Conversations are essential to our well-being. Psychologists are exploring the science of why they’re so powerful

Conversations hold immense power. They help us form new connections and deepen existing ones.

But conversations don’t always go well. Some lead to conflict, harm, or resentment—others never even start because we fear discomfort or rejection. And because of their inherent complexity, conversations have long remained a mystery to psychologists.

“Conversation is this ancient technology for aligning our brains so that we can be on the same page,” said Thalia Wheatley, PhD, a professor of psychological and brain sciences at Dartmouth College who studies neural, biological, and other markers of connection during conversations. “It’s the most ubiquitous social behavior that we do, yet we don’t really know much about it.”

In recent years, new theories and tools have started to unlock exciting insights about these critical social interactions, on everything from informal chats to political disagreements to negotiations between neighbors, businesses, and nations. Psychologists are using sophisticated approaches, such as natural language processing algorithms and a neuroimaging technique known as hyperscanning, where two or more brains are scanned simultaneously as they interact, to gain new perspectives. They’re also sharing those findings to help deepen everyday interactions between strangers, make disagreements more productive, and improve outcomes for mediators, suicide crisis negotiators, and more.

“People’s well-being is determined in large part by the quality of their social relationships, which rely heavily on conversations that they have with each other,” said Michael Kardas, PhD, an assistant professor of management at Oklahoma State University’s Spears School of Business who has studied why people avoid deep conversations. “Going a little deeper in conversation, as well as learning to navigate disagreements, can create the kinds of connections that leave people feeling happier.”

Why we avoid conversations

As our primary means of social interaction, conversations are essential to our well-being. But we tend to avoid some of the most fulfilling ones: deep talks and opportunities to build new connections with strangers.

One reason for this is that people incorrectly predict that conversations will become less enjoyable over time, said psychologist Juliana Schroeder, PhD, an associate professor at the University of California, Berkeley’s Haas School of Business. She, Kardas, and their colleagues have shown that people view conversations like other hedonic activities—just as the fifth chocolate bar isn’t as pleasurable as the first one, they assume conversation quality will decline over time. Even after having a good conversation, study participants would rather sit in silence for part of a 30-minute session than continue talking with a stranger the whole time, when given the choice. But those randomly assigned to talk for 30 minutes had a better experience overall than those who chose to end their conversations early (Journal of Personality and Social Psychology, Vol. 123, No. 4, 2022opens in new window; Trends in Cognitive Sciences, Vol. 26, No. 5, 2022opens in new window).

“People are ending conversations earlier than they might otherwise do because they’re expecting enjoyment will go down and that they’ll run out of things to say,” Schroeder said.

People also predict that deep conversations—especially with strangers—will be more awkward than they actually are. Kardas and his colleagues paired participants with strangers to discuss either shallow topics (for example: How often do you come here?) or deep ones (for example: Can you describe the last time you cried in front of another person?). Pre- and post-surveys indicated that deep conversations felt less awkward and led to more connectedness and happiness than participants expected (Journal of Personality and Social Psychology, Vol. 122, No. 3, 2022opens in new window).

Knowing that deep conversations carry less risk than people imagine could lower the barriers to starting one.

“Have a deep question in mind or take advantage of something shared with you during the conversation to ask something more personal of the other person,” Kardas said. “You don’t need to escalate the depth of the conversation as gradually as you might expect.”

Mismatched goals are another reason people may avoid conversations—or struggle to hold effective ones. People use conversations for a number of reasons—for example, to impress, flatter, help, hurt, persuade, understand, or deceive. Michael Yeomans, PhD, an assistant professor of strategy and organisational behaviour at Imperial College London, Alison Wood Brooks, PhD, of Harvard Business School, and their colleagues developed the “conversational circumplex” as a way of conceptualizing the varied and often competing goals that can motivate and shape conversations (Current Opinion in Psychology, Vol. 44, 2022opens in new window). It can help researchers understand how goals function in conversations and point to ways to make interactions more effective.

[Related: Conversations are powerful. Here are ways to embrace the awkward and deepen relationships]

Sometimes goals conflict within a person: For example, I may want to seek advice but avoid doing so for fear of looking incompetent. Research by Brooks has shown that on the contrary, seeking advice actually boosts perceptions of competence (Management Science, Vol. 61, No. 6, 2015opens in new window). Goals can also conflict interpersonally: I want to give honest feedback during a company meeting, but my bosses don’t want their ideas criticized in public.

“A lot of our conversations are about coordinating joint actions, which are negotiations of one kind or another,” Yeomans said.

Making conversations click

As psychologists dig deeper into the nuances of what makes social interaction fulfilling and effective, it’s clear that some of the conventional wisdom—for example, to ask questions—could be too vague and may even lower conversation quality.

In the case of “boomer asking,” where a speaker asks a question, then answers it themselves, Yeomans and his colleagues have shown that listeners find the question insincere (article under review). The speaker thinks asking a question first makes their conversation partner feel included, but those good intentions are executed in a way that actually alienates the listener.

Asking follow-up questions, on the other hand, does improve conversation quality. Speed daters who asked more follow-up questions were more likely to get a second date, according to research by Yeomans, Brooks, social psychologist Julia Minson, PhD, an associate professor of public policy at the Harvard Kennedy School, and their colleagues (Journal of Personality and Social Psychology, Vol. 113, No. 3, 2017opens in new window).

“Questions make you more likable, but the key is showing that you’re listening,” Minson said. “Asking questions is less effective if they have nothing to do with what your conversation partner just said.”

Effective listening tends to lead to short gaps during conversations, which is linked to higher satisfaction among participants, according to research by Wheatley, her postdoctoral fellow Emma Templeton, PhD, and their colleagues. They found that between both friends and strangers, short pauses between speakers are linked with more feelings of connectedness (Proceedings of the National Academy of Sciences, Vol. 119, No. 4, 2022opens in new window). Long pauses between strangers are awkward, but between friends, they can actually signal connectedness—for instance, if one friend shares something personal and the other takes a moment to reflect before answering (Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 378, No. 1875, 2023opens in new window).

“When someone responds too quickly to something very personal, or not quickly enough when you were expecting a fast back-and-forth, they weren’t ‘dancing’ with you in the right way,” Wheatley said.

Wheatley and others are now crossing a new frontier to delve deeper into how we converse. With hyperscanning, researchers can scan two brains at the same time while participants interact. At Princeton University, Diana Tamir, PhD, a professor of psychology, and her colleagues created a decoding model using years of neuroimaging data linking brain activity to various mental states that occur during conversations. They then collected functional neuroimaging data from participants during unstructured conversations, mapping out conversation patterns between strangers as well as patterns among friends.

Along with her postdoctoral researcher Sebastian Speer, PhD, and their colleagues, Tamir found that strangers’ neural patterns become more similar over the course of the conversation—they spend their time finding common ground, she said. But friends do something totally different. They start off with more similar mental states, then spend the conversation diverging, spending more and more time exploring new and different states. That pattern leads to better conversations, including when strangers have conversations where they diverge and their neural patterns look more like friends’ patterns (PsyArXiv Preprints, 2023opens in new window).

“Our main takeaway is that during a conversation, it’s important to know where the other person is, but it’s also important to take them into a new direction—to say something that’s going to surprise and interest them,” Tamir said.

[Related: Making new friends and keeping existing ones is hard. Here are some science-backed tips to help]

Navigating conflict

Effective conversations aren’t just ones that leave people feeling connected and happy. Sometimes, conflict occurs—but if navigated well, disagreements can also be effective.

“Disagreement is inevitable, and our success or failure in relationships is often based on whether we can navigate that disagreement successfully, and build trust in spite of it,” Yeomans said.

Conflict is tough to resolve if we use the same linguistic tools that work well for simple, factual corrections (“Taxes are due on April 15”) to discuss values and matters of opinion (“We should raise taxes”), said Yeomans. For that reason, he, Minson, and their colleagues are studying what they call conversational receptiveness, ways we can show our willingness to engage with opposing points of view during a discussion. The researchers used a natural language processing algorithm to analyze the words and phrases that predict receptiveness in a listener and have summarized their findings with the acronym HEAR—hedging, emphasizing agreement, acknowledging, and reframing (Organizational Behavior and Human Decision Processes, Vol. 160, 2020opens in new window).

Hedging statements involves pulling in enough uncertainty that your conversation partner doesn’t automatically react, Minson said. For instance, instead of “everyone has to get vaccinated for COVID,” try “most people should probably get vaccinated for COVID.” Emphasizing agreement involves finding even just a small bit of common ground—for example: “We both want to have a just world.”

“In any conversation, no matter how much we disagree, we agree on something,” Minson said.

Acknowledgment, which echoes standard family therapy advice, involves restating your counterpart’s position, with the intention of showing that you actually heard and understood them (“You just said…” or “I understand that…”). Finally, reframing statements using positive emotion words and avoiding negation words (such as “no,” “don’t,” and “won’t”) can help—instead of “I hate it when people interrupt me,” try “I really appreciate it when people let me finish my sentences.”

Small changes in wording can make a big difference, even in some of the highest-stakes situations. After analyzing 31 hours of suicide crisis negotiation talks, researchers found that certain words and phrases appeared to be more likely than others to keep the person at risk of suicide engaged in conversation. Saying “I’m just here to speak” got more traction than “I’m just here to talk,” while the word “sort” was more effective than “help” in statements like “Let’s sort this out together.” (Journal of Language and Social Psychology, online first publication, 2023opens in new window).

“This was fantastic because we were able to feed these insights directly into training programs,” said Elizabeth Stokoe, PhD, a professor of psychological and behavioural science at the London School of Economics who studies conversations in a variety of contexts and an author on the wording study. “Police later reported that they indeed got further faster if they used ‘sort’ and ‘speak’ rather than ‘help’ and ‘talk.’”

A focus on conversations may also prove useful in increasing the effectiveness of public health campaigns. Christin Scholz, PhD, a communication scientist and an assistant professor in persuasive communication at the University of Amsterdam, is studying how young adults experience and share anti-alcohol messaging, including during conversations with their peers, and how this may ultimately influence behavior. She calls this the “social life” of a public health message.

“Conversations hold a lot of power for influencing health behavior, but they can also backfire,” said Scholz.

For example, young adults tend to have positive conversations about alcohol, a behavior that’s linked with higher consumption (Health Psychology, Vol. 38, No. 7, 2019opens in new window). But some early experiments suggest that those conversations are subject to outside influence.

In one study currently under review, Scholz sent teens anti-alcohol messages (such as an image depicting a car crash paired with a warning about drunk driving), then instructed them to plan a conversation about what they saw. She found that conversations they later had with peers about alcohol were more negative. That result is encouraging, she said, but the intervention had no direct effect on alcohol consumption, indicating that more research is needed.

“So far, conversations aren’t really used to optimize the design of these campaigns, but social interaction is clearly a powerful part of what makes these messages effective,” she said.

New horizons for conversation science

Even when researchers use traditional methods, such as audio or video recordings of participants in the lab, studying conversations presents logistical challenges that many other experiments do not. For Minson, convincing two people to show up at the same time and discuss a topic they disagree on is the first hurdle. And once a conversation begins, there’s no telling how it will unfold.

“Conversation is essentially an infinitely complex space. No two conversations are the same,” said Yeomans. “There are an infinite number of words you can use, but also the order, timing, structure, and nonverbals can vary in infinite ways.”

For that reason, natural language processing is a major boon for conversation scientists because it can analyze lengthy transcripts to find patterns and associations. But guidance on using natural language processing in social science is almost entirely focused on single-author documents, such as news articles or product reviews, Yeomans said. He, Brooks, and their colleagues are building a guideopens in new window that shows why dialogue is qualitatively different—because things like word order, turn order, and timing suddenly matter. They also provide advice on how to record, transcribe, and analyze data, as well as how to share that data without violating people’s privacy.

Hyperscanning also promises to reveal new insights about how we communicate. Wheatley is collecting data on how brains align during an improvisational storytelling task, while Tamir plans to start exploring the neuroscience behind persuasion. For example: Does a skilled persuader bring another person’s neural activity toward their own? Do they find the perfect compromise in brain alignment? Or something else?

“With behavioral methods, you can look at people before and after they talk, but we can now get a continuous measure of where people are throughout an entire conversation,” Tamir said. “That just blows open our ability to look at the dynamics of things like conversation, persuasion, and compromise.”