In 1971 Ekman and Friesen published data from the South East Highlands of Papua New Guinea, among the Fore, a people who at that point had had almost no contact with the outside world. No cinema, no magazines, no Westerners in the villages beyond a handful of missionaries and researchers.
The method had to be adapted, because the Fore had no written language and the standard procedure of matching a face to a word could not be used. So the researchers told stories. A man's friends have come and he is happy. His child has died and he is sad. He is looking at something he dislikes intensely, or something that smells bad. Then they showed three photographs of Western faces and asked which one belonged to the story.
The Fore chose the predicted face at rates well above chance for happiness, anger, disgust, sadness. Fear and surprise were confused with each other. In the reverse test, Fore participants were asked to pose the expressions from the stories, and the resulting videotapes were shown to American college students, who identified them at similarly above-chance rates in the same direction.
The result became the empirical spine of basic emotion theory: a small set of emotions with evolved, hardwired expression programs, shared across the species and decodable across cultures.
It has been contested more or less continuously since, and any honest treatment has to include the objections. Lisa Feldman Barrett and colleagues, in a substantial 2019 review of the evidence, argue that the forced-choice format inflates agreement by supplying the categories, that emotion categories are constructed rather than natural kinds, and that the mapping between facial configurations and emotional states in the wild is much weaker and more variable than the laboratory studies suggest. Free-labeling studies produce lower agreement. Replications in other small-scale societies, including the Trobriand Islanders, have produced different readings of the same faces. There is a real and unresolved scientific dispute here, and the version taught in body-language seminars — that seven expressions are universal, full stop — is a simplification of one side of it.
But the concept from this literature that matters most in practice is not universality. It is Ekman's own qualification: display rules.
Display rules are culturally learned norms governing when an expression may be shown, to whom, and at what intensity. Ekman and Friesen demonstrated them directly by showing stress-inducing films to Japanese and American students. Alone, both groups' faces did the same things. With an authority figure present, the Japanese students masked negative expressions with smiles at much higher rates, and frame-by-frame analysis showed the negative expression beginning before the smile covered it.
That is the operationally important finding, and it is the one to carry into any cross-cultural setting. Whatever is universal is being filtered through local rules about what may be displayed. An expression you do not see may have been produced and covered. An intensity that reads as mild disagreement in one setting may be the maximum permissible display of fury in another.
Which means the naive cross-cultural read fails in both directions: you will miss suppressed signals, and you will over-read permitted ones. Calibration to the person and the context comes before interpretation, which is what the next chapter is about.
See Defense: cultural calibration before judgment