The Psychological Functions of Function Words
CINDY CHUNG and JAMES PENNEBAKER THE PSYCHOLOGICAL FUNCTIONS OF FUNCTION WORDS
anguage is the currency of most human social processes. We use words to convey our emotions and thoughts, to tell stories, and to understand the world. It issomewhat odd, then, that so few investigations in the social sciences actually focus on natural language use among people in the real world. There are many legitimate reasons for not studying what people say or write. Historically, the analysis of text was slow, complex, and costly. The purpose of this chapter is to suggest that social scientists in general and social psychologists in particular shouldreconsider the value of language studies. With recent advances in computer text analysis methods, we are now able to explore basic social processes in new and rich ways that could not have been done even a decade ago. When language has been studied at all within social psychology, it has usually relied on fairly rigorous experimental methods using an assortment of standardized human codingprocedures. These works are helping researchers to understand social attribution (Fiedler & Semin, 1992), intercultural communication (Hajek & Giles, 2003), and even how different cultures think about time (Boroditsky, 2001). When verbal samples have been collected, it has often been assumed that the best strategy is to not ask about one’s personal states directly. Instead, participants have been asked todescribe an ambiguous picture or tell a story, and the deep underlying meaning in the elicited statements has been interpreted (e.g. Schultheiss & Brunstein, 2001; Winter & McClelland, 1978) Over the last decade, a small group of researchers have adopted a somewhat different strategy. Their goal has been to understand how the words people use in their daily interactions reﬂect who they are andwhat they are doing. As detailed below, this strategy has also been method-driven. With the development of increasingly versatile computer programs and the availability of natural language
text on the internet, we are now standing at the gates of a new age of understanding the links between language and personality. It should be emphasized that this method-drivenapproach has also forced us to begin investigations by looking at word usage rather than exploring the broader meaning of language within a phrase or sentence (e.g. Semin, Rubini, & Fiedler, 1995), conversational turn (Tannen, 1993), or an entire narrative (McAdams, 2001). This chapter summarizes much of our own research that attempts to map and understand how word use can reﬂect basic social,personality, cognitive, and biological processes. Relying on computerized text analysis procedures, we are ﬁnding that the examination of often-overlooked “junk words” – more formally known as function words or particles – can provide powerful insight into the human psyche.
RECENT DEVELOPMENTS IN MEASUREMENT
It is beyond the scope of this paper to summarize the many computerized strategiesavailable to researchers (for a more comprehensive review see Pennebaker, Mehl, & Niederhoffer, 2003). Some methods, for example, simply count words related to particular themes (e.g., the DICTION program: Hart, Jarvis, Jennings, & Smith-Howell, 2005), whereas others look for words or phrases that reveal psychoanalytic concerns (Gottschalk, 1997) or themes related to drives or motives (e.g., the GeneralInquirer: Stone, Dunphy, & Smith, 1966). Various inductive methods have been evolving from the world of artiﬁcial intelligence. One such program, called Latent Semantic Analysis (LSA; Foltz, 1996), compares the similarity of any two texts in terms of their content. In our laboratory, we have been relying on a text analysis program that we developed called Linguistic Inquiry and Word Count, or...