Deliberative control in audiovisual sociolinguistic perception*

Cognitive models of sociolinguistics must support a wide range of goal-oriented behavior (e.g. Eckert, 2000a) without suggesting unrealistic levels of deliberative control on the part of speakers. The current study investigates the limits of deliberative control in audiovisual face-voice perception. Perceivers evaluated co-present recorded speech and static face pictures, rating the stimuli on the scales 'accented' and 'good-looking' in one of three conditions: as a combined voice and face; evaluating the face while ignoring the voice; and evaluating the voice while ignoring the face. Perceivers' ability to ignore social information from a face or voice upon instruction are taken as indicative of deliberative control in social evaluation. The results suggest that deliberative control and evaluative relevance both play a role in perception, but that available social information is difficult to ignore completely. They further suggest an asymmetry making voices more difficult to ignore than static faces and support a model of sociolinguistic perception and evaluation as a function of multiple competing processes under varying degrees of deliberative control.