Why do people suddenly agree on the same office habit, greeting style, slang word, or workplace routine? A new study suggests the answer may be less complicated than many theories have assumed. People may not simply copy the last thing they saw, and they may not run a careful mental calculation every time.
Instead, the research points to a more human pattern. People explore different options for a while, then lock in once a behavior feels consistent enough. In other words, a social norm may spread when the brain decides a pattern has crossed the line from “maybe” to “good enough.”
A hidden threshold
The new work was led by Douglas Guilbeault of Stanford University, Spencer Caplan of the CUNY Graduate Center, and Charles Yang of the University of Pennsylvania. Their study focuses on social conventions, which are shared habits or expectations that help people coordinate with one another.
A convention can be simple. Think of how coworkers greet each other, what people wear in a certain office, or how a group decides what counts as polite in a chat thread. Nobody has to write these things down for them to start feeling like rules.
The key idea in the study is called the Tolerance Principle. In plain English, it means people can accept a few exceptions and still treat a pattern as real. A child learns that “walked” and “talked” follow a common past-tense pattern, even while knowing that “go” becomes “went.”
Not just copying
For years, many models of social learning leaned on two big ideas. One said people mostly imitate others. The other said people try to optimize, meaning they pick the option that seems statistically most likely to work.
The new study found that neither view fully matched what people actually did in controlled experiments. Participants first behaved with some uncertainty, trying different choices, then became more stable once they had seen enough evidence.
“People often assume that social learning is about imitation or careful optimization,” Caplan said. “What we found is something more basic and more human.” He added that people explore options, then commit once a pattern crosses a “good enough” threshold.
The Name Game
To test this, the team looked at coordination experiments known as the Name Game. In these experiments, people were shown an unfamiliar face and had to settle on the same name while interacting through social networks.
Earlier Name Game research found that groups could move from confusion to agreement without a central leader. Players started with many different names, but in well-connected networks, a single shared choice could suddenly rise and spread.
That matters because real life often works this way. Nobody tells everyone in a school, workplace, or online group exactly which behavior will catch on. Still, patterns emerge. Sometimes quickly.
What the numbers showed
The researchers compared several computer models with human behavior. These models acted like different theories of the mind, including imitation, optimization, probability-based guessing, and the threshold approach.
After the early rounds of the Name Game, the threshold model predicted people’s next choices nearly 88 percent of the time. That was better than the imitation model, the optimizer model, and the probability-matching model described in the study.
The team also tested a newer experiment called the Mind-Reading Game. In it, 800 people were recruited, and 782 completed the task and passed the attention check. Participants had to infer a partner’s preferred color choice from noisy signals, much like trying to read a pattern in messy everyday behavior.
Why exceptions matter
Here is the interesting part. People did not need perfect consistency before they settled on a choice. They could tolerate some noise, the way you might accept that most people in a group use one greeting even if a few do something different.
That makes the Tolerance Principle useful because it explains rule learning without demanding perfection. Life is rarely clean. People miss messages, change their minds, misunderstand cues, or act differently when they are tired.
This is where the study’s angle feels practical. A norm may not need everyone on board at first. It may only need to feel regular enough for enough people to stop searching and start treating it like “the way we do things.”
Social change and tipping points
The findings also connect to the question of how old norms get replaced. A related University of Pennsylvania study published in 2018 found that when a committed minority reached about 25 percent of a group, it could trigger a shift in the wider social convention.
The new research adds a possible mental explanation for that kind of flip. If enough people repeatedly encounter a competing behavior, their old pattern may no longer feel solid. Then the brain starts sampling again.
That does not mean every campaign, workplace reform, or public health message can be reduced to a simple number. The authors note that real-world norms are also shaped by identity, status, power, and social pressure. The trouble is, real life is messier than a lab.
Why it matters
Still, the study gives researchers a sharper way to think about how shared behavior spreads. This could matter for public health campaigns, workplace culture, online communities, and any setting where people need to coordinate without being told exactly what to do.
Think about a workplace trying to change meeting habits. A new norm may not take hold just because one person models it once, or because it is logically better. People may need to see it often enough, from enough places, until it feels safe to commit.
At the end of the day, the study suggests that humans are not perfect calculators or blind copycats. We are pattern seekers. Once the pattern feels strong enough, we move.
The main study has been published in Proceedings of the National Academy of Sciences.
