Why Our Words Lean Toward Safety, Not Just Positivity, New Study Finds

A study from the University of Vermont, published in Science Advances, argues that the meaning of words is organized less around emotion than around power, danger and order. Its most striking result, according to ScienceDaily’s summary of the work, is that everyday language shows a consistent preference for safety over danger.

The finding challenges a framework that has shaped psychology and linguistics for about 70 years. Here is what the researchers did and what they say it shows.

What model does the study challenge?

Since the 1950s, much of the scientific thinking about word meaning has leaned on work by psychologist Charles Osgood and others. The resulting approach, known as the VAD model, describes meaning through three emotional dimensions: valence (positive versus negative), arousal (excited versus calm) and dominance (controlling versus submissive).

The model became influential in psychology, linguistics and artificial intelligence, especially in systems built to detect the emotional tone of written text.

What did the researchers do differently?

The team examined billions of individual uses of more than 20,000 words across a wide range of real-world texts. Using computational techniques, they concluded that the three VAD dimensions are not independent of one another, as had been assumed.

They propose a new approach, which they call “ousiometrics,” that measures the essential meaning of language mathematically. It describes meaning with three different dimensions: power (weak versus powerful), danger (safe versus dangerous) and structure (ordered versus chaotic).

According to the report, the new framework accounted for more than 90% of the variation in measured word meaning, compared with about 72% for the traditional VAD approach. The research was supported by the U.S. National Science Foundation, Google, MassMutual and other funders.

What is the “safety bias”?

When the researchers looked at how words appear in books, news reporting, social media and spoken language, words tied to safety showed up much more often than words tied to danger.

That offers a fresh explanation for the Pollyanna principle, the long-observed tendency of language to favor positive words over negative ones. Researchers have generally read that pattern as a preference for positive emotional expression. The new analysis suggests the cause is deeper: the apparent bias toward positivity largely reflects a preference for safety, the authors say. They describe positivity as a one-dimensional projection of an underlying safety bias.

Peter Dodds, director of the university’s Complex Systems Institute and the study’s senior author, called it a big observation and said expressions of safety are crucial to all language.

Could survival be behind it?

The researchers raise the possibility that the structure of language reflects evolutionary pressures tied to staying alive. In that view, words do more than convey feelings: they help people recognize threats, weigh uncertainty, establish trust and coordinate with others.

Julia Zimmerman, a postdoctoral researcher in the university’s Computational Story Lab and a co-author, said the three new dimensions reflect experiences that are fundamental to human life and relevant to every person who has ever lived.

The team also argues that other familiar tendencies, such as a preference for language associated with goodness and low aggression, may come from the same pattern. Many positive words, they suggest, may convey security, stability or predictability rather than simply happiness or approval.

Why does it matter?

The authors say the safety bias could change how researchers study language and human behavior. It could also affect the development of AI systems that interpret what people say and write. Large language models, automated content moderation and other technologies increasingly depend on understanding human communication, and the researchers say a model that better explains word meaning could help those tools.

The evolutionary explanation is a possibility the researchers raise, not a settled conclusion, and other scientists have yet to weigh in on the new framework in the coverage available.

Around Prague, based on reporting by ScienceDaily