| Is subliminal priming subject to the capacity limits and attentional modulations that constrain and empower conscious vision? In the experiments reported here, two subliminal primes are presented -- to the left and right of fixation -- followed by a central supraliminal target. Cues direct spatial attention to prime locations, allowing effects of cued and uncued subliminal primes to be assessed. Studies varying the cue type (exogenous/endogenous), the number (0-2) of the attentional cues, and the stimulus parameters (feature/conjunction tasks) explore respectively the cueing conditions, the capacity limits, and the depth of feature integration in the processing of attended but unconscious stimuli.;When a single prime location is exogenously cued, the time-course of subliminal processing parallels the time-course of conscious processing, with positive priming from the cued prime at short cue-to-prime SOA's (and from the uncued prime at long SOA's, suggestive of "inhibition of return"). While subliminal priming is modulated by exogenous cueing, it may not be subject to the steep capacity limits that constrain conscious processing -- when two primes are cued, both have effects equal in magnitude to those each has when cued alone. However, when neither location is effectively cued, neither prime shows a reliable effect. Notably, under effective endogenous cueing, both cued and uncued primes exhibit positive priming, possibly because of complex attention-related processing at both prime locations. These results suggest that the processing of subliminal primes is strongly cue-dependent but not sharply capacity limited.;When the two primes are different combinations of fixed color and shape features, and subjects classify targets into arbitrary conjunction categories, priming is found only when attention has been focused at one prime location. This priming does not reflect any synthesis of disparate subliminal features into a unified sub-threshold representation -- performance improves monotonically with increasing feature overlap between target and cued prime, without regard to the task-relevant conjunction categories. However, preliminary results suggest that subliminal color and form may be integrated in tasks involving meaningful classifications of naturalistic conjunctions. Subliminal features within the focus of attention prime performance, but the richness of implicit integration may vary with task parameters. |
