When Ex-cons Change Their Vocabulary, They Stay Out of Jail

Photo: Lawrence Migdale/Getty Images/Science Source

In psychology, a postulate underlying approximately a zillion approaches to therapeutic growth is that if you can change the way you think, you’ll change the way you live. The stakes are especially high if the way you lived used to include illicit-drug abuse, violence, or other crimes that can land you in prison. That’s why a clever new study out from the College of Social Work at Ohio State University is so promising: Using quantitative analysis, researchers found that the more prisoners had changed their linguistic habits by the time they were released, the more likely they’d stay out of trouble.

Co-authors Nathan Doogan and Keith Warren used a database of exchanges from therapeutic communities (think Alcoholics Anonymous, but residential and without the appeals to the divine) in three minimum-security facilities for felony offenders. In these communities, residents check one another’s behaviors through “pushups,” congratulatory notes that encourage positive behavior like tidily completing a cleaning task, and “pullups,” corrective notes to discourage negative behavior, like even just joking about doing anything violent. The pushups and pullups would be vetted by senior group members and staff, and if approved, would be read aloud at community meetings and entered into an electronic database. They tended to be a sentence, or maybe two, long — taken together, 2,342 participants gave 89,628 pullups and 177,238 pushups, collected over two to seven years, depending on the group.

That corpora was plugged into a “semantic analysis,” where the researchers treated the words and phrases prisoners used like a social network. Just like you hang out with some friends more than others, people tend to use words in familiar bundles. These arrangements of words, the authors contend, are representative of “schemas,” or the assumptions, depending on people’s personal histories, that they walk around with about how the world works.

Criminologists have contended that people who do crime have schemas that get them into trouble. In a 2011 sample of 700 youths, these “deviant schemas” included a hostile view of people and relationships, as indicated by agreeing with a statement like “When people are friendly, they usually want something from you,” and a preference for immediate rewards, measured by agreeing with “You have to have everything right away” and the like. They also had a cynical view of conventional norms, measured by how much they approved of kids their age having sex, smoking weed, and cheating on tests. The study, which followed fifth-graders into teenhood, found that community crime, deviant peers, and experiences of discrimination increased the belief in schemas, while supportive parenting decreased the belief in “criminogenic knowledge structure.” In the same way that narcissists are taught by their childhoods that they have to rely on themselves and dominate others, would-be criminals learn that they need to assert authority, command respect, and flaunt norms in order to thrive.

Doogan and Warren’s study took schemas in a new direction, quantitatively assessing whether gaining or losing linguistic tics would signal a change in mental schemas in the same way that gaining or losing friends changes one’s social structure. “Learning isn’t just about learning new concepts or new words, but changing the connections between your concepts,” Warren tells Science of Us. “What we thought is if the connections change, that was an index of learning, and learning ought to predict how well people do after they graduate.”

Holding with that theory, their analysis found that both gains or losses in people’s semantic patterns signaled a greater likelihood of avoiding re-offense or relapse up to a year after being released. Interestingly, having both gains and losses wasn’t better than having one or the other, since, Warren infers, either pruning away ways of thinking or getting more sophisticated both constitute change. If the patterns of speech stayed the same, the rate of recidivism was higher, no matter how large or small the starting vocabulary. Similarly, a raw increase in giving pushups or pullups over time didn’t reduce rates of reincarceration, since, after all, speaking more doesn’t necessarily mean that you’re changing mental habits.

The changes in connections between words were “essentially a measure of internalizing” the rules of living that the therapeutic community was trying to teach, Warren says, noting that “losing old connections was actually a good predictor of reduced hazard of reincarceration. You lose a lot of your old ideas.” Importantly, the researchers didn’t analyze sentiment, or the feelings evoked by the words — just the change in linguistic patterns themselves.

While the therapeutic communities are facilitated by professionals, most of the responsibility falls on the members themselves. As an intervention, this makes them super compelling, since they’re so cheap to run. Also, Warren says, the participants who are leading the therapy may be part of the secret sauce: All the way back in 1965, psychiatrist Frank Reissman proposed the Helper Therapy Principle, stating that part of what makes therapeutic communities, AA, and their kin work so well is not just that junior members are learn from senior members, but that the veterans gain from teaching. “People change their own cognition by helping,” Warren says, and “that includes people who are incarcerated.” To him, this research highlights engagement: You can skate through a prison sentence or a therapy practice or a university degree without changing that much. But if you really attend to your experience, your cognition really can change. And your behavior, too.

When Ex-cons Change Their Vocabulary, They Stay Out of Jail