| The visual system is inundated by a tremendous amount of sensory input that enters from across the retina at a rapid rate. However, only a limited amount of this enters our awareness or influences our behavior. For these reasons, there have long been questions about the amount that can be selected simultaneously and the factors that guide attention to select one subset over another. These issues are investigated using the attentional blink, which occurs when the detection of a target in a rapid stream of distractors leads to impaired detection of a second target that occurs shortly after the first.In Experiments 1-3, this paradigm is used to examine subitizing, the rapid and accurate enumeration of up to 3 or 4 objects. Some propose they can be enumerated preattentively (Trick & Pylyshyn, 1994) or selected by attention in parallel. Others suggest that more resources are necessary for enumerating each additional object within this subitizing range. The enumeration task was performed under conditions that were likely to produce an attentional blink for the to-be-enumerated targets. There were attentional blinks for the numerosities of 2 and 3, suggesting that differential resources are required for enumeration of numerosities within the subitizing range. However, no significant attentional blink was found for 0 or 1 objects, even when presented among distractors.Previous work by Joseph, Chun, and Nakayama (1997) used a similar paradigm and found that singleton detection was drastically impaired during the attentional blink. Since the enumeration of 0 and 1 objects among homogenous distractors is conceptually similar, the finding of preserved performance during the attentional blink seems contradictory. Experiments 4-6 investigated how our earlier findings could be reconciled with the work of Joseph et al. (1997). One particular difference between the studies was that knowledge of the upcoming target feature was available to our participants. This top-down guidance was found to increase the probability of singleton detection during the attentional blink. These findings are discussed with regard to a competitive model of processing, as well as the apparent contradiction of capacity-limited processing and the rich subjective experience of visual awareness. |
