One final observation. It’s interesting that Bush’s I-words never changed substantially once the war started. They temporarily went up a bit coinciding with his “Mission Accomplished” speech in May 2003, when he felt that the war was over. Subsequent drops in I-word use in the summers of 2003 and 2004 reflect his increasing focus on the war and, in 2004, his defensiveness with the press for its questioning his decisions about it.
THE POWER OF WORDS TO RETHINK HISTORY
I’ve tried to give a taste of the exciting possibilities that word analysis can bring to the study of politics and history. Wherever there is a word trail—no matter what language—computer text analysis methods can help interpret the psychology of the authors. Some of the studies in this area have practical applications. Other work is simply fun to do.
For example, my students, colleagues, and I have worked with several federal agencies to better understand the psychology of leaders of groups such as al-Qaeda to try to understand the relationships between their public messages and their sometimes-brutal actions. Can we understand their appeal by studying their language or the language of their followers? We have used these same methods to better understand extremist groups in the United States, ranging from far-right-wing neo-Nazis to far-left-wing Weathermen. In general, we are finding the violent groups have a very different linguistic fingerprint than nonviolent ones. Further, as they evolve from being nonviolent to increasingly violent, their language shifts accordingly.
We have also turned our language tools to the study of powerful, often despotic leaders of the past and present. For example, how have leaders such as Mao, Hitler, Castro, and others changed in their language use as they themselves changed from revolutionaries, to dictators, to sometimes-respected leaders in their countries? How has their language predicted changes in their countries and how have events changed their language?
Most important, language analyses can shed light on historical events in new ways. We’ve seen some possibilities with the Federalist Papers, relationships among poets and scholars, and even the Beatles. The historical questions that can be answered are limited only by the availability of language samples and the researchers’ imaginations. For example, did the Australian explorer Henry Hellyer really commit suicide or might he have been murdered? (Probably suicide based on his language in his diaries.) Did St. Paul really write all the letters attributed to him in the Bible? (Nope—not a chance.) Has Lady Gaga had an affair with Tom Cruise? (No idea. Hope not.)
Now it’s your turn.
USING PEOPLE’S WORDS TO PREDICT THEIR BEHAVIORS IN THE FUTURE
Can word analyses tell us if someone will eventually be a good president, a good spouse, a good employee or student? In fact, we do these calculations in our heads all the time. If a student writes to me and wants to work in my lab, I read her e-mail, her résumé, and her plans. Her words—which reflect her accomplishments, hopes for the future, and personality—will be the basis of my decision. People who rely on online dating sites may ultimately decide who their spouse will be based on his or her word use. And yes, we weigh political candidates’ faces and body language, but we also evaluate their platforms and plans, which are expressed in language.
Although we listen to and think about people’s words before voting for, hiring, or marrying them, it is impressive how frequently we err in our judgments. Would a language analysis program do a better job? Or would it help us in making better decisions? The jury is still out, but some interesting examples can be found among students planning to go to college and prisoners planning to find a normal life.
USING WORDS IN COLLEGE ADMISSIONS ESSAYS TO PREDICT COLLEGE GRADES
My linguist friend David Beaver and I were sitting in a bar talking about pronouns. (How many bar stories have you heard start off that way?) Wouldn’t it be great if there was some simple relationship between word use and how people later behave in life? We started to challenge each other about possible language samples we could get that could be linked to important real-world behaviors. And then I remembered Gary Lavergne.
Several years earlier, I had met Gary, the chief researcher in the University of Texas at Austin’s admissions office. Gary was not your usual statistician. He has published a series of nonfiction thrillers on mass murderers and, most recently, a book on the history of school desegregation in Texas. He was also interested in what factors predicted who would succeed in college. The University of Texas always has one of the largest student bodies of any campus in the United States. Although the school enrolls over seven thousand new first-year college students each year, the admission standards are surprisingly competitive. Part of the application process involves students writing two general essays.
Could the function words students use in their admissions essays predict their college grades? This was an appealing question for both David and me and, as it turned out, for Gary as well. To be clear, this was not a strategy to invent a new way to evaluate college essays to determine who should be admitted. Rather, we first wanted to learn if word use was related to academic performance and, if it was, whether we could influence the students to become better writers and thinkers in college.
We eventually analyzed over fifty thousand essays from twenty-five thousand students who had enrolled over a four-year period. The results were straightforward. Word use was indeed related to students’ grades over all four years of college. The word categories most strongly related to making good grades were:
• High rates of articles and concrete nouns
• High rates of big words
• Low rates of auxiliary and other verbs (especially present tense)
• Low rates of personal and impersonal pronouns
This constellation of words should look familiar to you. You might recall from earlier chapters that people differ in the degree to which they are categorical versus dynamic thinkers. A categorical thinker is someone who tends to focus on objects, things, and categories. The opposite end of this dimension are people who are more dynamic in their thinking. When thinking dynamically, people are describing action and changes. Often, dynamic thinkers devote much of their thinking to other people (which explains their high use of pronouns).
Does this mean that categorical thinkers are simply smarter than dynamic thinkers? Not at all. However, the American educational system is designed to test people concerning the ways they categorize objects and events.
Look at these two examples of college admissions essays that display categorical versus dynamic thinking. (The actual content of these essays has been changed considerably while keeping the rate of articles, nouns, large words, verbs, and pronouns intact.)
The categorical thinker
The concept of choice has played a prominent role in Western philosophy. One’s personality is polished to a more defined state by both conscious and unconscious considerations. The ultimate aim of liberty cannot be reached without a thorough control over the choices one makes. The divorce of my parents made me lead a double life. My partial withdrawal from reality had severe negative effects, including the inability to understand other viewpoints …
Notice how the writer’s sentences methodically define and categorize thoughts and experiences. The writing is structured and largely impersonal but, at the same time, ponderous. Compare the categorical thinker with a more dynamic one.
The dynamic thinker
I looked over at my brother, who was much older and wiser, only to see him crying. Before I knew it, I was crying too. I didn’t really know why, but if my brother thought it was bad, it was bad. Everyone moves, but it was the magnitude of my journey, a seven-hundred-mile trip from a small farming village to one of the biggest cities in America. It was going to be challenging but also an opportunity to grow. It involved giving up everything that is important to young children; family, friends, school.
The dynamic writer is far more personal and works to tell a story. The language is more informal and simple, using shorter words. Every sentence has m
ultiple verbs, which has the effect of making the story more alive.
Although both of these students came to college with virtually identical high school records and received liberal arts degrees, the categorical thinker had a much higher grade-point average every year in college. It wasn’t because the categorical thinker was a better writer. Rather, a categorical thinking style is more congruent with what we reward in college. Most exams, for example, ask students to break down complex problems into their component parts. At the same time, very few courses ask students to discuss ongoing events or to tell their own stories.
Most universities will never use word counting programs to decide who to admit to college. Once students discovered such a system was operative, their admissions essays would be a jumble of big words and articles, and practically verb-free. Instead, findings such as these point to ways we might think about training our students in high school and earlier. To the degree that categorical thinking is encouraged and rewarded in our educational systems, students should be explicitly trained in doing it.
Another argument is that we should explore whether dynamic thinking should be encouraged at the college level. Telling stories and tracking changes in people’s lives are skills that can serve people well. It also raises the question about how successful people are in the years after college. Is it possible that dynamic thinkers are better adjusted or happier? And finally, how flexible is thinking style? It is entirely possible that all of us occasionally need to think categorically and, at other times, dynamically.
USING WORDS TO PREDICT A BETTER LIFE AFTER PRISON
Drug and alcohol abuse takes a massive toll on society. One way that many states have attempted to curb and treat abuse is by establishing therapeutic communities—which are essentially treatment prisons where people are given the opportunity to undergo intensive drug and alcohol rehabilitation over several months after they have been convicted of a drug-related crime. If the participants successfully complete the program and stay drug- or alcohol-free for a specified time after they are released, their records are usually expunged. Most therapeutic communities require intensive group therapy along with writing exercises.
One of my former graduate students, Anne Vano, conducted an ambitious project to learn if the ways women wrote within a treatment facility might predict their lives once they were released from prison. Working with a single therapeutic community, Anne collected and transcribed writing samples from about 120 women. The writing samples she focused on were essays that the women wrote within about a week of their being released. The essays were expected to be personal and heartfelt. In the months afterward, Anne worked with the warden’s office to collect follow-up information, such as the women’s abilities to maintain jobs and whether or not they violated their parole or were re-arrested.
The stories the women wrote were powerful by any measure. They often described instances of being the victims of physical and sexual abuse and, at the same time, detailed their own deplorable behaviors toward others, such as their children. They often expressed great anxiety about their leaving the prison to return to an uncertain home life.
After leaving the prison, 15 percent of the 120 women were arrested and another 10 percent jumped parole four months after the program was completed. About 65 percent were holding down a steady job.
Interestingly, the way the women wrote in their final essay modestly predicted whether they were functioning effectively four months later. The two language dimensions that were most closely associated with therapeutic success were:
• a high social-emotional style, which includes use of personal pronouns and emotion words
• a high rate of positive emotion words
The tasks for the women on leaving the therapeutic community were to integrate into new jobs and into a functioning social network. Categorical and dynamic thinking were simply irrelevant dimensions for these women. To survive in their worlds outside prison they needed to be aware of others and themselves. It appears that the social-emotional and optimistic styles they exhibited in their writings were skills that could serve them well on the outside.
IT’S HARD TO imagine two studies more different than the college admissions and therapeutic community projects. Categorical thinking predicts better college grades for one group; social-emotional language predicts lower re-arrest rates in another. Different aspects of language are linked to different parts of our lives.
What I love about these two studies—and, in fact, all of the projects in this book—is that stealth words rearrange themselves in different configurations to predict a broad array of behaviors. For example, using language associated with high social-emotional style can help keep you out of prison and contribute to your being elected president and maybe provide some of the skills needed to write successful top-selling love songs.
Depending on the context, using I-words at high rates may signal insecurity, honesty, and depression proneness but also that you aren’t planning on declaring war any time in the near future. Using I-words at low rates, on the other hand, may get you into college and boost your grade-point average but may hurt your chances of making close friends.
It’s important to return to a theme that has bubbled up several times. The words related to social and psychological states are reflections of those states—not causes. They are telling us what is going on inside people’s heads. The people who use high rates of personal pronouns and emotion words just prior to their release from prison are approaching their writing topic in a social-emotional way. It’s unknown if the treatment program they were immersed in actually pushed them to think social-emotionally. It is also impossible to know if the words in their writing samples directly affected their behaviors once they were released. And it’s even more unlikely that if they had forced themselves to use these words in their essays (thinking it might be good for them) it would have influenced their lives outside the prison gates at all.
We are standing on the threshold of a new world. Think of the many applications that the computer analysis of function words has opened up. By analyzing inaugural speeches or ancestral diaries, we are able to know the influential writers or speakers of our past. We can also start to answer some of the burning psychological questions we have in our everyday lives. We can gain insight into how our online dating prospects view us, distinguish which rap artists are honest about being true gangsters, diagnose if our therapists are just as depressed as we are, or expose which of our colleagues secretly think they are highest in status.
Function words can help us know our worlds just a little better. From author identification that can help in catching criminals or in identifying historical authors, to understanding the thinking of presidents or tyrants, to predicting how people might behave in the future, function words are clues about the human psyche. Most promising, however, is that by looking at our own function words, we can begin to understand ourselves better.
Notes
In addition to the extensive comments and feedback from Cindy Chung, Sam Gosling, and Ruth Pennebaker, a number of other people read all or parts of the manuscript, including David Beaver, Molly Ireland, Jeff Hancock, Maureen O’Sullivan, Rich Slatcher, and Yla Tausczik. Thank you all.
PREFACE
ix Estimates of the number of words in the average person’s vocabulary range from thirty thousand to a hundred thousand. Part of the problem is in the definition of a word. For example, do you count singular and plural? Different spellings of the same word? How about a word that you might use that isn’t in the dictionary, such as “fancify” (as in, to make something fancy)? For a discussion of the number of function words, see Chung and Pennebaker (2007).
x One of the most exciting breakthroughs in language analysis is the recent release of the Google Books corpus. In a stunning article by Jean-Baptiste Michel and his colleagues in the journal Science, the authors analyzed the words from over five million books, or 4 percent of all the books that have ever been printed. Focusing on language use o
ver the last two hundred years, the authors were able to examine a wide range of historical trends. For example, how frequently has Freud or Darwin been mentioned? How long does fame last? How has language been evolving?
I urge you to try out Google Labs’ program Books Ngram Viewer at ngrams.googlelabs.com/. The study of history will never be the same.
xi Some of the best popular books on language that combine a rich knowledge of language with basic social and psychological questions include Goffman’s Forms of Talk, Lakoff and Johnson’s Metaphors We Live By, George Miller’s The Science of Words, Pinker’s The Language Instinct, Tannen’s You Just Don’t Understand, and Wierzbicka’s Understanding Cultures Through Their Key Words.
CHAPTER 1: DISCOVERING THE SECRET LIFE OF THE MOST FORGETTABLE WORDS
The Secret Life of Pronouns: What Our Words Say About Us Page 29