Abstract
Observers are typically unaware of masked stimuli, but they can become aware of them when they are"unmasked" by subsequent stimuli that mask their mask. In such cases the masked stimulus is considered "disinhibited." Here we show that a masked word can be disinhibited when it is semantically similar to a stimulus that follows its mask. A briefly-exposed word (50 ms) was preceded by a forward mask (300 ms) and followed by a 60-ms backward mask. The backward mask was followed by a silhouette (175 ms) depicting either a nameable real-world object or a novel object. There were two types of novel silhouettes: (1) experimental novel silhouettes suggested portions of real-world objects on their groundside (of which participants were unaware), and (2) control novel silhouettes that did not suggest a real-world object on their groundside. The masked words either matched (i.e, named) or mismatched the real-world objects portrayed by real world silhouettes or suggested on the groundside of experimental novel silhouettes. The words preceding the control novel silhouettes also named real-world objects. On each trial, participants reported whether they saw a word and if so, whether it named the subsequent silhouette. Participants were more likely to see words preceding real-world silhouettes in the match condition (%40) versus all other conditions (~%15; p<0.05). Thus, the masked word was more likely to be disinhibited by a semantically similar real world silhouette, supporting a dynamical view of masking in which semantic processes can play a role. In previous research we showed that semantics are accessed for real-world objects suggested (but not perceived) on the groundside of experimental novel silhouettes. Here semantic access for objects suggested by grounds did not interfere with masking, showing that access to semantics is weaker for real-world objects suggested on the groundside rather than the figure side of a border.
Meeting abstract presented at VSS 2013