of the research on liking. Total feeling = 7% verbal feeling + 38% vocal feeling + 55% facial feeling”   (1971, p. 44)

This text apparently establishes that verbal only explains seven percent of the expression of feeling, while nonverbal, tonal and facial together, account for 93%, the mythical number. The claim is based on two lab experiments with a similar method.

UCLA students were exposed to two channels of communication, either verbal and tonal (Mehrabian & Wiener, 1967) or tonal and facial expressions (Mehrabian & Ferris, 1967). The channels deliver different stances of the emotion of “likeness,” e.g., if the tone suggested liking, the words did not or if the tone was neutral, the facial expression was not. The students were asked about the meanings of the message, either separated by channel or combined. At the last paragraph of the second paper (Mehrabian & Ferris, 1967), a cumulative result is presented: “[t]he combined effect of simultaneous verbal, vocal, and facial attitude communications is a weighted sum of their independent effects— with the coefficients of .07, .38, and .55, respectively.” (Mehrabian & Ferris, 1967, p. 252)

This conclusion is problematic. Each paper presents self-disclosed limitations, and the combination is not well discussed. Moreover, Mehrabian’s argument is based on many generalizations: lab setting; limited undergraduate population; UCLA;^[1] only female speakers; one word (in Mehrabian & Ferris, 1967) or a single-uttered word (Mehrabian & Wiener, 1967); still photo (Mehrabian & Ferris, 1967); tonal expression (Mehrabian & Wiener, 1967); and one emotion, liking. Even then, Mehrabian only explains the variance in meaning, i.e. how much a channel explains regarding their inconsistency. This fact is usually overlooked. Worse still, even Mehrabian never generalized to all meanings or “Communication:” he studied emotions. Lastly, his analyses were faulty (Borgoon, 1994). Birdwhistell would vehemently reject every element of Mehrabian’s work.^[2] None of these shortcomings prevents textbooks from citing Mehrabian, and then associating him with Birdwhistell to promote the myth of how important nonverbal communication is.

The Alternative View

For ROLSI readers, presenting the alternative is almost not needed. Mehrabian took the outlier of inconsistency between the channels and made it the center. Birdwhistell and Goffman acknowledged that there could be misalignment between communication channels, but they assumed they usually work together. Following Birdwhistell, Adam Kendon posits: “detailed studies of how gesture and speech are interrelated (…) have shown that these two activities are so intimately connected that they appear to be governed by a single process….It has become clear that visible bodily action is often integrated with speech in such a way as to appear as if it is its partner and cannot be disregarded” (2004, p. 3). Goodwin (1979) suggested the intricate relations between gaze and talk and Streek (1993) demonstrated how hand gestures are coordinated with talk.

I’ve deliberately chosen to present such “older” research to demonstrate that the interdependence between verbal and nonverbal is well-established. However, this research is ignored in favor of the 60-93% myth in many interpersonal communication textbooks and mundane outlets. Why does the competitive view between verbal and nonverbal communication, in which the latter wins, continue? Alternatively, who gains from spreading this misinformation?

Conclusion: Agents of misinformation

There are competing views on communication. One starting point is the psychological side: communication resides in the individual, who controls and can change their communication if they choose to or are taught to. Communication can be broken into discrete channels, motions, and utterances and studied in a laboratory outside of context.

Using Birdwhistell (1970) for such a claim is an affront. Birdwhistell argued the opposite: “John does not communicate to Mary, and Mary does not communicate to John; Mary and John engage in communication.” (1970 p. 12) Nonverbal and verbal cannot be separated, and separating a discrete element is close to impossible.

The misinformation regarding nonverbal communication is hard to counter since interdependence is harder to sell. It becomes disinformation since many trainers will teach you how to change your nonverbal communication to win friends and influence people (Carnegie, 1934/2022). If you sign up for my workshop on nonverbal communication, your life will change forever because 93% of all communication is …. To uproot this misconception, we need to point it out, show its stupidity and fallacious roots, preach its alternative, and discuss its reasons. If I’ve tried to achieve that here, and failed, it would probably be because only seven percent of the text had any meaning, since 93% of communication is still considered by too many to be nonverbal…

References

Birdwhistell, R. L. (1970). Kinesics and context: Essays on body motion communication. University of Pennsylvania press.

Burgoon, J. K., (1994). Nonverbal signals. In M.L. Knapp & G. R. Miller (Eds.) Handbook of interpersonal communication (pp. 229-285). Sage.

Carnegie, D. (1934/2022). How to win friends and influence people. DigiCat. 

Goodwin, C. (1979). The interactive construction of a sentence in natural conversation. Everyday language: Studies in ethnomethodology, 97, 101-121.

Guerrero, L. K., & Floyd, K. (2006). Nonverbal communication in close relationships. Routledge.

Jones JR., R.G. (2024). Communication in the Real World (ver. 3). Flatworld Publication

Kendon, A. (2004). Gesture: Visible action as utterance. Cambridge University Press.

Mehrabian, A. (1971). Silent messages (Vol. 8, No. 152, p. 30). Belmont, CA: Wadsworth.

Mehrabian, A., & Ferris, S. R. (1967). Inference of attitudes from nonverbal communication in two channels. Journal of consulting psychology, 31(3), 248-252.

Mehrabian, A., & Wiener, M. (1967). Decoding of inconsistent communications. Journal of personality and social psychology, 6(1), 109-114.

Philpott, J. S. (1983). The relative contribution to meaning of verbal and nonverbal channels of communication: A meta-analysis. Unpublished MA thesis. The University of Nebraska.

Stokoe, E. (2018). Talk: The science of conversation. Hachette UK.

Streeck, J. (1993). Gesture as communication I: Its coordination with gaze and speech. Communications Monographs, 60(4), 275-299.

Wrench, J. S., Punyanunt-Carter, N. M., & Thweatt, K. S. (2020). Interpersonal communication: A mindful approach to relationships. Milne Open Textbooks.

---

^[1] I am a UCLA alumnus.

^[2] On labs Birdwhistell wrote: “We cannot study the social behavior of a fish by taking him out of the water. The child is a child in his world – the pieces he displays in a laboratory represent a very small and, perhaps, unrepresentative sample of his repertoire.” (1970, p. 6) Similarly he rejected a fixed meaning to a fixed gesture.

Groups of EM/CA analysts have sprung up all over the world to share expertise, pore over data together, bounce ideas off each other and provide a sense of shared community. Here Simon Stewart gives an enthusiastic account of recent developments of the group based on the south coast of England.

This post is intended to share with the CA community some of the resources and learning that have come from our group, CA Data Sessions South (CADSS), in its first 18 months.

Continue reading →

Explaining what we do to the general public can be a daunting exercise, but the rewards can be well worth it.  Marina Cantarutti, doing her doctoral research at the University of York, took on the task, and presented her work at a science fair of the kind that hosted Saul Albert and colleagues’ excellent CA Rollercoaster. She lived to tell the (happy) tale…

For some areas of linguistics, it may be a bit difficult to make your work accessible to the public without feeling you are betraying yourself, or your knowledge. The fear of trivialising is always at the back of one’s mind. Moreover, when you’re out there on your own, you are the sole representative of the discipline … daunting! Continue reading →