The reason people’s use of pronouns and other function words provides such a window into the mind may stem from their connection to social behavior at a neural level, Pennebaker suggests. Studies of brain-damaged patients have shown that the same brain area responsible for processing such words — a region in the left frontal lobe known as Broca’s area — is also involved in social tasks, such as recognizing emotional expressions.
Some research indicates this area also contains mirror neurons, specialized cells that may be involved in imitation and empathy. People with severe damage to Broca’s area lose both their social skills and the ability to call up prepositions, pronouns and other function words. A person with such damage might say, for example, “Want … see … movie … week.”
The apparent connection between function words and social behavior spurred Pennebaker to dig deeper. He suspected the psychological importance of these little-noticed words extends beyond the individual mind; perhaps, he thought, they also play a role in personal relationships. He assumed that connection could only work in one way: that the more two people like each other, the more closely their language styles would match. But when he began analyzing actual conversations, he found no correlation. “It just bugged the hell out of me,” Pennebaker says. If two people are speaking “as though their heads are connected,” how could that not be a reflection of how well they like each other?
The answer came to him as he was listening to the car radio one day. He heard a snippet of dialogue from a play about a couple in the midst of breaking up and noticed they were using language almost identically. Musing as he drove, Pennebaker thought of the way angry drivers hurl matching F-bombs.
He realized he’d been thinking of language style matching in “completely the wrong way. It’s not about liking — it’s about engagement.” In other words, such synchrony reflects not how much the speakers like each other, but how much they’re paying attention to each other’s mental and emotional states. In one test of this idea, he and a colleague analyzed conversations on the Nixon-era Watergate tapes. On the tapes, they found, language style matching occurred about equally in intensely positive and intensely antagonistic conversations.
Ireland, working in Pennebaker’s lab, was fascinated by the notion of such “antagonistic style matching,” which she regarded as a form of behavioral mimicry. “You see it everywhere — human boxers or cats circling each other before a fight, people shouting similar accusations at the same volume and pitch,” she says. Yet in part because people rarely pick fights during brief interactions in laboratory settings, making antagonistic exchanges tricky to study, antagonistic style matching “is almost completely ignored in psychology.”
Ireland’s chance to examine how language styles play out in antagonistic interactions came in 2008, when psychologist Marlone Henderson joined the UT psychology department. Henderson was studying negotiation, with a twist: Participants negotiated through instant messaging rather than face-to-face. Transcripts from such interactions, Ireland realized, were ideally suited to research on the dynamics of language styles because there is no body language or tone of voice to muddy the waters.
Ireland and Henderson decided to collaborate, mining the trove of data Henderson had already collected. In one experiment, Henderson had 128 college students take on the role of co-workers, giving them 20 minutes to negotiate travel plans for an upcoming business trip. To amp up the conflict, Henderson rigged the situation so participants’ negotiation goals were opposite those of their faux co-workers: If one wanted to fly, the other preferred to drive. If one sought a posh hotel, the other wanted cheaper digs.
When Ireland got her hands on Henderson’s data, she found just what she predicted: The more negotiators used similar styles of language, matching each other pronoun for pronoun, article for article, the more likely they were to reach an impasse. It wasn’t the number of function words people used that mattered; it was how closely their particular styles of using them matched.