Researchers at the University of Vermont have uncovered a pattern in human language that upends a scientific theory accepted for over 70 years. The findings, published in Science Advances, suggest that word meaning is shaped by something more fundamental than emotion: our understanding of power, danger, and order. The study introduces "ousiometrics," a mathematical approach to measuring the essential meaning of language, and its most striking revelation is that people consistently prefer safety over danger in their words.
Since the 1950s, word meaning has largely been understood through the VAD model, developed by psychologist Charles Osgood and others. It describes meaning via three emotional dimensions: valence (positive vs. negative), arousal (excited vs. calm), and dominance (controlling vs. submissive). This framework became influential in psychology, linguistics, and artificial intelligence - particularly in systems designed to detect emotional tone. But the Vermont team found evidence that this long-established model overlooks crucial 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, they identified a different underlying structure. The three dimensions in the traditional VAD model are not independent of one another, as previously assumed. Instead, 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.
One of the study's most unexpected discoveries emerged when 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. This finding offers a new explanation for the Pollyanna principle - the tendency for human language to lean toward positive rather than negative words. For decades, researchers interpreted this pattern as evidence that people favor positive emotional expressions. The new analysis suggests something more fundamental may be responsible.
According to the researchers, the apparent preference for positivity largely reflects an underlying preference for safety. "The Pollyanna principle's positivity bias is, in fact, a one-dimensional projection of an underlying safety bias," the study concludes. "This is a big observation that comes out of this work," said Peter Dodds, director of the UVM's Complex Systems Institute and senior author of the study. "Expressions of safety are crucial to all language."
The discovery raises an intriguing possibility: the structure of human language may reflect evolutionary pressures associated with staying alive. From this perspective, words do more than communicate feelings - they help people recognize threats, evaluate uncertainty, establish trust, and coordinate actions. Being able to distinguish between safety and danger is essential to survival, and the researchers suggest this basic human priority may influence how meaning is organized in language. Throughout daily life, people constantly exchange information about whether individuals, places, situations, and behaviors are threatening or safe. The new findings suggest these distinctions may be more central to communication than previously recognized.
They also offer a different interpretation of why positive language is so common. Rather than simply expressing happiness or approval, many positive words may convey a sense of security, stability, or predictability. Julia Zimmerman, a postdoctoral researcher in UVM's Computational Story Lab and a co-author of the study, said the three newly identified dimensions reflect experiences fundamental to human life. "Power, danger, and structure are relevant to every person that's ever lived," she said. Researchers have long recognized that people tend to favor language associated with goodness and low aggression. The team argues that these familiar tendencies may arise from the same deeper pattern. "We now understand" that these are "shadows of an underlying linguistic safety bias," the team writes. Although the results are consistent with a possible connection between language and survival, they do not establish that evolutionary pressures directly caused the safety bias.
The findings have implications extending well beyond linguistics. Many artificial intelligence systems use sentiment analysis to determine whether language expresses positive, negative, or neutral feelings. Some approaches also draw on the traditional VAD framework to measure emotional meaning. If that framework does not fully capture the underlying structure of language, systems built around it could overlook important distinctions in what people communicate. The researchers suggest that incorporating power, danger, and structure could eventually help AI models interpret language more accurately and explain their assessments more clearly - particularly useful when analyzing trust, risk, and decision-making.
In linguistics, the findings offer a different way to understand the foundations of meaning. Instead of treating emotional tone as the primary organizing principle, the new framework emphasizes distinctions related to strength, safety, and predictability. Psychology could also be affected: decades of studies have used the VAD model to investigate emotions, perception, and behavior. A different set of fundamental dimensions could encourage researchers to reconsider how some of those findings are interpreted. The discovery may also interest neuroscientists. The human brain is highly sensitive to threats and signals of safety, and the newly identified linguistic bias suggests these biological priorities may be reflected in language. This potential connection could help researchers explore relationships between brain activity and symbolic communication, although further research would be needed to establish them.
To investigate these patterns, the researchers created analytical methods designed to measure meaning across enormous collections of written and spoken language. One central innovation is the "ousiometer," a tool that rapidly evaluates the essential meaning of large bodies of text and produces an average meaning score. (The word "ouisa" comes from Ancient Greek and is a root for the English word "essence.") The ousiometer builds on an earlier tool developed by the team called the "hedonometer" (a happiness meter). While that instrument measures the overall emotional tone of language, the new approach examines broader patterns of meaning.
The researchers demonstrated that the ousiometer can analyze material from remarkably different sources, including Jane Austen's novels, Arthur Conan Doyle's Sherlock Holmes stories, the New York Times, Wikipedia, Twitter, and transcripts of talk radio programs. The tool can also track how meaning shifts as a story unfolds. For example, they analyzed an English translation of Victor Hugo's Les Misérables and mapped its changing meaning throughout the novel. They called this changing pattern an "ousiometric trajectory." The visualization resembles the twisting shape of a multicolored protein, with the narrative moving across a grid based on four contrasting pairs: dangerous/safe, weak/powerful, gentle/aggressive, and bad/good. By following this path, researchers can observe how the underlying meaning of the text changes from one section of the story to another.
Another important innovation involved distinguishing between words themselves and how often people actually use them. In linguistics, individual words are categories called "types," while each appearance of a word in text is known as a "token." (For example, as a category "apple" is a type, and every time the word "apple" is used in a sentence is a token.) This distinction matters because a word that appears millions of times contributes much more to everyday communication than one that is rarely used. Earlier research often gave each word equal importance, regardless of frequency. The Vermont team instead incorporated how often words appear in real communication, allowing patterns to emerge that might otherwise remain hidden. This approach was particularly important for identifying the widespread preference for safety-related language.
The ten-member research team was led by Peter Dodds and Chris Danforth, professors in UVM's College of Engineering and Mathematical Sciences. Their collaborators included researchers from the Santa Fe Institute, the Complexity Science Hub in Austria, Howard Hughes Medical Center, the University of California, Berkeley, the University of Adelaide, and MassMutual Data Science.
Understanding this hidden safety bias could have consequences for how researchers study communication across society. If words consistently reflect a preference for safety, that tendency may influence how stories are told, how information spreads, and how people interpret events and experiences. The findings could eventually inform research into political communication, mental health messaging, and AI systems designed to interact with people. More broadly, the study challenges the idea that human language can be understood primarily through emotions such as happiness, fear, and excitement. Instead, the researchers propose that meaning has a deeper structure connected to power, danger, and order. Their findings suggest that language may reflect the challenges humans face when navigating uncertain environments, maintaining relationships, and identifying threats. By revealing this previously overlooked structure, the team offers a new way to understand what words communicate and why certain kinds of meaning dominate everyday conversation. Human language, their research suggests, may be more closely connected to our fundamental need for safety than scientists have traditionally recognized.