This study explored whether cognitive dissonance reduction is accomplished primarily through self-deception. In a replication of Festinger and Carlsmith's (1959) study, participants in the experimental group (N=18) lied to a confederate while those in the control group (N=15) did not. Participants were then asked the four questions about their experience and their responses were recorded and subjected to a lexical analysis using LIWC (Pennebaker, Francis, & Booth, 2001). Consistent with self-deceptive speech act theory (Shapiro, 1989, 1996) and previous research on defensive speech (Barrett, Williams, & Fong, 2002; Glick & Schober, 2008), experimental participants were predicted to use more words in categories associated with self-deception. Although several predictions were not borne out, experimental participants reliably used a greater incidence of certainty, positive feeling, and optimism words in one of their responses. Discussion considered these findings as reflecting "conviction" of a fundamentally defensive nature (McGregor, Zanna, Holmes, & Spencer, 2001).