At the time, one prevailing theory within psychology said that, in order to understand others—and persuade them to listen to us—we should engage in what is known as perspective taking: We should try to see a situation from the other person’s perspective and show them we empathize. Psychology journals noted that “to communicate effectively, we must adopt the perspective of another person both while speaking and listening.” Textbooks taught that “perspective taking not only fosters greater interpersonal understanding” but also “constitutes a vital skill for very powerful negotiators.”

  When Epley looked back on his high school experiences, he realized that his parents, in their own way, had been trying to engage in perspective taking after his drunk driving near arrests. They had tried to put themselves in his shoes. They had tried to forge a connection by imagining the pressure he was under. They had hoped that demonstrating their empathy would convince him to listen to their advice.

  But, if anything, his parents’ perspective taking had made it clear to Epley, at that moment, how much they didn’t understand about him. When they tried to commiserate, when they shared stories of their own adolescent mistakes, all Epley heard was adults who had no idea what it was like to be a teenager nowadays.

  His parents had failed to connect with him because they hadn’t understood how he felt. And they didn’t understand because they had never asked. They had never inquired about his anger or uncertainty, had never asked why it had felt so necessary to prove himself by drinking all those beers. Even if they had asked, Epley wouldn’t have known what to say. He hadn’t understood, himself, what was going on inside his head, not until he had started talking to the counselor and she, instead of trying to put herself in his shoes, had simply asked him questions that elicited emotional replies: “Why are you making these choices?” “Is this who you want to be?” Then she had listened and had asked smart follow-up questions, and that, somehow, had inspired Epley to start listening to her, and then to himself, until he realized he needed to change.

  Now, as an adult, Epley wondered if the psychology textbooks had it wrong. Perhaps the correct approach wasn’t trying to put yourself in “someone else’s shoes.” That, after all, was impossible. Rather, maybe the best you can do is ask questions. Ask about people’s lives, about what they’re feeling, about their hopes and fears, and then listen for their struggles, disappointments, joys, and ambitions.

  Hearing people describe their emotional lives is important because when we talk about our feelings, we’re describing not just what has happened to us, but why we made certain choices and how we make sense of the world. “When you describe how you feel, you’re giving someone a map of the things you care about,” Epley said. “That’s why I connected with my parents, because I finally understood what mattered to them. I understood they were scared and worried and just wanted me to be safe.”

  This is why the How Do We Feel? conversation is so crucial. Every discussion is shaped by our emotions, and when we bring those feelings to the surface—when we share them and ask others to share with us—we begin to see how we might align.

  Epley began thinking there must be an alternative to perspective taking. Maybe there was a different technique to help people ask the kinds of questions that nudge emotions into the open? Perhaps, instead of perspective taking, we ought to be focused on perspective getting, on asking people to describe their inner lives, their values and beliefs and feelings, the things they care about most. Epley sensed there was something about asking questions—the right questions—that contained the seeds of real understanding.

  But which questions were the right ones?

  THE RIGHT QUESTIONS

  In 1995, Elaine and Arthur Aron, a husband-and-wife team of research psychologists at the State University of New York–Stony Brook, placed two chairs atop a bright orange rug in a windowless room and invited strangers, in pairs, to come in, sit down, and take turns asking each other a list of questions. None of the participants—who eventually numbered more than three hundred—knew one another prior to entering the room, and each session lasted only sixty minutes. The questions had been selected ahead of time by the researchers, and they ranged from the frivolous (“When did you last sing to yourself?”) to the profound (“If you were to die this evening with no opportunity to communicate with anyone, what would you most regret not having told someone?”).

  Afterward, each pair of participants went their separate ways; no one was instructed to stay in touch. However, when researchers followed up seven weeks later, they found that 57 percent of them had sought out their conversational partner in the days and weeks after the experiment. Thirty-five percent of participants had gotten together to socialize. One pair went to dinner, and then saw a few movies, and then started meeting over weekends, and then on holidays. About a year later, when they got married, they invited everyone in the psych lab to the ceremony. “The impact exceeded everyone’s expectations,” Arthur Aron told me. “Even now I’m surprised by it. We had no idea what this would become.”

  The Arons had launched their study to see if there was “a practical methodology for creating closeness,” a technique that might generate connection. In particular, they wanted to see if it was possible to make strangers into friends. Other experiments had revealed a long list of factors that had no impact whatsoever. Researchers had learned that simply because two people had experiences or beliefs in common—they both went to the same church and both smoked, or were both atheists who hated tobacco—these similarities, on their own, were not enough to foster camaraderie. Studies had shown that instructing people to make small talk, or do puzzles together, or tell each other jokes, were all ineffectual at creating a sense of closeness. Merely informing study participants “We have taken great care in matching partners” and “we expect that you and your partner will like one another” did not necessarily mean people would like each other at all.

  In fact, there was only one method the Arons tested that could reliably help strangers form a connection: A series of thirty-six questions that, as Elaine and Arthur Aron later wrote, elicited “sustained, escalating, reciprocal, personalistic self-disclosure.” These questions eventually became known as the Fast Friends Procedure, and grew famous among sociologists, psychologists, and readers of articles with headlines such as “The 36 Questions That Lead to Love.”[*1]

  What’s particularly interesting about the Fast Friends Procedure is that the thirty-six questions were chosen somewhat haphazardly, at least at first. Some of them came from a game named “The Ungame” that was popular among stoners and university students (a demographic that included more than a few of the Arons’ research assistants). Other questions were dreamed up during coffee breaks, or by whoever happened to be within earshot when everyone went to a bar. “There wasn’t a lot of what you would call ‘strict science’ in how we initially found the questions,” Ed Melinat, one of the Arons’ grad students, told me. “We must have come up with, I don’t know, two hundred of them, and then we would run tests until we figured out which ones worked best.”

  The researchers assumed the best approach was to start with shallow, safe questions (“Whom would you want as a dinner guest?”) and then slowly get to the deeper stuff. “It felt weird to ask people to start baring their souls right away,” Melinat said. “So we decided to start simple.”

  By question seven (“Do you have a secret hunch about how you will die?”) people were being asked to reveal their deepest anxieties. By question twenty-four (“How do you feel about your relationship with your mother?”) and twenty-nine (“Share with your partner an embarrassing moment”) participants were being asked to describe their closest relationships and most painful memories. Question thirty-five (“Of all the people in your family, whose death would you find most disturbing?”) felt so intimate that participants often asked, and answered it, in a near whisper. The last question was open-ended (“Share a personal problem and ask your partner’s advice”), and by then, one or both participants were frequently crying.

  THE IMPORTANCE OF VULNERABILITY

  As the Stony Brook team tried to figure out the best questions to use, they found themselves stumped by a seemingly simple problem: How do you distinguish emotional questions from unemotional ones?

  There were some questions—such as “Would you like to be famous?”—that could go either way. For some people, the answer was a simple yes or no. For others, it opened the floodgates to soul-baring confessions of bygone dreams and failed ambitions. So is that question a reliable invitation to emotional self-disclosure, or an example of small talk?

  Eventually, the researchers figured out how to gauge if a question was likely to spark an emotional reply: Questions that asked about everyday experiences or uncontroversial opinions—such as “How did you celebrate last Halloween?” or “What is the best gift you ever received?”—tended to yield answers that were reliably unemotional.

  In contrast, questions that pushed people to describe their beliefs, values, or meaningful experiences tended to result in emotional replies, even if the questions themselves didn’t seem all that emotional. These kinds of questions were powerful because they often prompted people to reveal vulnerabilities. When someone asks “What do you value most in a friendship?” (question sixteen), it might not seem particularly emotionally probing, but it frequently draws unexpectedly revealing replies about past incidents of hurt or betrayal, or expressions of love for friends, or other anxieties or pleasures. Such questions make ever-deepening follow-ups (“What did you say after he broke up with you?”) easy to ask.

  Put another way, the difference between a shallow question and one that sparks an opportunity for emotional connection is vulnerability. And vulnerability is what makes How Do We Feel? so powerful.

  EMOTIONAL CONTAGION

  To the Arons, the idea that vulnerability was important made perfect sense, in part because it lined up with a well-documented psychological phenomenon known as “emotional contagion.” In the early 1990s, a series of experiments had shown that humans typically “synchronize their own emotions with the emotions expressed by those around them.” This synchronization is sometimes deliberate, like when we choose to empathize with another person; more often it is automatic, happening outside of our consciousness, causing us to tear up or get angry or proud on someone else’s behalf, whether we want to or not.

  This contagion is at the root of the How Do We Feel? conversation, and it explains why emotions influence our dialogues, even when we don’t recognize them. “Emotional contagion is a fairly primitive process,” a study published in 2010 observed. “Men and women tend to ‘catch’ expressions of joy, love, anger, fear, and sadness.” Emotional contagion, scholars believe, evolved because it helped humans form bonds with other people. It begins almost at birth: One study found that “10-week-old infants could and would imitate their mothers’ facial expressions of happiness, sadness, and anger.” This instinct has evolved within our brains so that we’ll feel good when we connect with other people—and thus become more likely to build alliances and friendships, families and societies.

  However, emotional contagion must be triggered by something, and one of the most reliable triggers is vulnerability. We become more prone to emotional contagion when we hear someone else express—or when we reveal our own—deeply held beliefs and values, or when we describe past experiences that were meaningful to us, or when we expose something else that opens us to others’ judgments. These are the same factors the Arons used to distinguish deep questions from shallow ones.

  In other words, we become more susceptible to emotional contagion, and more emotionally contagious ourselves, when we share something that feels raw, something that another person might judge. We might not care about their judgment, we might forget it as soon as we hear it, but the act of exposing ourselves to someone’s scrutiny engenders a sense of intimacy. To get deep, we have to make an offering of our vulnerability. “The louder the emotion, the more likely that contagion will occur,” Amit Goldenberg, a Harvard psychology researcher, told me. “And vulnerability is one of our loudest emotions. We’re hardwired to notice it.”

  This explains why the Fast Friends Procedure is so effective, and it illuminates which kinds of questions are most likely to help people emotionally align. There is a cycle: Asking deep questions about feelings, values, beliefs, and experiences creates vulnerability. That vulnerability triggers emotional contagion. And that, in turn, helps us connect.

  As the Arons continued exploring these kinds of phenomena, they discovered another interesting detail: The Fast Friends Procedure worked only if participants took turns asking each other questions. In a separate experiment, each participant was instructed to answer all thirty-six questions in a row while their partner listened, and then trade places. Volunteers said the experience was awkward and boring. No one felt close afterward. But when the Arons, in their experiment, told people to go back and forth and “share your answer with your partner, then let him or her share their answer to the same question with you,” people started to bond. “Reciprocity is critical,” Arthur Aron told me. “It’s one of the most powerful forces in the world. If you don’t have reciprocity, then people aren’t matching each other’s emotional ups and downs.”

  Once again, this is the matching principle—which says that communication requires recognizing what kind of conversation is occurring, and then matching it—at work. These thirty-six questions are effective because they help people match each other emotionally, and going back and forth encourages everyone to offer, and then reciprocate, vulnerability. This also demonstrates why mimicry isn’t enough. “Reciprocity is nuanced,” said Margaret Clark, a psychology professor at Yale. If someone reveals something devastating, like a scary diagnosis or the death of a parent, it doesn’t bring us together if we use that as an excuse to talk about our own health, or a family member who died long ago. “You don’t want to grab the spotlight,” Clark told me. Rather, reciprocity means thinking about how to show empathy. Sometimes it requires simply acknowledging someone’s emotions and showing them you care. “It’s being responsive to others’ needs,” Clark said.

  What’s more, vulnerability can mean different things in different settings. For instance, scientists have found a troubling double standard within some workplaces: When men express emotions like anger or impatience, it is commonly viewed as a sign of self-confidence, even good leadership. When a man cries at work, it is evidence of how much he cares. But when women express emotions such as anger or sadness, “they are more likely to suffer negative social and professional consequences,” found one study from 2016. “Women incur social and economic penalties for expressing masculine-typed emotions…. At the same time, when women express female-typed emotions, they are judged as overly emotional and lacking emotional control, which ultimately undermines women’s competence and professional legitimacy.” These kinds of unequal standards can make some displays of vulnerability feel unsafe.[*2]

  However, despite these complexities, the insights of the Fast Friends Procedure, paired with Epley’s research, are useful because they provide us with a framework for emotional connection: If you want to connect with someone, ask them what they are feeling, and then reveal your own emotions. If others describe a painful memory or a moment of joy, and we reveal our own disappointments or what makes us proud, it provides a chance to harness the neurochemicals that have evolved to help us feel closer. It creates an opportunity for emotional contagion.

  The How Do We Feel? conversation is a tool that functions by inviting others to reveal their vulnerabilities, and then being vulnerable in return.

  These are useful insights, but that doesn’t make them practical advice. It’s easy, after all, to ask deep questions when you are inside a laboratory and a scientist has handed you a list to work from. But how do you get deep in the real world?

  GETTING DEEP FAST

  Imagine that you’ve just met someone. A friend of a friend, or a recently hired coworker, or maybe you’re on a blind date. You both introduce yourselves, say a bit about your backgrounds. You get What’s This Really About? out of the way. Then there’s a pause, an expectant silence.

  What should you say next?

  The Fast Friends Procedure suggests asking a question. But you can’t progress through all thirty-six of them, not here. So maybe you jump to question three: “Before making a telephone call, do you ever rehearse what you are going to say?” Or, given that time is short, go deeper with question eighteen: “What is your most terrible memory?”

  You don’t need a PhD in psychology to realize that’s not a great plan. Asking a stranger these kinds of questions, outside of a psych lab, will ensure you’ll spend the evening alone. In the real world, the thirty-six questions are of little real help.

  What kinds of questions, then, should we ask?

  In 2016, a group of scientists from Harvard began wondering the same thing. They scrutinized hundreds of conversations that had been recorded during events such as speed-dating meetups, and gauged which conversations were successful (as measured by people saying they wanted to go on a real date), and which weren’t (people indicated they didn’t want to follow up). They found that during successful conversations, people tended to ask each other the kinds of questions that drew out replies where people expressed their “needs, goals, beliefs [and] emotions,” as the researchers later wrote. In unsuccessful conversations, people talked mostly about themselves, or they asked shallow questions, the kinds of inquiries that didn’t reveal anything about how their partners felt.