On January 27, 2008, Penny Boudreau’s twelve-year-old daughter, Karissa, went missing in her hometown of Bridgewater, Canada. That afternoon, mother and daughter had had a fight in a grocery-store parking lot. They’d been having a “heart-to-heart” about “typical teen-age things,” Boudreau said. At 7:30 P.M., Boudreau, worried, called a few friends and teachers—none had heard a thing—and notified the police. By the following day, Karissa was still unaccounted for and the Bridgewater police began notifying other precincts. They issued a media alert and began a full search effort.
On January 29th, the police station held a press conference. Penny, distraught, pleaded for her daughter to return, in a widely televised appeal. On February 1st, she repeated her plea. Anyone with any knowledge of her daughter’s whereabouts, she begged, should make immediate contact. The search parties widened, and local residents joined law enforcement to help track the girl they were now calling “Bridgewater’s daughter.” Still, Karissa remained missing.
On February 9th, two weeks after her disappearance, Karissa was finally found. A woman had pulled over along the side of Highway 331—the road that makes its way along the LaHave River, through Bridgewater—to let her nine-year-old son go to the bathroom out of sight of the passing cars. She rushed to his side when he began screaming: out of the snow, right on the bank of the river, he could see a set of human toes. Bridgewater police soon positively identified the body as Karissa Boudreau. She was naked from the waist down, her body frozen solid by the winter cold, a pair of white Winnie-the-Pooh underwear down below her knees. Five days later, the police announced that the missing-person case had officially become a homicide investigation.
* * *
People lie all the time. According to the psychologist Robert Feldman, who has spent more than four decades studying the phenomenon, we lie, on average, three times during a routine ten-minute conversation with a stranger or casual acquaintance. Hardly anyone refrains from lying altogether, and some people report lying up to twelve times within that time span. I might open a conversation, for instance, by saying how nice it is to meet someone—when I’m really not at all happy about it. I might go on to say that I grew up in Boston—a lie, technically, since I really grew up in a small town about forty minutes outside the city. I could say that the person’s work sounds fascinating, when it’s no such thing, or compliment him on his (drab) tie or his (awful) shirt. And if the person mentions loving a certain downtown restaurant where I’ve had a terrible experience? I’m likely to just smile and nod and say, Yes, great place. Trust me: we often lie without giving it so much as a second thought.
We lie in most any context—Feldman’s work has turned up frequent lies in relationships ranging from the most intimate (marriage) to the completely casual. Some lies are small (“You look like you’ve lost a bit of weight”) and some bigger (“I did not have sex with that woman”). Sometimes they are harmless, and sometimes they are not.
Many of us believe that we can tell when someone else is lying, and, over the years, a folklore has developed around the facial and physical cues that can give someone away. Liars don’t look you straight in the eye. When someone is lying, he looks up and to the side, as if searching for something. A liar fidgets and seems somehow nervous. Sometimes, he’ll scratch or pull his ear. He’ll hesitate, as if he’s not sure he wants to tell you something. These, however, are all “old wives’ tales,” Leanne ten Brinke, a psychologist at the University of California at Berkeley whose work focusses on detecting deception, told me. “The empirical literature just doesn’t bear that out.”
The mismatch between our conception of a liar and the reality—that there’s no “Pinocchio’s nose,” as ten Brinke put it—is surely one reason that, despite our confidence, our ability to tell a lie from the truth is hardly different from chance. The psychologist Paul Ekman, professor emeritus at U.C. San Francisco, has spent more than half a century studying nonverbal expressions of emotion and deception. Over the years, he has had more than fifteen thousand subjects watch video clips of people either lying or telling the truth about topics ranging from emotional reactions to witnessing amputations to theft, from political opinions to future plans. Their success rate at identifying honesty has been approximately fifty-five per cent. The nature of the lie—or truth—doesn’t even matter.
Over time, Ekman did find that one particular characteristic could prove useful—microexpressions, or incredibly fast facial movements that last, on average, somewhere between one-fifteenth and one-twentieth of a second and are exceedingly difficult to control consciously. Those, however, were too fleeting and complex for any kind of un-trained expert to spot: out of Ekman’s fifteen thousand subjects, only fifty people could consistently point them out.
To ten Brinke, something about the existing narrative of deception didn’t quite make sense. Why would we be so bad at something that was so necessary? If the only predictive signs of deception took so much time and energy to learn, that wouldn’t make them of much use. “It didn’t fit well with our evolutionary perspective of human development,” ten Brinke says. “Wouldn’t it have been helpful for us to be able to detect lies and cheats?”
Maybe the untrained “us” wasn’t so bad at lie detection as all that. It could instead be the case that researchers had simply been asking the wrong questions. It wasn’t conscious lie detection, a forced, yes-or-no judgment, that mattered. Maybe, instead, our ability lay in our unconscious perception: in our sensing something if we weren’t looking for it, something that might disappear if we tried to probe it head on. “For lie detection to be an adaptive skill, that helps us to avoid liars and befriend truth-tellers, it doesn’t have to be conscious alarm bells. It could be more subtle,” ten Brinke says. “More of a feeling that you don’t really want to lend this person twenty dollars, that you’re not excited to go on a second date with this guy.” Ten Brinke and her colleagues decided to focus their efforts on finding evidence for unconscious lie detection.
In a series of studies, out this month in the journal Psychological Science, the Berkeley team had students watch a video of a possible criminal who was being questioned about stealing a hundred dollars. As in an actual interrogation, the suspect responded to both baseline questions (“What are you wearing?” “What’s the weather like outside?”) and target questions (“Did you steal the money?” “Are you lying to me right now?”). Half of the potential criminals were lying; half were telling the truth. Each participant watched one truthful and one deceptive video.
Next, the students completed a simple assessment: Were the pleaders in the videos telling the truth? Just as in prior studies, ten Brinke’s subjects, when asked direct questions, did no better than chance at determining who was truthful and who wasn’t.
But then the students participated in one of two unconscious lie-detection tasks. In each, they saw still photos of the two pleaders alongside words that were associated with either truth, such as “honest” and “genuine,” or lies, such as “deceitful” and “dishonest.” Their goal was to categorize the words as indicative of either truth or lies, as quickly and accurately as possible, regardless of the face they saw along with it. If “genuine” flashed on the screen, they would press a button to classify it as a truth-category word as soon as possible.
When the researchers dug deeper, they saw that the participants’ unconscious instinct fared far better: in both studies, they were significantly faster at properly categorizing lie- and truth-related concepts when those concepts were presented with the lying or truthful face, respectively, from the video. Seeing a liar’s face made people faster at classifying lie-related words than truth-related words—and seeing a truth-teller had the opposite effect. “When you see a liar’s face, the concept of deception is activated in your mind even if you’re not consciously aware of it,” ten Brinke hypothesizes. “It’s still unclear just how high a percentage of lies our unconscious mind is able to sense accurately, but discrimination is definitely occurring.”
Unconscious discrimination seems to play out in more life-like scenarios, too. In a series of prior studies, conducted by an unrelated group at the University of Manheim, the psychologist Marc-André Reinhard and his colleagues found that the ability of student judges to detect deception improved drastically if they were given time to think—but only if, in that time frame, they thought about something other than the case they were judging. If they had to make an immediate judgment, they did no better than chance. The same was true if they were allowed to deliberate consciously. But when they were kept from consciously deliberating, by, for example, completing a demanding word-search puzzle, their accuracy improved significantly. Reinhard concluded that, in the unconscious-deliberation condition, the brain had had time to integrate the subtle cues that our conscious mind can’t quite perceive into a more complete judgment.
* * *
In 2008, ten Brinke was still working on her master’s degree, at Dalhousie University, in Halifax, Nova Scotia, about an hour from Bridgewater. For the past year, she had been researching signs of deception in people who turned to television for help in finding a loved one. She had collected dozens of tapes, along with the eventual outcomes of their cases—approximately half the time, it turned out that the pleader was, in fact, guilty of the crime in question. She and her adviser, Steven Porter, had been working on compiling the characteristics of those who were genuinely emotionally distressed—and of those who were faking it. Over time, they had been able to build up a list of behavioral signals that differed consistently in the two groups.
On the afternoon of January 29, 2008, ten Brinke was alone in her office, a small, windowless room tucked in the back corner of the psychology department. She turned to the local news, and there, in the middle of her screen, was Penny Boudreau, pleading tearfully for her daughter’s safe return. Ten Brinke frowned. Something felt—off. “Something is not right,” ten Brinke recalls thinking. “Penny raised some red flags for me.”
Concerned, ten Brinke shared her observation with her adviser. He agreed—but their data was still incomplete and largely unproven. Ten Brinke had been going on a subtle hunch and not much else; she didn’t feel at all confident in her intuition. They waited to see how the case would progress.
On June 14th, after four months of intensive investigation, Bridgewater Police Chief Brent Crowhurst announced that, at long last, the department had more news: Penny Boudreau had been arrested and charged with first-degree murder.
Today, Boudreau is serving a life sentence. She confessed to murdering Karissa—she wanted to save her faltering relationship with her boyfriend, she told the court, and he’d told her that she would have to choose between him and the child—and described in quiet detail her daughter’s final moments. (“Mommy, don’t,” Boudreau says, were Karissa’s last words, as Boudreau straddled her chest and strangled her with twine.)
Would ten Brinke have gone to the police with her hunch today? “I would have much more data to back up my assessment now,” she told me. “And yes, it might be helpful to share with police to aid them in directing their investigation. But it isn’t a silver bullet.” All cues of lying are just that: cues. They are fallible, and they are misleading. No matter how much data you have and how many thousands of hours of training you acquire, you will never have the certainty of a Pinocchio to guide you. And the worst thing you can do? Become so confident in your ability to tell a lie that your conscious certainty gets in the way of your unconscious perceptions. We can tell who’s lying—as long as we don’t think about it too much.