Abstract
Viewers’ attention in film is strongly driven by stimulus features (e.g., motion), while comprehension processes (e.g., the viewer’s event model), have relatively minimal influence, which we call the "Tyranny of Film". Conversely, in picture stories, viewers’ event models have a larger impact on attentional selection. Thus, does removing motion, but maintaining the other filmic features (e.g., shot composition), attenuate the "Tyranny of Film", thereby increasing the influence of comprehension processes on visual selective attention? Or, is removing motion insufficient to allow viewers’ event models to guide attentional selection?
This study used the opening scene of "Touch of Evil". The scene shows a time-bomb put in a car, a couple unknowingly gets in the car and drives away, and the scene ends just before the bomb explodes. Previous work using this clip showed few effects of viewers’ event models on attention--supporting the "Tyranny of Film". This eye-tracking study removed motion by presenting the narrative as a slideshow of single frames from the film, while maintaining the narrative content and viewing duration. Context participants watched the full slideshow. No-context participants didn’t see the bomb or the couple in the car, instead started later when another couple was shown walking down the street. At the end of the slideshow, we asked participants to predict what would happen next.
Context participants were more likely to infer the bomb would explode, showing a critical difference in their event models. Critically, Context participants were more likely to fixate the car throughout the shared viewing period, breaking the "Tyranny of Film". Nevertheless, there were no significant effects of inference generation on this effect. Thus, the Context participants’ greater looks at the car were likely due to treating the couple in the car as protagonists/agents, rather than maintaining the bomb in their event models.