The Grammar of Persuasion: A Meta-Analytic Review Disconfirming the Role of Nouns as Linguistic Cues of Subsequent Behavior

Abstract

Recent discoveries suggest that nouns, as a word class, exhibit a greater capacity to model real-life behavior. In the studies by [Proc Natl Acad Sci USA 108, 12653–12656 (2011)] participants prompted with a noun label rather than a semantically corresponding verb (be a voter vs. vote) were more likely to vote. However, many attempts to replicate the original results have shown unsuccessful [Proc Natl Acad Sci USA 113, 7112–7117 (2016)]. To shed light on the effectiveness of subtle linguistic manipulations, we conducted a meta-analysis of 28 effect sizes (N = 13,376), being an exact or conceptual replications of Bryan et al. Although the average effect was significant (g = 0.12), the concurrent diagnostics indicated a low trust in the replicability or stability of the effect. Based on that, we cannot conclude that the greater effectiveness of nouns as cues has been reliably demonstrated. We discuss theoretical prospects involving the verb-action link as well as the self-verification motives.