Third International Conference on Spoken Language Processing (ICSLP 94)
An experiment is reported in which, on each trial, an ambiguous fragment of auditory speech was delivered on one of an array of hidden loudspeakers, and a face, either upright or upside-down, was shown on a centrally located screen, and articulated one of two different utterances or remained still. Subjects pointed to the apparent origin of the speech sounds and reported what had been said. Identification responses were strongly influenced by the nature of the seen movements (McGurk effect ) and localisation of laterally heard items was shifted toward the centrally located video monitor on trials with a moving face compared to those with a still face (ventriloquism [2,3]). Degree of McGurk interference was practically independent of spatial separation between auditory and visual source. Face inversion had no effect on ventriloquism, but it reduced the McGurk effect. This experimental dissociation suggests that the two phenomena originate in different components of the cognitive architecture.
Bibliographic reference. Bertelson, Paul / Vroomen, Jean / Wiegeraad, Geert / Gelder, Beatrice de (1994): "Exploring the relation between mcgurk interference and ventriloquism", In ICSLP-1994, 559-562.