Researchers at the University of Vermont have discovered a surprising pattern in human language that challenges a scientific theory widely accepted for more than 70 years. Their findings suggest that the meaning of words is shaped by something more fundamental than emotion: our understanding of power, danger, and order. The study, published in Science Advances, introduces a new approach called "ousiometrics," which uses mathematical methods to measure the essential meaning of language.
Its most striking finding is that the words people use show a consistent and widespread preference for safety over danger. This hidden safety bias could transform how researchers study language, understand human behavior, and develop artificial intelligence systems that interpret what people say and write. Scientists Discover a Hidden Safety Bias in Human Language Since the 1950s, much of the scientific understanding of word meaning has relied on a framework developed through pioneering research by psychologist Charles Osgood and others.
Known as the VAD model, it describes meaning through three emotional dimensions: valence (positive vs. negative), arousal (excited vs. calm), and dominance (controlling vs. submissive). This approach became influential in psychology, linguistics, and artificial intelligence, particularly in systems designed to identify the emotional tone of written language. However, the University of Vermont researchers found evidence that this long-established model overlooks important aspects of how meaning is organized.
Their analysis examined billions of individual uses of more than 20,000 words across a wide range of real-world texts. Using advanced computational techniques, the team identified a different underlying structure. They found that the three dimensions in the traditional VAD model are not independent of one another, as previously assumed.
Instead, their results indicate that word meaning can be more effectively described through three separate dimensions: power (weak vs. powerful), danger (safe vs. dangerous), and structure (ordered vs. chaotic). The research, supported by the US National Science Foundation, Google, MassMutual, and other funders, found that this new framework accounted for more than 90% of the variation in measured word meaning. The traditional VAD approach explained approximately 72%.
That improvement could be especially important as large language models, automated content moderation, and other AI technologies increasingly depend on understanding human communication. Why Human Language Favors Safety Over Danger One of the study's most unexpected discoveries emerged when the researchers examined how words actually appear in everyday communication. Across books, news reporting, social media, and spoken language, words associated with safety appeared much more frequently than those associated with danger.
Extract — continue reading at the source.
