traditional niqab, a veil covering all but her eyes. The judge asked her to take it off. She refused. So the judge dismissed her case. He didn’t think he could fairly adjudicate a disagreement between two parties when he couldn’t see one of them. He told her:
One of the things that I need to do as I am listening to testimony is I need to see your face and I need to see what’s going on. And unless you take that off, I can’t see your face and I can’t tell whether you’re telling me the truth or not, and I can’t see certain things about your demeanor and temperament that I need to see in a court of law.2
Do you think the judge was right? I’m guessing many of you do. We wouldn’t spend as much time as we do looking at people’s faces if we didn’t think there was something valuable to be learned. In novels, we read that “his eyes widened in shock” or “her face fell in disappointment,” and we accept without question that faces really do fall and eyes really do widen in response to the feelings of shock and disappointment. We can watch Ross’s 4C + 5D + 7C + 10E + 16E + 25E + 26E and know what it means—with the sound off—because thousands of years of evolution have turned 4C + 5D + 7C+ 10E + 16E + 25E + 26E into the expression human beings make when filled with shock and anger. We believe someone’s demeanor is a window into their soul. But that takes us back to Puzzle Number Two. Judges in bail hearings have a window into the defendant’s soul. Yet they are much worse at predicting who will reoffend than Sendhil Mullainathan’s computer, which has a window into no one’s soul.
If real life were like Friends, judges would beat computers. But they don’t. So maybe real life isn’t like Friends.
4.
The cluster of islands known as the Trobriands lies 100 miles east of Papua New Guinea, in the middle of the Solomon Sea. The archipelago is tiny, home to 40,000 people. It’s isolated and tropical. The people living there fish and farm much as their ancestors did thousands of years ago, and their ancient customs have proven remarkably durable, even in the face of the inevitable encroachments of the 21st century. In the same way that carmakers take new models to the Arctic to test them under the most extreme conditions possible, social scientists sometimes like to “stress test” hypotheses in places such as the Trobriands. If something works in London or New York and it works in the Trobriands, you can be pretty sure you’re onto something universal—which is what sent two Spanish social scientists to the Trobriand Islands in 2013.
Sergio Jarillo is an anthropologist. He had worked in the Trobriands before and knew the language and culture. Carlos Crivelli is a psychologist. He spent the earliest part of his career testing the limits of transparency. Once he examined dozens of videotapes of judo fighters who had just won their matches to figure out when, exactly, they smiled. Was it at the moment of victory? Or did they win, then smile? Another time he watched videotapes of people masturbating to find out what their faces looked like at the moment of climax. Presumably an orgasm is a moment of true happiness. Is that happiness evident and observable in the moment? In both cases, it wasn’t—which didn’t make sense if our emotions are really a billboard for the heart. These studies made Crivelli a skeptic, so he and Jarillo decided to put Darwin to the test.
Jarillo and Crivelli started with six headshots of people looking happy, sad, angry, scared, and disgusted—with one final picture of someone with a neutral expression. Before they left for the Trobriands, the two men took their pictures to a primary school in Madrid and tried them out on a group of children. They put all six photos before a child and asked, “Which of these is the sad face?” Then they went to the second child and asked, “Which of these is the angry face?” and so on, cycling through all six pictures over and over again. Here are the results. The children had no difficulty with the exercise:
Then Jarillo and Crivelli flew to the Trobriand Islands and repeated the process.
The Trobrianders were friendly and cooperative. They had a rich, nuanced language, which made them an ideal test case for a study of