Your face says it all? Not so fast

The only problem with this con­cept, according to North­eastern Uni­ver­sity Dis­tin­guished Pro­fessor of Psy­chology Lisa Feldman Bar­rett, is that it isn't true at all.

For nearly two decades, Bar­rett has been tracking down the research that estab­lished this mis­con­cep­tion and wouldn't rest until she actu­ally per­formed the exper­i­ments to dis­prove it.

In two research papers, recently and soon to be pub­lished in the jour­nals Psy­cho­log­ical Sci­ence and Emo­tion, respectively, she's finally done exactly that. The new research calls into ques­tion the very foun­da­tions of emo­tion sci­ence. As Bar­rett found, "Emo­tions are not uni­ver­sally per­ceived. Every­thing that's pred­i­cated on that is a mistake."

Here's how the fal­sity came to be under­stood as fact. In the 1970s, a young psy­chol­o­gist named Paul Ekman trav­eled to Papua New Guinea to test whether emo­tions were uni­ver­sally expe­ri­enced and expressed as he sus­pected. To test his hypoth­esis, he looked at whether people rec­og­nized the same emo­tions in facial expres­sions around the world. Was a scowling face always clas­si­fied as angry regard­less of the observer's cul­tural back­ground? A pouting face as sad?

He showed Amer­i­cans, as well as people in the remote south seas island who'd had little expo­sure to Western cul­ture, a series of pho­tographs depicting car­i­ca­tured expres­sions and asked his sub­jects to match the faces to one of six emo­tion words or sto­ries depicting emo­tional sce­narios. No matter where they came from, Ekman's sub­jects saw the same emo­tions reflected in the same photographs.

But Bar­rett knew from her own research that con­text plays an enor­mous role in the way we per­ceive each other's facial expres­sions. She won­dered whether the con­straints that Ekman put on his subjects -- asking them to match images to finite cat­e­gories and rich sto­ries about emo­tional events rather than freely sort them at will -- might in fact create the result he expected to find.

Enter Maria Gen­dron, a post-​​doctoral researcher in Barrett's lab. In the fall of 2011, Gen­dron and a few other mem­bers of the team boarded a plain to Namibia, then hopped in a Toyota 4x4 for an hours long, off-​​road ride to one of the most remotely sit­u­ated tribes on the con­ti­nent. The Himba, Gen­dron said, were as little accli­mated to Western cul­ture as she could find.

Par­tic­i­pants in the Namibian Himba tribe did not rec­og­nize the same emo­tions in facial expres­sions and vocal­iza­tions as Amer­ican par­tic­i­pants. Photo cour­tesy of Maria Gendron.

She spent the next 18 days -- and then another 20 during the spring of last year -- sleeping in a tent atop the car by night and searching for uni­versal emo­tions by day. She didn't find any.

Gen­dron looked at both facial expres­sions and vocal­iza­tions, hypoth­e­sizing that if emo­tion truly is uni­ver­sally rec­og­niz­able, the medium of expres­sion shouldn't matter.

First Gen­dron gave her sub­jects 36 photos of faces (six people posing each of six expres­sions) and asked them to freely sort the photos into piles based upon sim­ilar facial expression.

"A uni­versal solu­tion would be six piles labeled with emo­tion words," Bar­rett said. "This is not what we saw." Instead the par­tic­i­pants cre­ated many more than six piles and used very few emo­tion words to describe them. The same photo would end up in var­ious piles, which the sub­jects labeled as "happy," "laughing," or "kumisa," a word that roughly trans­lates to wonder.

The vocal­iza­tions fared no better. This time, Gen­dron asked people to freely label the sounds. Again, few emo­tion words were used. The same sounds seemed gleeful to some sub­jects and dev­as­tated to others.

Finally, Gen­dron and Bar­rett repeated the exper­i­ment back in Boston, so they could com­pare the results to a group living in Western cul­ture. The results were sig­nif­i­cantly dif­ferent. "The par­tic­i­pants in Boston were able to label the expres­sions with the expected terms but fared better when the words were pro­vided as part of the task," Gen­dron said. This indi­cates that what were assumed to be "psy­cho­log­ical uni­ver­sals" may in fact be "Western" -- or per­haps even "American" -- cultural cat­e­gories, she said.