Abstract
Philosophers and psychologists' alike have long debated whether causality can be directly perceived or has to be inferred (Hume, 1740; Rips, 2011; White, 2012; Michotte, 1946). Albert Michotte provided an important contribution to this discussion by presenting evidence that events can be parametrically varied such that they sometimes elicit a causal percept and sometimes do not. Based on this series of experiments, Michotte argued that human observers perceive causality, and that causality is a primary visual property such as color or motion. In this work, we asked whether causal events entered awareness faster than non-causal events, a result that would provide additional evidence for the perceptual nature of causality. In our first experiment, causal (launching) and non-causal (passing) events were presented to human observers while these events were rendered invisible through continuous flash suppression (Tsuchiya & Koch, 2005). We measured the time it took for observers to detect any part of the event. Our results indicated that launch events entered awareness faster than passing events in nearly all observers. In our second experiment, we aimed at replicating the effect observed in the first experiment and we added a control event to address differences in local motion saliency for the launch and pass event. We succeeded in replicating the suppression time difference between launch and pass events. Moreover, launch events also entered awareness faster than control events, which entered awareness equally fast as pass events. Our results therefore imply that causal percepts are at least partially constructed at early stages of visual processing, giving further weight to Michotte's perceptual account of causality perception.
Meeting abstract presented at VSS 2016