James Pennebaker studies words. Originally interested in the beneficial effect of writing about personal trauma, he and his students developed software to analyze this writing. Their investigation soon expanded to include spoken conversations, emails, political speeches, and other language samples. They discovered that much can be learned from the short "stealth words" that we barely notice, but that make up more than half of our speech. "Pronouns (such as I, you, we, and they), articles (a, an, the), prepositions (e.g., to, for, over), and other stealth words broadcast the kind of people we are."
Pennebaker summarizes his trauma research, noting that "people who benefit from writing express more optimism, acknowledge negative events, are constructing a meaningful story of their experience, and have the ability to change perspective as they write." Searching for reliable linguistic indicators of these processes identified writing style rather than more substantive content words. The resulting LIWC software works well regardless of a text's content.
Using both research findings and representative everyday examples, Pennebaker reviews what he has learned. Topics addressed include gender, status and social class, personality, leadership style, deception, interpersonal attraction, and group solidarity. The author not only presents conclusions from his own research, but links to supporting findings using non-linguistic methods. Specific findings include:
- LIWC correctly identifies an author's gender 72% of the time using writing style. This increases to 76% when content words are included. (Human guesses range from 55 to 65%.)
- On detecting depression: "Sadness generally causes people to focus inwardly. Pronouns tend to track people's focus of attention, and when in great emotional or physical pain, they tend to use I-words at high rates. Sadness, unlike most other emotions, is associated with looking back into the past and into the future. In other words, people tend to use past- and future-tense verbs more when they are sad or depressed compared to other strong emotions."
- "No system has ever been shown to reliably catch liars at rates higher than 65 percent. And even those with hit rates in that neighborhood (including me) have done so in highly controlled and artificial circumstances."
- "Linguistic style matching" across nine categories of function words occurs within the first 15 to 30 seconds of an attentive conversation. It is generally beyond conscious awareness. LSM profiles can predict a number of things better than chance, including whether a "speed dating" couple will pursue a further relationship after their initial four-minute discussion.
Pennebaker clearly wants to share, not just his insights, but the methods used to achieve them. Much of his research was done collaboratively, not just with students and fellow researchers, but with public figures, professionals in other fields, and anyone else with interesting documents. Readers are pointed to web sites that let them experiment with Pennebaker's techniques and a version of his LIWC software is available for more in-depth investigations. An appendix includes "A Handy Guide for Spotting and Interpreting Function Words in the Wild."
This book is an accessible summary of James Pennebaker's work with helpful citations of similar research by others. It serves as a guide to more technical discussions of text analysis through an extensive Bibliography and References section--and pointers to downloadable research reports from the author's web site. Interested readers might also enjoy Roderick Hart's
Campaign Talk or one of the other related books the author mentions.
