Parafoveal-on-foveal priming during eye movements: Interaction between visual and semantic effects

T. Baccino1, F. Lavigne1, F. Vitu2

1Université de Nice Sophia-Antipolis, Laboratoire de Psychologie Expérimentale, 24 Av. des Diables Bleus, F-06000 Nice, France (e-mail:baccino@unice.fr);
2Labo. Psycho. Expérimentale, Centre H.Pieron, Université René Descartes, 28 rue Serpente, F-75006 Paris, France

The interaction between visual and semantic properties of a target word, presented in the parafovea, were investigated in an eye tracking experiment. Participants were required to fixate a 'prompt' word before to make a saccade to a 'target' word and to judge about their semantic association. Three conditions of associations (semantic factor) and two polarities were manipulated (visual factor). The parafoveal 'target' word might be semantically related or unrelated with the 'prompt' word, or might be associated with a 'non-word'. Each pair of 'prompt-target' words were counterbalanced across polarity, either in positive polarity (Dark letters on bright background) or in negative polarity (Bright letters on dark background). Gaze duration on the prompt showed a sensitivity to a semantic pre-processing of the 'target' word only in positive polarity. In this visual condition, 'prompt' word gaze duration was reliably shorter when the parafoveal word was semantically associated. No effect arose in negative polarity. More detailed analyses on the mean fixation duration and single fixations cases confirmed this tendency. Further analyses on saccade extent and landing position on the target will be given later. The results reported here manifestly showed that properties of the parafoveal target word can influence foveal gaze duration and that influence is tightly related to the visual quality of the 'target' word. That effect can be related to well-known facilitation effects showed by priming experiments onto visual degraded target (Stanovitch & West, 1981). Such parafoveal-on-foveal effects are incompatible with models of reading allocating the attention sequentially to successive words. The data suggests rather that foveal and parafoveal information occurs in parallel. The specific interaction between visual/semantic parafoveal effects shed also into light some reading problems occurring on display.

New address:
Francoise Vitu-Thibault
Laboratoire de Psychologie Experimentale,
Centre Universitaire de Boulogne,
Institut de Psychologie,
71 av. Edouard Vaillant,
92774 Boulogne-Billancourt.