Social responses to synthesized speech: Theory and application

The present dissertation examines social responses to synthesized speech in an e-commerce context. Synthesized speech is defined as doubly disembodied language based on two types of disembodiments embedded within synthesized speech. In Experiment 1, participants (N = 40) heard five positive reviews about a book either by five different synthetic voices or by a single synthetic voice. Participants hearing multiple synthetic voices felt greater social presence of the virtual speakers, evaluated the reviewed books more positively, predicted more favorable public reactions to the books, and trusted the web site more. The positive effect of the multiple voice presentation on persuasion was mediated by the feeling of social presence. All the results indicate that people respond socially to synthesized speech. Experiment 2 (N = 40) shows that the social responses observed in Experiment 1 do not disappear even when participants explicitly know the artificiality of synthetic voices. This result provides strong evidence that social responses to synthesized speech cannot be eliminated by technological literacy or awareness. In Experiment 3 (N = 40), participants exhibited the exactly same social responses observed in Experiment 1 and Experiment 2, even when they consciously imagined the actual source of synthesized speech. This result indicates that conscious cognitive imagining of the actual source of the speech does not weaken the effects of automatic perceptual responses to synthetic voice on social presence and social responses. The analyses of the combined data of Experiment 1 and Experiment 3 suggest that automatic imagination of sources based on perceptual responses to voices and conscious imagination of sources based on cognitive instruction work independently for the mental construction of virtual speakers and persuasion. Based on the results of the three experiments, I conclude that doubly disembodied language is understood in the same way as disembodied language, which in turn is understood in the same way as face-to-face communication. Theoretical as well as practical implications of the current dissertation are discussed.