Recent studies of attentional orienting to objects in real-world scenes have suggested that object semantics partly determine which objects are most likely to draw attention. However, experimental findings have been quite contradictory. When the semantic consistency of objects in their contextual scene is manipulated, some studies find attentional priority for inconsistent objects and others find attentional priority for consistent objects.

In the present paper, we report performance and eye-movement data from three different paradigms: overt object selection during free scene exploration, covert object selection in an attentional-cueing paradigm and overt object selection under forced-choice conditions.

Based on our findings, we present a working model of the interaction between semantic object-scene consistency and voluntary versus involuntary attention shifts.