PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES Words of Wisdom: Language Use Over the Life Span James W. Pennebaker and Lori D. Stone University of Texas at Austin Two projects explored the links between language use and aging. In the first project, written or spoken text samples from disclosure studies from over 3,000 research participants from 45 different studies representing 21 laboratories in 3 countries were analyzed to determine how people change in their use of 14 text dimensions as a function of age. A separate project analyzed the collected works of 10 well-known novelists, playwrights, and poets who lived over the last 500 years. Both projects found that with increasing age, individuals use more positive and fewer negative affect words, use fewer self- references, use more future-tense and fewer past-tense verbs, and demonstrate a general pattern of increasing cognitive complexity. Implications for using language as a marker of personality among current and historical texts are discussed. Ask any two people to tell the same story, and they will tell it differently. The facts may be virtually identical, but the words and word orders are bound to be unique. These differences in word use can reflect subtle differences in perspective or interpretation of the story. They may also point to differences in sex, social class, life experience, or basic personality. Word use can tell us far more than just the story; it is a window into the narrator’s world. That language is potentially diagnostic about a person’s psycho- logical state is not a new idea. Freud (1915/1989) argued that parapraxes, or slips of the tongue, betrayed people’s deeper needs and motives. Other psychoanalytically oriented researchers, such as Gottschalk and Gleser (1969), posited that the phrasing of emotion-related thoughts provided valuable information about people’s anxieties and conflicts. Sociolinguists and those who study semantics in linguistics, anthropology, and sociology have long believed that word use, inflection, accent, and other features of natural language convey people’s social, economic, and psy- chological worlds (Eckert, 1999; Giles & Wiemann, 1993; Lakoff, 1987). The genesis of the linguistic approach discussed in this article began over 15 years ago when it was discovered that having people write or talk about emotional upheavals in the laboratory could ultimately affect their physical health. Across a large number of studies, participants who wrote about traumatic experiences for 3–5 days for 15–20 min per day evidenced better health than controls who were asked to write about superficial topics. These findings have been replicated across an impressive number of labs, cultures, and contexts (Lepore & Smyth, 2002; Pennebaker & Graybeal, 2001). A computerized text analysis program called Linguistic Inquiry and Word Count (LIWC; Pennebaker, Francis, & Booth, 2001) was developed to determine what features of writing might predict health improvements. Like earlier approaches to computerized content analysis (Rosenberg, Schnurr, & Oxman, 1990; P. J. Stone, Dunphy, Smith, & Ogilvy, 1966), LIWC uses a word-count strat- egy whereby it searches for over 2,000 words or word stems within any given text file. The search words have previously been cate- gorized by independent judges into over 70 linguistic dimensions, including standard language categories (e.g., articles, prepositions, pronouns), psychological processes (e.g., positive and negative emotion categories, cognitive processes such as use of causation words, self-discrepancies), relativity-related words (e.g., time, verb tense, motion, space), and traditional content dimensions (e.g., sex, death, occupation). The LIWC approach has fulfilled its original promise. Among people who write about emotional topics, certain word categories reliably predict subsequent improvements in physical health. For example, individuals who use an increasing number of words suggesting causal thinking and self-reflection have visited physi- cians for illness at lower rates after participating in the experiment than those who do not use these word categories (Klein & Boals, 2001; Pennebaker, Mayne, & Francis, 1997). More recently, a study by Pennebaker and King (1999) found that the ways students use words in class assignments predict physical health, health- James W. Pennebaker and Lori D. Stone, Department of Psychology, University of Texas at Austin. Lori D. Stone is now at Harris Interactive, Rochester, New York. James W. Pennebaker is codeveloper and owner of the Linguistic Inquiry and Word Count text analysis program. Preparation of this article was aided by National Institutes of Health Grant MH52391. Thanks are extended to Tom Lay, who organized the data for the Author Project. We are particularly grateful for the dozens of researchers who have provided text samples from their own research projects over the last several years. Thanks are also extended to Bill Swann, who provided comments on an earlier version of the article as a colleague and not as a consulting editor for the journal. Correspondence concerning this article should be addressed to James W. Pennebaker, Department of Psychology, University of Texas, Austin, Texas 78712. E-mail: pennebaker@psy.utexas.edu Journal of Personality and Social Psychology, 2003, Vol. 85, No. 2, 291–301 Copyright 2003 by the American Psychological Association, Inc. 0022-3514/03/$12.00 DOI: 10.1037/0022-3514.85.2.291 291