People who are prone to conspiratorial thinking don’t necessarily write fully formed conspiracy theories when asked to interpret ambiguous events, but their worldview still leaks into their writing through certain word choices. Recent research published in journals Pro Swan provides evidence that although these people may not spontaneously construct complex conspiracy plans, they consistently use a recognizable vocabulary of suspicion and power.
Conspiracy theories are alternative narratives that explain major events as the secret work of a powerful and malicious group. Believing in these stories tends to have negative social consequences, including decreased trust in science and decreased compliance with social norms. Belief in conspiracy theories goes beyond harmless speculation on the internet and is associated with less support for public health initiatives and less participation in environmental protection activities. Scientists want to understand the psychological roots of these beliefs in order to better address their impact on society.
Previous research has investigated why people believe in conspiracies, often pointing to a psychological need for certainty and control in an unpredictable world. People with a strong conspiratorial worldview often have a low tolerance for ambiguity and tend to see meaningful connections between completely random events. They tend to assign ulterior motives to the world around them and act as if nothing happens by chance.
However, relatively little attention has been paid to how these stories are actually produced by individuals. Scientists conducted this study to see if a general tendency toward conspiratorial thinking, often called conspiracism, causes people to spontaneously generate conspiracy theories when faced with unexplained situations.
Alessandro Miani, a researcher at the Department of Psychology at the University of Friborg in Switzerland, wanted to test this exact association. “After building LOCO, the largest corpus of conspiracy theories available, we became interested in testing whether conspiracy theories and conspiratorial language are actually related,” Miani said.
“While most previous studies have analyzed the language of conspiracy websites and social media at the textual level, no studies have tested whether individual psychological differences, such as conspiracy mentality (the tendency to interpret events as the product of a secret plan), are reflected in how people actually use language and write.”
To test this idea, the authors designed an experiment centered around the apocalyptic psychological thriller “Leave the World Behind.” The film was chosen because it features an ambiguous, open-ended plot with a disjointed timeline that leaves viewers guessing about what is happening behind the scenes. 285 Italian university students were asked to watch the film and write a short essay explaining their personal interpretation of its meaning.
The researchers measured each participant’s level of conspiratorial thinking using a standard psychological questionnaire called the Conspiracies Scale. To evaluate the essays, they used an extensive language model. This is a type of artificial intelligence that is programmed to understand and generate human language. Artificial intelligence evaluated each essay based on how strongly it characterized a conspiratorial narrative.
Contrary to the researchers’ expectations, scores representing conspiratorial thinking did not match the level of conspiratorial narrative in the essays. Those with a higher tendency to believe in conspiracies were less likely to write outright conspiracy theories than those with lower scores.
Suspecting that certain psychological questionnaires might not be capturing the right nuances for Italian participants, the authors conducted a second study. This time, we recruited 100 different university students. The procedure was exactly the same, but the researchers replaced the original questionnaire with a general conspiracist belief scale.
This alternative measure asks more direct questions about government cover-ups, secret organizations, and hidden knowledge. The results remained the same when using this more direct measurement tool. The researchers again found no meaningful connection between participants’ conspiratorial thinking and the overarching narrative of their written interpretations.
“We were surprised that conspiratorial narratives didn’t emerge as much as we expected,” Miani told Cypost. “We pre-registered the hypothesis that people with high levels of conspiracism would ‘fill in the gaps with conspiratorial interpretations’ of ambiguous movies and conducted two studies with two different conspiracy belief measures. In both cases, the expected association between conspiracy theories and conspiratorial story content was simply not present.”
To understand what was going on, the scientists combined data from both studies and sampled a total of 385 participants. They decided to take a closer look at the specific words used by the participants, rather than the overall narrative structure. They utilized a custom dictionary of conspiracy-related terms such as deception, government, and elite to see if participants were using conspiracy vocabulary even if they were not writing full-blown conspiracy theories.
This detailed analysis revealed that participants with higher levels of conspiratorial thinking actually used more conspiracy-related words. “When asked to interpret an ambiguous event (in our case, the apocalyptic thriller ‘Leave the World Behind’), people with a strong conspiracy theory mindset tend to use language that is consistent with language found on conspiracy websites (using words like ‘deception,’ ‘government,’ ‘world,’ etc.),” Miani said.
But Miani pointed out that the effect is vocabulary-specific. “There’s a caveat: They haven’t really constructed a conspiratorial narrative as a consistent story of a malign elite secretly orchestrating harm against the population,” he said. “Rather, conspiratorial thinking seems to be associated with word choice.”
The participants also exhibited a linguistic feature that researchers refer to as “megalaria.” Megallia is the disproportionate use of highly sophisticated words in simple sentences. The authors suggest that this may be an attempt to sound more authoritative or uniquely intelligent.
To measure how advanced the participants’ vocabulary was, the researchers used a new method to calculate vocabulary sophistication. They programmed artificial intelligence to estimate the age of acquisition for thousands of Italian words, based on the principle that words learned later in life are generally perceived as more complex. Applying this measure to essays, we found that participants high in conspiratorial tendencies actually had lower overall vocabulary sophistication, despite their occasional habit of unusually long words.
Furthermore, these participants tended to write sentences that were more complex in syntax, meaning that their sentence structures were more complex and hierarchical. This combination of complex grammar and suspiciously specialized vocabulary suggests that conspiratorial thinking shapes the stylistic components of one’s writing.
Miani emphasized how to view these results in context. “The main practical lesson is that while there are measurable individual differences in language use associated with conspiratorial thinking, they are small and operate at the level of lexical choices rather than the level of structured narratives,” Miani said. “The association we found was modest.”
The scientists compared these findings to those typically found in online spaces dedicated to conspiracy theories. In naturalistic online environments, conspiracy-related texts tend to be very long and feature a sparse structure that jumps between unrelated topics. This fragmentary style is often used to overwhelm the reader with supposed evidence to persuade them.
However, in the laboratory setting, participants did not write long essays and did not display the distributed structure typical of online forums. Researchers believe this difference highlights the role of the environment in shaping how people express their beliefs. Online, people actively try to recruit others to their way of thinking.
“When interpreting ambiguous events, such as the movies we showed, participants with high levels of conspiracism did not write conspiratorial stories as we expected. Instead, they used vocabulary that matched words found on conspiracy websites,” Miani said. “It’s also worth noting that the setting is important; this was a lab study for course credit, and unlike people posting on online conspiracy forums, the participants may have had little incentive to persuade anyone.”
Miani also cautioned against taking the lack of narrative findings out of context. “I would caution that ‘no connection to conspiratorial narratives’ should not be interpreted as ‘conspiracy theories have no connection to language,'” Miani said. “It is related to language; it is not a fully constructed narrative, but only appears in the choice of words. And this association must be interpreted in the light of our limitations.”
Another limitation is the reliance on artificial intelligence to grade conspiratorial narrative essays. Large-scale language models are likely to recognize conspiracy theories only if the text is highly structured and logically connected. If participants expressed their beliefs in a fragmented or disjointed manner, the artificial intelligence may have missed the intent of the conspiracy. A simple dictionary method that only counts specific words circumvents this problem and successfully identifies the underlying worldview.
Future research should investigate how conspiracy theories are constructed in group settings rather than isolated individuals. The authors suggest that complex conspiracy theories may not be fully formed in one person’s head. Instead, they can develop cumulatively, as many different people exchange ideas, share certain vocabularies, and slowly build shared narratives over time.
“When our results are replicated in different settings, it raises an interesting question: Who is actually creating the conspiracy theories that are circulating online?” Miani said. This series of studies represents the next step for the research team.
Scientists may investigate whether the most prominent online conspiracy theories are written by genuine believers or by actors who intentionally spread misinformation for political or economic gain. Understanding the difference between organic beliefs and strategic manipulation can help society better deal with the spread of misinformation.
The study, “Leave the world (perspective) behind, leave the words behind: The impact of conspiracism on writing”, was authored by Alessandro Miani, Ines Adornetti, Daniela Altavilla, Valentina Deriu, Alessandra Chiera, and Francesco Ferretti.
