Abstract
Conscious perception’s role in semantic processing has been in the focus of research for decades. In such research, one typically infers that unconscious semantic processing is possible if invisible masked primes are found to affect responses to visible targets (i.e., response priming) on a categorization task involving semantic processing. Such research has shown that the number of possible targets affects processing level, inducing semantic processing with large target sets and shallow perceptual processing with small sets. Recently, we have explored the role of conscious awareness in semantic processing, using the liminal-prime paradigm. This paradigm allows comparing processing of a liminal masked prime when this prime is consciously perceived vs. missed. We found that semantic processing associated with digits depended, but was not contingent, on conscious perception, whereas low-level perceptual processing was independent of conscious perception. Here, we tested whether these conclusions generalize to broader semantic categories. We manipulated task demands by varying target-set size: four (Exp 1) and forty (Exp 2). Participants categorized target animal names based on the animal’s size. Masked primes included target and non-target stimuli (used and novel primes, respectively). When target-set size was small, used primes generated response priming regardless of conscious perception, but only consciously perceived novel primes generated response priming. By contrast, when target-set was large, novel primes elicited response priming even when these primes were invisible, albeit to a lesser extent compared to visible primes. These findings suggest that task demands modulate the extent of unconscious, but not of conscious, processing.