Official newspaper of The University of Texas at Austin

The Daily Texan

Official newspaper of The University of Texas at Austin

The Daily Texan

Official newspaper of The University of Texas at Austin

The Daily Texan

Advertise in our classifieds section
Your classified listing could be here!
October 4, 2022

UT professor creates program to identify people by word use

UT psychology professor James Pennebaker, author of “The Secret Life of Pronouns: What Our Words Say About Us,” developed a program with his research team called Linguistic Inquiry and Word Count. The program scans text for its diction and produces a report that Pennebaker says uncovers the powerful implications of what he calls our “most forgettable words.”

Surprisingly, Pennebaker hadn’t run his own book through the program. When asked, he spun his chair around to his desk, entered “The Secret Life of Pronouns” into the LIWC, checked his email and within minutes, the program finished. Pennebaker interpreted the findings, scrolling down a table of percentages.

“We can see my article use is pretty high, which probably means I’m an older guy. I use a high rate of social words, so I’m interested in other human beings,” he said.

Pennebaker has spent the last 15 years dissecting language digitally. He has run diary entries, blog posts, eHarmony profiles and great works of literature through the LIWC. For him, the link between the program’s statistics and the human mind is self-evident. However, for most of us, even experts, pronoun use is an area where our intuition fails.

For example, after the 2008 election news analysts, critics, English scholars and certain presidential speech writers began commenting on what one media critic dubbed President Obama’s “inordinate fondness” for the pronoun “I.” His frequent use of the word showed his over-confidence and his distance from people, critics claimed. The only problem was that until Pennebaker, no one bothered to crunch the numbers. It turns out that Obama sets a record low “I” use for presidents going back through Truman. But according to Pennebaker’s research, low “I” use and Obama’s language in general do “suggest self-assurance and emotional distance.” Critics and analysts weren’t off in their assessment of his personality, but they completely misunderstood his language style.

When we talk with someone, we are sometimes not as interested in what they say as how they say it. Realizing this was a breakthrough for Pennebaker in developing LIWC, which distinguishes between content words such as “table,” “apple” and “jump” — and style words such as “I,” “the,” “up,” “have” and “don’t.” While content words tell what someone is reading or eating that day, style words often reveal fundamental aspects of our personalities.

“Style may not tell us much about where a person is walking, how hungry they are or their preference in fruit,” Pennebaker wrote. “But it is a meaningful window into people’s lives, attitudes and social worlds.”

To figure out how to interpret style words in different contexts, Pennebaker compared word-use trends for people of different genders, ages and social classes.

In a section called “The Sound of Power,” Pennebaker explains how two general word groups repeatedly emerged. Men, older people and those in higher social classes used more articles, nouns, prepositions and “big words.” On the other hand, women, young people and people in lower social classes used more pronouns and auxiliary and cognitive verbs. By applying these general trends, the LIWC was able to assess gender, truthfulness and social relationships with surprising accuracy. For example, the program correctly identified gender 72 percent of the time, proved a better lie detector than policemen, psychologists or federal interrogation officers and was a reliable judge for relationship compatibility.

Though Pennebaker’s program can outperform our intuitions in certain contexts, it has definite limitations. Though the LIWC has the advantage of being able to process massive amounts of data, it can miss context clues which would be obvious to a human.

“Don’t be misled. It’s a remarkably stupid program,” Pennebaker said.

The program can’t replace our intuition, but it is Pennebaker’s hope that if we pay attention to the cues he’s uncovered, the program might help us improve how we think.

“The effects aren’t going to be transformative, but it provides another way for us to think about relationships with others,” he said.

Printed on September 12, 2011 as: UT professor links word use to identifying personalities

More to Discover
Activate Search
UT professor creates program to identify people by word use