Abstract editorial illustration of personal narratives and language analysis, with
How Psychopathy Traits Are Linked to Language in Personal Narratives
Distinct word-use patterns linked to three psychopathy traits
A narrative-psychology study from the University of Mississippi examined how scores on three psychopathy-related dimensions—boldness, meanness, and disinhibition—were associated with word use in participants’ written life stories. The researchers found different patterns across the dimensions, with some associations pointing in opposite directions.
These are group-level associations between questionnaire scores and word-category frequencies, the kind of pattern also examined in Computational Analysis of Patient Language Identifies Suicidal Risk. The study does not show that writing can diagnose psychopathy or reveal an individual’s personality.
How the study was conducted
First author Ky Bray and senior author Laura Drislane conducted the study at the University of Mississippi. The reported group counts are 307 university students and 144 formerly incarcerated adults. Those figures sum to 451, while the available study coverage also describes the sample as 450. Because the reported sources do not resolve that discrepancy, the combined sample size should be treated as uncertain rather than stated as a confirmed total.
Participants completed the Triarchic Psychopathy Measure (TriPM), a questionnaire measuring boldness, meanness, and disinhibition, and wrote narratives about a high point, a low point, and a turning point in their lives, following a life-story interview format developed by Dan McAdams.
The researchers analyzed the narratives with Linguistic Inquiry and Word Count (LIWC), a dictionary-based program that tallies words belonging to predefined categories, such as emotion, achievement, social connection, power, anxiety, and reward. They used linear mixed models to examine associations between word-category frequencies, trait scores, narrative context, and covariates including sample, gender, and age.
LIWC counts words in categories; it does not interpret a story as a person would. The traits were measured by questionnaire, not diagnosed from the text.
The three dimensions are:
- Boldness: social dominance, emotional resilience, and low fear or anxiety.
- Meanness: callousness, antagonism, and reduced concern for others.
- Disinhibition: impulsivity, poor behavioral control, and instability.
Context mattered: high points, low points, and turning points
Several overall associations depended on narrative context: a trait’s relationship with a word category sometimes appeared only in a particular kind of memory.
Higher disinhibition was associated with more negative-emotion words in low-point stories, and with fewer references to friends and achievement in high-point stories. Higher meanness was associated with more positive emotion, affiliation, power, achievement, and reward words especially in turning-point narratives, and with fewer negative-emotion words especially in high-point narratives. Boldness had no significant overall main effect on emotion language, but was associated with fewer anxiety and fear words in low-point stories.
These context-specific patterns do not describe a fixed writing style shared by everyone with a particular trait. They are associations observed across different kinds of personal narratives.
What each trait was associated with
Disinhibition was associated with more negative-emotion words and fewer mentions of friends and accomplishments in the specified narrative contexts. “In contrast, those higher in disinhibition seemed to tell more dysphoric stories about their lives than those lower in disinhibition,” Bray said.
Meanness was associated with more positive-emotion, achievement, status, and reward-related language, and less negative-emotion language. The study establishes associations between questionnaire scores and word-category frequencies; it does not establish why participants with higher scores used these words.
Boldness was associated with fewer fear- and distress-related words, particularly in accounts of a low point. “Bold individuals may present as composed, with reduced verbal expressions of anxiety, even when recounting their most difficult experiences,” Bray said.
Drislane summarized the contrast: “The most surprising finding is how differently, and sometimes in opposite directions, the three traits were related to language. Meanness, for instance, was associated with less negative emotion, more positive emotion, and more achievement-related language, while disinhibition showed almost the reverse pattern.”
Suggestions the researchers make, but the study did not test
Drislane suggested that the findings might eventually have implications for forensic settings, therapy, or risk assessment. That is a forward-looking suggestion, not a result tested by the study. The researchers did not test whether software could identify traits from a new essay, predict behavior, or improve real-world decisions. The reported word counts describe group-level patterns; they do not establish a writer’s motives or justify decisions about an individual.
The study’s contribution is a more specific account of how the three measured traits related to life-story language. Meanness and disinhibition had contrasting emotional-language associations, a finding that would be obscured by treating psychopathy as a single undifferentiated score. “What this tells us is the story we tell about our life has meaning,” Drislane said. “It helps people make sense of their experiences, develop a worldview and a coherent sense of who they are. What we're starting to see is that in psychopathy, there are traits that relate to differences in meaning-making.”
What this study can—and cannot—say about text analysis
The study used a dictionary-based word-counting method and compared its counts with participants’ self-reported trait scores. It did not test whether an automated system could infer traits from writing, nor whether such inferences would be accurate for individuals.
That distinction matters: an association between a questionnaire score and word-category frequency is not evidence that the words reveal a diagnosis. The findings should be read as descriptions of patterns across narratives, not as a method for assessing a particular writer.
The limits the authors flag
The researchers identify limitations including the sample, its mix of students and formerly incarcerated adults, and reliance on self-reported trait measures. The unresolved discrepancy between the reported group counts and combined sample figure is another reason not to overstate the sample total. The results do not support inferences about any single writer; using few fear words in a difficult story is not a psychological assessment.
The takeaway
The three measured traits showed different associations with word use, and those patterns varied with the kind of life event being described. The method was a word counter, not a mind reader: the findings describe group-level relationships between questionnaire scores and language, not a way to identify or diagnose individuals.
Sources: Ky Bray and Laura Drislane, “Investigating interpersonal and affective features of psychopathy through language use,” Personality and Individual Differences; University of Mississippi News; Neuroscience News.
psychopathy traits are linked to language in personal