Object substitution masking and object-token individuation

A recently reported form of visual masking - known as Object Substitution Masking (OSM) - occurs when a minimal mask (e.g., four single-pixel dots) is presented simultaneously with the target (e.g., a Landolt-C). When the mask lingers in the display following target offset, severe masking can occur (e.g., Enns & DiLollo, 1997). This is true despite there being little opportunity for contour interference between the mask and target. Using apparent motion to manipulate the representation of the object-token that is associated with the target, we have shown that it is the lingering of that object token that is critical to the masking effect, rather than the lingering of any particular stimulus at the target location. In fact, a mask need never appear at the target location and can be as minimal as a single dot that is much smaller than the target itself (Lleras & Moore, submitted). As long as the spatio-temporal characteristics of the mask are consistent with it being the same object as the target, OSM occurred. These and related results suggest that OSM may be caused by a failure to individuate the mask and target with separate object tokens, before the target offsets. If true, then OSM may provide an index of those processes that normally serve to link previous states of an object with its current state as it moves through space and time. Under this view, the relationship of OSM to this updating process is like that of visual illusions to normal perceptual processes; it is a limitation that occurs under unusual conditions and that can reveal characteristics of the system as it functions under normal conditions. Here we present experiments supporting this interpretation of OSM and its potential as a tool for studying updating processes in object perception.