Leonardo the robot may not be able to feel, but he sure can smile. And wink. And nod. Indeed, one of the reasons why people who interact with Leonardo find it so hard to think of him as a machine is that Leonardo expresses emotions that he doesn’t actually have. An emotional expression is an observable sign of an emotional state, and both robots and people are programmed to make them.

Emotions can be expressed by the tone of our speech, the direction of our gaze, and even the rhythm of our gait. But no part of the body is more emotionally expressive than the face. The muscles of the human face can make 46 distinct patterns known as “action units” (Ekman, 1965; Ekman & Friesen, 1971), and combinations of different action units are reliably related to specific emotional states (Davidson et al., 1990). For example, when people feel happy, their zygomatic major muscles pull up their lip corners while their obicularis oculi muscles crinkle the outside edges of their eyes. Psychologists refer to the resulting expression as “Action Units 6+12” but the rest of the world just calls it smiling.

Why do faces express emotion? In 1872, Charles Darwin published The Expression of the Emotions in Man and Animals, in which he speculated about the evolutionary significance of emotional expression. Darwin noticed that human and nonhuman animals share certain facial and postural expressions, and he suggested that these expressions were meant to communicate information about internal states. It’s not hard to see how such communications could be useful (Shariff & Tracy, 2011). For example, if a dominant animal can bare its teeth and communicate the message, “I am angry at you,” and if a subordinate animal can lower its head and communicate the message, “I am afraid of you,” then the two can establish a pecking order without actually spilling any blood. In this sense, emotional expressions are a bit like the words of a nonverbal language.

254

The Universality of Expression

Of course, a language only works if everybody speaks the same one, which is why Darwin advanced the universality hypothesis, which suggests that all human beings naturally make and understand the same emotional expressions. There is some evidence for this hypothesis. For example, people who have never seen a human face make the same facial expressions as those who have. Congenitally blind people smile when they are happy (Galati, Scherer, & Ricci-Bitt, 1997; Matsumoto & Willingham, 2009), and 2-day-old infants make a “disgust face” when bitter chemicals are put in their mouths (Steiner, 1973, 1979). In addition, people are fairly accurate when judging the emotional expressions of members of other cultures (Ekman & Friesen, 1971; Elfenbein & Ambady, 2002; Frank & Stennet, 2001; Haidt & Keltner, 1999). Not only do Chileans, Americans, and Japanese all recognize a smile as a sign of happiness and a frown as a sign of sadness, but so do members of preliterate cultures. In the 1950s, researchers took photographs of Westerners expressing anger, disgust, fear, happiness, sadness, and surprise (see FIGURE 8.7) and showed them to members of the South Fore, a people who lived a Stone Age existence in the highlands of Papua New Guinea and who at that point had had little contact with the modern world. Researchers asked these participants to match each photograph to a word (such as “happy” or “afraid”) and discovered that the South Fore made matches that were essentially the same as those made by Americans. Evidence of this sort has convinced many psychologists that facial displays of at least six emotions—anger, disgust, fear, happiness, sadness, and surprise—are universal. And a few other emotions—embarrassment, amusement, guilt, shame, and pride—may have universal patterns of facial expression as well (Keltner, 1995; Keltner & Buswell, 1996; Keltner & Haidt, 1999; Keltner & Harker, 1998; Tracy et al., 2013).

The Cause and Effect of Expression

It seems obvious that our emotional experiences cause our emotional expressions. What’s less obvious is that it also works the other way around. The facial feedback hypothesis (Adelmann & Zajonc, 1989; Izard, 1971; Tomkins, 1981) suggests that emotional expressions can cause emotional experiences. And they can! People feel happier when they are asked to make the sound of a long e or to hold a pencil in their teeth (both of which cause contraction of the zygomatic major muscle) than when they are asked to make the sound of a long u or to hold a pencil in their lips (Strack, Martin, & Stepper, 1988; Zajonc, 1989; see FIGURE 8.7). Similarly, when people are instructed to arch their brows, they find facts more surprising; and when instructed to wrinkle their noses, they find odors less pleasant (Lewis, 2012). These things happen because facial expressions and emotional states become strongly associated with each other over time, and eventually, each can bring about the other. These effects are not limited to the face. For example, people feel more assertive when instructed to make a fist (Schubert & Koole, 2009) and rate others as more hostile when instructed to extend their middle fingers (Chandler & Schwarz, 2009).

255

256

The fact that emotional expressions can cause the emotional experiences they signify may help explain why people are generally so good at recognizing the emotional expressions of others. Many studies show that people unconsciously mimic other people’s body postures and facial expressions (Chartrand & Bargh, 1999; Dimberg, 1982). When we see someone smile (or even when we read about someone smiling), our zygomatic major muscle contracts ever so slightly—as yours almost surely is right now (Foroni & Semin, 2009). Because facial expressions can cause the emotions they signify, mimicking another person’s facial expression causes us to feel what they are feeling which allows us to identify their emotions. That’s why people find it difficult to identify another person’s emotions when they are unable to make facial expressions of their own, for example, if their facial muscles are paralyzed with Botox (Niedenthal et al., 2005). People also find it difficult to identify another person’s emotions when they are unable to experience emotions of their own (Hussey & Safford, 2009; Pitcher et al., 2008). For example, people with amygdala damage don’t normally feel fear and anger, and they are typically poor at recognizing the expressions of those emotions in others (Adolphs, Russell, & Tranel, 1999).

Our emotional expressions can communicate our feelings truthfully—or not. When a friend makes a sarcastic remark about our haircut, we truthfully express our contempt with an arched brow and a reinforcing hand gesture; but when our boss makes the same remark, we swallow hard and fake a pained smile. Our knowledge that it is permissible to show contempt for a peer but not a superior is a display rule, which is a norm for the appropriate expression of emotion (Ekman, 1972; Ekman & Friesen, 1968).

People in different cultures have different display rules. For example, in one study, Japanese and American college students watched an unpleasant video of car accidents and amputations (Ekman, 1972; Friesen, 1972). When the students didn’t know that the experimenters were observing them, Japanese and American students made similar expressions of disgust, but when they realized that they were being observed, the Japanese students (but not the American students) masked their disgust with pleasant expressions. In many Asian countries, it is considered rude to display negative emotions in the presence of a respected person, and so citizens of these countries tend to neutralize their expressions. The fact that different cultures have different display rules may help explain why we are good at recognizing the facial expressions of people from other cultures but really really really good at recognizing the facial expressions of people from our own cultures (Elfenbein & Ambady, 2002).

Like most rules, display rules are sometimes difficult to obey. Anyone who has ever watched the loser of a beauty pageant congratulate the winner knows that no matter how hard they try, people can’t always hide their emotional states. Even when people smile bravely to mask their disappointment, for example, their faces tend to express small bursts of disappointment that last just 1/25 to 1/5 of a second – so fast that they are almost impossible to detect with the naked eye (Porter & ten Brinke, 2008). In addition, some facial muscles resist conscious control. For example, people can easily control the zygomatic major muscle that raises the corners of their mouths to make a smile, but most can’t easily control the obicularis oculi muscle that crinkles the corners of their eyes. This fact allows trained observers to tell when a smile is or isn’t genuine (see FIGURE 8.8).

257

Our faces don’t always tell the truth–and neither do our mouths! (DePaulo et al., 2003). When people tell lies, they tend to speak more slowly, take longer to respond to questions, and respond in less detail than they do when telling the truth. Liars are also less fluent, less engaging, more uncertain, more tense, and less pleasant than truth-tellers. Oddly enough, one of the telltale signs of a liar is that his or her performances tend to be just a bit too good. A liar’s speech lacks the little imperfections that are typical of truthful speech, such as superfluous detail (“I noticed that the robber was wearing the same shoes that I saw on sale last week at Bloomingdale’s and I found myself wondering what he paid for them”), spontaneous correction (“He was six feet tall … well, no, actually more like six-two”), and expressions of self-doubt (“I think he had blue eyes, but I’m really not sure”).

Given the reliable differences between liars and truth-tellers, you might think that people would be quite good at detecting lies. In fact, under most circumstances, people are barely better than chance (DePaulo, Stone, & Lassiter, 1985; Ekman, 1992; Zuckerman, DePaulo, & Rosenthal, 1981; Zuckerman & Driver, 1985). One reason is that people have a strong bias toward believing that others are sincere, which explains why people tend to mistake liars for truth-tellers more often than they mistake truth-tellers for liars (Gilbert, 1991). A second reason is that people don’t seem to know what information they should consider and what information they should ignore (Vrij et al., 2011). For instance, people think that fast talking is a sign of lying when actually it isn’t, and that slow talking is not a sign of lying when actually it is. Not only are people bad lie detectors, but they don’t even know how bad they are! The correlation between a person’s ability to detect lies and the person’s confidence in that ability is essentially zero (DePaulo et al., 1997).

When people can’t do something well, such as adding large numbers or moving large rocks, they typically turn the job over to a machine (see FIGURE 8.9). Can machines detect lies better than we can? The answer is yes, though that’s not saying very much. The most widely used lie detection machine is the polygraph, which measures the physiological responses that are associated with stress, which people often feel when they are afraid of being caught in a lie. A polygraph can detect lies at a rate that is better than chance, but its error rate is still remarkably high. The fact is that neither people nor machines are particularly good at lie detection, which is why lying remains such a popular sport among humans.

258