The participants I was coding—UC Berkeley undergraduates—sat alone in a room staring at a monitor. The experimenter asked the participant to relax and wait for the next task. The participant appeared to lapse into a dissociative state, drifting into thoughts about what the Oedipal complex really means, what the schizophrenic poet on the street corner was yelling about, or whether the day’s hot temperature is another sign of global warming. And then—BAM!—an unexpected 120-decibel blast of white noise, as piercing as a pistol shot. There indeed was the startle response in all of its technicolor glory. People’s faces clenched as they flinched uncontrollably, some almost lurching out of their chairs.
And then I noticed something unexpected. In the first frame after the startle response, people look purified, cleansed, as if their body and mind had been shut down for a second and then turned on—the orienting function of the startle. And then in the next frame their gaze shifted to the side. A knowing, abashed look washed over their faces. People looked as if they had been goosed, or whispered to of something lewd. And then a flicker of a nonverbal display that Darwin had actually missed. Participants averted their gaze downwards, they turned their head and body away, they showed an awkward, self-conscious smile. Some blushed. Some touched their cheeks or noses with a finger or two.
Hastily I took videoclips of six of these participants to Ekman in his office downstairs. As we reviewed these two-or 3-second snippets, Ekman shook his head, first side to side, and then up and down in brief staccato bursts, smiling. He had seen these expressions in the Foré of New Guinea. He knew the contours of an emotional signal whose evolutionary story one could tell. He turned to me with a gleam in his eye. There was a signal of emotion there, one that the field had ignored.
CHARTING CRIMSON FACES
My first step was to embarrass people, a task that has given license to a more mischievous side of researchers’ imaginations. To produce embarrassment in a laboratory, researchers have had college students suck on pacifiers in front of friends. Students have modeled bathing suits for an experimenter, taking notes, clipboard in hand. Young children have been overpraised by adults avidly snapping photographs of the child (prior to eighteen months, children absorb this flood of attention with the aplomb that the removal of a bib might prompt; after eighteen months of age, they show embarrassment). In perhaps the most mortifying experiment, participants had to sing Morris Albert’s song “Feelings” using dramatic hand gestures. At a later date, they returned to the lab to watch a film clip with other students, which turned out to be of their performance of that cloying song.
Before I began my study of embarrassment, Paul Rozin of the University of Pennsylvania recommended two new paradigms, which I never had the temerity to employ. In a first, participants would ride alone in an elevator. Just prior to a group of people entering into the elevator at the next stop, I would surreptitiously plant the smell of a fart, and watch the participant squirm with embarrassment as new individuals entered with eyebrows raised in disbelief. In a second, I would give the participant a handkerchief filled with gobs of mucous. Then another individual, planted by me in the experiment, would ask the participant for the use of the handkerchief.
As appealing as these methods were to my more theatrical inclinations, my production of embarrassment needed to be constrained in certain ways. I had to choose a task in which participants’ heads were relatively stationary so that I could code their facial actions in frame-by-frame analysis (head and body movements can reduce the visible traces of facial muscle actions to impressionistic blurs). I had to ensure that participants did not move their facial muscles after the embarrassing episode, so I could isolate the actions only accompanying embarrassment. In light of these constraints, I had individuals follow muscle-by-muscle instructions to achieve a difficult facial expression, guided by a martinet of an experimenter, all the while being videotaped. The instructions were as follows (try it if no one is looking):
Raise your eyebrows
Close one eye
Pucker your lips
Puff out your cheeks
The experimenter quickly noted, with drill sergeant precision, participants’ deviations from the instructions (“keep your eyebrows raised” “your eyes are fluttering, please just keep one eye closed” “now close your mouth, and don’t press your lips together, pucker them” “remember to puff out your cheeks, and don’t stick out your tongue”). Typically, after a valiant, thirty-second struggle, participants achieved the expression and then were asked to hold it for ten seconds. As participants’ facial muscles quivered and they tried to hold their smiles at bay, they showed visible signs, a furtive glance askance, of imagining what their appearance was, permanently recorded on videotape: They looked like Popeye, drunk on ale, puckering up for a smooch from Olive Oyl, sure to be rebuffed. They were part of some weird joke or an act of absurdist theater. After ten seconds of this pose, participants were asked to rest. It was in the milliseconds after resting that I saw my quarry: the embarrassment display.
With these videos in hand, I spent much of a summer in the coding room of Ekman’s lab, with its cream-colored walls and drawers bursting with electrical plugs, wires, and videotapes. Each fifteen-second snippet of behavior required about half an hour to code, as I charted each twenty-millisecond shift in gaze and discerned the specific muscle actions that define the awkward, embarrassed smile. At the time, most scientists assumed that the display of embarrassment was a jumble of confused actions. In real time my participants did appear rather shaken, uncertain, and disorganized.
Yet with careful frame-by-frame analysis a different picture emerged, and one in line with Darwin-inspired analyses of emotional displays as involuntary, truthful signs of our commitments to particular courses of actions. Our facial expression of anger, for example, signals to others likely aggressive actions, and prompts actions in others that prevent costly aggressive encounters. Within this school of thought, emotional displays are highly coordinated, stereotyped patterns of behavior, honed by thousands of generations of evolution and the beneficial effects displays have on social interactions. Evolved displays unfold briefly, typically between two and three seconds. The brevity of emotional displays is, in part, due to limits on the time that certain facial muscles can fire. Emotional displays are brief, as well, because of the pressing needs facial expressions are attuned to—the approaching predator, the child catapulting toward danger, the flickering signs of interest shown by a potential mate amid many suitors. Involuntary displays of emotion have different temporal dynamics than nonemotional displays: They are gradual in their onset and offset. More voluntary displays, in contrast, like polite smiles, pouts, dramatic glares, or provocative puckers, can come on the face in milliseconds, and remain on the face for minutes, hours, days, or, for some regrettable souls, a lifetime.